Selfdocs 600dpi A4

Complete Listing of CWP Free
Program Self-Documentations
generated by GENDOCS,
a shell script by John Stockwell
Center for Wave Phenomena
Colorado School of Mines
April 10, 2007
Center for Wave Phenomena

Names and Short descriptions of the Codes
CWPROOT = /usr/local/cwp
Mains:
In CWPROOT/src/cwp/main:
* CTRLSTRIP - Strip non-graphic characters
* DOWNFORT - change Fortran programs to lower case, preserving strings
* FCAT - fast cat with 1 read per file
* ISATTY - pass on return from isatty(2)
* MAXINTS - Compute maximum and minimum sizes for integer types
* PAUSE - prompt and wait for user signal to continue
* T - time and date for non-military types
* UPFORT - change Fortran programs to upper case, preserving strings
In CWPROOT/src/par/main:
A2B - convert ascii floats to binary
B2A - convert binary floats to ascii
CSHOTPLOT - convert CSHOT data to files for CWP graphers
DZDV - determine depth derivative with respect to the velocity ",
FARITH - File ARITHmetic -- perform simple arithmetic with binary files
FTNSTRIP - convert a file of binary data plus record delimiters created
FTNUNSTRIP - convert C binary floats to Fortran style floats
GRM - Generalized Reciprocal refraction analysis for a single layer
H2B - convert 8 bit hexidecimal floats to binary
HTI2STIFF - convert HTI parameters alpha, beta, d(V), e(V), gamma
HUDSON - compute effective parameters of anisotropic solids
KAPERTURE - generate the k domain of a line scatterer for a seismic array
LINRORT - linearized P-P, P-S1 and P-S2 reflection coefficients
MAKEVEL - MAKE a VELocity function v(x,y,z)

MKPARFILE - convert ascii to par file format
MRAFXZWT - Multi-Resolution Analysis of a function F(X,Z) by Wavelet
PRPLOT - PRinter PLOT of 1-D arrays f(x1) from a 2-D function f(x1,x2)
RANDVEL3D - Add a random velocity layer (RVL) to a gridded
RAYT2DAN -- P- and SV-wave raytracing in 2D anisotropic media
RAYT2D - traveltime Tables calculated by 2D paraxial RAY tracing
RECAST - RECAST data type (convert from one data type to another)
REFREALAZIHTI - REAL AZImuthal REFL/transm coeff for HTI media
REFREALVTI - REAL REFL/transm coeff for VTI media and symmetry-axis
REGRID3 - REwrite a [ni3][ni2][ni1] GRID to a [no3][no2][no1] 3-D grid
RESAMP - RESAMPle the 1st dimension of a 2-dimensional function f(x1,x2)
1
SMOOTH2 --- SMOOTH a uniformly sampled 2d array of data, within a user-
SMOOTH3D - 3D grid velocity SMOOTHing by the damped least squares
SMOOTHINT2 --- SMOOTH non-uniformly sampled INTerfaces, via the damped
STIFF2VEL - Transforms 2D elastic stiffnesses to (vp,vs,epsilon,delta)
SUBSET - select a SUBSET of the samples from a 3-dimensional file
SWAPBYTES - SWAP the BYTES of various data types
THOM2HTI - Convert Thompson parameters V_p0, V_s0, eps, gamma,
THOM2STIFF - convert Thomsen’s parameters into (density normalized)
TRANSP - TRANSPose an n1 by n2 element matrix
TVNMOQC - Check tnmo-vnmo pairs; create t-v column files
UNIF2ANISO - generate a 2-D UNIFormly sampled profile of elastic
UNIF2 - generate a 2-D UNIFormly sampled velocity profile from a layered
UNISAM2 - UNIformly SAMple a 2-D function f(x1,x2)
UNISAM - UNIformly SAMple a function y(x) specified as x,y pairs
UTMCONV - CONVert longitude and latitude to UTM, and vice versa
VEL2STIFF - Transforms VELocities, densities, and Thomsen or Sayers
VELCONV - VELocity CONVersion
VELPERTAN - Velocity PERTerbation analysis in ANisotropic media to
VELPERT - estimate velocity parameter perturbation from covariance
VTLVZ -- Velocity as function of Time for Linear V(Z);
WKBJ - Compute WKBJ ray theoretic parameters, via finite differencing
XY2Z - converts (X,Y)-pairs to spike Z values on a uniform grid
Z2XYZ - convert binary floats representing Z-values to ascii
In CWPROOT/src/psplot/main:
PSBBOX - change BoundingBOX of existing PostScript file
PSCONTOUR - PostScript CONTOURing of a two-dimensional function f(x1,x2)
PSCUBE - PostScript image plot with Legend of a data CUBE
PSCCONTOUR - PostScript Contour plot of a data CUBE
PSEPSI - add an EPSI formatted preview bitmap to an EPS file
PSGRAPH - PostScript GRAPHer
PSIMAGE - PostScript IMAGE plot of a uniformly-sampled function f(x1,x2)
PSLABEL - output PostScript file consisting of a single TEXT string
PSMANAGER - printer MANAGER for HP 4MV and HP 5Si Mx Laserjet
PSMERGE - MERGE PostScript files
PSMOVIE - PostScript MOVIE plot of a uniformly-sampled function f(x1,x2,x3)
PSWIGB - PostScript WIGgle-trace plot of f(x1,x2) via Bitmap
PSWIGP - PostScript WIGgle-trace plot of f(x1,x2) via Polygons
In CWPROOT/src/xplot/main:
* LCMAP - List Color Map of root window of default screen
* LPROP - List PROPerties of root window of default screen of display
* SCMAP - set default standard color map (RGB_DEFAULT_MAP)
2
XCONTOUR - X CONTOUR plot of f(x1,x2) via vector plot call
* XESPB - X windows display of Encapsulated PostScript as a single Bitmap
* XEPSP - X windows display of Encapsulated PostScript
XIMAGE - X IMAGE plot of a uniformly-sampled function f(x1,x2)
XPICKER - X wiggle-trace plot of f(x1,x2) via Bitmap with PICKing
* XPSP - Display conforming PostScript in an X-window
XWIGB - X WIGgle-trace plot of f(x1,x2) via Bitmap
In CWPROOT/src/Xtcwp/main:
XGRAPH - X GRAPHer
XMOVIE - image one or more frames of a uniformly sampled function f(x1,x2)
XRECTS - plot rectangles on a two-dimensional grid
In CWPROOT/src/Xmcwp/main:
* FFTLAB - Motif-X based graphical 1D Fourier Transform
In CWPROOT/src/su/graphics/psplot:
SUPSCONTOUR - PostScript CONTOUR plot of a segy data set
SUPSCUBE - PostScript CUBE plot of a segy data set
SUPSCUBECONTOUR - PostScript CUBE plot of a segy data set
SUPSGRAPH - PostScript GRAPH plot of a segy data set
SUPSIMAGE - PostScript IMAGE plot of a segy data set
SUPSMAX - PostScript of the MAX, min, or absolute max value on each trace
SUPSMOVIE - PostScript MOVIE plot of a segy data set
SUPSWIGB - PostScript Bit-mapped WIGgle plot of a segy data set
SUPSWIGP - PostScript Polygon-filled WIGgle plot of a segy data set
In CWPROOT/src/su/main:
BHEDTOPAR - convert a Binary tape HEaDer file to PAR file format
DT1TOSU - Convert ground-penetrating radar data in the Sensors & Software
LAS2SU - convert las2 format well log curves to su traces
SEGYCLEAN - zero out unassigned portion of header
char* sdoc[] = {
SEGYHDRS - make SEG-Y ascii and binary headers for segywrite
SEGYREAD - read an SEG-Y tape
SEGYWRITE - write an SEG-Y tape
SETBHED - SET the fields in a SEGY Binary tape HEaDer file, as would be
SU3DCHART - plot x-midpoints vs. y-midpoints for 3-D data
SUABSHW - replace header key word by its absolute value
SUACOR - auto-correlation
char* sdoc[] = {
SUADDNOISE - add noise to traces
3
SUADDSTATICS - ADD random STATICS on seismic data
SUALFORD - trace by trace Alford Rotation of shear wave data volumes
SUAMP - output amp, phase, real or imag trace from
SUASCII - print non zero header values and data in various formats
SUATTRIBUTES - instantaneous trace ATTRIBUTES
SUAZIMUTH - compute trace AZIMUTH, offset, and midpoint coordinates
SUBFILT - apply Butterworth bandpass filter
SUCCWT - Complex continuous wavelet transform of seismic traces
SUCDDECON - DECONvolution with user-supplied filter by straightforward
SUCENTSAMP - CENTRoid SAMPle seismic traces
SUCHART - prepare data for x vs. y plot
SUCHW - Change Header Word using one or two header word fields
SUCLIPHEAD - Clip header values
char* sdoc[] = {
SUCOMMAND - pipe traces having the same key header word to command
SUCONV - convolution with user-supplied filter
SUCOUNTKEY - COUNT the number of unique values for a given KEYword.
SUCWT - generates Continous Wavelet Transform amplitude, regularity
SUDATUMFD - 2D zero-offset Finite Difference acoustic wave-equation
SUDATUMK2DR - Kirchhoff datuming of receivers for 2D prestack data
SUDATUMK2DS - Kirchhoff datuming of sources for 2D prestack data
SUDIPDIVCOR - Dip-dependent Divergence (spreading) correction
SUDIPFILT - DIP--or better--SLOPE Filter in f-k domain
SUDIVCOR - Divergence (spreading) correction
SUDIVSTACK - Diversity Stacking using either average power or peak
SUDMOFK - DMO via F-K domain (log-stretch) method for common-offset gathers
SUDMOFKCW - converted-wave DMO via F-K domain (log-stretch) method for
SUDMOTIVZ - DMO for Transeversely Isotropic V(Z) media for common-offset
SUDMOTX - DMO via T-X domain (Kirchhoff) method for common-offset gathers
SUDMOVZ - DMO for V(Z) media for common-offset gathers
SUDUMPTRACE - print selected header values and data.
SUEA2DF - SU version of (an)elastic anisotropic 2D finite difference
char* sdoc[] = {
SUEIPOFI - EIgenimage (SVD) based POlarization FIlter for
SUFCTANISMOD - Flux-Corrected Transport correction applied to the 2D
SUFDMOD1 - Finite difference modelling (1-D 1rst order) for the
SUFDMOD2 - Finite-Difference MODeling (2nd order) for acoustic wave equation
SUFDMOD2_PML - Finite-Difference MODeling (2nd order) for acoustic wave
SUFFT - fft real time traces to complex frequency traces
SUFILTER - applies a zero-phase, sine-squared tapered filter
SUFLIP - flip a data set in various ways
SUFNZERO - get Time of First Non-ZERO sample by trace
SUFRAC -- take general (fractional) time derivative or integral of
4
SUFXDECON - random noise attenuation by FX-DECONvolution
SUGABOR - Outputs a time-frequency representation of seismic data via
SUGAIN - apply various types of gain to display traces
SUGAUSSTAPER - Multiply traces with gaussian taper
SUGAZMIGQ - SU version of Jeno GAZDAG’s phase-shift migration
SUGET - Connect SU program to file descriptor for input stream.
SUGETHW - sugethw writes the values of the selected key words
SUGOUPILLAUD - calculate 1D impulse response of
SUGOUPILLAUDPO - calculate Primaries-Only impulse response of a lossless
SUHARLAN - signal-noise separation by the invertible linear
SUHILB - Hilbert transform
char *sdoc[] = {
SUHROT - Horizontal ROTation of three-component data
SUHTMATH - do unary arithmetic operation on segy traces with
SUIFFT - fft complex frequency traces to real time traces
SUILOG -- time axis inverse log-stretch of seismic traces
SUIMP2D - generate shot records for a line scatterer
SUIMP3D - generate inplane shot records for a point
SUIMPEDANCE - Convert reflection coefficients to impedances.
SUINTERP - interpolate traces using automatic event picking
SUINTVEL - convert stacking velocity model to interval velocity model
SUINVXZCO - Seismic INVersion of Common Offset data for a smooth
SUINVZCO3D - Seismic INVersion of Common Offset data with V(Z) velocity
SUK1K2FILTER - symmetric box-like K-domain filter defined by the
SUKDMIG2D - Kirchhoff Depth Migration of 2D poststack/prestack data
SUKDSYN2D - Kirchhoff Depth SYNthesis of 2D seismic data from a
SUKEYCOUNT - sukeycount writes a count of a selected key
SUKFILTER - radially symmetric K-domain, sin^2-tapered, polygonal
SUKFRAC - apply FRACtional powers of i|k| to data, with phase shift
SUKILL - zero out traces
SULOG -- time axis log-stretch of seismic traces
char *sdoc[] = {
SUMEAN - get the mean values of data traces ",
SUMEDIAN - MEDIAN filter about a user-defined polygonal curve with
SUMIGFD - 45-90 degree Finite difference depth migration for
SUMIGFFD - Fourier finite difference depth migration for
SUMIGGBZOAN - MIGration via Gaussian beams ANisotropic media (P-wave)
SUMIGGBZO - MIGration via Gaussian Beams of Zero-Offset SU data
SUMIGPREFD --- The 2-D prestack common-shot 45-90 degree
SUMIGPREFFD - The 2-D prestack common-shot Fourier finite-difference
SUMIGPREPSPI --- The 2-D PREstack commom-shot Phase-Shift-Plus
SUMIGPRESP - The 2-D prestack common-shot split-step Fourier ",
5
SUMIGPS - MIGration by Phase Shift with turning rays
SUMIGPSPI - Gazdag’s phase-shift plus interpolation depth migration
SUMIGPSTI - MIGration by Phase Shift for TI media with turning rays
SUMIGSPLIT - Split-step depth migration for zero-offset data.
SUMIGTK - MIGration via T-K domain method for common-midpoint stacked data
SUMIGTOPO2D - Kirchhoff Depth Migration of 2D postack/prestack data
SUMIX - compute weighted moving average (trace MIX) on a panel
SUMIXGATHERS - mix two gathers
SUMUTE - mute above (or below) a user-defined polygonal curve with ",
SUNAN - remove NaNs & Infs from the input stream
SUNHMOSPIKE - generates SPIKE test data set with a choice of several
SUNMO - NMO for an arbitrary velocity function of time and CDP
SUNORMALIZE - Trace balancing by rms, max, or median ",
SUNULL - create null (all zeroes) traces
SUOCEXT - smaller Offset EXTrapolation via Offset Continuation
SUOLDTONEW - convert existing su data to xdr format
SUOP2 - do a binary operation on two data sets
SUOP - do unary arithmetic operation on segys
SUPACK1 - pack segy trace data into chars
SUPACK2 - pack segy trace data into 2 byte shorts
SUPASTE - paste existing SEGY headers on existing data
SUPEF - Wiener predictive error filtering
SUPERMUTE - permute or transpose a 3d datacube
SUPGC - Programmed Gain Control--apply agc like function
SUPICKAMP - pick amplitudes within user defined and resampled window
SUPLANE - create common offset data file with up to 3 planes
SUPOFILT - POlarization FILTer for three-component data
SUPOLAR - POLarization analysis of three-component data
SUPUT - Connect SU program to file descriptor for output stream.
SUPWS - Phase stack or phase-weighted stack (PWS) of adjacent traces
SUQUANTILE - display some quantiles or ranks of a data set
SURADON - compute forward or reverse Radon transform or remove multiples
SURAMP - Linearly taper the start and/or end of traces to zero.
SURANDSPIKE - make a small data set of RANDom SPIKEs
SURANDSTAT - Add RANDom time shifts STATIC errors to seismic traces
SURANGE - get max and min values for non-zero header entries
SURECIP - sum opposing offsets in prepared data (see below)
SUREDUCE - convert traces to display in reduced time ",
SUREFCON - Convolution of user-supplied Forward and Reverse
SURELANAN - REsiduaL-moveout semblance ANalysis for ANisotropic media
SURELAN - compute residual-moveout semblance for cdp gathers based
SURESAMP - Resample in time
SURESSTAT - Surface consistent source and receiver statics calculation
6
SUSHAPE - Wiener shaping filter
SUSHIFT - shifted/windowed traces in time
SUSHW - Set one or more Header Words using trace number, mod and
SUSORT - sort on any segy header keywords
SUSORTY - make a small 2-D common shot off-end
SUSPECFK - F-K Fourier SPECtrum of data set
SUSPECFX - Fourier SPECtrum (T -> F) of traces
SUSPECK1K2 - 2D (K1,K2) Fourier SPECtrum of (x1,x2) data set
SUSPIKE - make a small spike data set
SUSPLIT - Split traces into different output files by keyword value
SUSTACK - stack adjacent traces having the same key header word
SUSTATIC - Elevation static corrections, apply corrections from
SUSTATICRRS - Elevation STATIC corrections, apply corrections from
SUSTKVEL - convert constant dip layer interval velocity model to the
SUSTOLT - Stolt migration for stacked data or common-offset gathers
SUSTRIP - remove the SEGY headers from the traces
SUSWAPBYTES - SWAP the BYTES in SU data to convert data from big endian
SUSYNCZ - SYNthetic seismograms for piecewise constant V(Z) function
SUSYNLV - SYNthetic seismograms for Linear Velocity function
SUSYNLVCW - SYNthetic seismograms for Linear Velocity function
SUSYNLVFTI - SYNthetic seismograms for Linear Velocity function in a
SUSYNVXZ - SYNthetic seismograms of common offset V(X,Z) media via
SUSYNVXZCS - SYNthetic seismograms of common shot in V(X,Z) media via
SUTAB - print non zero header values and data for non-graphic terminals
SUTAPER - Taper the edge traces of a data panel to zero.
SUTAUP - forwared and inverse T-X and F-K global slant stacks
SUTIFOWLER VTI constant velocity prestack time migration
SUTIHALEDMO - TI Hale Dip MoveOut (based on Hale’s PhD thesis)
SUTIVEL - SU Transversely Isotropic velocity table builder
SUTSQ -- time axis time-squared stretch of seismic traces
SUTTOZ - resample from time to depth
SUTVBAND - time-variant bandpass filter (sine-squared taper)
SUTXTAPER - TAPER in (X,T) the edges of a data panel to zero.
SUUNPACK1 - unpack segy trace data from chars to floats
SUUNPACK2 - unpack segy trace data from shorts to floats
SUUTM - UTM projection of longitude and latitude in SU trace headers
SUVCAT - append one data set to another, with or without an ",
SUVEL2DF - compute stacking VELocity semblance for a single time in
SUVELAN - compute stacking velocity semblance for cdp gathers
SUVELAN_NCCS - compute stacking VELocity panel for cdp gathers
SUVELAN_NSEL - compute stacking VELocity panel for cdp gathers
SUVELAN_UCCS - compute stacking VELocity panel for cdp gathers
SUVELAN_USEL - compute stacking velocity panel for cdp gathers
7
SUVIBRO - Generates a Vibroseis sweep (linear, linear-segment,
SUVLENGTH - Adjust variable length traces to common length
SUWAVEFORM - generate a seismic wavelet
SUWEIGHT - weight traces by header parameter, such as offset
SUWELLRF - convert WELL log depth, velocity, density data into a
SUWIND - window traces by key word
SUXCOR - correlation with user-supplied filter
SUXEDIT - examine segy diskfiles and edit headers
SUZERO -- zero-out (or set constant) data within a time window
SUZTOT - resample from depth to time
In CWPROOT/src/su/graphics/psplot:
In CWPROOT/src/su/graphics/xplot:
SUXCONTOUR - X CONTOUR plot of Seismic UNIX tracefile via vector plot call
SUXGRAPH - X-windows GRAPH plot of a segy data set
SUXIMAGE - X-windows IMAGE plot of a segy data set
SUXMAX - X-windows graph of the MAX, min, or absolute max value on
SUXMOVIE - X MOVIE plot of a 2D or 3D segy data set
SUXPICKER - X-windows WIGgle plot PICKER of a segy data set
SUXWIGB - X-windows Bit-mapped WIGgle plot of a segy data set
In CWPROOT/src/tri/main:
GBBEAM - Gaussian beam synthetic seismograms for a sloth model
NORMRAY - dynamic ray tracing for normal incidence rays in a sloth model
TRI2UNI - convert a TRIangulated model to UNIformly sampled model
TRIMODEL - make a triangulated sloth (1/velocity^2) model
TRIRAY - dynamic RAY tracing for a TRIangulated sloth model
TRISEIS - Gaussian beam synthetic seismograms for a sloth model
UNI2TRI - convert UNIformly sampled model to a TRIangulated model
In CWPROOT/src/xtri:
SXPLOT - X Window plot a triangulated sloth function s(x1,x2)
8
In CWPROOT/src/tri/graphics/psplot:
SPSPLOT - plot a triangulated sloth function s(x,z) via PostScript
In CWPROOT/src/comp/dct/main:
DCTCOMP - Compression by Discrete Cosine Transform
DCTUNCOMP - Discrete Cosine Transform Uncompression
ENTROPY - compute the ENTROPY of a signal
WPTCOMP - Compression by Wavelet Packet Compression
WPTUNCOMP - Uncompress WPT compressed data
WTCOMP - Compression by Wavelet Transform
WTUNCOMP - UNCOMPression of WT compressed data
In CWPROOT/src/comp/dwpt/1d/main:
WPC1COMP2 --- COMPress a 2D seismic section trace-by-trace using
WPC1UNCOMP2 --- UNCOMPRESS a 2D seismic section, which has been
In CWPROOT/src/comp/dwpt/2d/main:
WPCCOMPRESS --- COMPRESS a 2D section using Wavelet Packets
WPCUNCOMPRESS --- UNCOMPRESS a 2D section
Shells:
In CWPROOT/src/cwp/shell:
# ARGV - give examples of dereferencing char **argv
# COPYRIGHT - insert CSM COPYRIGHT lines at top of files in working directory
# CPALL , RCPALL - for local and remote directory tree/file transfer
# CWPFIND - look for files with patterns in CWPROOT/src/cwp/lib
# DIRTREE - show DIRectory TREE
# FILETYPE - list all files of given type
# Grep - recursively call egrep in pwd
# NEWCASE - Changes the case of all the filenames in a directory, dir
# OVERWRITE - copy stdin to stdout after EOF
# PRECEDENCE - give table of C precedences from Kernighan and Ritchie
# REPLACE - REPLACE string1 with string2 in files
# THIS_YEAR - print the current year
# TIME_NOW - prints time in ZULU format with no spaces
# TODAYS_DATE - prints today’s date in ZULU format with no spaces
# USERNAMES - get list of all login names
# VARLIST - list variables used in a Fortran program
# WEEKDAY - prints today’s WEEKDAY designation
# ZAP - kill processes by name
In CWPROOT/src/par/shell:
9
# GENDOCS - generate complete list of selfdocs in latex form
# STRIPTOTXT - put files from $CWPROOT/src/doc/Stripped into a new
# UPDATEDOCALL - put self-docs in ../doc/Stripped
# UPDATEDOC - put self-docs in ../doc/Stripped and ../doc/Headers
# UPDATEHEAD - update ../doc/Headers/Headers.all
In CWPROOT/src/psplot/shell:
# MERGE2 - put 2 standard size PostScript figures on one page
# MERGE4 - put 4 standard size PostScript plots on one page
In CWPROOT/src/su/shell:
# LOOKPAR - show getpar lines in SU code with defines evaluated
# MAXDIFF - find absolute maximum difference in two segy data sets
# RECIP - sum opposing (reciprocal) offsets in cdp sorted data
# RMAXDIFF - find percentage maximum difference in two segy data sets
# SUAGC - perform agc on SU data
# SUBAND - Trapezoid-like Sin squared tapered Bandpass filter via SUFILTER
# SUDIFF, SUSUM, SUPROD, SUQUO, SUPTDIFF, SUPTSUM,
# SUDOC - get DOC listing for code
# SUENV - Instantaneous amplitude, frequency, and phase via: SUATTRIBUTES
# SUFIND - get info from self-docs
# SUFIND - get info from self-docs
# SUGENDOCS - generate complete list of selfdocs in latex form
# SUHELP - list the CWP/SU programs and shells
# SUKEYWORD -- guide to SU keywords in segy.h
# SUNAME - get name line from self-docs
# UNGLITCH - zonk outliers in data
Libs:
In CWPROOT/src/cwp/lib:
ABEL - Functions to compute the discrete ABEL transform:
AIRY - Approximate the Airy functions Ai(x), Bi(x) and their respective
ALLOC - Allocate and free multi-dimensional arrays
ANTIALIAS - Butterworth anti-aliasing filter
AXB - Functions to solve a linear system of equations Ax=b by LU
BIGMATRIX - Functions to manipulate 2-dimensional matrices that are too big
BUTTERWORTH - Functions to design and apply Butterworth filters:
COMPLEX - Functions to manipulate complex numbers
COMPLEXD - Functions to manipulate double-precision complex numbers
COMPLEXF - Subroutines to perform operations on complex numbers.
COMPLEXFD - Subroutines to perform operations on double complex numbers.
CONVOLUTION - Compute z = x convolved with y
10
CUBICSPLINE - Functions to compute CUBIC SPLINE interpolation coefficients
DBLAS - Double precision Basic Linear Algebra subroutines
DGE - Double precision Gaussian Elimination matrix subroutines adapted
PFAFFT - Functions to perform Prime Factor (PFA) FFT’s, in place
*CWP_Exit - exit subroutine for CWP/SU codes
FRANNOR - functions to generate a pseudo-random float normally distributed
FRANUNI - Functions to generate a pseudo-random float uniformly distributed
HANKEL - Functions to compute discrete Hankel transforms
HILBERT - Compute Hilbert transform y of x
HOLBERG1D - Compute coefficients of Holberg’s 1st derivative filter
INTCUB - evaluate y(x), y’(x), y’’(x), ... via piecewise cubic interpolation
INTL2B - bilinear interpolation of a 2-D array of bytes
INTLIN - evaluate y(x) via linear interpolation of y(x[0]), y(x[1]), ...
INTLINC - evaluate complex y(x) via linear interpolation of y(x[0]), y(x[1]), ...
INTSINC8 - Functions to interpolate uniformly-sampled data via 8-coeff. sinc
INTTABLE8 - Interpolation of a uniformly-sampled complex function y(x)
LINEAR_REGRESSION - Compute linear regression of (y1,y2,...,yn) against
MKDIFF - Make an n-th order DIFFerentiator via Taylor’s series method.
MKHDIFF - Compute filter approximating the bandlimited HalF-DIFFerentiator.
MKSINC - Compute least-squares optimal sinc interpolation coefficients.
MNEWT - Solve non-linear system of equations f(x) = 0 via Newton’s method
POLAR - Functions to map data in rectangular coordinates to polar and vise versa
PRINTERPLOT - Functions to make a printer plot of a 1-dimensional array
QUEST - Functions to ESTimate Quantiles:
RESSINC8 - Functions to resample uniformly-sampled data via 8-coefficient sinc
RFWTVA - Rasterize a Float array as Wiggle-Trace-Variable-Area.
RFWTVAINT - Rasterize a Float array as Wiggle-Trace-Variable-Area, with
SBLAS - Single precision Basic Linear Algebra Subroutines
SCAXIS - compute a readable scale for use in plotting axes
SGA - Single precision general matrix functions adapted from LINPACK FORTRAN:
SHFS8R - Shift a uniformly-sampled real-valued function y(x) via
SINC - Return SINC(x) for as floats or as doubles
SORT - Functions to sort arrays of data or arrays of indices
SQR - Single precision QR decomposition functions adapted from LINPACK FORTRAN:
STOEP - Functions to solve a symmetric Toeplitz linear system of equations
STRSTUFF -- STRing manuplation subs
SWAPBYTE - Functions to SWAP the BYTE order of binary data
SYMMEIGEN - Functions solving the eigenvalue problem for symmetric matrices
TEMPORARY_FILENAME - Creates a file name in a user-specified directory.
TRIDIAGONAL - Functions to solve tridiagonal systems of equations Tu=r for u.
VANDERMONDE - Functions to solve Vandermonde system of equations Vx=b
WAVEFORMS Subroutines to define some wavelets for modeling of seismic
11
XCOR - Compute z = x cross-correlated with y
XINDEX - determine index of x with respect to an array of x values
YCLIP - Clip a function y(x) defined by linear interpolation of the
YXTOXY - Compute a regularly-sampled, monotonically increasing function x(y)
ZASC - routine to translate ncharacters from ebcdic to ascii
ZEBC - routine to translate ncharacters from ascii to ebcdic
In CWPROOT/src/par/lib:
ATOPKGE - convert ascii to arithmetic and with error checking
DOCPKGE - Function to implement the CWP self-documentation facility
EALLOC - Allocate and free multi-dimensional arrays with error reports.
ERRPKGE - routines for reporting errors
FILESTAT - Functions to determine and output the type of a file from file
GETPARS - Functions to GET PARameterS from the command line. Numeric
LINCOEFF - subroutines to create linearized reflection coefficients
MINFUNC - routines to MINimize FUNCtions
MODELING - Seismic Modeling Subroutines for SUSYNLV and clones
REFANISO - Reflection coefficients for Anisotropic media
RKE - integrate a system of n-first order ordinary differential equations
SMOOTH - Functions to compute smoothing of 1-D or 2-D input data
SUBCALLS - routines for system functions with error checking
SYSCALLS - routines for SYSTEM CALLs with error checking
TAUP - Functions to perform forward and inverse taup transforms (radon or
UPWEIK - Upwind Finite Difference Eikonal Solver
VND - large out-of-core multidimensional block matrix transpose
WTLIB - Functions for wavelet transforms
In CWPROOT/src/su/lib:
fgethdr - get segy tape identification headers from the file by file pointer
FGETTR - Routines to get an SU trace from a file
FPUTTR - Routines to put an SU trace to a file
HDRPKGE - routines to access the SEGY header via the hdr structure.
TABPLOT - TABPLOT selected sample points on selected trace
VALPKGE - routines to handle variables of type Value
In CWPROOT/src/psplot/lib:
BASIC - Basic C function interface to PostScript
PSAXESBOX3 - Functions draw an axes box via PostScript, estimate bounding box
PSAXESBOX - Functions to draw PostScript axes and estimate bounding box
PSCAXESBOX - Draw an axes box for cube via PostScript
PSCONTOUR - draw contour of a two-dimensional array via PostScript
PSDRAWCURVE - Functions to draw a curve from a set of points
PSLEGENDBOX - Functions to draw PostScript axes and estimate bounding box
12
PSWIGGLE - draw wiggle-trace with (optional) area-fill via PostScript
In CWPROOT/src/xplot/lib:
AXESBOX - Functions to draw axes in X-windows graphics
COLORMAP - Functions to manipulate X colormaps:
DRAWCURVE - Functions to draw a curve from a set of points
IMAGE - Function for making the image in an X-windows image plot
LEGENDBOX - draw a labeled axes box for a legend (i.e. colorscale)
RUBBERBOX - Function to draw a rubberband box in X-windows plots
WINDOW - Function to create a window in X-windows graphics
XCONTOUR - draw contour of a two-dimensional array via X vectorplot calls
In CWPROOT/src/Xtcwp/lib:
AXES - the Axes Widget
FX - Functions to support floating point coordinates in X
MISC - Miscellaneous X-Toolkit functions
RESCONV - general purpose resource type converters
In CWPROOT/src/Xmcwp/lib:
RADIOBUTTONS - convenience functions creating and using radio buttons
SAMPLES - Motif-based Graphics Functions
In CWPROOT/src/tri/lib:
CHECK - CHECK triangulated models
CIRCUM - define CIRCUMcircles for Delaunay triangulation
COLINEAR - determine if edges or vertecies are COLINEAR in triangulated
CREATE - create model, boundary edge triangles, edge face, edge vertex, add
DELETE - DELETE vertex, model, edge, or boundary edge from triangulated model
DTE - Distance to Edge
FIXEDGES - FIX or unFIX EDGES between verticies
INSIDE - Is a vertex or point inside a circum circle, etc. of a triangulated
NEAREST - NEAREST edge or vertex in triangulated model
PROJECT - project to edge in triangulated model
READWRITE - READ or WRITE a triangulated model
In CWPROOT/src/cwputils:
CPUSEC - return cpu time (UNIX user time) in seconds
CPUTIME - return cpu time (UNIX user time) in seconds using ANSI C built-ins
WALLSEC - Functions to time processes
WALLTIME - Function to show time a process takes to run
13
In CWPROOT/src/comp/dct/lib:
DCT1 - 1D Discreet Cosine Transform Routines
DCT2 - 2D Discrete Cosine Transform Routines
DCTALLOC - ALLOCate space for transform tables for 1D DCT
GETFILTER - GET wavelet FILTER type
LCT1 - functions used to perform the 1D Local Cosine Transform (LCT)
LPRED - Lateral Prediction of Several Plane Waves
WAVEPACK1 - 1D wavelet packet transform
WAVEPACK2 - 2D Wavelet PACKet transform
WAVEPACK1 - 1D wavelet packet transform
WAVETRANS2 - 2D wavelet transform by tensor-product of two 1D transforms
In CWPROOT/src/comp/dct/libutil:
BUFFALLOC - routines to ALLOCate/initialize and free BUFFers
HUFFMAN - Routines for in memory Huffman coding/decoding
PENCODING - Routines to en/decode the quantized integers for lossless
QUANT - QUANTization routines
RLE - routines for in memory silence en/decoding
In CWPROOT/src/comp/dwpt/1d/lib:
WBUFFALLOC - routines to allocate/initialize and free buffers in wavelet
WPC1 - routines for compress a single seismic trace using wavelet packets
WPC1CODING - routines for encoding the integer symbols in 1D WPC
wpc1Quant - quantize
WPC1TRANS - routines to perform a 1D wavelet packet transform
In CWPROOT/src/comp/dwpt/2d/lib:
WPCBUFFAL - routines to allocate/initialize and free buffers
WPC - routines for compress a 2D seismic section using wavelet packets
WPCCODING - Routines for in memory coding of the quantized coeffiecients
WPCENDEC - Wavelet Packet Coding, Encoding and Decoding routines
WPCHUFF -Routines for in memory Huffman coding/decoding
WPCPACK2 - routine to perform a 2D wavelet packet transform
WPCQUANT - quantization routines for WPC
WPCSILENCE - routines for in memory silence en/decoding
To search on a program name fragment, type:

suname name_fragment <CR>
For more information type: program_name <CR>
Items labeled with an asterisk (*) are C programs that may

or may not have this self documentation feature.
14
Items labeled with a pound sign (#) are shell scripts that may,
or may not have the self documentation feature.
15
Self Documentations
Mains:
______
CTRLSTRIP - Strip non-graphic characters
ctrlstrip <dirtyfile >cleanfile
______
DOWNFORT - change Fortran programs to lower case, preserving strings
Usage: downfort < infile.f > outfile.f
Credits:
Brian Sumner c. 1984
______
FCAT - fast cat with 1 read per file
Usage: fcat file1 file2 ... > file3
Credits:
Shuki
This program belongs to the Center for Wave Phenomena

/
______
ISATTY - pass on return from isatty(2)
Usage: isatty filedes
See: man isatty for further information
16
Credits:
CWP: Shuki

______
MAXINTS - Compute maximum and minimum sizes for integer types
(quick and dirty)
Usage: maxints
Note: These results will be in limits.h on most systems
Credits:
CWP: Jack K. Cohen

______
PAUSE - prompt and wait for user signal to continue
Usage: pause [optional arguments]
Note:
Default prompt is "press return key to continue" which is *evoked
by calling pause with no arguments. The word,
"continue", is replaced by any optional arguments handed to pause.
Thus, the command "pause do plot" will evoke the prompt,
"press return key to do plot".
______
T - time and date for non-military types
17
Usage: t
Credit: Jack
______
UPFORT - change Fortran programs to upper case, preserving strings
Usage: upfort < infile.f > outfile.f
Reverse of: downfort
18
A2B - convert ascii floats to binary
a2b <stdin >stdout outpar=/dev/null
Required parameters:
none
Optional parameters:
n1=2 floats per line in input file
outpar=/dev/null output parameter file, contains the

number of lines (n=)
other choices for outpar are: /dev/tty,
/dev/stderr, or a name of a disk file
Credits:
CWP: Jack K. Cohen, Dave Hale
Hans Ecke 2002: Replaced line-wise file reading via gets() with
float-wise reading via fscanf(). This makes it
much more robust: it does not impose a specific
structure on the input file.
19
B2A - convert binary floats to ascii
b2a <stdin >stdout
none
n1=2 floats per line in output file
outpar=/dev/tty output parameter file, contains the

Credits:
CWP: Jack K. Cohen
20
CSHOTPLOT - convert CSHOT data to files for CWP graphers
cshotplot <cshot1plot [optional parameter file]
none
Optional parameter:
outpar=/dev/tty output parameter file, contains:
number of plots (n2=)
points in each plot (n1=)
colors for plots (linecolor=)
21
DZDV - determine depth derivative with respect to the velocity ",
parameter, dz/dv, by ratios of migrated data with the primary
amplitude and those with the extra amplitude
dzdv <infile afile=afile dfile=dfile>outfile [parameters]
Required Parameters:
infile= input of common image gathers with primary amplitude
afile= input of common image gathers with extra amplitude
dfile= output of imaged depths in common image gathers
outfile= output of dz/dv at the imaged points
nx= number of migrated traces
nz= number of points in migrated traces
dx= horizontal spacing of migrated trace
dz= vertical spacing of output trace
fx= x-coordinate of first migrated trace
fz= z-coordinate of first point in migrated trace
off0= first offset in common image gathers
noff= number of offsets in common image gathers
doff= offset increment in common image gathers
cip=x1,z1,r1,..., cip=xn,zn,rn description of input CIGS
x x-value of a common image point
z z-value of a common image point at zero offset
r r-parameter in a common image gather
Optional Parameters:
nxw, nzw=0 window widths along x- and z-directions in
which points are contributed in solving dz/dv.
Notes:
This program is used as part of the velocity analysis technique developed
by Zhenyue Liu, CWP:1995.
Author: CWP: Zhenyue Liu, 1995
Reference:
Liu, Z. 1995, "Migration Velocity Analysis", Ph.D. Thesis, Colorado
School of Mines, CWP report #168.
22
FARITH - File ARITHmetic -- perform simple arithmetic with binary files
farith <infile >outfile [optional parameters]
in=stdin input file
out=stdout output file
in2= second input file (required for binary operations)
if it can’t be opened as a file, it might be a scalar
n=size_of_in, fastest dimension (used only for op=cartprod is set)
isig= index at which signum function acts (used only for
op=signum)
scale= value to scale in by, used only for op=scale)
bias= value to bias in by, used only for op=bias)
op=noop noop for out = in

neg for out = -in
abs for out = abs(in)
scale for out = in *scale
bias for out = in + bias
exp for out = exp(in)
log for out = log(in)
sqrt for out = (signed) sqrt(in)
sqr for out = in*in
pinv for out = (punctuated) 1 / in
pinvsqr for out = (punctuated) 1 /in*in
pinvsqrt for out = (punctuated signed) 1 /sqrt(in)
add for out = in + in2
sub for out = in - in2
mul for out = in * in2
div for out = in / in2
cartprod for out = in x in2
requires: n=size_of_in, fastest dimension in output
signum for out[i] = in[i] for i< isig and
= -in[i] for i>= isig
requires: isig=point where signum function acts
Seismic operations:
slowp for out = 1/in - 1/in2 Slowness perturbation
slothp for out = 1/in^2 - 1/in2^2 Sloth perturbation
Notes:
op=sqrt takes sqrt(x) for x>=0 and -sqrt(ABS(x)) for x<0 (signed sqrt)
23
op=pinv takes y=1/x for x!=0, if x=0 then y=0. (punctuated inverse)
The seismic operations assume that in and in2 are wavespeed profiles.
"Slowness" is 1/wavespeed and "sloth" is 1/wavespeed^2.
Use "suop" and "suop2" to perform unary and binary operations on
data in the SU (SEGY trace) format.
The options "pinvsq" and "pinvsqrt" are also useful for seismic
computations involving converting velocity to sloth and vice versa.
The option "cartprod" (cartesian product) requires also that the

parameter n=size_of_in be set. This will be the fastest dimension
of the rectangular array that is output.
The option "signum" causes a flip in sign for all values with index
greater than "isig" (really -1*signum(index)).
For file operations on SU format files, please use: suop, suop2
AUTHOR: Dave Hale, Colorado School of Mines, 07/07/89

Zhaobo Meng added scale and cartprod, 10/01/96
Zhaobo Meng added signum, 9 May 1997
Tony Kocurko added scalar operations, August 1997
John Stockwell added bias option 4 August 2004
24
FTNSTRIP - convert a file of binary data plus record delimiters created
via Fortran to a file containing only binary values (as created via C)
ftnstrip <ftn_data >c_data
Caveat: this code assumes the conventional Fortran format of header

and trailer integer containing the number of byte in the
record. This is overwhelmingly common, but not universal.
Credits:
CWP: Jack K. Cohen
25
FTNUNSTRIP - convert C binary floats to Fortran style floats
ftnunstrip <stdin >stdout
none
n1=1 floats per line in output file

Notes: This program assumes that the record length is constant

throughout the input and output files.
In fortran code reading these floats, the following implied
do loop syntax would be used:
DO i=1,n2
READ (10) (someARRAY(j), j=1,n1)
END DO
Here n1 is the number of samples per record, n2 is the number
of records, 10 is some default file (fort.10, for example), and
someArray(j) is an array dimensioned to size n1
Credits:
CWP: John Stockwell, Feb 1998,
based on ftnstrip by: Jack K. Cohen
26
GRM - Generalized Reciprocal refraction analysis for a single layer
grm <stdin >stdout [parameters]
nt= Number of arrival time pairs
dx= Geophone spacing (m)
v0= Velocity in weathering layer (m/s)
abtime= If set to 0, use last time as a-b, else give time (ms)
XY= Value of XY if you want to override the optimum XY
algorithm in the program. If it is not an integer multiple of
dx, then it will be converted to the closest
one.
XYmax Maximum offset distance allowed when searching for
optimum XY (m) (Default is 2*dx*10)
depthres Size of increment in x during verical depth search(m)
(Default is 0.5m)
Input file:
4 column ASCII - x,y, forward time, reverse time
Output file:
1) XYoptimum
2) apparent refractor velcocity
3) x, y, z(x,y), y-z(x,y)
z(x,y) = calculated (GRM) depth below (x y)
y-z(x,y) = GRM depth subtracted from y - absolute depth
.............
4) x, y, d(x,y), y-d(x,y), (error)
d(x,y) = dip corrected depth estimate below (x,y)
y-d(x,y) = dip corrected absolute depth
error = estimated error in depth due only to the inexact
matching of tangents to arcs in dip estimate.
If the XY calculation is bypassed and XY specified, the values

used will precede 1) above. XYoptimum will still be calculated
and displayed for reference.
Notes:
Uses average refactor velocity along interface.
Credits:
27
CWP: Steven D. Sheaffer
D. Palmer, "The Generalized Reciprocal Method of Seismic

Refraction Interpretation", SEG, 1982.
28
H2B - convert 8 bit hexidecimal floats to binary
h2b <stdin >stdout outpar=/dev/tty
none
Note: this code may be used to recover binary data from PostScript
bitmaps. To do this, strip away all parts of the PSfile that
are not the actual hexidecimal bitmap and run through h2b.
Note: that the binary file may need to be transposed using

"transp" to appear to be the same as input data.
Note: output will be floats with the values 0-255
29
HTI2STIFF - convert HTI parameters alpha, beta, d(V), e(V), gamma
into stiffness tensor
hti2stiff [optional parameter] (output is to outpar)
Optional Parameters
alpha=2 isotropy-plane p-wave velocity
beta=1 fast isotropy-plan s-wave velocity
ev=0 e(V)
dv=0 d(V)
gamma=0 shear-wave splitting parameter
rho=1 density
sign sign of c13+c55 ( for most materials sign=1)
outpar=/dev/tty output parameter file
Output:
c_ijkl stiffness components for x1=symmetry axis
x3= vertical
Credits: Andreas Rueger, CWP Aug 01, 1996
Reference: Andreas Rueger, P-wave reflection coefficients for

transverse isotropy with vertical and horizontal axis
of symmetry, GEOPHYSICS
30
HUDSON - compute effective parameters of anisotropic solids
using Hudson’s crack theory.
Required paramters: <none>
Optional parameters
vp=4.5 p-wave velocity uncracked solid

vs=2.53 s-wave velocity uncracked solid
rho=2.8 density
cdens=0 crack density
aspect=0 aspect ratio
fill=0 gas filled cracks
=1 water filled
outpar =/dev/tty output file
Notes:
The cracks are assumed to be vertically aligned, penny-shaped and the
matrix is isotropic. The resulting anisotropic solid is of HTI symmetry.
Output:
Computes(a) stiffness elements
(b) density normalized stiffness components
(c) generic Thomsen parameters (vp0,vs0,eps,delta,gamma)
(d) equivalent VTI parameters (alpha,beta,ev,dv,gv)
AUTHOR:: Andreas Rueger, Colorado School of Mines, 10/10/96
Additional notes:
The routine can be easily modified to allow for any
filling adding attenuation is not trivial
Technical Reference:
Hudson’s theory: Hudson, 1981: Wave speed and attenuation of elastic
waves in material containing cracks.
Geophys. J. R. astr. Soc 64, 133-150
Crampin, 1984: Effective anisotropic elastic constants
for waves propagating through cracked
solids:
31
Geophys. J. R. astr. Soc 76, 135-145
Equivalent VTI : Rueger, 1996: Reflection coefficients in transversely
isotropic media with vertical and
horizontal axis of symmetry: Geophysics
32
KAPERTURE - generate the k domain of a line scatterer for a seismic array
kaperture [optional parameters] >stdout
Optional parameters
x0=1000 point scatterer location
z0=1000 point scatterer location
nshot=1 number of shots
sxmin=0 first shot location
szmin=0 first shot location
dsx=100 x-steps in shot location
dsz=0 z-steps in shot location
ngeo=1 number of receivers
gxmin=0 first receiver location
gzmin=0 first receiver location
dgx=100 x-steps in receiver location
dgz=0 z-steps in receiver location
fnyq=125 Nyquist frequency (Hz)
fmax=125 maximum frequency (Hz)
fmin=5 minimum frequency (Hz)
nfreq=2 number of frequencies
both=0 = 1 gives negative freqs too
nstep=60 points on Nyquist circle
c=5000 speed
xmin, xmax, ymin, ymax
and npairs (needed for psgraph or xgraph)
Notes:
nfreq=1 produces fmin
nstep=0 suppresses the Nyquist circle
and npairs
Examples:
Default case: both=0 nfreq=2
kaperture nshot=NSHOT ngeo=NGEO nstep=NSTEP |

psgraph n=NPAIRS,NSTEP mark=1,0 marksize=1,0 linewidth=0,1 |...
WHERE: NPAIRS=NSHOT*NGEO
Other cases:
33
both=0 nfreq=NFREQ > 2
kaperture both=0 nfreq=NFREQ nshot=NSHOT ngeo=NGEO nstep=NSTEP |
WHERE: NPAIRS=NFREQ*NSHOT*NGEO
both=1 nfreq=NFREQ > 2

kaperture both=1 nfreq=NFREQ nshot=NSHOT ngeo=NGEO nstep=NSTEP |
WHERE: NPAIRS=NFREQ*NSHOT*NGEO*2
When in doubt to the size of NPAIRS, redirect output of kaperture to

/dev/tty the first time to get npairs=:
kaperture [optional parameters] > /dev/tty
34
LINRORT - linearized P-P, P-S1 and P-S2 reflection coefficients
for a horizontal interface separating two of any of the
following halfspaces: ISOTROPIC, VTI, HTI and ORTHORHOMBIC.
linrort [optional parameters]
hspace1=ISO medium type of the incidence halfspace:

=ISO ... isotropic
=VTI ... VTI anisotropy
=HTI ... HTI anisotropy
=ORT ... ORTHORHOMBIC anisotropy
for ISO:
vp1=2 P-wave velocity, halfspace1
vs1=1 S-wave velocity, halfspace1
rho1=2.7 density, halfspace1
for VTI:
vp1=2 P-wave vertical velocity (V33), halfspace1
vs1=1 S-wave vertical velocity (V44=V55), halfspace1
eps1=0 Thomsen’s generic epsilon, halfspace1
delta1=0 Thomsen’s generic delta, halfspace1
gamma1=0 Thomsen’s generic gamma, halfspace1 ",
for HTI:
vs1=1 "fast" S-wave vertical velocity (V44), halfspace1
eps1_v=0 Tsvankin’s "vertical" epsilon, halfspace1
delta1_v=0 Tsvankin’s "vertical" delta, halfspace1
gamma1_v=0 Tsvankin’s "vertical" gamma, halfspace1 ",
for ORT:
vs1=1 x2-polarized S-wave vertical velocity (V44), halfspace1
eps1_1=0 Tsvankin’s epsilon in [x2,x3] plane, halfspace1
delta1_1=0 Tsvankin’s delta in [x2,x3] plane, halfspace1
gamma1_1=0 Tsvankin’s gamma in [x2,x3] plane, halfspace1
eps1_2=0 Tsvankin’s epsilon in [x1,x3] plane, halfspace1
gamma1_2=0 Tsvankin’s gamma in [x1,x3] plane, halfspace1
35
hspace2=ISO medium type of the reflecting halfspace (the same
convention as above)
medium parameters of the 2nd halfspace follow the same convention

as above:
vp2=2.5 vs2=1.2 rho2=3.0

eps2=0 delta2=0
eps2_v=0 delta2_v=0 gamma2_v=0
eps2_1=0 delta2_1=0 gamma2_1=0
eps2_2=0 delta2_2=0 gamma2_2=0
delta2_3=0
(note you do not need "gamma2" parameter for evaluation

of weak-anisotropy reflection coefficients)
a_file=-1 the string ’-1’ ... incidence and azimuth angles are
generated automatically using the setup values below
a_file=file_name ... incidence and azimuth angles are
read from a file "file_name"; the program expects a
file of two columns [inc. angle, azimuth]
in the case of a_file=-1:

fangle=0 first incidence phase angle
langle=30 last incidence angle
dangle=1 incidence angle increment
fazim=0 first azimuth (in deg)
lazim=0 last azimuth (in deg)
dazim=1 azimuth increment (in deg)
kappa=0. azimuthal rotation of the lower halfspace2 (e.t. a

symmetry axis plane for HTI, or a symmetry plane for
ORTHORHOMBIC) with respect to the x1-axis
out_inf=info.out information output file

out_P=Rpp.out file with Rpp reflection coefficients
out_S=Rps.out file with Rps reflection coefficients
out_SVSH=Rsvsh.out file with SV and SH projections of reflection
coefficients
out_Error=error.out file containing error estimates evaluated during
the computation of the reflection coefficients;
36
Output:
out_P:
inc. phase angle, azimuth, reflection coefficient; for a_file=-1, the
inc. angle is the fast dimension
out_S:
inc. phase angle, azimuth, Rps1, Rps2, cos(PHI), sin(PHI); for
a_file=-1, the inc. angle is the fast dimension ",
out_SVSH:
inc. phase angle, azimuth, Rsv, Rsh, cos(PHI), sin(PHI); for
a_file=-1, the inc. angle is the fast dimension
out_Error:
error estimates of Rpp, Rpsv and Rpsh approximations; global error is
analysed as well as partial contributions to the error due to the
isotropic velocity contrasts, and due to anisotropic upper and lower
halfspaces. The error file is self-explanatory, see also descriptions
of subroutines P_err_2nd_order, SV_err_2nd_order and SH_err_2nd_order.
Adopted Convention:
The right-hand Cartesian coordinate system with the x3-axis pointing

upward has been chosen. The upper halfspace (halfspace1)
contains the incident P-wave. Incidence angles can vary from <0,PI/2),
azimuths are unlimited, +azimuth sense counted from x1->x2 axes
(azimuth=0 corresponds to the direction of x1-axis). In the current
version, the coordinate system is attached to the halfspace1 (e.t.
the symmetry axis plane of HTI halfspace1, or one of symmetry planes
of ORTHORHOMBIC halfspace1, is aligned with the x1-axis), however, the
halfspace2 can be arbitrarily rotated along the x3-axis with respect
to the halfspace1. The positive weak-anisotropy polarization of the
reflected P-P wave (e.t. positive P-P reflection coefficient) is close
to the direction of isotropic slowness vector of the wave (pointing
outward the interface). Similarly, weak-anisotropy S-wave reflection
coefficients are described in terms of "SV" and "SH" isotropic
polarizations, "SV" and "SH" being unit vectors in the plane
perpendicular to the isotropic slowness vector. Then, the positive
"SV" polarization vector lies in the incidence plane and points
towards the interface, and positive "SH" polarization vector is
perpendicular to the incidence plane, aligned with the positive
x2-axis, if azimuth=0. Rotation angle "PHI", characterizing a
rotation of "the best projection" of the S1-wave polarization
vector in the isotropic SV-SH plane in the incidence halfspace1, is
37
counted in the positive sense from "SV" axis (PHI=0) towards the
"SH" axis (PHI=PI/2). Of course, S2 is perpendicular to S1, and
the projection of S1 and S2 polarizations onto the SV-SH plane
coincides with SV and SH directions, respectively, for PHI=0.
The units for velocities are km/s, angles I/O are in degrees
Additional Notes:
The coefficients are computed as functions of phase incidence
angle and azimuth (determined by the incidence slowness vector).
Vertical symmetry planes of the HTI and
ORTHORHOMBIC halfspaces can be arbitrarily rotated along the
x3-axis. The linearization is based on the assumption of weak ",
contrast in elastic medium parameters across the interface,
and the assumption of weak anisotropy in both halfspaces.
See the "Adopted Convention" paragraph below for a proper
input.
Author: Petr Jilek, CSM-CWP, December 1999.
38
LORENZ - compute the LORENZ attractor
lorenz > [stdout]
Required Parameters: none

rho=28.0 parameter for lorenz equations
sigma=10.0 parameter for lorenz equations
eta=1.6666667 parameter for lorenz equations
y0=1.0 initial value of y[0]
y1=-1.0 initial value of y[1]
y2=1.0 initial value of y[2]
h=.01 increment in time
tol=1.e-08 error tolerance
stepmax=500 maximum number of steps to compute
mode=xy xy-pairs, =yz yz-pairs, =xz xz-pairs,
=xyz xyz-triplet
Notes:
This program is really just a demo showing how to use the
differential equation solver rke_solve written by Francois
Pinard, based on a modified form of the 4th order Runge-Kutta
method, which employs the error checking method of R. England
1969.
The output consists of unformated C-style binary floats, of

either pairs or triplets as specified by the "mode" paramerter.
Examples:
lorenz stepmax=1000 mode=xy | xgraph n=1000 &
lorenz stepmax=1000 mode=yz | xgraph n=1000 &
lorenz stepmax=1000 mode=xz | xgraph n=1000 &
The lorenz equations describe a simplified model of a convection

cell, and are given by the autonomous system of ODE’s
x’(t) = sigma * ( y - x )
y’(t) = x * ( rho - z ) - y
z’(t) = x * y - eta * z
Author: CWP: Aug 2004: John Stockwell
39
40
MAKEVEL - MAKE a VELocity function v(x,y,z)
makevel > outfile nx= nz= [optional parameters]
nx= number of x samples (3rd dimension)
nz= number of z samples (1st dimension)
ny=1 number of y samples (2nd dimension)
dx=1.0 x sampling interval
fx=0.0 first x sample
dy=1.0 y sampling interval
fy=0.0 first y sample
dz=1.0 z sampling interval
fz=0.0 first z sample
v000=2.0 velocity at (x=0,y=0,z=0)
dvdx=0.0 velocity gradient with respect to x
dvdy=0.0 velocity gradient with respect to y
dvdz=0.0 velocity gradient with respect to z
vlens=0.0 velocity perturbation in parabolic lens
tlens=0.0 thickness of parabolic lens
dlens=0.0 diameter of parabolic lens
xlens= x coordinate of center of parabolic lens
ylens= y coordinate of center of parabolic lens
zlens= z coordinate of center of parabolic lens
vran=0.0 standard deviation of random perturbation
vzfile= file containing v(z) profile
vzran=0.0 standard deviation of random perturbation to v(z)
vzc=0.0 v(z) chirp amplitude
z1c=fz z at which to begin chirp
z2c=fz+(nz-1)*dz z at which to end chirp
l1c=dz wavelength at beginning of chirp
l2c=dz wavelength at end of chirp
exc=1.0 exponent of chirp
41
MKPARFILE - convert ascii to par file format
mkparfile <stdin >stdout
string1="par1" first par string
string2="par2" second par string
This is a tool to convert values written line by line to parameter

vectors in the form expected by getpar. For example, if the input
file looks like:
t0 v0
t1 v1
...
then
mkparfile <input >output string1=tnmo string2=vnmo
yields:
tnmo=t0,t1,...
vnmo=v0,v1,...
42
MRAFXZWT - Multi-Resolution Analysis of a function F(X,Z) by Wavelet
Transform. Modified to perform different levels of resolution
analysis for each dimension and also to allow to transform
back only the lower level of resolution.
mrafxzwt [parameters] < infile > mrafile
n1= size of first (fast) dimension
n2= size of second (slow) dimension
p1= maximum integer such that 2^p1 <= n1
p2= maximum integer such that 2^p2 <= n2
order=6 order of Daubechies wavelet used (even, 4<=order<=20)
mralevel1=3 maximum multi-resolution analysis level in dimension 1
mralevel2=3 maximum multi-resolution analysis level in dimension 2
trunc=0.0 truncation level (percentage) of the reconstruction
verbose=0 =1 to print some useful information
reconfile= reconstructed data file to write
reconmrafile= reconstructed data file in MRA domain to write
dfile= difference between infile and reconfile to write
dmrafile= difference between mrafile and reconmrafile to write
dconly=0 =1 keep only dc component of MRA
if (n1 or n2 is not integer powers of 2) specify the following:
nc1=n1/2 center of trimmed image in the 1st dimension
nc2=n2/2 center of trimmed image in the 2nd dimension
trimfile= if given, output the trimmed file
Notes:
This program performs multi-resolution analysis of an input function
f(x,z) via the wavelet transform method. Daubechies’s least asymmetric
wavelets are used. The smallest wavelet coefficient retained is given
by trunc times the absolute maximum size coefficient in the MRA.
The input dimensions of the data must be expressed by (p1,p2) which
Author: Zhaobo Meng, 11/25/95, Colorado School of Mines *

Modified: Carlos E. Theodoro, 06/25/97, Colorado School of Mines *
Included options for: *
- different level of resolutionf or each dimension; *
- transform back the lower level of resolution, only. *
43
*
Reference: *
Daubechies, I., 1988, Orthonormal Bases of Compactly Supported *
Wavelets, Communications on Pure and Applied Mathematics, Vol. XLI, *
909-996. *
44
PRPLOT - PRinter PLOT of 1-D arrays f(x1) from a 2-D function f(x1,x2)
prplot <infile >outfile [optional parameters]
n1=all number of samples in 1st dimension
d1=1.0 sampling interval in 1st dimension
f1=d1 first sample in 1st dimension
n2=all number of samples in 2nd dimension
d2=1.0 sampling interval in 2nd dimension
f2=d2 first sample in 2nd dimension
label2=Trace label for 2nd dimension
45
RANDVEL3D - Add a random velocity layer (RVL) to a gridded
v(x,y,z) velocity model
randvel3d <infile n1= n2= >outfile [parameters]
n1= number of samples along 1st dimension
n2= number of samples along 2nd dimension
n3=1 number of samples along 3rd dimension
mode=1 add single layer populated with random vels

=2 add nrvl layers of random thickness and vel
seed=from_clock random number seed (integer)
---->New layer geometry info

i1beg=1 1st dimension beginning sample
i1end=n1/5 1st dimension ending sample
i2beg=1 2nd dimension beginning sample
i2end=n2 2nd dimension ending sample
i3beg=1 3rd dimension beginning sample
i3end=n3 3rd dimension ending sample
---->New layer velocity info
vlsd=v/3 range (std dev) of random velocity in layer,
where v=v(0,0,i1) and i1=(i1beg+i1end)/2
add=1 add random vel to original vel (v_orig) at that point
=0 replace vel at that point with (v_orig+v_rand)
how=0 random vels can be higher or lower than v_orig
=1 random vels are always lower than v_orig
=2 random vels are always higher than v_orig
cvel=2000 layer filled with constant velocity cvel
(overides vlsd,add,how params)
---->Smoothing parameters (0 = no smoothing)
r1=0.0 1st dimension operator length in samples
r2=0.0 2nd dimension operator length in samples
r3=0.0 3rd dimension operator length in samples
slowness=0 =1 smoothing on slowness; =0 smoothing on velocity
nrvl=n1/10 number of const velocity layers to add

pdv=10. percentage velocity deviation (max) from input model
46
Notes:
1. Smoothing radii usually fall in the range of [0,20].
2. Smoothing radii can be used to set aspect ratio of random velocity
anomalies in the new layer. For example (r1=5,r2=0,r3=0) will
result in vertical vel streaks that mimick vertical fracturing.
3. Smoothing on slowness works better to preserve traveltimes relative
to the unsmoothed case.
4. Default case is a random velocity (+/-30%) near surface layer whose
thickness is 20% of the total 2D model thickness.
5. Each layer vel is a random perturbation on input model at that level.
6. The depth dimension is assumed to be along axis 1.
Example:
1. 2D RVL with no smoothing
makevel nz=250 nx=200 | randvel3d n1=250 n2=200 | ximage n1=250
2. 3D RVL with no smoothing
makevel nz=250 nx=200 ny=220 |
randvel3d n1=250 n2=200 n3=220 |
xmovie n1=250 n2=200
Author: Saudi Aramco: Chris Liner Jan/Feb 2005

Based on smooth3d (CWP: Zhenyue Liu March 1995)
47
RAYT2DAN -- P- and SV-wave raytracing in 2D anisotropic media
rayt2dan > ttime parameterfiles= nt= nx= nz= [optional parameters]
VP0file= name of file containing VP0(x,z)
nt= number of time samples for each ray
nx= number of samples (x) for the parameter fields
nz= number of samples (z) for the parameter fields
SV=0 for P-waves, SV=1 for Shear waves
Parameters defining the velocity field

dt=0.008 time sampling interval ",
fx=0 first lateral sample (x) in parameter field
fz=0 first lateral sample (z) in parameter field
dx=100.0 sample spacing (x) for the parameter fields
dz=100.0 sample spacing (z) for the parameter fields
Parameters defining the takeoff angle of a ray at a source position ",

fa=-60 first take-off angle of rays (degrees)
na=61 number of rays ",
da=2 increment of take-off angle
amin=0 minimum emergence angle; must be > -90 degrees
amax=90 maximum emergence angle; must be < 90 degrees
Parameters defining the output traveltime table

fxo=fx first lateral sample in traveltime table
nxo=nx number of later samples in traveltime table
dxo=dx lateral interval in traveltime table
fzo=fz first depth sample in traveltime table
nzo=nz number of depth samples in traveltime table
dzo=dz depth interval in traveltime table
fac=0.01 factor to determine the radius of extrap.
Parameters defining the source positions

fsx=fx x-coordinate of first source
nsx=1 number of sources
dsx=2*dxo x-coordinate increment of sources
aperx=0.5*nx*dx ray tracing aperature in x-direction
Files for general anisotropic parameters confined to a vertical plane:
48
a1111file name of file containing a1111(x,z)
VS0file name of file containing VS0(x,z)
For transversely isotropic media Thomsen’s parameters could be used:

deltafile name of file containing delta(x,z)
epsilonfile name of file containing epsilon(x,z)
if anisotropy parameters are not given the program considers ",

isotropic media. ",
Credits:
Debashish Sarkar
Cerveny, V., 1972, Seismic rays and ray intensities

in inhomogeneous anisotropic media:
Geophys. J. R. Astr. Soc., 29, 1--13.
Hangya, A., 1986, Gaussian beams in anisotropic elastic media:

Gajewski, D. and Psencik, I., 1987, Computation of high frequency

seismic wavefields in 3-D laterally inhomogeneous anisotropic
media: Geophys. J. R. Astr. Soc., 91, 383-411.
49
RAYT2D - traveltime Tables calculated by 2D paraxial RAY tracing
rayt2d vfile= tfile= [optional parameters]
vfile=stdin file containning velocity v[nx][nz]
tfile=stdout file containning traveltime tables
t[nxs][nxo][nzo]
Optional parameters
dt=0.008 time sample interval in ray tracing
nt=401 number of time samples in ray tracing
fz=0 first depth sample in velocity

nz=101 number of depth samples in velocity
dz=100 depth interval in velocity
fx=0 first lateral sample in velocity
nx=101 number of lateral samples in velocity
dx=100 lateral interval in velocity
fzo=fz first depth sample in traveltime table

nzo=nz number of depth samples in traveltime table
dzo=dz depth interval in traveltime table
fxo=fx first lateral sample in traveltime table
nxo=nx number of lateral samples in traveltime table
dxo=dx lateral interval in traveltime table
surf="0,0;99999,0" Recording surface "x1,z1;x2,z2;x3,z3;...

fxs=fx x-coordinate of first source
nxs=1 number of sources
dxs=2*dxo x-coordinate increment of sources
aperx=0.5*nx*dx ray tracing aperature in x-direction
fa=-60 first take-off angle of rays (degrees)

na=61 number of rays
da=2 increment of take-off angle
amin=0 minimum emergence angle
amax=90 maximum emergence angle
fac=0.01 factor to determine radius for extrapolation

ek=1 flag of implementing eikonal in shadow zones
ms=10 print verbal information at every ms sources
restart=n job is restarted (=y yes; =n no)
50
npv=0 flag of computing quantities for velocity analysis
if npv>0 specify the following three files
pvfile=pvfile input file of velocity variation pv[nxo][nzo]
tvfile=tvfile output file of traveltime variation tables
tv[nxs][nxo][nzo]
csfile=csfile output file of cosine tables cs[nxs][nxo][nzo]
Notes:
1. Each traveltime table is calculated by paraxial ray tracing; then
traveltimes in shadow zones are compensated by solving eikonal
equation.
2. Input velocity is uniformly sampled and smooth one preferred.
3. Traveltime table and source ranges must be within velocity model.
4. Ray tracing aperature can be chosen as sum of migration aperature
plus half of maximum offset.
5. Memory requirement for this program is about
[nx*nz+4*mx*nz+3*nxo*nzo+na*(nx*nz+mx*nz+3*nxo*nzo)]*4 bytes
where mx = min(nx,2*(1+aperx/dx)).
Author: Zhenyue Liu, 10/11/94, Colorado School of Mines
Trino Salinas, 01/01/96 included the option to handle nonflat

reference surfaces.
Subroutines from Dave Hale’s modeling library were adapted in
this code to define topography using cubic splines.
References:
Beydoun, W. B., and Keho, T. H., 1987, The paraxial ray method:
Geophysics, vol. 52, 1639-1653.
Cerveny, V., 1985, The application of ray tracing to the numerical

modeling of seismic wavefields in complex structures, in Dohr, G.,
ED., Seismic shear waves (part A: Theory): Geophysical Press,
Number 15 in Handbook of Geophysical Exploration, 1-124.
51
RECAST - RECAST data type (convert from one data type to another)
recast <stdin [optional parameters] >stdout
none
in=float input type (float)
=double (double)
=int (int)
=char (char)
=uchar (unsigned char)
=short (short)
=long (long)
=ulong (unsigned long)
out=double output type (double)

=float (float)
=int (int)
=char (char)
=uchar (unsigned char)
=short (short)
=long (long)

number of values (n1=)
Notes: Converting bigger types to smaller is hazardous. For float

or double conversions to integer types, the results are
rounded to the nearest integer.
Credits:
CWP: John Stockwell, Jack K. Cohen
52
REFREALAZIHTI - REAL AZImuthal REFL/transm coeff for HTI media
refRealAziHTI [optional parameters] >coeff.data
vp1=2 p-wave velocity medium 1 (with respect to symm.axes)
vs1=1 s-wave velocity medium 1 (with respect to symm.axes)
eps1=0 epsilon medium1
delta1=0 delta medium 1
gamma1=0 gamma medium 1
rho1=2.7 density medium 1
vp2=2 p-wave velocity medium 2 (with respect to symm.axes)
vs2=1 s-wave velocity medium 2 (with respect to symm.axes)
eps2=0 epsilon medium 2
gamma2=0 gamma medium 2
modei=0 incident mode is qP
=1 incident mode is qSV
=2 incident mode is SP
modet=0 scattered mode
rort=1 reflection(1) or transmission (0)
azimuth=0 azimuth with respect to x1-axis (clockwise)
fangle=0 first incidence angle
langle=45 last incidence angle
dangle=1 angle increment
iscale=0 default: angle in degrees
=1 angle-axis in rad
=2 axis horizontal slowness
=3 sin^2 of incidence angle
ibin=1 binary output
=0 Ascci output
outparfile =outpar parameter file for plotting
coeffile =coeff.data coefficient-output file
test=1 activate testing routines in code
info=0 output intermediate results
Notes:
Axes of symmetry have to coincide in both media. This code computes
all 6 REAL reflection/transmissions coefficients on the fly. However,
the set-up is such Real reflection/transmission coefficients in
HTI-media with coinciding symmetry axes.
However, the set-up is such that currently only one coefficient is
53
dumped into the output. This is easily changed. The solution of the
scattering problem is obtained numerically and involves the Gaussian
elimination of a 6X6 matrix. ",

original name of code <graebnerTIH.c>
modified, extended version of this code <refTIH3D>
Technical references:
Sebastian Geoltrain: Asymptotic solutions to direct

and inverse scattering in anisotropic elastic media;
CWP 082.
Graebner, M.; Geophysics, Vol 57, No 11:
Plane-wave reflection and transmission coefficients
for a transversely isotropic solid.
Cerveny, V., 1972, Seismic rays and ray intensities in inhomogeneous
anisotropic media: Geophys. J. R. astr. Soc., 29, 1-13.
.. and some derivations by Andreas Rueger.
If propagation is perpendicular or
parallel to the symmetry axis, the solution is analytic (see
graebner2D.c and rtRealIso.c
54
REFREALVTI - REAL REFL/transm coeff for VTI media and symmetry-axis
planes of HTI media
refRealVTI [Optional parameters]
vp1=2 p-wave velocity medium 1 (along symm.axes)
vs1=1 s-wave velocity medium 1 (along symm.axes)
eps1=0 Thomsen’s epsilon medium 1
delta1=0 Thomsen’s delta medium 1
axis1=0 medium 1 is VTI
=1 medium 1 is HTI
vp2=2.5 p-wave velocity medium 2 (along symm.axes)
vs2=1.2 s-wave velocity medium 2 (along symm.axes)
eps2=0 epsilon medium2
axis2=0 medium 2 is VTI
=2 medium 2 is HTI
modei=0 incident mode is qP
=1 incident mode is qSV
modet=0 scattered mode
rort=1 reflection(1) or transmission (0)
fangle=0 first angle
langle=45 last angle
dangle=1 angle increment
iscale=0 =1 angle-axis in rad
=2 axis horizontal slowness
=3 sin^2 of incidence angle
ibin=1 binary output
=0 Ascci output
outparfile =outpar parameter file for plotting
coeffile =coeff.data coefficient-output file
Notes:
Coefficients are based on Graebner’s 1992 Geophysics paper. Note the
mistype in the equation for K1. The algorithm can be used for VTI
and HTI media on the incidence and scattering side.
55
original name of algorithm: graebner1.c
Technical reference: Graebner, M.; Geophysics, Vol 57, No 11:

Plane-wave reflection and transmission coefficients
for a transversely isotropic solid.
Rueger, A.; Geophysics 1996 (accepted):
P-wave reflection coefficients ...
56
REGRID3 - REwrite a [ni3][ni2][ni1] GRID to a [no3][no2][no1] 3-D grid
regrid3 < oldgrid > newgrid [parameters]
ni1=1 fastest (3rd) dimension in input grid
ni2=1 second fastest (2nd) dimension in input grid
ni3=1 slowest (1st) dimension in input grid
no1=1 fastest (3rd) dimension in output grid
no2=1 second fastest (2nd) dimension in output grid ",
no3=1 slowest (1st) dimension in output grid
verbose=0 =1 print some useful information
Notes:
REGRID3 can be used to span a 1-D grid to a 2-D grid, or a 2-D grid
to a 3-D grid; or to change grid parameters within the dimensions.
Together with MUL and UNIF3, most 3-D velocity model can be
Credits:
CWP: Zhaobo Meng, 1996, Colorado School of Mines
57
RESAMP - RESAMPle the 1st dimension of a 2-dimensional function f(x1,x2)
resamp <infile >outfile [optional parameters]
n1=all number of samples in 1st (fast) dimension
n2=all number of samples in 2nd (slow) dimension
n1r=n1 number of samples in 1st dimension after resampling
d1r=d1 sampling interval in 1st dimension after resampling
f1r=f1 first sample in 1st dimension after resampling
NOTE: resamp currently performs NO ANTI-ALIAS FILTERING before resampling!
58
SMOOTH2 --- SMOOTH a uniformly sampled 2d array of data, within a user-
defined window, via a damped least squares technique
smooth2 < stdin n1= n2= [optional parameters ] > stdout
n1= number of samples in the 1st (fast) dimension
n2= number of samples in the 2nd (slow) dimension
r1=0 smoothing parameter in the 1 direction
win=0,n1,0,n2 array for window range
rw=0 smoothing parameter for window function
efile= =efilename if set write relative error(x1) to
efilename
Notes:
Larger r1 and r2 result in a smoother data. Recommended ranges of r1 ",
and r2 are from 1 to 20.
The file verror gives the relative error between the original velocity
and the smoothed one, as a function of depth. If the error is
between 0.01 and 0.1, the smoothing parameters are suitable. Otherwise,
consider increasing or decreasing the smoothing parameter values.
Smoothing can be implemented in a selected window. The range of 1st

dimension for window is from win[0] to win[1]; the range of 2nd
dimension is from win[2] to win[3].
Smoothing the window function (i.e. blurring the edges of the window)
may be done by setting a nonzero value for rw, otherwise the edges
of the window will be sharp.
Credits:
CWP: Zhen-yue Liu,
adapted for par/main by John Stockwell 1 Oct 92
Windowing feature added by Zliu on 16 Nov 1992
59
SMOOTH3D - 3D grid velocity SMOOTHing by the damped least squares
smooth3d <infile >outfile [parameters]
n1= number of samples along 1st dimension
n2= number of samples along 2nd dimension
n3= number of samples along 3rd dimension
Smoothing parameters (0 = no smoothing)

r1=0.0 operator length in 1st dimension
r2=0.0 operator length in 2nd dimension
r3=0.0 operator length in 3rd dimension
Sample intervals:
d1=1.0 1st dimension
d2=1.0 2nd dimension
d3=1.0 3rd dimension
iter=2 number of iteration used
time=0 which dimension the time axis is (0 = no time axis)
depth=1 which dimension the depth axis is (ignored when time>0)
mu=1 the relative weight at maximum depth (or time)
verbose=0 =1 for printing minimum wavelengths
slowness=0 =1 smoothing on slowness; =0 smoothing on velocity
vminc=0 velocity values below it are cliped before smoothing
vmaxc=99999 velocity values above it are cliped before smoothing
Notes:
1. The larger the smoothing parameters, the smoother the output velocity.
These parameters are lengths of smoothing operators in each dimension.
2. iter controls the orders of derivatives to be smoothed in the output
velocity. e.g., iter=2 means derivatives until 2nd order smoothed.
3. mu is the multipler of smoothing parameters at the bottom compared to
those at the surface.
4. Minimum wavelengths of each dimension and the total may be printed
for the resulting output velocity is. To compute these parameters for
the input velocity, use r1=r2=r3=0.
5. Smoothing directly on slowness works better to preserve traveltime.
So the program optionally converts the input velocity into slowness ",
and smooths the slowness, then converts back into velocity.
60
Author: CWP: Zhenyue Liu March 1995
Reference:
Liu, Z., 1994,"A velocity smoothing technique based on damped least squares
in Project Reveiew, May 10, 1994, Consortium Project on
Seismic Inverse Methods for Complex Stuctures.
61
SMOOTHINT2 --- SMOOTH non-uniformly sampled INTerfaces, via the damped
least-squares technique
smoothint2 <input ninf= >output [optional parameters]
<input file containing original interfaces
>output file containing smoothed interfaces
ninf=5 number of interfaces
r=100 smoothing parameter
npmax=101 maximum number of points in interfaces
Notes:
The input file is an ASCII file. Each interface is represented by pairs
(non-uniform sampling) of x and z values, with one pair of values on
each line, separated by spaces or tabs. Each interface is separated with
an entry with a large negative z value for example: 1.0 -9999.
There is no entry for the surface. The surface is assumed to be flat
with z=0.
This is similar to a CSHOT model file without a surface entry and
without comments.
The smoothing method is analogous to a moving window averaging process

(but not the same) with the parameter "r" being analogous to the "width
of the window. Thus, the size of "r" must be chosen to by compatible
with the scale (wavelengths) of the variations of the interfaces in the
model being smoothed.
Example using the test data set generated by unif2:

unif2 tfile=tfilename
Compare the unsmoothed interface model:
unif2 < tfilename method=interpolation_method |
psimage n1=100 n2=100 d1=10 d2=10 | ...
To the smoothed interface model:
smoothint2 r=100 < tfilename | unif2 method=interpolation_method | ",
psimage n1=100 n2=100 d1=10 d2=10 | ...
Credits:
CWP: Zhenyue Liu, Jan 1994
62
Reference:
Liu, Zhenyue, 1994, Velocity smoothing: theory and implementation,
Project Review, 1994, Consortium Project on Seismic Inverse Methods
for Complex Stuctures (in review)
63
STIFF2VEL - Transforms 2D elastic stiffnesses to (vp,vs,epsilon,delta)
stiff2vel nx= nz= [optional parameters]
nx= number of x samples (2nd dimension)
rho_file=’rho.bin’ input file containing rho(x,z)
c11_file=’c11.bin’ input file containing c11(x,z)
vp_file=’vp.bin’ output file containing P-wave velocities
vs_file=’vs.bin’ output file containing S-wave velocities
rho_file=’rho.bin’ output file containing densities
eps_file=’eps.bin’ output file containing Thomsen epsilon
delta_file=’delta.bin’ output file containing Thomsen delta
Notes:
1. All quantities in MKS units
2. Isotropy implied by c11(x,z)=c33(x,z)=0
3. Vertical symmetry axis is assumed.
Coded:
Aramco: Chris Liner 9/25/2005
(based on vel2stiff.c)
64
SUBSET - select a SUBSET of the samples from a 3-dimensional file
subset <infile >outfile [optional parameters]
n1=nfloats number of samples in 1st dimension
n2=nfloats/n1 number of samples in 2nd dimension
n3=nfloats/(n1*n2) number of samples in 3rd dimension
id1s=1 increment in samples selected in 1st dimension
if1s=0 index of first sample selected in 1st dimension
n1s=1+(n1-if1s-1)/id1s number of samples selected in 1st dimension
ix1s=if1s,if1s+id1s,...indices of samples selected in 1st dimension
id2s=1 increment in samples selected in 2nd dimension
if2s=0 index of first sample selected in 2nd dimension
n2s=1+(n2-if2s-1)/id2s number of samples selected in 2nd dimension
ix2s=if2s,if2s+id2s,...indices of samples selected in 2nd dimension
id3s=1 increment in samples selected in 3rd dimension
if3s=0 index of first sample selected in 3rd dimension
n3s=1+(n3-if3s-1)/id3s number of samples selected in 3rd dimension
ix3s=if3s,if3s+id3s,...indices of samples selected in 3rd dimension
For the 1st dimension, output is selected from input as follows:

output[i1s] = input[ix1s[i1s]], for i1s = 0 to n1s-1
Likewise for the 2nd and 3rd dimensions.
65
SWAPBYTES - SWAP the BYTES of various data types
swapbytes <stdin [optional parameters] >stdout
none
in=float input type (float)
=double (double)
=short (short)
=ushort (unsigned short)
=long (long)
=int (int)

number of values (n1=)
Notes:
The byte order of the mantissa of binary data values on PC’s and DEC’s
is the reverse of so called "big endian" machines (IBM RS6000, SUN,etc.)
hence the need for byte-swapping capability. The subroutines in this code
have been tested for swapping between PCs and big endian machines, but
have not been tested for DEC products.
Caveat:
2 byte short, 4 byte long, 4 byte float, 4 byte int,
and 8 bit double assumed.
Credits:
CWP: John Stockwell (Jan 1994)
Institut fur Geophysik, Hamburg: Jens Hartmann supplied byte swapping
subroutines
66
THOM2HTI - Convert Thompson parameters V_p0, V_s0, eps, gamma,
to the HTI parameters alpha, beta, epsilon(V), delta(V),
gamma
thom2hti [optional parameter]
vp=2 symm.axis p-wave velocity

vs=1 symm.axis s-wave velocity
eps=0 Thomsen’s (generic) epsilon
gamma=0 Thomsen’s generic gamma
weak=1 compute weak approximation
Outputs:
alpha, beta, e(V), d(V), gamma
Notes:
Output is dumped to the screen and, if selected to outpar
Code can be used to find models that satisfy the constraints

that are imposed on HTI models caused by vertically fractured
layers. For definition and use of the HTI parameter set see CWP-235.
Credits: Andreas Rueger, CWP

For definition and use of the HTI parameter set see CWP-235.
67
THOM2STIFF - convert Thomsen’s parameters into (density normalized)
stiffness components for transversely isotropic
material with in-plane-tilted axis of symmetry
thom2stiff [optional parameter]
vp=2 symm.axis p-wave velocity

vs=1 symm.axis s-wave velocity
eps=0 Thomsen’s (generic) epsilon
delta=0 Thomsen’s (generic) delta
gamma=0 Thomsen’s (generic) gamma
rho=1 density
phi=0 angle(DEG) vertical --> symmetry axes (clockwise)
sign=1 sign of c11+c44 (generally sign=1)
Output:
aijkl,cijkl (density normalized) stiffness components
Author: CWP: Andreas Rueger 1995
68
TRANSP - TRANSPose an n1 by n2 element matrix
transp <infile >outfile n1= [optional parameters]
n1 number of elements in 1st (fast) dimension of matrix
n2=all number of elements in 2nd (slow) dimension of matrix
nbpe=sizeof(float) number of bytes per matrix element
verbose=0 =1 for diagnostic information
69
TVNMOQC - Check tnmo-vnmo pairs; create t-v column files
tvnmoqc [parameters] cdp=... tnmo=... vnmo=...
Example:
tvnmoqc mode=1 \\
cdp=15,35 \\
tnmo=0.0091,0.2501,0.5001,0.7501,0.9941 \\
vnmo=1497.0,2000.0,2500.0,3000.0,3500.0 \\
tnmo=0.0082,0.2402,0.4902,0.7402,0.9842 \\
vnmo=1495.0,1900.0,2400.0,2900.0,3400.0
Required Parameter:
prefix= Prefix of output t-v file(s)
Required only for mode=2
Optional Parameter:
mode=1 1=qc: check that tnmo values increase
2=qc and output t-v files
mode=1
TVNMOQC checks that there is a tnmo and vnmo series for each CDP
and it checks that each tnmo series increases in time.
mode=2
TVNMOQC does mode=1 checking, plus ...
TVNMOQC converts par (MKPARFILE) values written as:
cdp=15,35,...,95 \\
tnmo=t151,t152,...,t15n \\
vnmo=v151,v152,...,v15n \\
tnmo=t351,t352,...,t35n \\
vnmo=v351,v352,...,v35n \\
tnmo=... \\
vnmo=... \\
tnmo=t951,t952,...,t95n \\
vnmo=v951,v952,...,v95n \\
to column format. The format of each output file is:
t1 v1
t2 v2
70
...
tn vn
One file is output for each input pair of tnmo-vnmo series.
A CDP VALUE MUST BE SUPPLIED FOR EACH TNMO-VNMO ROW PAIR.
Prefix of each output file is the user-supplied value of

parameter PREFIX.
Suffix of each output file is the cdp value.
For the example above, output files names are:
PREFIX.15 PREFIX.35 ... PREFIX.95
Credits:
MTU: David Forel (adapted from SUNMO)
71
UNIF2ANISO - generate a 2-D UNIFormly sampled profile of elastic
constants from a layered model.
unif2aniso < infile [Parameters]
none
nx=100 number of x samples (2nd dimension)
nz=100 number of z samples (1st dimension)
dx=10 x sampling interval
dz=10 z sampling interval
npmax=201 maximum number of points on interfaces

x0=0.0,0.0,..., distance x at which vp00 and vs00 are specified

z0=0.0,0.0,..., depth z at which vp00 and vs00 are specified
vp00=1500,2000,..., P-velocity at each x0,z0 (m/sec)

vs00=866,1155..., S-velocity at each x0,z0 (m/sec)
rho00=1000,1100,..., density at each x0,z0 (kg/m^3)
eps00=0,0,0..., Thomsen or Sayers epsilon

delta00=0,0,0..., Thomsen or Sayers delta
gamma00=0,0,0..., Thomsen or Sayers gamma
drdx=0.0,0.0,..., x-derivative of density (d rho/dx)

drdz=0.0,0.0,..., z-derivative of density (d rho/dz)
dvpdx=0.0,0.0,..., x-derivative of P-velocity (dvp/dx)

dvpdz=0.0,0.0,..., z-derivative of P-velocity (dvs/dz)
dvsdx=0.0,0.0,..., x-derivative of S-velocity (dvs/dx)

dvsdz=0.0,0.0,..., z-derivative of S-velocity (dvs/dz)
dedx=0.0,0.0,..., x-derivative of epsilon (de/dx)

dedz=0.0,0.0,..., z-derivative of epsilon with depth z (de/dz)
72
dddx=0.0,0.0,..., x-derivative of delta (dd/dx)
dddz=0.0,0.0,..., z-derivative of delta (dd/dz)
dgdz=0.0,0.0,..., x-derivative of gamma (dg/dz)

dgdx=0.0,0.0,..., z-derivative of gamma (dg/dx)
phi00=0,0,..., rotation angle(s) in each layer
...output filenames
c11_file=c11_file output filename for c11 values
rho_file=rho_file output filename for density values
paramtype=1 =1 Thomsen parameters, =0 Sayers parameters(see below)

method=linear for linear interpolation of interface
=mono for monotonic cubic interpolation of interface
=akima for Akima’s cubic interpolation of interface
=spline for cubic spline interpolation of interface
tfile= =testfilename if set, a sample input dataset is

output to "testfilename".
Notes:
The input file is an ASCII file containing x z values representing a
piecewise continuous velocity model with a flat surface on top.
The surface and each successive boundary between media is represented

by a list of selected x z pairs written column form. The first and
last x values must be the same for all boundaries. Use the entry
1.0 -99999 to separate the entries for successive boundaries. No
boundary may cross another. Note that the choice of the method of
interpolation may cause boundaries to cross that do not appear to
cross in the input data file.
The number of interfaces is specified by the parameter "ninf". This
73
number does not include the top surface of the model. The input data
format is the same as a CSHOT model file with all comments removed.
The algorithm works by transforming the P-wavespeed , S-wavespeed,

density and the Thomsen or Sayers parameters epsilon, delta, and gamma
into elastic stiffness coefficients. Furthermore, the user can specify
rotations, phi, to the elasticity tensor in each layer.
Common ranges of Thomsen parameters are

epsilon: 0.0 -> 0.5
delta: -0.2 -> 0.4
gamma: 0.0 -> 0.4
If only P-wave, S-wave velocities and density is given as input,

the model is, by definition, isotropic.
If files containing Thomsen/Sayers parameters are given, the model

will be assumed to have VTI symmetry.
Example using test input file generating feature:

unif2aniso tfile=testfilename produces a 5 interface demonstration model
unif2aniso < testfilename
ximage < c11_file n1=100 n2=100
ximage < rho_file n1=100 n2=100
Credits:
CWP: John Stockwell, April 2005.
CWP: based on program unif2 by Zhenyue Liu, 1994
74
UNIF2 - generate a 2-D UNIFormly sampled velocity profile from a layered
model. In each layer, velocity is a linear function of position.
unif2 < infile > outfile [parameters]
none
nx=100 number of x samples (2nd dimension)
nz=100 number of z samples (1st dimension)
dx=10 x sampling interval
npmax=201 maximum number of points on interfaces

x0=0.0,0.0,..., distance x at which v00 is specified

z0=0.0,0.0,..., depth z at which v00 is specified
v00=1500,2000,2500..., velocity at each x0,z0 (m/sec)
dvdx=0.0,0.0,..., derivative of velocity with distance x (dv/dx)
dvdz=0.0,0.0,..., derivative of velocity with depth z (dv/dz)
method=linear for linear interpolation of interface

tfile= =testfilename if set, a sample input dataset is

output to "testfilename".
Notes:
The input file is an ASCII file containing x z values representing a
piecewise continuous velocity model with a flat surface on top. The surface
and each successive boundary between media are represented by a list of
selected x z pairs written column form. The first and last x values must
be the same for all boundaries. Use the entry 1.0 -99999 to separate
entries for successive boundaries. No boundary may cross another. Note
that the choice of the method of interpolation may cause boundaries
to cross that do not appear to cross in the input data file.
75
The number of interfaces is specified by the parameter "ninf". This
number does not include the top surface of the model. The input data
format is the same as a CSHOT model file with all comments removed.
Example using test input file generating feature:

unif2 tfile=testfilename produces a 5 interface demonstration model
unif2 < testfilename | psimage n1=100 n2=100 d1=10 d2=10 | ...
Credits:
CWP: Zhenyue Liu, 1994
CWP: John Stockwell, 1994, added demonstration model stuff.
76
UNISAM2 - UNIformly SAMple a 2-D function f(x1,x2)
unisam2 [optional parameters] <inputfile >outputfile
none
x1= array of x1 values at which input f(x1,x2) is sampled
... Or specify a unform linear set of values for x1 via:
nx1=1 number of input samples in 1st dimension
dx1=1 input sampling interval in 1st dimension
fx1=0 first input sample in 1st dimension
...
n1=1 number of output samples in 1st dimension
d1= output sampling interval in 1st dimension
f1= first output sample in 1st dimension
x2= array of x2 values at which input f(x1,x2) is sampled
... Or specify a unform linear set of values for x2 via:
nx2=1 number of input samples in 2nd dimension
dx2=1 input sampling interval in 2nd dimension
fx2=0 first input sample in 2nd dimension
...
n2=1 number of output samples in 2nd dimension
d2= output sampling interval in 2nd dimension
f2= first output sample in 2nd dimension
...
method1=linear =linear for linear interpolation
=mono for monotonic bicubic interpolation
=akima for Akima bicubic interpolation
=spline for bicubic spline interpolation
method2=linear =linear for linear interpolation
=mono for monotonic bicubic interpolation
=akima for Akima bicubic interpolation
=spline for bicubic spline interpolation
NOTES:
The number of input samples is the number of x1 values times the
number of x2 values. The number of output samples is n1 times n2.
The output sampling intervals (d1 and d2) and first samples (f1 and f2)
default to span the range of input x1 and x2 values. In other words,
d1=(x1max-x1min)/(n1-1) and f1=x1min; likewise for d2 and f2.
Interpolation is first performed along the 2nd dimension for each
77
value of x1 specified. Interpolation is then performed along the
1st dimension.
AUTHOR: Dave Hale, Colorado School of Mines, 01/12/91\n"
78
UNISAM - UNIformly SAMple a function y(x) specified as x,y pairs
unisam xin= yin= nout= [optional parameters] >binaryfile

... or ...
unisam xfile= yfile= npairs= nout= [optional parameters] >binaryfile
... or ...
unisam xyfile= npairs= nout= [optional parameters] >binaryfile
xin=,,, array of x values (number of xin = number of yin)
yin=,,, array of y values (number of yin = number of xin)
... or
xfile= binary file of x values
yfile= binary file of y values
... or
xyfile= binary file of x,y pairs
npairs= number of pairs input (active only if xfile= and yfile=
or xyfile= are set)
nout= number of y values output to binary file
dxout=1.0 output x sampling interval
fxout=0.0 output first x
method=linear =linear for linear interpolation (continuous y)
=mono for monotonic cubic interpolation (continuous y’)
=akima for Akima’s cubic interpolation (continuous y’)
=spline for cubic spline interpolation (continuous y’’)
isint=,,, where these sine interpolations to apply
amp=,,, amplitude of sine interpolations
phase0=,,, starting phase (defaults: 0,0,0,...,0)
totalphase=,,, total phase (default pi,pi,pi,...,pi.)
nwidth=0 apply window smoothing if nwidth>0

Zhaobo Meng, Colorado School of Mines,
added sine interpolation and window smoothing, 09/16/96
CWP: John Stockwell, added file input options, 24 Nov 1997
Remarks: In interpolation, suppose you need 2 pieces of

sine interpolation before index 3 to 4, and index 20 to 21
then set: isint=3,20. The sine interpolations use a sine
79
function with starting phase being phase0, total phase
being totalphase (i.e. ending phase being phase0+totalphase
for each interpolation).
80
UTMCONV - CONVert longitude and latitude to UTM, and vice versa
utmconv <stdin >stdout [optional parameters]
idx=23 reference ellipsoid index (default is WGS 1984)
format=%.3f output number format (printf style for one float)
a=(from idx) user-specified semimajor axis of ellipsoid
f=(from idx) user-specified flattening of ellipsoid
letter=0 =1: use UTM letter designator for latitude/Northing
invert=0 =0: convert latitude and longitude to UTM
=1: convert UTM to latitude and longitude
verbose=0 =1: echo parameters and number of converted coords
lon0= central meridian for TM projection in degrees

(default uses the 60 standard UTM longitude zones)
xoff=500000 false Easting (default: UTM)
ysoff=10000000 false Northing, southern hemisphere (default: UTM)
ynoff=0 false Northing, northern hemisphere (default: UTM)
Notes:
Universal Transverse Mercator (UTM) coordinates are defined between
latitudes 80S (-80) and 84N (84). Longitude values must be between
-180 degrees (west) and 179.999... degrees (east).
Input and output is in ASCII format. For a conversion from lon/lat

to UTM (invert=0), input is a two-column table of longitude and
latitude in decimal degrees. Output is a three-column table of
UTM Easting, UTM Northing, and UTM zone. The zone is given either
by the zone number only (default, negative on southern hemisphere)
or by the positive zone number plus a letter designator (letter=1).
Example:
Convert 40.822N, 14.125E to UTM with zone number and letter,
output values rounded to nearest integer:
echo 14.125 40.822 | utmconv letter=1 format=%.0f
The output is "426213 4519366 33T" (Easting, Northing, UTM zone).
Reference ellipsoids:
An ellipsoid may be specified by its semimajor axis a and its
flattening f, or one of the following ellipsoids may be selected
by its index idx (semimajor axes in meters):
0 Sphere with radius of 6371000 m
81
1 Airy 1830
2 Australian National 1965
3 Bessel 1841 (Ethiopia, Indonesia, Japan, Korea)
4 Bessel 1841 (Namibia)
5 Clarke 1866
6 Clarke 1880
7 Everest (Brunei, E. Malaysia)
8 Everest (India 1830)
10 Everest (Pakistan)
11 Everest (W. Malaysia, Singapore 1948)
12 Everest (W. Malaysia 1969)
13 Geodetic Reference System 1980 (GRS 1980)
14 Helmert 1906
15 Hough 1960
16 Indonesian 1974
17 International 1924 / Hayford 1909
18 Krassovsky 1940
19 Modified Airy
20 Modified Fischer 1960
21 South American 1969
22 World Geodetic System 1972 (WGS 1972)
23 World Geodetic System 1984 (WGS 1984) / NAD 1983
UTM grid:
The Universal Transverse Mercator (UTM) system is a world wide
coordinate system defined between 80S and 84N. It divides the
Earth into 60 six-degree zones. Zone number 1 has its central
meridian at 177W (-177 degrees), and numbers increase eastward.
Within each zone, an Easting of 500,000 m is assigned to its

central meridian to avoid negative coordinates. On the northern
hemisphere, Northings start at 0 m at the equator and increase
northward. On the southern hemisphere a false Northing of
10,000,000 m is applied, i.e. Northings start at 10,000,000 m at
the equator and decrease southward. Letters are sometimes used
to identify different zones of latitude. The letters C-M
indicate zones on the southern and the letters N-X zones on
the northern hemisphere.
Author:
Nils Maercklin, RISSC, University of Naples, Italy, March 2007
82
References:
NIMA (2000). Department of Defense World Geodetic System 1984 -
its definition and relationships with local geodetic systems.
Technical Report TR8350.2. National Imagery and Mapping Agency,
Geodesy and Geophysics Department, St. Louis, MO. 3rd edition.
J. P. Snyder (1987). Map Projections - A Working Manual.
U.S. Geological Survey Professional Paper 1395, 383 pages.
U.S. Government Printing Office.
83
VEL2STIFF - Transforms VELocities, densities, and Thomsen or Sayers
parameters to elastic STIFFnesses
vel2stiff [Required parameters] [Optional Parameters] > stdout
vpfile= file with P-wave velocities
vsfile= file with S-wave velocities
rhofile= file with densities
epsfile= file with Thomsen/Sayers epsilon
deltafile= file with Thomsen/Sayers delta
gammafile= file with Thomsen/Sayers gamma
phi_file= angle of axis of symmetry from vertical (radians)

paramtype=1 (1) Thomsen parameters, (0) Sayers parameters(see below)
nx=101 number of x samples 2nd (slow) dimension

nz=101 number of z samples 1st (fast) dimension
Notes:
Transforms velocities, density and Thomsen/Sayers parameters
epsilon, delta, and gamma into elastic stiffness coefficients.
If only P-wave, S-wave velocities and density is given as input,

the model is assumed to be isotropic.
If files containing Thomsen/Sayers parameters are given, the model

will be assumed to have VTI symmetry.
All input files vpfile, vsfile, rhofile etc. are assumed to consist
only of C style binary floating point numbers representing the
corresponding material values of vp, vs, rho etc. Similarly, the output
84
files consist of the coresponding stiffnesses as C style binary floats.
If the output files are to be used as input for a modeling program,
such as suea2df, then further, the contents are assumed be arrays of
floating point numbers of the form of Array[n2][n1], where the fast
dimension, dimension 1, represents depth.
Author:
CWP: Sverre Brandsberg-Dahl 1999
Extended:
CWP: Stig-Kyrre Foss 2001
- to include the option to use the parameters by Sayers (1995)
instead of the Thomsen parameters
Technical reference:
Sayers, C. M.: Simplified anisotropy parameters for transversely
isotropic sedimentary rocks. Geophysics 1995, pages 1933-1935.
85
VELCONV - VELocity CONVersion
velconv <infile >outfile intype= outtype= [optional parameters]
intype= input data type (see valid types below)
outtype= output data type (see valid types below)
Valid types for input and output data are:

vintt interval velocity as a function of time
vrmst RMS velocity as a function of time
vintz velocity as a function of depth
zt depth as a function of time
tz time as a function of depth
nt=all number of time samples
dt=1.0 time sampling interval
ft=0.0 first time
nz=all number of depth samples
dz=1.0 depth sampling interval
fz=0.0 first depth
nx=all number of traces
Example: "intype=vintz outtype=vrmst" converts an interval velocity

function of depth to an RMS velocity function of time.
Notes: nt, dt, and ft are used only for input and output functions
of time; you need specify these only for vintt, vrmst, orzt.
Likewise, nz, dz, and fz are used only for input and output
functions of depth.
The input and output data formats are C-style binary floats.
86
VELPERTAN - Velocity PERTerbation analysis in ANisotropic media to
determine the model update required to flatten image gathers",
velpertan boundary= par=cig.par refl1= refl2= npicks1=

npicks2= cdp1= cdp2= vfile= efile= dfile= nx= dx= fx=
ncdp= dcdp= fcdp= off0= noff= doff= >outfile [parameters]
refl1= file with picks on the 1st reflector
refl2= file with picks on the 2nd reflector
vfile= file defining VP0 at all grid points from prev. iter. ",
efile= file defining eps at all grid points from prev. iter. ",
dfile= file defining del at all grid points from prev. iter. ",
boundary= file defining the boundary above which
parameters are known; update is done below this ",
boundary ",
npicks1= number of picks on the 1st reflector
npicks2= number of picks on the 2nd reflector
ncdp= number of cdp’s
dcdp= cdp spacing
fcdp= first cdp
off0= first offset in common image gathers
noff= number of offsets in common image gathers
doff= offset increment in common image gathers
cip1=x1,r1,r2,..., cip=xn,r1n,r2n description of input CIGS
cip2=x2,r1,r2,..., cip=xn,r1n,r2n description of input CIGS
x x-value of a common image point
r1 hyperbolic component of the residual moveout
r2 non-hyperbolic component of residual moveout
method=akima for linear interpolation of the interface
VP0=2000 Starting value for vertical velocity at a point in the
target layer
x00=0.0 x-coordinate at which VP0 is defined
z00=0.0 z-coordinate at which VP0 is defined
eps=0.0 Starting value for Thomsen’s parameter epsilon
del=0.0 Starting value for Thomsen’s parameter delta
kz=0.0 Starting value for the vertical gradient in VP0
kx=0.0 Starting value for the lateral gradient in VP0
nx=100 number of nodes in the horizontal direction for the
87
velocity grid
nz=100 number of nodes in the vertical direction for the
velocity grid
dx=10 horizontal grid increment
dz=10 vertical grid increment
fx=0 first horizontal grid point
fz=0 first vertical grid point
dt=0.008 traveltime increment
nt=500 no. of points on the ray
amax=360 max. angle of emergence
amin=0 min. angle of emergence
Smoothing parameters:
win=0,n1,0,n2 array for window range
rw=0 smoothing parameter for window function
nbound=2 number of points picked on the boundary
tol=0.1 tolerance in computing the offset (m)
Notes:
This program is used as part of the velocity analysis technique developed
by Debashish Sarkar, CWP:2003.
Notes:
The output par file contains the coefficients describing the residual
moveout. This program is used in conjunction with surelanan.
Author: CSM: Debashish Sarkar, December 2003

based on program: velpert.c written by Zhenuye Liu.
88
VELPERT - estimate velocity parameter perturbation from covariance
of imaged depths in common image gathers (CIGs)
verpert <dfile dzfile=dzfile >outfile [parameters]
dfile input of imaged depths in CIGs
dzfile=dzfile input of dz/dv at the imaged depths in CIGs
outfile output of the estimated parameter
noff number of offsets
ncip number of common image gathers
moff=noff number of first offsets used in velocity estimation
Notes:
1. This program is part of Zhenyue Liu’s velocity analysis technique.
The input dzdv values are computed using the program dzdv.
2. For given depths, using moff smaller than noff may avoid poor
values of dz/dv at far offsets. However, a too small moff used
will the sensitivity of velocity error to the imaged depth.
3. Outfile contains three parts:
dlambda correction of the velocity paramter. dlambda plus
the initial parameter (used in migration) will be
the updated one.
deviation to measure how close imaged depths are to each other
in CIGs. Old deviation corresponds to the initial
parameter; new deviation corresponds to the updated one.
sensitivity to predict how sensitive the error in the estimated
parameter is to an error in the measurement of imaged
depths.
error of parameter <= sensitivity * error of depth.
Reference:
Liu, Z. 1995, "Migration Velocity Analysis", Ph.D. Thesis, Colorado
School of Mines, CWP report #168.
89
VTLVZ -- Velocity as function of Time for Linear V(Z);
writes out a vector of velocity = v0 exp(a t/2)
vtlvz > velfile nt= dt= v0= a=
Required parameters
nt= number of time samples
dt= time sampling interval
v0= velocity at the surface
a= velocity gradient
90
WKBJ - Compute WKBJ ray theoretic parameters, via finite differencing
wkbj <vfile >tfile nx= nz= xs= zs= [optional parameters]
<vfile file containing velocities v[nx][nz]
xs= x coordinate of source
zs= z coordinate of source
sfile=sfile file containing sigmas sg[nx][nz]
bfile=bfile file containing incident angles bet[nx][nz]
afile=afile file containing propagation angles a[nx][nz]
Notes:
Traveltimes, propagation angles, sigmas, and incident angles in WKBJ
by finite differences in polar coordinates. Traveltimes are calculated
by upwind scheme; sigmas and incident angles by a Crank-Nicolson scheme.
Credits:
CWP: Zhenyue Liu, Dave Hale, pre 1992.
91
XY2Z - converts (X,Y)-pairs to spike Z values on a uniform grid
xy2z < stdin npairs= [optional parameters] >stdout
Required parameter:
npairs= number of pairs input
Optional parameter:
scale=1.0 value to scale spikes by
nx1=100 dimension of first (fast) dimension of output array
nx2=100 dimension of second (slow) dimension of output array
x1pad=2 zero padding in x1 dimension
x2pad=2 zero padding in x2 dimension
yx=0 assume (x,y) pairs
=1 assume (y,x) pairs
Notes:
Converts ordered (x,y) pairs to spike x1values, of height=scale on a
uniform grid.
Credits:
CWP: John Stockwell, Nov 1995
92
Z2XYZ - convert binary floats representing Z-values to ascii
form in X Y Z ordered triples
z2xyz <stdin >stdout
n1= number of floats in 1st (fast) dimension
Notes: This program is useful for converting panels of float

data (representing evenly spaced z values) to the x y z
ordered triples required for certain 3D plotting packages.
Example of NXplot3d usage on a NeXT:

suplane | sufilter | z2xyz n1=64 > junk.ascii
Now open junk.ascii as a mesh data file with NXplot3d.

(NXplot3d is a NeXTStep-only utility for viewing 3d data sets
Credits:
CWP: John Stockwell based on "b2a" by Jack Cohen
93
BHEDTOPAR - convert a Binary tape HEaDer file to PAR file format
bhedtopar < stdin outpar=parfile
Required parameter:
none
swap=0 =1 to swap bytes
outpar=/dev/tty =parfile name of output param file
Notes:
This program dumps the contents of a SEGY binary tape header file, as
would be produced by segyread and segyhdrs to a file in "parfile" format.
A "parfile" is an ASCII file containing entries of the form param=value.
Here "param" is the keyword for the binary tape header field and
"value" is the value of that field. The parfile may be edited as
any ASCII file. The edited parfile may then be made into a new binary tape
header file via the program setbhed.
See sudoc setbhed for examples
Credits:
CWP: John Stockwell 11 Nov 1994
94
DT1TOSU - Convert ground-penetrating radar data in the Sensors & Software
X.dt1 GPR format to SU format.
dt1tosu < gpr_data_in_dt1_format > stdout
ns=from header number of samples per trace
dt=.8 time sample interval (see below)
swap=1 perform byte swapping (big endian machines)
=0 don’t swap bytes (little endian machines)
verbose=0 silent
=1 S & S header values from first trace
sent to outpar
=2 S & S header values from all traces
sent to outpar
list=0 silent
=1 list explaining labels used in verbose
is printed to stderr
Caution: An incorrect ns field will munge subsequent processing.
Notes:
For compatiblity with SEGY header, apparent dt is set to
.8 ms (800 microsecs). Actual dt is .8 nanosecs.
Using TRUE DISTANCES, this scales velocity
and frequency by a factor of 1 million.
Example: v_air = 9.83X10^8 ft/s (real)
v_air = 983 ft/s (apparent for su)
Example: fnyquist = 625 MHz (real)
fnyquist = 625 Hz (apparent for su)
IBM RS6000, NeXT, SUN are examples of big endian machines

PC’s and DEC are examples of little endian machines
Caveat:
This program has not been tested on DEC, some modification of the
byte swapping routines may be required.
Credits:
CWP: John Stockwell, Jan 1994 Based on a code "sugpr" by
UTULSA: Chris Liner & Bill Underwood (Dec93)
95
modifications permit S & S dt1 header information to be transferred
directly to SU header
Trace header fields set: ns, tracl, tracr, dt, delrt, trid,
hour, minute, second
96
LAS2SU - convert las2 format well log curves to su traces
las2su <stdin nskip= ncurve= >stdout [optional params]
none
ncurve=automatic number of log curves in the las file
dz=0.5 input depth sampling (ft)
m=0 output (d1,f1) in feet
=1 output (d1,f1) in meters
ss=0 do not subsample (unless nz > 32767 )
=1 pass every other sample
verbose=0 =1 to echo las header lines to screen
outhdr=las_header.asc name of file for las headers
Notes:
1. It is recommended to run LAS_CERTIFY available from CWLS
at http://cwls.org.
2. First log curve MUST BE depth.
3. If number of depth levels > 32767 (segy NT limit)
then input is subsampled by factor of 2 or 4 as needed
4. Logs may be isolated using tracl header word (1,2,...,ncurve)
tracl=1 is depth
Credits:
* Chris Liner based on code by Ferhan Ahmed and a2b.c (June 2005)
* CWP: John Stockwell 31 Oct 2006, combining lasstrip and
* CENPET: lasstrip 2006 by Werner Heigl
*
* Ideas for improvement:
* add option to chop off part of logs (often shallow
* portions are not of interest
* cross check sampling interval from header against
* values found from first log curve (=depth)
*
97
SEGYCLEAN - zero out unassigned portion of header
segyclean <stdin >stdout
Since "foreign" SEG-Y tapes may use the unassigned portion

of the trace headers and since SU now uses it too, this
program zeros out the fields meaningful to SU.
Example:
segyread trmax=200 | segyclean | suximage
Credits:
CWP: Jack Cohen
98
char* sdoc[] = {
SEGYHDRMOD - replace the text header on a SEGY file
segyhdrmod text=file data=file
text= name of file containing new 3200 byte text header

data= name of file containing SEGY data set
====================================================================*\
sgyhdrmod - replace the text header on a SEGY data file in place
This program only reads and writes 3200 bytes
Reginald H. Beardsley [email protected]
\*====================================================================*/
99
SEGYHDRS - make SEG-Y ascii and binary headers for segywrite
segyhdrs [ < sudata ] [optional parameters] [ > copy of sudata ]
ns= if no input trace header
dt= if no input trace header
ns=tr.ns from header number of samples on input traces
dt=tr.dt from header sample rate (microseconds) from traces
bfile=binary name of file containing binary block
hfile=header name of file containing ascii block
Some binary header fields are set:
jobid=1 job id field
lino=1 line number (only one line per reel)
reno=1 reel number
format=1 data format
All other fields are set to 0, by default.

To set any binary header field, use sukeyword to find out
the appropriate keyword, then use the getpar form:
keyword=value to set keyword to value
The header file is created as ascii and is translated to ebcdic

by segywrite before being written to tape. Its contents are
formal but can be edited after creation as long as the forty
line format is maintained.
Caveat: This program has not been tested under XDR for machines
not having a 2 byte unsigned short integral data type.
Credits:
CWP: Jack K. Cohen, John Stockwell

MOBIL: Stew Levin
100
SEGYREAD - read an SEG-Y tape
segyread > stdout tape=
or
SEG-Y data stream ... | segyread tape=- > stdout
Required parameter:
tape= input tape device or seg-y filename (see notes)
buff=1 for buffered device (9-track reel tape drive)
=0 possibly useful for 8mm EXABYTE drives
verbose=0 silent operation
=1 ; echo every ’vblock’ traces
vblock=50 echo every ’vblock’ traces under verbose option
hfile=header file to store ebcdic block (as ascii)
bfile=binary file to store binary block
over=0 quit if bhed format not equal 1, 2, 3, 5, or 8
= 1 ; override and attempt conversion
format=bh.format if over=1 try to convert assuming format value
conv=1 convert data to native format
= 0 ; assume data is in native format
ns=bh.hns number of samples (use if bhed ns wrong)
trcwt=1 apply trace weighting factor (bytes 169-170)
=0, do not apply. (Default is 0 for formats 1 and 5)
trmin=1 first trace to read
trmax=INT_MAX last trace to read
endian=1 set =0 for little-endian machines(PC’s,DEC,etc.)
errmax=0 allowable number of consecutive tape IO errors
remap=...,... remap key(s)
byte=...,... formats to use for header remapping
Notes:
Traditionally tape=/dev/rmt0. However, in the modern world tape device
names are much less uniform. The magic name can often be deduced by
"ls /dev". Likely man pages with the names of the tape devices are:
"mt", "sd" "st". Also try "man -k scsi", " man mt", etc.
Sometimes "mt status" will tell the device name.
For a SEG-Y diskfile use tape=filename.
101
Remark: a SEG-Y file is not the same as an su file. A SEG-Y file
consists of three parts: an ebcdic header, a binary reel header, and
the traces. The traces are (usually) in 32 bit IBM floating point
format. An SU file consists only of the trace portion written in the
native binary floats.
Formats supported:
1: IBM floating point, 4 byte (32 bits)
2: two’s complement integer, 4 byte (32 bits)
5: IEEE floating point, 4 byte (32 bits)
tape=- read from standard input. Caveat, under Solaris, you will
need to use the buff=1 option, as well.
Header remap:
The value of header word remap is mapped from the values of byte
Map a float at location 221 to sample spacing d1:

segyread <data >outdata remap=d1 byte=221f
Map a long at location 225 to source location sx:

segyread <data >outdata remap=sx byte=225l
Map a short at location 229 to gain constant igc:

segyread <data >outdata remap=igc byte=229s
Or all combined:
segyread <data >outdata remap=d1,sx,igc byte=221f,225l,229s
Segy header words are accessed as Xt where X denotes the byte number
starting at 1 in correspondance with the SEGY standard (1975)
Known types include: f float (4 bytes)
l long int (4 bytes)
s short int (2 bytes)
b byte (1 bytes)
type: sudoc segyread for further information
Note:
102
If you have a tape with multiple sequences of ebcdic header,
binary header,traces, use the device that
invokes the no-rewind option and issue multiple segyread
commands (making an appropriate shell script if you
want to save all the headers). Consider using >> if
you want a single trace file in the end. Similar
considerations apply for multiple reels of tapes,
but use the standard rewind on end of file.
Note: For buff=1 (default) tape is accessed with ’read’, for buff=0
tape is accessed with fread. We suggest that you try buff=1
even with EXABYTE tapes.
Caveat: may be slow on an 8mm streaming (EXABYTE) tapedrive
Warning: segyread or segywrite to 8mm tape is fragile. Allow sufficient
time between successive reads and writes.
Warning: may return the error message "efclose: fclose failed"
intermittently when segyreading/segywriting to 8mm (EXABYTE) tape
even if actual segyread/segywrite is successful. However, this
error message may be returned if your tape drive has a fixed
block size set.
Caution: When reading or writing SEG-Y tapes, the tape
drive should be set to be able to read variable block length
tape files.
Credits:
SEP: Einar Kjartansson
CWP: Jack K. Cohen, Brian Sumner, Chris Liner
: John Stockwell (added 8mm tape stuff)
conv parameter added by:
Tony Kocurko
Department of Earth Sciences
Memorial University of Newfoundland
St. John’s, Newfoundland
read from stdin via tape=- added by Tony Kocurko
bhed format = 2,3 conversion by:
Remco Romijn (Applied Geophysics, TU Delft)
J.W. de Bruijn (Applied Geophysics, TU Delft)
bhed format = 8 conversion by: John Stockwell
header remap feature added by:
Matthias Imhof, Virginia Tech
--------------------------
Additional Notes:
103
Brian’s subroutine, ibm_to_float, which converts IBM floating
point to IEEE floating point is NOT portable and must be
altered for non-IEEE machines. See the subroutine notes below.
A direct read by dd would suck up the entire tape; hence the

dancing around with buffers and files.
104
SEGYWRITE - write an SEG-Y tape
segywrite <stdin tape=
tape= tape device to use (see sudoc segyread)
Optional parameter:
verbose=0 silent operation
=1 ; echo every ’vblock’ traces
vblock=50 echo every ’vblock’ traces under verbose option
buff=1 for buffered device (9-track reel tape drive)
=0 possibly useful for 8mm EXABYTE drive
conv=1 =0 don’t convert to IBM format
hfile=header ebcdic card image header file
bfile=binary binary header file
trmin=1 first trace to write
trmax=INT_MAX last trace to write
endian=1 =0 for little-endian machines (PC’s, DEC,etc...)
errmax=0 allowable number of consecutive tape IO errors
format= override value of format in binary header file
Note: The header files may be created with ’segyhdrs’.
Note: For buff=1 (default) tape is accessed with ’write’, for buff=0
tape is accessed with fwrite. Try the default setting of buff=1
for all tape types.
Caveat: may be slow on an 8mm streaming (EXABYTE) tapedrive
Warning: segyread or segywrite to 8mm tape is fragile. Allow time
between successive reads and writes.
Precaution: make sure tapedrive is set to read/write variable blocksize
tapefiles.
For more information, type: sudoc <segywrite>
Warning: may return the error message "efclose: fclose failed"

intermittently when segyreading/segywriting to 8mm EXABYTE tape,
even if actual segyread/segywrite is successful. However, this
may indicate that your tape drive has been set to a fixed block
size. Tape drives should be set to variable block size before reading
105
or writing tapes in the SEG-Y format.
Credits:
CWP: Jack, Brian, Chris
: John Stockwell (added EXABYTE functionality)
Notes:
Brian’s subroutine, float_to_ibm, for converting IEEE floating
point to IBM floating point is NOT portable and must be
altered for non-IEEE machines. See the subroutine notes below.
On machines where shorts are not 2 bytes and/or ints are not
4 bytes, routines to convert SEGY 16 bit and 32 bit integers
will be required.
The program, segyhdrs, can be used to make the ascii and binary
files required by this code.
106
SETBHED - SET the fields in a SEGY Binary tape HEaDer file, as would be
produced by segyread and segyhdrs
setbhed par= [optional parameters]
Required parameter:
none
bfile=binary output binary tape header file
par= =parfile
Set field by field, if desired:
jobid= job id field
lino= line number (only one line per reel)
reno= reel number
format= data format
... etc....
To set any binary header field, use sukeyword to find out
the appropriate keyword, then use the getpar form:
keyword=value to set keyword to value
Notes:
As with all other programs in the CWP/SU package that use getpars,
(GET PARameters from the command line) a file filled with such
statments may be included via option par=parfile. In particular, a
parfile created by "bhedtopar" may be used as input for the program
"setbhed".
Caveat: This program breaks if a "short" isn’t 2 bytes since

the SEG-Y standard demands a 2 byte integer for ns.
Credits:
CWP: John Stockwell 11 Nov 1994
107
SU3DCHART - plot x-midpoints vs. y-midpoints for 3-D data
su3dchart <stdin >stdout
outpar=null name of parameter file
degree=0 =1 convert seconds of arc to degrees
The output is the (x, y) pairs of binary floats
Example:
su3dchart <segy_data outpar=pfile >plot_data
psgraph <plot_data par=pfile \\
linewidth=0 marksize=2 mark=8 | ...
rm plot_data
su3dchart <segy_data | psgraph n=1024 d1=.004 \\

Note: sx, etc., are declared double because float has only 7
significant numbers, that’s not enough, for example,
when tr.scalco=100 and coordinates are in second of arc
and located near 30 degree latitude and 59 degree longitude
Credits:
CWP: Shuki Ronen
Toralf Foerster
Trace header fields accessed: sx, sy, gx, gy, counit, scalco.
108
SUABSHW - replace header key word by its absolute value
suabshw <stdin >stdout key=offset
none
Optional parameter:
key=offset header key word
Credits:
CWP: Jack K. Cohen
109
SUACOR - auto-correlation
suacor <stdin >stdout [optional parms]
ntout=101 odd number of time samples output
norm=1 if non-zero, normalize maximum absolute output to 1
sym=1 if non-zero, produce a symmetric output from
lag -(ntout-1)/2 to lag +(ntout-1)/2
Credits:
CWP: Dave Hale
Trace header fields accessed: ns

Trace header fields modified: ns and delrt
110
SUADDEVENT - add a linear or hyperbolic moveout event to seismic data
suaddevent <stdin >stdout [optional parameters]
none
type=nmo =lmo for linear event
t0=1.0 zero-offset intercept time IN SECONDS
vel=3000. moveout velocity in m/s
amp=1. amplitude
dt= must provide if 0 in headers (seconds)
Typical usage:
sunull nt=500 dt=0.004 ntr=100 | sushw key=offset a=-1000 b=20 \\
| suaddevent v=1000 t0=0.05 type=lmo | suaddevent v=1800 t0=0.8 \
| sufilter f=8,12,75,90 | suxwigb clip=1 &
Credits:
Gary Billings, Talisman Energy, May 1996, Apr 2000, June 2001
Note: code is inefficient in that to add a single "spike", with sinc

interpolation, an entire trace is generated and added to
the input trace. In fact, only a few points needed be created
and added, but the current coding avoids the bookkeeping re
which are the relevant points!
111
char* sdoc[] = {
SUADDHEAD - put headers on bare traces and set the tracl and ns fields
suaddhead <stdin >stdout ns= ftn=0
Required parameter:
ns=the number of samples per trace
Optional parameter:
ifdef SU_LINE_HEADER
head= file to read headers in
not supplied -- will generate headers
given -- will read in headers and attach
floating point arrays to form
traces ",
(head can be created via sustrip program)
endif
ftn=0 Fortran flag
0 = data written unformatted from C
1 = data written unformatted from Fortran
tsort=3 trace sorting code:
1 = as recorded (no sorting)
2 = CDP ensemble
3 = single fold continuous profile
4 = horizontally stacked ",
ntrpr=1 number of data traces per record
if tsort=2, this is the number of traces per cdp",
Trace header fields set: ns, tracl

Use sushw/suchw to set other needed fields.
Caution: An incorrect ns field will munge subsequent processing.

Note: n1 and nt are acceptable aliases for ns.
Example:
suaddhead ns=1024 <bare_traces | sushw key=dt a=4000 >segy_traces
This command line adds headers with ns=1024 samples. The second part
of the pipe sets the trace header field dt to 4 ms.
112
SUADDNOISE - add noise to traces
suaddnoise <stdin >stdout sn=20 noise=gauss seed=from_clock
if any of f=f1,f2,... and amp=a1,a2,... are specified by the user
and if dt is not set in header, then dt is mandatory
sn=20 signal to noise ratio
noise=gauss noise probability distribution
=flat for uniform; default Gaussian
f=f1,f2,... array of filter frequencies (as in sufilter)
amps=a1,a2,... array of filter amplitudes
dt= (from header) time sampling interval (sec)
verbose=0 =1 for echoing useful information
tmpdir= if non-empty, use the value as a directory path

prefix for storing temporary files; else if the
the CWP_TMPDIR environment variable is set use
its value for the path; else use tmpfile()
Notes:
Output = Signal + scale * Noise
scale = (1/sn) * (absmax_signal/sqrt(2))/sqrt(energy_per_sample)
If the signal is already band-limited, f=f1,f2,... and amps=a1,a2,...

can be used, as in sufilter, to bandlimit the noise traces to match
the signal band prior to computing the scale defined above.
Examples of noise bandlimiting:

low freqency: suaddnoise < data f=40,50 amps=1,0 | ...
high freqency: suaddnoise < data f=40,50 amps=0,1 | ...
near monochromatic: suaddnoise < data f=30,40,50 amps=0,1,0 | ...
with a notch: suaddnoise < data f=30,40,50 amps=1,0,1 | ...
bandlimited: suaddnoise < data f=20,30,40,50 amps=0,1,1,0 | ...
Credits:
CWP: Jack Cohen, Brian Sumner, Ken Larner
John Stockwell (fixed filtered noise option)
113
Notes:
At S/N = 2, the strongest reflector is well delineated, so to
see something 1/nth as strong as this dominant reflector
requires S/N = 2*n.
Trace header field accessed: ns
114
SUADDSTATICS - ADD random STATICS on seismic data
suaddstatics required parameters [optional parameters] > stdout
shift= the static shift will be generated
randomly in the interval [+shift,-shif] (ms)
sources= number of source locations
receivers= number of receiver locations
cmps= number of common mid point locations
maxfold= maximum fold of input data
datafile= name and COMPLETE path of the input file
dt=tr.dt time sampling interval (ms)
seed=getpid() seed for random number generator
verbose=0 =1 print useful information
Notes:
SUADDSTATICS applies static time shifts in a surface consistent way on
seismic data sets. SUADDSTATICS writes the static time shifts in the
header field TSTAT. To perform the actual shifts the user should use
the program SUSTATIC after SUADDSTATICS. SUADDSTATICS outputs the
corrupted data set to stdout.
Header field used by SUADDSTATICS: cdp, sx, gx, tstat, dt.
Credits: CWP Wences Gouveia, 11/07/94, Colorado School of Mines
115
SUALFORD - trace by trace Alford Rotation of shear wave data volumes
sualford inS11=file1 inS22=file2 inS12=file3 inS21=file4

outS11=file5 outS22=file6 outS12=file7 outS21=file8 [optional
parameters]
inS11= input data volume for the 11 component
outS11= output data volume for the 11 component
angle_inc= sets the increment to the angle by which
the data sets are rotated. The minimum is
set to be 1 degree and default is 5.
Az_key= to set the header field storing the azimuths
for the fast shear wave on the output volumes
Q_key= to set the header field storing the quality
factors of performed optimum rotations
lag_key= to set the header field storing the lag in
miliseconds the fast and slow shear components
xcorr_key= to set the header field containing the maxi-
mum normalized cross-correlation between the
and slow shear waves.
taper= 2*taper+1 is the length of the sample overlap
between the unrotated data with the rotated
data on the traces. The boundary between them
is defined by time windowning.
taper = -1, for no-overlap
taper = 0, for overlap of one sample
taper =>1, for use of cosine scale to
to interpolate between the
unrotated and rotated parts
of the traces
taperwin= another taper used to taper the data within

the window of analysis to diminish the effect
of data near the window edges.In this way one
116
can focus on a given reflector. Also given in
number of samples
maxlag= maximum limit in ms for the lag between fast

and slow shear waves. If this threshold is
attained or surpassed, the quality factor for
the rotation is zeroed as well as all the
parameters found for that certain rotation
ntraces= number of traces to be used per computation

ntraces=3 will use three adjacent traces to
compute the angle of rotation "
Notes:
The Alford Rotation is a method to rotate the four components

of a shear wave survey into its natural coordinate system, where
the fast and slow shear correspond to the inline to inline shear (S11)
and xline to xline (S22) volumes, respectively.
This Alford Rotation code tries to maximize the energy in the

diagonal volumes, i.e., S11 and S22, while minimizing the energy
in the off-diagonals, i.e., in volumes S12 and S21, in a trace by
trace manner. It then returns the new rotated volumes, saving the
the quality factor for the rotation and azimuth angle of the fast
shear wave direction for each trace headers of the new rotated S11
volume.
The fields in the header containing the Azimuth and Quality factor
and the sample lag between fast and slow shear are otrav, grnolf and
grnofr, respectively, by default. The values are multiplied by ten in
the case of the angles and by a thousand for quality factors. To
change this defaults use the optional parameters Az_key, Q_key and
lag_key
modified header fields:

the ones specified by Az_key, Q_key, lag_key and xcorr_key. By default
these are otrav, grnlof, tstat and grnors, respectively.
Credits:
CWP: Rodrigo Felicio Fuck
117
Code translated and adapted from original version in Fortran
by Ted Schuck (1993)
Schuck, E. L. , 1993, Multicomponent, three dimensional seismic

characterization of a fractured coalbed methane reservoir,
Cedar Hill Field, San Juan County, New Mexico, Ph.D. Thesis,
118
SUAMP - output amp, phase, real or imag trace from
(frequency, x) domain data
suamp <stdin >stdout mode=amp
none
Optional parameter:
mode=amp output flag
amp = output amplitude traces
phase = output phase traces
real = output real parts
imag = output imag parts
Note:
The trace returned is half length from 0 to Nyquist.
Example:
sufft <data | suamp >amp_traces
Example:
sufft < data > complex_traces
suamp < complex_traces mode=real > real_traces
suamp < complex_traces mode=imag > imag_traces
Note: the inverse of the above operation is:

suop2 real_traces imag_traces op=zipper > complex_traces
Credits:
CWP: Shuki, Jack
Notes:
If efficiency becomes important consider inverting main loop
and repeating extraction code within the branches of the switch.
Trace header fields accessed: ns, trid

Trace header fields modified: ns, trid
119
SUASCII - print non zero header values and data in various formats
suascii <stdin >ascii_file
Optional parameter:
bare=0 print headers and data
=1 print only data
=2 print headers only
=3 print data in print data in .csv format, e.g. for Excel
=4 print data as tab delimited .txt file, e.g. for GnuPlot
=5 print data as .xyz file, e.g. for plotting with GMT
ntr=50 maximum number of output traces (bare=3 or bare=4 only)

index=0 don’t include time/depth index in ascii file (bare=4)
=1 include time/depth index in ascii file
key= if set, name of keyword containing x-value

in .xyz output (bare=5 only)
sep= if set, string separating traces in .xyz output
(bare=5; default is no separation)
verbose=0 =1 for detailed information
Notes:
The programs suwind and suresamp provide trace selection and
subsampling, respectively.
With bare=0 and bare=1 traces are separated by a blank line.
With bare=3 a maximum of ntr traces are output in .csv format

("comma-separated value"), e.g. for import into spreadsheet
applications like Excel.
With bare=4 a maximum of ntr traces are output in as tab delimited

columns. Use bare=4 for plotting in GnuPlot.
With bare=5 traces are written as "x y z" triples as required

by certain plotting programs such as the Generic Mapping Tools
(GMT). If sep= is set, traces are separated by a line containing
the string provided, e.g. sep=">" for GMT multisegment files.
"option=" is an acceptable alias for "bare=".
Related programs: sugethw, sudumptrace
120
Credits:
CWP: Jack K. Cohen c. 1989
CENPET: Werner M. Heigl 2006 - bug fixes & extensions
RISSC: Nils Maercklin 2006
Trace header field accessed: ns, dt, delrt, d1, f1, trid
121
SUATTRIBUTES - instantaneous trace ATTRIBUTES
suattributes <stdin >stdout mode=amp
none
Optional parameter:
mode=amp output flag
=amp envelope traces
=phase phase traces
=freq frequency traces
=freqw Frequency Weighted Envelope
=thin Thin-Bed (inst. freq - average freq)
=bandwith Instantaneous bandwidth
=normamp Normalized Phase (Cosine Phase)
=fdenv 1st envelope traces derivative
=sdenv 2nd envelope traces derivative
=q Ins. Q Factor
unwrap= default unwrap=0 for mode=phase
default unwrap=1 for mode=freq
dphase_min=PI/unwrap
wint= windowing for freqw
windowing for thin
default=1
o--------o--------o
data-1 data data+1
Notes:
This program performs complex trace attribute analysis. The first three
attributes, amp,phase,freq are the classical Taner, Kohler, and
Sheriff, 1979.
The unwrap parameter is active only for mode=freq and mode=phase. The
quantity dphase_min is the minimum change in the phase angle taken to be
the result of phase wrapping, rather than natural phase variation in the
data. Setting unwrap=0 turns off phase-unwrapping altogether. Choosing
unwrap > 1 makes the unwrapping function more sensitive to phase changes.
Setting unwrap > 1 may be necessary to resolve higher frequencies in
data (or sample data more finely). The phase unwrapping is crude. The
differentiation needed to compute the instantaneous frequency
122
freq(t)= d(phase)/dt is a simple centered difference.
Examples:
suvibro f1=10 f2=50 t1=0 t2=0 tv=1 | suattributes2 mode=amp | ...
suvibro f1=10 f2=50 t1=0 t2=0 tv=1 | suattributes2 mode=phase | ...
suvibro f1=10 f2=50 t1=0 t2=0 tv=1 | suattributes2 mode=freq | ...
suplane | suattributes mode=... | supswigb |...
Credits:
CWP: Jack Cohen
CWP: John Stockwell (added freq and unwrap features)
UGM (Geophysics Students): Agung Wiyono
email:[email protected] (others)
Algorithm:
c(t) = hilbert_tranform_kernel(t) convolved with data(t)
amp(t) = sqrt( c.re^2(t) + c.im^2(t))

phase(t) = arctan( c.im(t)/c.re(t))
freq(t) = d(phase)/dt
Reference: Taner, M. T., Koehler, A. F., and Sheriff R. E.

"Complex seismic trace analysis", Geophysics, vol.44, p. 1041-1063, 1979

Trace header fields modified: d1, trid
123
SUAZIMUTH - compute trace AZIMUTH, offset, and midpoint coordinates
and set user-specified header fields to these values
suazimuth <stdin >stdout [optional parameters]
none
key=otrav header field to store computed azimuths in
scale=1.0 value(key) = scale * azimuth
az=0 azimuth convention flag
0: 0-179.999 deg, reciprocity assumed
1: 0-359.999 deg, points from receiver to source
-1: 0-359.999 deg, points from source to receiver
sector=1.0 if set, defines output in sectors of size
sector=degrees_per_sector, the default mode is
the full range of angles specified by az
offset=0 if offset=1 then set offset header field

offkey=offset header field to store computed offsets in
cmp=0 if cmp=1, then compute midpoint coordinates and

set header fields for (cmpx, cmpy)
mxkey=ep header field to store computed cmpx in
mykey=cdp header field to store computed cmpy in
Notes:
All values are computed from the values in the coordinate fields
sx,sy (source) and gx,gy (receiver).
The output field "otrav" for the azimuth was chosen arbitrarily as
an example of a little-used header field, however, the user may
choose any field that is convenient for his or her application.
Setting the sector=number_of_degrees_per_sector sets key field to

sector number rather than an angle in degrees.
For az=0, azimuths are measured from the North, however, reciprocity
is assumed, so azimuths go from 0 to 179.9999 degrees. If sector
option is set, then the range is from 0 to 180/sector.
For az=1, azimuths are measured from the North, with the assumption
that the direction vector points from the receiver to the source.
124
For az=-1, the direction vector points from the source to the
receiver. No reciprocity is assumed in these cases, so the angles go
from 0 to 359.999 degrees.
If the sector option is set, then the range is from 0 to 360/sector.
Caveat:
This program honors the value of scalco in scaling the values of
sx,sy,gx,gy. Type "sukeyword scalco" for more information.
Type "sukeyword -o" to see the keywords and descriptions of all

header fields.
To plot midpoints, use: su3dchart
Credits:
based on suchw, su3dchart
CWP: John Stockwell and UTulsa: Chris Liner, Oct. 1998
UTulsa: Chris Liner added offset option, Feb. 2002
cll: fixed offset option and added cmp option, May 2003
RISSC: Nils Maercklin added key options for offset and
midpoints, and added azimuth direction option, Sep. 2006
Algorithms:
offset = osign * sqrt( (gx-sx)*(gx-sx) + (gy-sy)*(gy-sy) )
with osign = sgn( min((sx-gx),(sy-gy)) )
midpoint x value xm = (sx + gx)/2

midpoint y value ym = (sy + gy)/2
Azimuth will be defined as the angle, measured in degrees,

turned from North, of a vector pointing to the source from the midpoint,
or from the midpoint to the source. Azimuths go from 0-179.000 degrees
or from 0-180.0 degrees.
value(key) = scale*[90.0 - (180.0/PI)*(atan((sy - ym)/(sx - xm))) ]

or
value(key) = scale*[180.0 - (180.0/PI)*(atan2((ym - sy),(xm - sx)) ]
Trace header fields accessed: sx, sy, gx, gy, scalco.

Trace header fields modified: (user-specified keys)
125
SUBFILT - apply Butterworth bandpass filter
subfilt <stdin >stdout [optional parameters]
if dt is not set in header, then dt is mandatory
Optional parameters: (nyquist calculated internally)

zerophase=1 =0 for minimum phase filter
locut=1 =0 for no low cut filter
hicut=1 =0 for no high cut filter
fstoplo=0.10*(nyq) freq(Hz) in low cut stop band
astoplo=0.05 upper bound on amp at fstoplo
fpasslo=0.15*(nyq) freq(Hz) in low cut pass band
apasslo=0.95 lower bound on amp at fpasslo
fpasshi=0.40*(nyq) freq(Hz) in high cut pass band
apasshi=0.95 lower bound on amp at fpasshi
fstophi=0.55*(nyq) freq(Hz) in high cut stop band
astophi=0.05 upper bound on amp at fstophi
verbose=0 =1 for filter design info
dt = (from header) time sampling interval (sec)
Credits:
CWP: Dave for bf.c subs and test drivers
CWP: Jack for su wrapper
Caveat: zerophase will not do good if trace has a spike near

the end. One could make a try at getting the "effective"
length of the causal filter, but padding the traces seems
painful in an already expensive algorithm.
Trace header fields accessed: ns, dt, trid
126
SUCCWT - Complex continuous wavelet transform of seismic traces
succwt < tdata.su > tfdata.su [optional parameters]
None
noct=5 Number of octaves (int)
nv=10 Number of voices per octave (int)
fmax=Nyq Highest frequency in transform
p=-0.5 Power of scale value normalizing CWT
=0 for amp-preserved spec. decomp.
c=1/(2*fmax) Time-domain inverse gaussian damping parameter
(bigger c means more wavelet oscillations,
default gives minimal oscillations)
k=1 Use complex Morlet as wavelet transform kernel
=2 use Fourier kernel ... Exp[i 2 pi f t]
fs=1 Use dyadic freq sampling (CWT standard, honors
noct, nv)
=2 use linear freq sampling (Fourier standard)
df=1 Frequency sample interval in Hz (used only for fs=2)
NOTE: not yet implimented (hardwired to df=1)
dt=(from tr.dt) Sample interval override (in secs, if time data)
verbose=0 Run silent, except echo c value. (=1 for more info)
Examples:
This generates amplitude spec of the CWT impulse response (IR).
suspike ntr=1 ix1=1 nt=125 | succwt | suamp | suximage &
Real part of Fourier IR with linear freq sampling:
suspike ntr=1 ix1=1 nt=125 | succwt k=2 fs=2 | suamp mode=real | suximage &
Real part of Fourier IR with dyadic freq sampling:
suspike ntr=1 ix1=1 nt=125 | succwt k=2 | suamp mode=real | suximage &
Inverse CWT: (within a constant scale factor)

... | succwt p=-1 | suamp mode=real | sustack key=cdp > inv.su
Notes:
1. Total number of scales: nscale = noct*nv
2. Each input trace spawns nscale complex output traces
3. Lowest frequency in the transform is fmax/( 2^(noct-1/nv) )
4. Header field (cdp) used as cwt spectrum counter
5. Header field (cdpt) used as scale counter within cwt spectrum
127
6. Header field (gut) holds number of cwt scales ‘na’
7. Header field (unscale) holds CWT scale ‘a’
Header fields set: tracl, cdp, cdpt, unscale, gut
Copyright (c) University of Tulsa, 2003-4.

All rights reserved.
Author: UTulsa: Chris Liner, SEP: Bob Clapp
todo:
fix fs=2 case to allow df not equal to 1
History:
6/18/04
major overhaul by Clapp, including fourier implementation.
Speedup ~ 41 times (4100 %)
2/20/04
made p=-0.5 default
2/16/04
added p option to experiment with CWT normalization
2/12/04
replace fb (bandwidth parameter) with c (t-domain gaussian damping const.)
2/10/04 --- in sync with EAS paper in prep
changed morlet scaling (c = 1) to preserve time-domain peak amplitude
changed morlet exp sign to std CWT definition (conjugate) and
mathematica result that only gives positive freq gaussian with neg exp
1/26/04
added linear frequency sampling option
1/23/04
figured out fb and made it a getpar
key: Look at real ccwt output and determine fb by number of
oscillations desired: Default gives -+-+-+-
1/20/04
beefed up verbose output
dimension wavelet to length 2*nt and change correlation call
... this is done to avoid conv edge effects
1/19/04
added fourier wavlet option for comparison with Fourier Transform action
1/17/04
complex morlet amp scaling now set to preserve first scale amp with IR
1/16/04
added dt getpar to handle depth input properly
128
preserves first tracl so tracl is ok after spice
11/11/03
initial version
Trace header fields set: tracl, cdp, cdpt, unscale, gut
129
SUCDDECON - DECONvolution with user-supplied filter by straightforward
Complex Division in the frequency domain
sucddecon <stdin >stdout [optional parameters]
filter= ascii filter values separated by commas
...or...
sufile= file containing SU traces to use as filter
(must have same number of traces as input data
for panel=1)
panel=0 use only the first trace of sufile as filter
=1 decon trace by trace an entire gather
pnoise=0.001 white noise factor for stabilizing results
(see below)
centered=0 not centered, =1 center the output on each trace
verbose=0 silent, =1 chatty
Notes:
For given time-domain input data I(t) (stdin) and deconvolution
filter F(t), the frequency-domain deconvolved trace can be written as:
I(f) I(f) * complex_conjugate[F(f)]

D(f) = ----- ===> D(f) = ------------------------
F(f) |F(f)|^2 + delta
The real scalar delta is introduced to prevent the resulting deconvolved

trace to be dominated by frequencies at which the filter power is close
to zero. As described above, delta is set to some fraction (pnoise) of
the mean of the filter power spectra. Time sampling rate must be the
same in the input data and filter traces. If panel=1 the two input files
must have the same number of traces. Data and filter traces don’t need to
necessarily have the same number of samples, but the filter trace length
length be always equal or shorter than the data traces.
Trace header fields accessed: ns, dt

Trace header fields modified: none
Credits:
CWP: Ivan Vasconcelos
some changes by John Stockwell
130
CAVEATS:
In the option, panel=1 the number of traces in the sufile must be
the same as the number of traces on the input.
Trace header fields accessed: ns,dt

131
SUCENTSAMP - CENTRoid SAMPle seismic traces
sucentsamp <stdin [optional parameters] >sdout
none
dt=from header sampling interval
verbose=1 =0 to stop advisory messages
Notes:
This program takes seismic traces as input, and returns traces
consisting of spikes of height equal to the area of each lobe
of each oscillation, located at the centroid of the lobe in
question. The height of each spike equal to the area of the
corresponding lobe.
Caveat: No check is made that the data are real time traces!
Credits:
Providence Technologies: Tom Morgan
132
SUCHART - prepare data for x vs. y plot
suchart <stdin >stdout key1=sx key2=gx
none
key1=sx abscissa
key2=gx ordinate
outpar=null name of parameter file
The output is the (x, y) pairs of binary floats
Examples:
suchart < sudata outpar=pfile >plot_data
psgraph <plot_data par=pfile title="CMG" \\
rm plot_data
suchart < sudata | psgraph n=1024 d1=.004 \\

fold chart:
suchart < stacked_data key1=cdp key2=nhs |
psgraph n=NUMBER_OF_TRACES d1=.004 \\
linewidth=0 marksize=2 mark=8 > chart.ps
Credits:
CWP: Jack K. Cohen
Notes:
The vtof routine from valpkge converts values to floats.
133
SUCHW - Change Header Word using one or two header word fields
suchw <stdin >stdout [optional parameters]
none
key1=cdp,... output key(s)
key2=cdp,... input key(s)
key3=cdp,... input key(s)
a=0,... overall shift(s)
b=1,... scale(s) on first input key(s)
c=0,... scale on second input key(s)
d=1,... overall scale(s)
e=1,... exponent on first input key(s)
f=1,... exponent on second input key(s)
The value of header word key1 is computed from the values of

key2 and key3 by:
val(key1) = (a + b * val(key2)^e + c * val(key3)^f) / d
Examples:
Shift cdp numbers by -1:
suchw <data >outdata a=-1
Add 1000 to tracr value:

suchw key1=tracr key2=tracr a=1000 <infile >outfile
We set the receiver point (gx) field by summing the offset

and shot point (sx) fields and then we set the cdp field by
averaging the sx and gx fields (we choose to use the actual
locations for the cdp fields instead of the conventional
1, 2, 3, ... enumeration):
suchw <indata key1=gx key2=offset key3=sx b=1 c=1 |

suchw key1=cdp key2=gx key3=sx b=1 c=1 d=2 >outdata
Do both operations in one call:
suchw<indata key1=gx,cdp key2=offset,gx key3=sx,sx b=1,1 c=1,1 d=1,2 >outdata
134
Credits:
CWP: Jack K. Cohen
CWP: John Stockwell, 7 July 1995, added array of keys feature
Delphi: Alexander Koek, 6 November 1995, changed calculation so
headers of different types can be expressed in each other
135
SUCLIPHEAD - Clip header values
sucliphead <stdin >stdout [optional parameters]
none
key=cdp,... header key word(s) to clip
min=0,... minimum value to clip
max=ULONG_MAX,ULONG_MAX,... maximum value to clip
Credits:
Geocon: Garry Perratt
136
char* sdoc[] = {
SUCMP - CoMPare two seismic data sets, returns 0 to the shell ",
if the same and 1 if different
sucmp file_A file_B
none
limit=1.0e-4 normalized difference threshold value
Notes:
This program is the seismic equivalent of the Unix cmp(1)
command. However, unlike cmp(1), it understands seismic data
and will consider files which have only small numerical
differences to be the same.
Sucmp first checks that the number of traces and number of samples
are the same. It then compares the trace headers bit for bit.
Finally it checks that the fractional difference of A & B is
less than limit.
This program is intended as an aid in regression testing changes to

seismic processing programs.
Expected usage is in shell scripts or Makefiles, e.g.

#!/bin/sh
#-------------------------------------------------------
# Run a test data set and verify the result is correct
# If the data doesn’t match show the data on the screen.
#-------------------------------------------------------
./fubar par=tst1.par
sucmp tst1.su ref/tst1.su
if [ $? ]
then
suxwigb <tst1.su &
suxwigb <ref/tst1.su & "
fi
137
Author: Reginald H. Beardsley
[email protected]
sucmp - compare two seismic files in CWP/SU format to see if they

are the same within the user specified limit.
Algorithm:
Loop over both input files comparing data values. To be

considered the same files must have:
- same number of traces

- same number of samples per trace
- trace values within limits of each other
Note that the program exits as soon as the files fail to match.
For readability, explicit temporary variables are used which the

compiler will optimize away. Braces are used on all conditionals
so that a breakpoint can be set to stop the debugger only if the
condition is true.
Because of the overloading of trace header fields in CWP/SU, the

headers are compared bit for bit.
138
SUCOMMAND - pipe traces having the same key header word to command
sucommand <stdin >stdout [Optional parameters]
none
verbose=0 wordy output
key=cdp header key word to pipe on
command="suop nop" command piped into
dir=0 0: change of header key
-1: break in monotonic decrease of header key
+1: break in monotonic increase of header key
Notes:
This program permits limited parallel processing capability by opening
pipes for processes, signaled by a change in key header word value.
Credits:
VT: Matthias Imhof
Note:
The "valxxx" subroutines are in su/lib/valpkge.c. In particular,
"valcmp" shares the annoying attribute of "strcmp" that
if (valcmp(type, val, valnew) {
...
}
will be performed when val and valnew are different.
139
SUCONV - convolution with user-supplied filter
suconv <stdin >stdout filter= [optional parameters]
Required parameters: ONE of

sufile= file containing SU trace to use as filter
filter= user-supplied convolution filter (ascii)
none

Trace header fields modified: ns
Notes: It is quietly assumed that the time sampling interval on the

single trace and the output traces is the same as that on the traces
in the input file. The sufile may actually have more than one trace,
but only the first trace is used.
Examples:
suplane | suwind min=12 max=12 >TRACE
suconv<DATA sufile=TRACE | ...
Here, the su data file, "DATA", is convolved trace by trace with the
the single su trace, "TRACE".
suconv<DATA filter=1,2,1 | ...

Here, the su data file, "DATA", is convolved trace by trace with the
the filter shown.
Credits:
CWP: Jack K. Cohen, Michel Dietrich
CAVEATS: no space-variable or time-variable capacity.

The more than one trace allowed in sufile is the
beginning of a hook to handle the spatially variant case.
Modification:
CWP: Huub Douma (HD), 14/01/2005
Implemented correct use of delrt

140
141
SUCOUNTKEY - COUNT the number of unique values for a given KEYword.
sucountkey < input.su key=[sx,gx,cdp,...]

Required parameter:
key= array of SU header keywords being counted
verbose=1 chatty, =0 just print keyword number
Example:
suplane | sucountkey key=tracl,tracr,offset
Credits: Baoniu Han, [email protected], Nov, 2000
Globals variables*/
segy tr;
internal structure
typedef struct {
int num;
Value *val;
} my_Count;
int main (int argc, char **argv)

{
cwp_String key[SU_NKEYS]; /* array of keywords */
cwp_String type[SU_NKEYS]; /* array of keywords */
int index[SU_NKEYS]; /* name of type of getparred key*/
my_Count count_table[SU_NKEYS]; /* count table for keywords */
int verbose; /* verbose flag */
int ikey; /* key counter */

int nkeys; /* number of header fields set */
int i; /* count index */
Value val0; /* value of key field */
long buf_size;
/* Initialize
initargs(argc, argv);
requestdoc(1);
142
/* Get "key" values
if ((nkeys=countparval("key"))!=0) {
getparstringarray("key",key);
} else {
key[0]="cdp";
}
if (!getparint("verbose", &verbose)) verbose = 1;
/* you can increase this number to change the buffer size

/* for each key word
buf_size=100000L;
/* get types and indexes corresponding to the keys

for (ikey=0; ikey<nkeys; ++ikey) {
type[ikey]=hdtype(key[ikey]);
index[ikey]=getindex(key[ikey]);
count_table[ikey].num=0;
count_table[ikey].val=(Value *) malloc(sizeof(Value)*buf_size);
}
/* get info from first trace

if (!gettr(&tr)) err("can’t get first trace");
for(ikey=0;ikey<nkeys;++ikey){
count_table[ikey].num=1;
gethval(&tr, index[ikey], &count_table[ikey].val[0]);
}
/* loop over traces

while (gettr(&tr)) {
for (ikey=0; ikey<nkeys; ++ikey) {

gethval(&tr,index[ikey],&val0);
for(i=0;i<count_table[ikey].num;i++)
if (valcmp(type[ikey], val0, count_table[ikey].val[i])==0) break;
if (i==count_table[ikey].num){
gethval(&tr, index[ikey], &count_table[ikey].val[count_table[ikey].num]);
count_table[ikey].num++;
}
143
}
if (verbose==1)
warn("The counting of key word is finished, the total number for specified header is
for (ikey=0;ikey<nkeys;++ikey){
if (verbose==1)
warn(" key %s has %d unique entries in this data", key[ikey], count_table[ikey].num)
else
fprintf(stdout,"%s %d ", key[ikey], count_table[ikey].num);
return(CWP_Exit());
}
144
SUCWT - generates Continous Wavelet Transform amplitude, regularity
analysis in the wavelet basis
sucwt < stdin [Optional parameters ] > stdout
none
base=10 Base value for wavelet transform scales
first=-1 First exponent value for wavelet transform scales
expinc=0.01 Exponent increment for wavelet transform scales
last=1.5 Last exponent value for wavelet transform scales
Wavelet Parameters:
wtype=0 2nd derivative of Gaussian (Mexican hat)
=1 4th derivative of Gaussian (witch’s hat)
=2 6th derivative of Gaussian (wizard’s hat)
nwavelet=1024 number of samples in the wavelet
xmin=-20 minimum x value wavelet is computed
xcenter=0 center x value wavelet is computed
xmax=20 maximum x value wavelet is computed
sigma=1 sharpness parameter ( sigma > 1 sharper)

holder=0 =1 compute Holder regularity estimate
divisor=1.0 a floating point number >= 1.0 (see notes)
Notes:
This is the CWT version of the time frequency analysis notion that is
applied in sugabor.
The parameter base is the base of the power that is applied to scale
the wavelet. Some mathematical literature assume base 2. Base 10 works
well here.
Default option yields an output similar to that of sugabor. With the

parameter holder=1 an estimate of the instantaneous Holder regularity
(the Holder exponent) is output for each input data value. The result
is a Holder exponent trace for each corresponding input data trace.
The strict definition of the Holder exponent is the maximum slope of

the rise of the spectrum in the log(amplitude) versus log(scale) domain:
145
divisor=1.0 means the exponent is computed simply by fitting a line
through all of the values in the transform. A value of divisor>1.0
indicates that the Holder exponent is determined as the max of slopes
found in (total scales)/divisor length segments.
Some experimentation with the parameters nwavelet, first, last, and

expinc may be necessary before a desirable output is obtained. The
most effective way to proceed is to perform a number of tests with
holder=0 to determine the range of first, last, and expinc that best
represents the data in the wavelet domain. Then experimentation with
holder=1 and values of divisor>=1.0 may proceed.
Credits:
CWP: John Stockwell, Nov 2004
inspired in part by "bhpcwt" in the BHP_SU package, code written by
BHP: Michael Glinsky, c. 2002, based loosely on a Matlab CWT function
References:
Li C.H., (2004), Information passage from acoustic impedence to

seismogram: Perspectives from wavelet-based multiscale analysis,
Journal of Geophysical Research, vol. 109, B07301, p.1-10.
Mallat, S. and W. L. Hwang, (1992), Singularity detection and

processing with wavelets, IEEE Transactions on information, v 38,
March 1992, p.617 - 643.
146
SUDATUMFD - 2D zero-offset Finite Difference acoustic wave-equation
DATUMing
sudatumfd <stdin > stdout [optional parameters]
nt= number of time samples on each trace

nx= number of receivers per shot gather
nsx= number of shot gathers
nz= number of downward continuation depth steps
dz= depth sampling interval (in meters)
mx= number of horizontal samples in the velocity model
mz= number of vertical samples in the velocity model
vfile1= velocity file used for thin-lens term
vfile2= velocity file used for diffraction term
dx= horizontal sampling interval (in meters)
dt=.004 time sampling interval (in seconds)

buff=5 number of zero traces added to each side of each
shot gather as a pad
tap_len=5 taper length (in number of traces)
x_0=0.0 x coordinate of leftmost position in velocity model
Notes:
The algorithm is a 45-degree implicit-finite-difference scheme based
on the one-way wave equation. It works on poststack (zero-offset)
data only. The two velocity files, vfile1 and vfile2, are binary
files containing floats with the format v[ix][iz]. There are two
potentially different velocity files for the thin-lens and
diffraction terms to allow for the use of a zero-velocity layer
which allows for datuming from an irregular surface.
Source and receiver locations must be set in the header values in

order for the datuming to work properly. The leftmost position of
of the velocity models given in vfile1 and vfile2 must also be given.
Author: Chris Robinson, 10/16/00, CWP, Colorado School of Mines
147
References:
Beasley, C., and Lynn, W., 1992, The zero-velocity layer: migration
from irregular surfaces: Geophysics, 57, 1435-1443.
Claerbout, J. F., 1985, Imaging the earth’s interior: Blackwell

Scientific Publications.
148
SUDATUMK2DR - Kirchhoff datuming of receivers for 2D prestack data
(shot gathers are the input)
sudatumk2dr infile= outfile= [parameters]
infile=stdin file for input seismic traces
outfile=stdout file for common offset migration output
ttfile= file for input traveltime tables
The following 9 parameters describe traveltime tables:
fzt= first depth sample in traveltime table
nzt= number of depth samples in traveltime table
dzt= depth interval in traveltime table
fxt= first lateral sample in traveltime table
nxt= number of lateral samples in traveltime table
dxt= lateral interval in traveltime table
fs= x-coordinate of first source
ns= number of sources
ds= x-coordinate increment of sources
fxi= x-coordinate of the first surface location

dxi= horizontal spacing on surface
nxi= number of input surface locations
sgn= Sign of the datuming process (up=-1 or down=1)
dt= or from header (dt) time sampling interval of input data
ft= or from header (ft) first time sample of input data
surf="0,0;99999,0" The first surface defined the recording surface
surf="0,0;99999,0" and the second one, the new datum.
"x1,z1;x2,z2;x3,z3;...
fzo=fzt z-coordinate of first point in output trace
dzo=0.2*dzt vertical spacing of output trace
nzo=5*(nzt-1)+1 number of points in output trace ",
fxso=fxt x-coordinate of first shot
dxso=0.5*dxt shot horizontal spacing
nxso=2*(nxt-1)+1 number of shots
fxgo=fxt x-coordinate of first receiver
dxgo=0.5*dxt receiver horizontal spacing
nxgo=nxso number of receivers per shot
fmax=0.25/dt frequency-highcut for input traces
offmax=99999 maximum absolute offset allowed in migration
aperx=nxt*dxt/2 migration lateral aperature
149
angmax=60 migration angle aperature from vertical
v0=1500(m/s) reference velocity value at surface
dvz=0.0 reference velocity vertical gradient
antiali=1 Antialiase filter (no-filter = 0)
jpfile=stderr job print file name
mtr=100 print verbal information at every mtr traces
ntr=100000 maximum number of input traces to be migrated
Notes:
1. Traveltime tables were generated by program rayt2d (or other ones)
on relatively coarse grids, with dimension ns*nxt*nzt. In the
datuming process, traveltimes are interpolated into shot/gephone
positions and output grids.
2. Input traces must be SU format and organized in common rec. gathers
3. If the offset value of an input trace is not in the offset array
of output, the nearest one in the array is chosen.
4. Amplitudes are computed using the reference velocity profile, v(z),
specified by the parameters v0= and dvz=.
5. Input traces must specify source and receiver positions via the header
fields tr.sx and tr.gx. Offset is computed automatically.
Author: Trino Salinas, 05/01/96, Colorado School of Mines
This code is based on sukzmig2d.c written by Zhenyue Liu, 03/01/95.

This code implements a Kirchhoff extraplolation operator that allows to

transfer data from one reference surface to another. The formula used in
this application is a far field approximation of the Berryhill’s original
formula (Berryhill, 1979). This equation is the result of a stationary
phase analysis to get an analog asymptotic expansion for the two-and-one
half dimensional extrapolation formula (Bleistein, 1984).
The extrapolation formula permits the downward continuation of upgoing

waves and upward continuation of downgoing waves. For upward conti-
nuation of upgoing waves and downward continuation of downgoing waves,
the conjugate transpose of the equation is used (Bevc, 1993).
References :
150
Berryhill, J.R., 1979, Wave equation datuming: Geophysics,
44, 1329--1344.
_______________, 1984, Wave equation datuming before stack

(short note) : Geophysics, 49, 2064--2067.
Bevc, D., 1993, Data parallel wave equation datuming with

irregular acquisition topography : 63rd Ann. Internat.
Mtg., SEG, Expanded Abstracts, 197--200.
Bleistein, N., 1984, Mathematical methods for wave phenomena,

Academic Press Inc. (Harcourt Brace Jovanovich Publishers),
New York.
151
SUDATUMK2DS - Kirchhoff datuming of sources for 2D prestack data
(input data are receiver gathers)
sudatumk2ds infile= outfile= [parameters]
fxi= x-coordinate of the first surface location

dxi= horizontal spacing on surface
nxi= number of input surface locations
sgn= Sign of the datuming process (up=-1 or down=1)
surf="0,0;99999,0" The first surface defined the recording surface
surf="0,0;99999,0" and the second one, the new datum.
"x1,z1;x2,z2;x3,z3;...
fxso=fxt x-coordinate of first shot
dxso=0.5*dxt shot horizontal spacing
nxso=2*(nxt-1)+1 number of shots
fxgo=fxt x-coordinate of first receiver
exgo=fxgo+(nxgo-1)*dxgo x-coordinate of the last receiver
dxgo=0.5*dxt receiver horizontal spacing
nxgo=nxso number of receivers per shot
152
antiali=1 Antialiase filter (no-filter = 0)
Notes:
migration process, traveltimes are interpolated into shot/gephone
2. Input traces must be SU format and organized in common shot gathers
Author: Trino Salinas, 05/01/96, Colorado School of Mines
This code is based on sukzmig2d.c written by Zhenyue Liu, 03/01/95.

This code implements a Kirchhoff extraplolation operator that allows to

transfer data from one reference surface to another. The formula used in
this application is a far field approximation of the Berryhill’s original
formula (Berryhill, 1979). This equation is the result of a stationary
phase analysis to get an analog asymptotic expansion for the two-and-one
half dimensional extrapolation formula (Bleistein, 1984).
The extrapolation formula permits the downward continuation of upgoing

waves and upward continuation of downgoing waves. For upward conti-
nuation of upgoing waves and downward continuation of downgoing waves,
the conjugate transpose of the equation is used (Bevc, 1993).
References :
153
Berryhill, J.R., 1979, Wave equation datuming: Geophysics,
44, 1329--1344.
_______________, 1984, Wave equation datuming before stack

(short note) : Geophysics, 49, 2064--2067.
Bevc, D., 1993, Data parallel wave equation datuming with

irregular acquisition topography : 63rd Ann. Internat.
Mtg., SEG, Expanded Abstracts, 197--200.
Bleistein, N., 1984, Mathematical methods for wave phenomena,

Academic Press Inc. (Harcourt Brace Jovanovich Publishers),
New York.
154
SUDIPDIVCOR - Dip-dependent Divergence (spreading) correction
sudipdivcor <stdin >stdout [optional parms]
dxcdp distance between sucessive cdps in meters
np=50 number of slopes
tmig=0.0 times corresponding to interval velocities in vmig
vmig=1500.0 interval velocities corresponding to times in tmig
vfile=binary (non-ascii) file containing velocities vmig(t)
conv=0 =1 to apply the conventional divergence correction
trans=0 =1 to include transmission factors
verbose=0 =1 for diagnostic print
Notes:
The tmig, vmig arrays specify an interval velocity function of time.
Linear interpolation and constant extrapolation is used to determine
interval velocities at times not specified. Values specified in tmig
must increase monotonically.
Alternatively, interval velocities may be stored in a binary file

containing one velocity for every time sample. If vfile is
specified, then the tmig and vmig arrays are ignored.
The time of the first sample is assumed to be constant, and is
taken as the value of the first trace header field delrt.
Whereas the conventional divergence correction (sudivcor) is

valid only for horizontal reflectors, which have zero reflection
slope, the dip-dependent divergence correction is valid for any
reflector dip or reflection slope. Only the conventional
correction will be applied to the data if conv=1 is specified.
Note that the conventional correction over-amplifies
reflections from dipping beds
The transmission factor should be applied when the divergence

corrected data is to be migrated with a reverse time migration
based on the constant density wave equation.
Trace header fields accessed: ns, dt, delrt
155
SUDIPFILT - DIP--or better--SLOPE Filter in f-k domain
sudipfilt <infile >outfile [optional parameters]
dt=(from header) if not set in header then mandatory
dx=(from header, d1) if not set in header then mandatory
slopes=0.0 monotonically increasing slopes
amps=1.0 amplitudes corresponding to slopes
bias=0.0 slope made horizontal before filtering
verbose=0 verbose = 1 echoes information

Notes:
d2 is an acceptable alias for dx in the getpar
Slopes are defined by delta_t/delta_x, where delta

means change. Units of delta_t and delta_x are the same
as dt and dx. It is sometimes useful to fool the program
with dx=1 dt=1, thus avoiding units and small slope values.
Linear interpolation and constant extrapolation are used to

determine amplitudes for slopes that are not specified.
Linear moveout is removed before slope filtering to make
slopes equal to bias appear horizontal. This linear moveout
is restored after filtering. The bias parameter may be
useful for spatially aliased data. The slopes parameter is
compensated for bias, so you need not change slopes when you
change bias.
Credits:
CWP: Dave (algorithm--originally called slopef)

Jack (reformatting for SU)
156
Trace header fields accessed: ns, dt, d2
157
SUDIVCOR - Divergence (spreading) correction
sudivcor <stdin >stdout [optional parms]
none
trms=0.0 times corresponding to rms velocities in vrms
vrms=1500.0 interval velocities corresponding to times in trms
vfile= binary (non-ascii) file containing velocities vrms(t)
Notes:
The trms, vrms arrays specify an rms velocity function of time.
interval velocities at times not specified. Values specified in trms
Alternatively, rms velocities may be stored in a binary file

containing one velocity for every time sample. If vfile is specified,
then the trms and vrms arrays are ignored.
The time of the first sample is assumed to be constant, and is taken

as the value of the first trace header field delrt.
Credits:
CWP: Jack K. Cohen, Francesca Fazarri
158
SUDIVSTACK - Diversity Stacking using either average power or peak
power within windows
none
key=tracf key header word to stack on
winlen=0.064 window length in seconds.
typical choices: 0.064, 0.128, 0.256,
0.512, 1.024, 2.048, 4.096
if not specified the entire trace is used
peak=1 peak power option default is average power
Notes:
Diversity stacking is a noise reduction technique used in the
summation of duplicate data. Each trace is scaled by the inverse
of its average power prior to stacking. The composite trace is
then renormalized by dividing by the sum of the scalers used.
This program stacks adjacent traces having the same key header
word, which can be specified by the key parameter. The default
is "tracf" (trace number within field record).
For more information on key header words, type "sukeyword -o".
Examples:
For duplicate field shot records:
susort < field.data tracf | sudivstack > stack.data
For CDP ordered data:
sudivstack < cdp.data key=cdp > stack.data
Author: Mary Palen-Murphy,

Masters’ candidate, Colorado School of Mines,
Geophysics Department, 1994
Implementation of "key=" option: Nils Maercklin,

GeoForschungsZentrum (GFZ) Potsdam, Germany, 2002.
References:
159
Embree, P.,1968, Diversity seismic record stacking method and system:
U.S. patent 3,398,396.
Gimlin, D. R., and Smith, J. W., 1980, A comparison of seismic trace
summing techniques: Geophysics, vol. 45, pages 1017-1041.
Trace header fields accessed: ns, dt, key=keyword

Trace header fields modified: tracl
160
SUDMOFKCW - converted-wave DMO via F-K domain (log-stretch) method for
common-offset gathers
sudmofkcw <stdin >stdout cdpmin= cdpmax= dxcdp= noffmix= [...]
cdpmin minimum cdp (integer number) for which to apply DMO
cdpmax maximum cdp (integer number) for which to apply DMO
dxcdp distance between adjacent cdp bins (m)
noffmix number of offsets to mix (see notes)
tdmo=0.0 times corresponding to rms velocities in vdmo (s)
vdmo=1500.0 rms velocities corresponding to times in tdmo (m/s)
gamma=0.5 velocity ratio, upgoing/downgoing
ntable=1000 number of tabulated z/h and b/h (see notes)
sdmo=1.0 DMO stretch factor; try 0.6 for typical v(z)
flip=0 =1 for negative shifts and exchanging s1 and s2
(see notes below)
fmax=0.5/dt maximum frequency in input traces (Hz)
Notes:
Input traces should be sorted into common-offset gathers. One common-
offset gather ends and another begins when the offset field of the trace
headers changes.
The cdp field of the input trace headers must be the cdp bin NUMBER, NOT
the cdp location expressed in units of meters or feet.
The number of offsets to mix (noffmix) should typically equal the ratio of
the shotpoint spacing to the cdp spacing. This choice ensures that every
cdp will be represented in each offset mix. Traces in each mix will
contribute through DMO to other traces in adjacent cdps within that mix.
The tdmo and vdmo arrays specify a velocity function of time that is
used to implement a first-order correction for depth-variable velocity.
The times in tdmo must be monotonically increasing. The velocity function
is assumed to have been gotten by traditional NMO.
For each offset, the minimum time at which a non-zero sample exists is
used to determine a mute time. Output samples for times earlier than this",
mute time will be zeroed. Computation time may be significantly reduced
161
if the input traces are zeroed (muted) for early times at large offsets.
z/h is horizontal-reflector depth normalized to half source-reciver offset

h. Normalized shift of conversion point is b/h. The code now does not
support signed offsets, therefore it is recommended that only end-on data,
not split-spread, be used as input (of course after being sorted into
common-offset gathers). z/h vs b/h depends on gamma (see Alfaraj’s Ph.D.
thesis, 1993).
Flip factor = 1 implies positive shift of traces (in the increasing CDP
bin number direction). When processing split-spread data, for example,
if one side of the spread is processed with flip=0, then the other side
of the spread should be processed with flip=1. The flip factor also
determines the actions of the factors s1 and s2, i.e., stretching or
squeezing.
Trace header fields accessed: nt, dt, delrt, offset, cdp.
Credits:
CWP: Mohamed Alfaraj
Dave Hale
Transformation to zero offset for mode-converted waves
Mohammed Alfaraj, Ph.D. thesis, 1993, Colorado School of Mines
Dip-Moveout Processing - SEG Course Notes

Dave Hale, 1988
162
SUDMOFK - DMO via F-K domain (log-stretch) method for common-offset gathers
sudmofk <stdin >stdout cdpmin= cdpmax= dxcdp= noffmix= [...]
tmpdir= if non-empty, use the value as a directory path prefix
for storing temporary files; else if the CWP_TMPDIR
environment variable is set use its value for the path;
else use tmpfile()
Notes:
headers changes.
The times in tdmo must be monotonically increasing.
163
Credits:
CWP: Dave
Dave Hale, 1988
Trace header fields accessed: ns, dt, delrt, offset, cdp.
164
SUDMOTIVZ - DMO for Transeversely Isotropic V(Z) media for common-offset
gathers
sudmotivz <stdin >stdout cdpmin= cdpmax= dxcdp= noffmix= [...]
vnfile= binary (non-ascii) file containing NMO interval
velocities (m/s)
vfile= binary (non-ascii) file containing interval velocities (m/s)
etafile= binary (non-ascii) file containing eta interval values (m/s)
tdmo=0.0 times corresponding to interval velocities in vdmo (s)
vndmo=1500.0 NMO interval velocities corresponding to times in tdmo (m/s)
vdmo=vndmo interval velocities corresponding to times in tdmo (m/s)
etadmo=1500.0 eta interval values corresponding to times in tdmo (m/s)
smute=1.5 stretch mute used for NMO correction
speed=1.0 twist this knob for speed (and aliasing)
Notes:
headers changes.
vnfile, vfile and etafile should contain the regularly sampled interval
values of NMO velocity, velocity and eta respectivily as a
function of time. If, for example, vfile is not supplied, the interval
velocity function is defined by linear interpolation of the values in the
tdmo and vdmo arrays. The times in tdmo must be monotonically increasing.
165
If vfile or vdmo are not given it will be equated to vnfile or vndmo.
For each offset, the minimum time to process is determined using the
smute parameter. The DMO correction is not computed for samples that
have experienced greater stretch during NMO.
166
SUDMOTX - DMO via T-X domain (Kirchhoff) method for common-offset gathers
sudmotx <stdin >stdout cdpmin= cdpmax= dxcdp= noffmix= [optional parms]
dxcdp distance between successive cdps
offmax=3000.0 maximum offset
tmute=2.0 mute time at maximum offset offmax
vrms=1500.0 RMS velocity at mute time tmute
else use tmpfile()
Notes:
headers changes.
The defaults for offmax and vrms are appropriate only for metric units.
If distances are measured in feet, then these parameters should be
specified explicitly.
offmax, tmute, and vrms need not be specified precisely.

If these values are unknown, then one should overestimate offmax
and underestimate tmute and vrms.
No muting is actually performed. The tmute parameter is used only to
167
determine parameters required to perform DMO.
Credits:
CWP: Dave Hale
A non-aliased integral method for dip-moveout
Dave Hale
submitted to Geophysics, June, 1990
168
SUDMOVZ - DMO for V(Z) media for common-offset gathers
sudmovz <stdin >stdout cdpmin= cdpmax= dxcdp= noffmix= [...]
vfile= binary (non-ascii) file containing interval velocities (m/s)
tdmo=0.0 times corresponding to interval velocities in vdmo (s)
vdmo=1500.0 interval velocities corresponding to times in tdmo (m/s)
smute=1.5 stretch mute used for NMO correction
speed=1.0 twist this knob for speed (and aliasing)
Notes:
headers changes.
vfile should contain the regularly sampled interval velocities as a

function of time. If vfile is not supplied, the interval velocity
function is defined by linear interpolation of the values in the tdmo
and vdmo arrays. The times in tdmo must be monotonically increasing.
For each offset, the minimum time to process is determined using the
smute parameter. The DMO correction is not computed for samples that
have experienced greater stretch during NMO.
169
SUDUMPTRACE - print selected header values and data.
Print first num traces.
Use SUWIND to skip traces.
sudumptrace < stdin [> ascii_file]
num=4 number of traces to dump
key=key1,key2,... key(s) to print above trace values
hpf=0 header print format is float
=1 print format is exponential
Examples:
sudumptrace < inseis.su PRINTS: 4 traces, no headers
sudumptrace < inseis.su key=tracf,offset
sudumptrace < inseis.su num=7 key=tracf,offset > info.txt
sudumptrace < inseis.su num=7 key=tracf,offset hpf=1 > info.txt
Related programs: suascii, sugethw
Credits:
MTU: David Forel, Jan 2005
Trace header field accessed: nt, dt, delrt
170
SUEA2DF - SU version of (an)elastic anisotropic 2D finite difference
forward modeling
suea2df > outfile c11file= c55file [optional parameters]
c11file=c11_file c11 voigt elasticity parameter filename
rhofile=rho_file density filename
Anisotropy parameters:
aniso=0 =1 - include anisotropy parameters
mode=0 =0 output particle velocity, =1 output stresses
(snapshots only)
... the next 3 parameters become active only when aniso=1....

dt=0.001 time sampling interval (s)

ft=0.0 first time (s)
lt=1.0 last time (s)
nx=200 number of values in slow (x-direction)

dx=10.0 spatial sampling interval (m) x-coor
fx=-1000 first x coor (m)
nz=100 number of values in fast (z)-dimension

dz=dx spatial sampling interval (m) z-coor
fz=0 firstz coor (m)
Source parameters:
sx=0 source x position (m)
sz=500 source location (m)
stype=’p’ source type
p: P-wave
v: velocity
pw: P plane-wave
171
sang=0 for stype=’pw’: plane wave angle
wtype=’dg’ wave type
dg: Gaussian derivative
ga: Gaussian
ri: Ricker
sp: spike, sp2: double spike
ts=0.05 source duration (s)
favg=50 source average frequency
Attenuation parameters:
ql= model q at xl,zl
qsw=0 switch to include attenuation =1 - include
Boundary condition parameters:

bc=10,10,10,10 Top,left,bottom,right boundary condition
=0 none
=1 symmetry
=2 free surface (top only)
>2 absorbing (value indicates width of absorbing
layer
bc_a=0.95; bc initial taper value for absorbing boundary
bc_r=0.; bc exponential factor for absorbing boundary
variables are scaled by bc_a*pow(i,-bc_r)
Output parameters:
sofile= name of source file
snfile= name of file containing for snapshots
snaptime= times of snapshots i.e. snaptime=0.1,0.2,0.3
vsx= x coordinate of vertical line of seismograms

hsz= z coordinate of horizontal line of seismograms
vrslfile="vsp.su" output file for vertical line of seismograms[nz][nt]
hsfile="hs.su" output file for horizontal line of seismograms[nx][nt]
tsw=0 switch to use shear stress only in non-fluid
media - may help reduce dispersion tsw=1. If
tsw=0 then standard calculation
verbose=0 =1 to print progress on screen
Notes:
1) The outfile contains information generated by the input parameters,
such as memory allocation, stability, etc. If your input file does
not work, check this file first.
172
The model is specified as binary files of stiffness parameters and
densities. These may be created any way the user desires. The program
unif2 or makevel may be used to generate densities, and the program
unif2aniso may be used to generate the stiffnesses. You will need to
transpose these files, as the input format for suea2df assumes that the
fast dimesion is the horizontal or the x-dimension. You may do this via
transp n1=nz < c11_file > transp_c11_file
If qsw=1 a ql array specifies the values of the quality factor Q, to

specify the attenuation of the medium.
If aniso=1 then the program will expect the additional stiffnesss files.
The epl and dsl are epsilon and delstar in Thomson (1986, Geophys.)
anl is the angle the anisotropy makes with the horizontal.
Output files (always generated)

hsfile
vrslfile
hsfile.chd - header for hsfile
vrslfile.chd - header for vrslfile
hsfile.mod - model file
Output files (if requested)

sofile - ascii source file
snfile - su format snapshots file
Credits: UU GEOPHY Chris Juhlin 15 May 1999

Copyright (c) Uppsala University, 1998.
Parts of program use Seismic Unix Package - CSM
Changes - C. Juhlin
1. Fixed upgrading of stresses. There was an error in the coding for

the Tzz term, c15 was being used instead of c35. This only caused
problems for dipping anisotropic layers
2. Added some header information for hutput of snapshots.
173
3. 2001-01-30: Added option to set absorbing bc constants bc_a and bc_r
4. 2001-02-23: Corrected bug in outputting model boundaries to standard

output in
routine get_econst
5. 2001-04-26: Added option for updating velocities to only use

shear stress if material is non-fluid, this appears to reduce dispersion at
near grazing angles for fluid-solid boundary. Set tsw=1 to invoke
6. 2001-05-14: Changed loop in free-surface boundary condition for velocties

Thanks goes to Mike Holzrichter for pointing out this problem and the wrong
scaling factor in the updating.
7. 2001-05-16: Changed set_layers function to avoid negative indexing.

Thanks goes to Mike Holzrichter for pointing out this problem
8. 2001-05-17: Modified make_seis to take into account VSP geometry

correctly and not store unnecessary data.
9. 2001-08-21: Fixed set_layers so mode fills properly in depth. Earlier

versions were accessing incorrect array locations at last defined depth.
10. 2003-04-21: Fixed boundary conditions.
11. 2003-05-02: Extended the model area by half the grid spacing on the RHS.
This makes the model area symmetric allowing a plane wave source to be
introduced into the model (stype=pw). The w, txx and tzz grids contain now
one more column than the u and txz grids.
12. 2003-05-02: Added routines to allow plane waves to be introduced at a

specified angle (sang) into the model with functions add_pw_source_V and
add_pw_source_S.
13. 15 Oct 2005 -- tossed all the model building stuff. Read models
from binary files made by unif2aniso
Alogrithm based on Juhlin (1995, Geophys. Prosp.)

and Levander (1988, Geophysics)
Attenuation included as in Graves (1996, BSSA)
174
175
char* sdoc[] = {
SUEDIT - examine segy diskfiles and edit headers
suedit diskfile (open for possible header modification if writable)

suedit <diskfile (open read only)
The following commands are recognized:

number read in that trace and print nonzero header words
<CR> go to trace one step away (step is initially -1)
+ read in next trace (step is set to +1)
- read in previous trace (step is set to -1)
dN advance N traces (step is set to N)
% print some percentiles of the trace data
r print some ranks (rank[j] = jth smallest datum)
p [n1 [n2]] tab plot sample n1 to n2 on current trace
g [tr1 tr2] ximage plot the trace [traces tr1 to tr2]
w [tr1 tr2] xwigb plot the trace [traces tr1 to tr2]
f [tr1 tr2] ximage plot the amplitude spectra of the trace
u [tr1 tr2] apply user pipeline to specified traces
! key=val change a value in a field (e.g. ! tracr=101)
? print help file
q quit
NOTE: sample numbers are 1-based (first sample is 1).
’u 1000000 1000100 suwind >subset.su’ will quickly extract a few

traces from the middle of a large dataset
Credits:
SEP: Einar Kjartansson, Shuki Ronen, Stew Levin
CWP: Jack K. Cohen
Unocal: Reg Beardsley
Trace header fields modified: ntr (only for internal plotting)
176
SUEIPOFI - EIgenimage (SVD) based POlarization FIlter for
three-component data
sueipofi <stdin >stdout [optional parameters]
none
dt=(from header) time sampling intervall in seconds
wl=0.1 SVD time window length in seconds
pwr=1.0 exponent of filter weights
interp=cubic interpolation between initially calculated
weights, choose "cubic" or "linear
verbose=0 1 = echo additional information
file=polar base name for additional output file(s) of

filter weights (see flags below)
rl1=0 1 = rectilinearity along first principal axis
rl2=0 1 = rectilinearity along second principal axis
pln=0 1 = planarity
Notes:
Three adjacent traces are considered as one three-component
dataset.
The filter is the sum of the first two eigenimages of the singular
value decomposition (SVD) of the signal matrix (time window).
Weighting functions depending on linearity and planarity of the
signal are applied, additionally. To avoid edge effects, these are
interpolated linearily or via cubic splines between initially
calculated values of non-overlapping time windows.
The algorithm is based on the assumption that the particle motion
trajectory is essentially 2D (elliptical polarization).
Caveat:
Cubic spline interpolation may result in filter weights exceeding
the set of values of initial weights. Weights outside the valid
interval [0.0, 1.0] are clipped.
177
Author: Nils Maercklin,
E-mail: [email protected]
References:
Franco, R. de, and Musacchio, G., 2000: Polarization Filter with
Singular Value Decomposition, submitted to Geophysics and
published electronically in Geophysics online (www.geo-online.org).
Jurkevics, A., 1988: Polarization analysis of three-comomponent
array data, Bulletin of the Seismological Society of America,
vol. 78, no. 5.
Press, W. H., Teukolsky, S. A., Vetterling, W. T., and Flannery, B. P.
1996: Numerical Recipes in C - The Art of Scientific Computing,
Cambridge University Press, Cambridge.

178
SUFCTANISMOD - Flux-Corrected Transport correction applied to the 2D
elastic wave equation for finite difference modeling in
anisotropic media
sufctanismod > outfile [optional parameters]

outfile is the wavefield snapshot x-component
x-component of wavefield snapshot is in snapshotx.data
y-component of wavefield snapshot is in snapshoty.data
z-component of wavefield snapshot is in snapshotz.data
Optional Output Files:

reflxfile= reflection seismogram file name for x-component
no output produced if no name specified
reflyfile= reflection seismogram file name for y-component
reflzfile= reflection seismogram file name for z-component
vspxfile= VSP seismogram file name for x-component
vspyfile= VSP seismogram file name for y-component
vspzfile= VSP seismogram file name for z-component
suhead=1 To get SU-header output seismograms (else suhead=0)
New parameter:
receiverdepth=0 depth of horizontal receivers (in gridpoints)
dofct=1 1 do the FCT correction
0 do not do the FCT correction
FCT Related parameters:
eta0=0.03 diffusion coefficient
typical values ranging from 0.008 to 0.06
about 0.03 for the second-order method
about 0.012 for the fourth-order method
eta=0.04 anti-diffusion coefficient
typical values ranging from 0.008 to 0.06
about 0.04 for the second-order method
about 0.015 for the fourth-order method
fctxbeg=0 x coordinate to begin applying the FCT correction
179
fctzbeg=0 z coordinate to begin applying the FCT correction
fctxend=nx x coordinate to stop applying the FCT correction
fctzend=nz z coordinate to stop applying the FCT correction
deta0dx=0.0 gradient of eta0 in x-direction d(eta0)/dx

deta0dz=0.0 gradient of eta0 in z-direction d(eta0)/dz
detadx=0.0 gradient of eta in x-direction d(eta)/dx
detadz=0.0 gradient of eta in z-direction d(eta)/dz
General Parameters:
order=2 2 second-order finite-difference
4 fourth-order finite-difference
nt=200 number of time steps

dt=0.004 time step
nx=100 number of grid points in x-direction

nz=100 number of grid points in z-direction
dx=0.02 spatial step in x-direction

dz=0.02 spatial step in z-direction
sx=nx/2 source x-coordinate (in gridpoints)

sy=nz/2 source z-coordinate (in gridpoints)
fpeak=20 peak frequency of the wavelet
wavelet=1 1 AKB wavelet

2 Ricker wavelet
3 impulse
4 unity
isurf=2 1 absorbing surface condition

2 free surface condition
3 zero surface condition
source=1 1 point source

2 sources are located on a given refelector ",
(two horizontal and one dipping reflectors)
3 sources are located on a given dipping refelector ",
sfile= the name of input source file, if no name specified then

use default source location. (source=1 or 2)
180
Density and Elastic Parameters:
dfile= the name of input density file,
if no name specified then
assume a linear density profile with ...
rho00=2.0 density at (0, 0)
drhodx=0.0 density gradient in x-direction d(rho)/dx
drhodz=0.0 density gradient in z-direction d(rho)/dz
afile= name of input elastic param. (c11) aa file, if no name

specified then, assume a linear profile with ...
aa00=2.0 elastic parameter at (0, 0)
daadx=0.0 parameter gradient in x-direction d(aa)/dx
daadz=0.0 parameter gradient in z-direction d(aa)/dz
cfile= name of input elastic param. (c33) cc file, if no name

cc00=2.0 elastic parameter at (0, 0)
dccdx=0.0 parameter gradient in x-direction d(cc)/dx
dccdz=0.0 parameter gradient in z-direction d(cc)/dz
ffile= name of input elastic param. (c13) ff file, if no name

ff00=2.0 elastic parameter at (0, 0)
dffdx=0.0 parameter gradient in x-direction d(ff)/dx
dffdz=0.0 parameter gradient in z-direction d(ff)/dz
lfile= name of input elastic param. (c44) ll file, if no name

ll00=2.0 elastic parameter at (0, 0)
dlldx=0.0 parameter gradient in x-direction d(ll)/dx
dlldz=0.0 parameter gradient in z-direction d(ll)/dz
nfile= name of input elastic param. (c66) nn file, if no name

nn00=2.0 elastic parameter at (0, 0)
dnndx=0.0 parameter gradient in x-direction d(nn)/dx
dnndz=0.0 parameter gradient in z-direction d(nn)/dz
Optimizations:
The moving boundary option permits the user to restrict the computations
of the wavefield to be confined to a specific range of spatial coordinates.
The boundary of this restricted area moves with the wavefield
181
movebc=0 0 do not use moving boundary optimization
1 use moving boundaries
If movebc=1 then specify:
mbx1=0 initial left side of moving boundary
mbz1=0 initial top of moving boundary
mbx2=nx initial right side of moving boundary
mbz2=nz initial bottom of moving boundary
Author: Tong Fei, Center for Wave Phenomena,

Colorado School of Mines, Dec 1993
Some additional features by: Stig-Kyrre Foss, CWP
Colorado School of Mines, Oct 2001
New features (Oct 2001):
- setting receiver depth
- outputfiles with SU-headers
- additional commentary
Notes:
This program performs seismic modeling for elastic anisotropic
media with vertical axis of symmetry.
The finite-difference method with the FCT correction is used.
Stability condition: vmax*dt /(sqrt(2)*min(dx,dz)) < 1
Two major stages are used in the algorithm:

(1) conventional finite-difference wave extrapolation
(2) followed by an FCT correction
References:
The detailed algorithm description is given in the article
"Elimination of dispersion in finite-difference modeling
and migration" in CWP-137, project review, page 155-174.
Original reference to the FCT method:

Boris, J., and Book, D., 1973, Flux-corrected transport. I.
SHASTA, a fluid transport algorithm that works:
Journal of Computational Physics, vol. 11, p. 38-69.
182
SUFDMOD1 - Finite difference modelling (1-D 1rst order) for the
acoustic wave equation "
sufdmod1 <vfile >sfile nz= tmax= sz= [optional parameters]
Required parameters :
<vfile or vfile= binary file containing velocities[nz]
>sfile or sfile= SU file containing seimogram[nt]
nz= number of z samples
tmax= maximum propagation time
sz= z coordinate of source
Optional parameters :
fz=0.0 first depth sample
rz=1 coordinate of receiver
sz=1 coordinate of source
dfile= binary input file containing density[nz]
wfile= output file for wave field (snapshots in a SU trace panel)
abs=0,1 absorbing conditions on top and bottom
styp=0 source type (0: gauss, 1: ricker 1, 2: ricker 2)
freq=15.0 approximate source center frequency (Hz)
nt=1+tmax/dt number od time samples (dt determined for numerical
stability)
zt=1 trace undersampling factor for trace and snapshots
zd=1 depth undersampling factor for snapshots
press=1 to record the pressure field; 0 records the particle
velocity
verbose=0 =1 for diagnostic messages
Notes :
This program uses a first order explicit velocity/pressure finite
difference equation.
The source function is applied on the pressure component.
If no density file is given, constant density is assumed
Wavefield can be easily viewed with suximage, user must provide f2=0
to the ximage program in order to get correct time labelling
Seismic trace is shifted in order to get a zero phase source
Source begins and stop when it’s amplitude is 10^-4 its maximum
Time and depth undersampling only modify the output trace and snapshots.
These parameters are useful for keeping snapshot file small and
the number of samples under SU_NFLTS.
183
NULL };
float source (float t, int styp, float dt, float dz, float t0, float alpha);
int main (int argc, char **argv)

{
float *rv; /* array of rock velocity from cfile
float *rd; /* array of rock density from dfile on p knots
float *rd1_5; /* array of rock density from dfile on v knots
float *p; /* pressure
float *v; /* particle velocity
float tmax, dt, t0; /* maximum time , time step, */
/* time delay for near causal source */
float vmax; /* maximum rock velocity */
int verbose; /* is verbose? */
int nz, nt; /* number of z samples, time samples */
float fz, dz; /* first sample depth spatial depth */
float sz; /* source coordinate */
int abs[2]; /* array of absorbing conditions */
int isz; /* source location index */
float rz; /* receiver depth
int irz; /* zcoordinate (in samples) of the source
int iz, it, itsis; /* counter
int ies; /* end of source index
int press; /* to choose between pressure or particle
/* velocity
float t; /* time
int td, zd; /* time and depth decimation
segy snapsh, sismo; /* recording of the seismic field,
/* seismogram
char *dfile=""; /* density file name
char *wfile=""; /* seismogram file name
char *velfile=""; /* velocity file name
char *sfile=""; /* velocity file name
float freq=0.0; /* source center freq
float alpha=0.0; /* source exp */
float epst0=0.0; /* source first amp ratio
int styp; /* source type
FILE *seisfp=stdout; /* pointer to seismic trace output file

FILE *wavefp=NULL; /* pointer to wave field output file
FILE *velocityfp=stdin; /* pointer to input velocity file
FILE *densityfp=NULL; /* pointer to input density file
184
/* hook up getpar to handle the parameters
initargs (argc, argv);
requestdoc(0);
/* verbose
if (!getparint ("verbose",&verbose)) verbose=0;
/* get required parameters

if (!getparint ("nz",&nz)) err("must specify nz ! ");
if (verbose) warn("nz= %d", nz);
if (!getparfloat ("tmax",&tmax)) err("must specify tmax ! ");

if (verbose) warn("tmax= %f", tmax);
if (!getparfloat("sz", &sz)) err ("must specify sz ! ");

if (verbose) warn("sz= %f", sz);
/* get optional parameters

if (!getparint ("nt", &nt)) nt=0;
if (verbose) warn("nt= %d", nt);
if (!getparint ("styp", &styp)) styp=0;
if (verbose) warn("styp= %d ", styp);
if (!getparfloat ("dz", &dz)) dz=1;
if (!getparfloat ("fz", &fz)) fz=0.0;
/* source coordinates to samples

isz=NINT((sz-fz)/dz);
if (verbose) warn( "source on knot number %d ", isz);
if (!getparfloat("rz", &rz)) rz=0.0;

irz = NINT ((rz-fz)/dz);
if (verbose) warn("receiver depth : %f on knot # %d ", rz, irz);
if (!getparfloat("freq", &freq)) freq=15.0;

if (verbose) warn("frequency : %f Hz",freq);
getparstring ("velfile", &velfile);

if (verbose) {
185
if (*velfile != ’\0’ ) warn("Velocity file : %s ",velfile);
else warn("Velocity file supplied via stdin");
}
getparstring ("sfile", &sfile);

if (verbose) {
if (*sfile != ’\0’ ) warn("Output trace file : %s ",sfile);
else warn("Output trace via stdout");
}
getparstring ("dfile", &dfile);

if (verbose) {
if (*dfile != ’\0’ ) warn("Density file : %s ",dfile);
else warn("No density file supplied ");
}
getparstring ("wfile", &wfile);

if (verbose) {
if (*wfile != ’\0’ ) warn("Wave file : %s ",wfile);
else warn("No wave file requested ");
}
if ( NINT((float) nz/((float) zd)) + 1 >SU_NFLTS) {

warn ("Too many depth points : impossible to output wave field. Increase zd ?");
*wfile=’\0’;
}
/* get absorbing conditions

if (!getparint("abs",abs)) { abs[0]=0; abs[1]=1; }
if (verbose) {
if (abs[0]==1) warn("absorbing condition on top ");
if (abs[1]==1) warn("absorbing condition on bottom ");
}
/* get decimation coefficients
if (!getparint("td",&td)) td=1 ;
if (verbose) warn("time decimation ccoefficent: %d ",td);
if (!getparint("zd",&zd)) zd=1 ;
if (verbose) warn("depth decimation ccoefficent: %d ",zd);
/* choose pressure or particle velocity

if (!getparint("press", &press)) press=1 ;
if ((press != 0) && (press != 1)) err ("press must equal 0 or 1");
186
if (verbose) {
if (press==1) warn( "program will output pressure values");
else if (press==0) warn( "program will output particle velocity values");
}
/* allocate space
p=alloc1float(nz);
v=alloc1float(nz);
rv=alloc1float(nz);
rd=alloc1float(nz);
rd1_5=alloc1float(nz);
/* read velocity file

if (*velfile != ’\0’ ) {
if ((velocityfp=fopen(velfile,"r"))==’\0’) err("cannot open velfile=%s ",velfile);
}
if (efread (rv, sizeof(float), nz, velocityfp)!=nz)
err("cannot read %d velocity values ", nz);
/* read density file and linearly inderpolate on corrrect location

if (*dfile != ’\0’) {
if ((densityfp=fopen(dfile,"r"))==’\0’) err("cannot open dfile=%s ",dfile);
if (fread(rd,sizeof(float), nz, densityfp)!=nz) err("error reading dfile %s",dfile);
fclose(densityfp);
}
else for (iz=0; iz<nz; iz++) rd[iz]=2500;
for (iz=0; iz<nz-1; iz++) rd1_5[iz]=(rd[iz]+rd[iz+1])/2;
rd1_5[nz-1]=rd[nz-1];
/* time step computation

vmax=0;
for (iz=0; iz<nz; iz++) if (rv[iz]>vmax) vmax=rv[iz];if (verbose) warn( "vmax= %f ",
dt=dz/1.414/vmax/2; if (verbose) warn( "time step dt= %f ", dt);
/* maximum number of iterations

if (nt==0) nt=1+tmax/dt;
if (verbose) warn( "number of time steps nt= %d ", nt);
if (NINT( (float) nt/((float)td))+1>SU_NFLTS) err("too many time steps. Increase td
/* source parameter computation

alpha=2*9.8696*freq*freq;
187
/* time shift to get a t0 centered source
if ((styp==0) || (styp == 2)) epst0=fabs(source (0, styp, dt, dz, 0, alpha) / 1e4);
else if (styp==1) epst0=fabs(source (1/sqrt(2*alpha), styp, dt, dz, 0, alpha)) / 1e4
if (verbose) warn( "epst0 = %f ", epst0);
t=tmax+dt;
do t=t-dt; while (fabs(source(t, styp, dt, dz, 0, alpha))<epst0);
t0=t;
ies=2*t/dt;
if (verbose) warn("time shift t0 = %f s", t0);
array initialization
for (iz=0; iz<nz; iz++) { v[iz]=0; p[iz]=0; }
if (*wfile != ’\0’) {
wavefp=fopen (wfile,"w");
snapsh.d1=dz*zd; snapsh.f1=fz ; snapsh.ns=nz/zd+1; snapsh.d2=dt*td; snapsh.f2=0;
/* snapsh.f2=0 is useless since 0 is the "no value" code for SU headers
}
/* propagation computation
itsis=0;
for (it=0; it<=nt; it++) {
t=it*dt;
if (abs[0]==1) p[0]=(p[0]*(1-rv[0]*dt/dz)+2*rd[0]*rv[0]*rv[0]*dt/dz*v[0])/(1+rv[0]*d
else p[0]=0;
for (iz=1; iz<nz; iz++) p[iz]=p[iz]+rd[iz]*rv[iz]*rv[iz]*dt/dz*(v[iz]-v[iz-1]);
if (abs[1]!=1) p[nz-1]=0;
if (it<ies) {
p[isz]=p[isz]+source(t, styp, dt, dz, t0, alpha);
}
for (iz=0; iz<nz-1; iz++) v[iz]=v[iz]+dt/rd1_5[iz]/dz*(p[iz+1]-p[iz]);
if (abs[1] != 1) v[nz-1]=0;
else
v[nz-1]=((rd1_5[nz-1]*dz-dt*rd[nz-1]*rv[nz-1])*v[nz-1]-2*dt*p[nz-1])/(rd1_5[nz-1]*dz
if (it % td == 0) {
if (press==1)
sismo.data[itsis]=p[irz];
else
188
sismo.data[itsis]=v[irz];
itsis++;
}
if ((*wfile!=’\0’) && (it % td == 0)) {

if (press==1)
for (iz=0; iz<nz/zd; ++iz) snapsh.data[iz]=p[iz*zd];
else
for (iz=0; iz<nz/zd; ++iz) snapsh.data[iz]=v[iz*zd];
fputtr(wavefp, &snapsh);
}
if (*wfile!=’\0’) fclose (wavefp);
sismo.dt=td*dt*1e6; sismo.ns=itsis; sismo.delrt=-t0*1000; sismo.trid=1; sismo.tracl=
if (*sfile != ’\0’) seisfp=efopen(sfile,"w");

fputtr (seisfp, &sismo);
return(CWP_Exit());
}
float source (float t, int styp, float dt, float dz, float t0, float alpha)
{
float x=0.0, sou=0.0;
x=-alpha*(t-t0)*(t-t0);
if (x>-40) {
if (styp==0) sou=exp(x);
else if (styp==1) sou=-2*alpha*(t-t0)*exp(x);
else if (styp==2) sou=2*alpha*(1+2*x)*exp(x);
}
else sou=0;
return sou/dz*dt*1e8;
}
189
SUFDMOD2_PML - Finite-Difference MODeling (2nd order) for acoustic wave
equation with PML absorbing boundary conditions.
Caveat: experimental PML absorbing boundary condition version,
may be buggy!
sufdmod2_pml <vfile >wfile nx= nz= tmax= xs= zs= [optional parameters]
<vfile file containing velocity[nx][nz]
>wfile file containing waves[nx][nz] for time steps
xs= x coordinates of source
zs= z coordinates of source
tmax= maximum time
nt=1+tmax/dt number of time samples (dt determined for stability)
mt=1 number of time steps (dt) per output time step

fmax = vmin/(10.0*h) maximum frequency in source wavelet

fpeak=0.5*fmax peak frequency in ricker wavelet
dfile= input file containing density[nx][nz]

vsfile= output file for vertical line of seismograms[nz][nt]
hsfile= output file for horizontal line of seismograms[nx][nt]
ssfile= output file for source point seismograms[nt]
verbose=0 =1 for diagnostic messages, =2 for more
abs=1,1,1,1 Absorbing boundary conditions on top,left,bottom,right

sides of the model.
=0,1,1,1 for free surface condition on the top
...PML parameters....
pml_max=1000.0 PML absorption parameter
pml_thick=0 half-thickness of pml layer (0 = do not use PML)
190
Notes:
This program uses the traditional explicit second order differencing
method.
Two different absorbing boundary condition schemes are available. The

first is a traditional absorbing boundary condition scheme created by
Hale, 1990. The second is based on the perfectly matched layer (PML)
method of Berenger, 1995.
Authors: CWP:Dave Hale

CWP:modified for SU by John Stockwell, 1993.
CWP:added frequency specification of wavelet: Craig Artley, 1993
TAMU:added PML absorbing boundary condition:
Michael Holzrichter, 1998
CWP/WesternGeco:corrected PML code to handle density variations:
Greg Wimpey, 2006
References: (Hale’s absobing boundary conditions)

Clayton, R. W., and Engquist, B., 1977, Absorbing boundary conditions
for acoustic and elastic wave equations, Bull. Seism. Soc. Am., 6,
1529-1540.
Clayton, R. W., and Engquist, B., 1980, Absorbing boundary conditions

for wave equation migration, Geophysics, 45, 895-904.
Hale, D., 1990, Adaptive absorbing boundaries for finite-difference

modeling of the wave equation migration, unpublished report from the
Center for Wave Phenomena, Colorado School of Mines.
Richtmyer, R. D., and Morton, K. W., 1967, Difference methods for

initial-value problems, John Wiley & Sons, Inc, New York.
Thomee, V., 1962, A stable difference scheme for the mixed boundary problem
for a hyperbolic, first-order system in two dimensions, J. Soc. Indust.
Appl. Math., 10, 229-245.
Toldi, J. L., and Hale, D., 1982, Data-dependent absorbing side boundaries,
Stanford Exploration Project Report SEP-30, 111-121.
References: (PML boundary conditions)
191
Jean-Pierre Berenger, ‘‘A Perfectly Matched Layer for the Absorption of
Electromagnetic Waves,’’ Journal of Computational Physics, vol. 114,
pp. 185-200.
Hastings, Schneider, and Broschat, ‘‘Application of the perfectly

matched layer (PML) absorbing boundary condition to elastic wave
propogation,’’ Journal of the Accoustical Society of America,
November, 1996.
Allen Taflove, ‘‘Electromagnetic Modeling: Finite Difference Time

Domain Methods’’, Baltimore, Maryland: Johns Hopkins University Press,
1995, chap. 7, pp. 181-195.
Trace header fields set: ns, delrt, tracl, tracr, offset, d1, d2,
sdepth, trid
192
SUFDMOD2 - Finite-Difference MODeling (2nd order) for acoustic wave equation
sufdmod2 <vfile >wfile nx= nz= tmax= xs= zs= [optional parameters]
<vfile file containing velocity[nx][nz]
>wfile file containing waves[nx][nz] for time steps
xs= x coordinates of source, or, alternatively, the name
of a file that contains the x- and z-coordinates,
with the number of pairs as the first record and
the actual pairs of (x,z) locations following.
zs= z coordinates of source
sstrength=1.0 strength of source
tmax= maximum time
nt=1+tmax/dt number of time samples (dt determined for stability)
mt=1 number of time steps (dt) per output time step

fmax = vmin/(10.0*h) maximum frequency in source wavelet

fpeak=0.5*fmax peak frequency in ricker wavelet
dfile= input file containing density[nx][nz]

vsfile= output file for vertical line of seismograms[nz][nt]
hsfile= output file for horizontal line of seismograms[nx][nt]
ssfile= output file for source point seismograms[nt]
verbose=0 =1 for diagnostic messages, =2 for more
abs=1,1,1,1 Absorbing boundary conditions on top,left,bottom,right

sides of the model.
=0,1,1,1 for free surface condition on the top
Notes:
193
This program uses the traditional explicit second order differencing
method.
The input velocity and density files contain C-style binary floating
numbers with the z-direction being the fast dimension in the file.
Here v[ix][iz] means that this is a v(z,x) velocity function.
Authors: CWP:Dave Hale

CWP:modified for SU by John Stockwell, 1993.
Trace header fields set: sx, gx, ns, delrt, tracl, tracr, offset, d1, d2,
sdepth, trid
Modifications: Tony Kocurko (TK:)

Memorial University in Newfoundland and Labrador
- Allow user to supply the name of a file containing
shot point locations, rather than supplying them
as values to the xs= and zs= command line arguments.
- Correct the calculation of izs[is].
194
SUFFT - fft real time traces to complex frequency traces
suftt <stdin >sdout sign=1
none
sign=1 sign in exponent of fft
dt=from header sampling interval
verbose=1 =0 to stop advisory messages
Notes: To facilitate further processing, the sampling interval

in frequency and first frequency (0) are set in the
output header.
sufft | suifft is not quite a no-op since the trace

length will usually be longer due to fft padding.
Caveats:
No check is made that the data IS real time traces!
Output is type complex. To view amplitude, phase or real, imaginary

parts, use suamp
Examples:
sufft < stdin | suamp mode=amp | ....
sufft < stdin | suamp mode=phase | ....
sufft < stdin | suamp mode=real | ....
sufft < stdin | suamp mode=imag | ....
Credits:
CWP: Shuki Ronen, Chris Liner, Jack K. Cohen

CENPET: Werner M. Heigl - added well log support
Note: leave dt set for later inversion
Trace header fields accessed: ns, dt, d1, f1

Trace header fields modified: ns, d1, f1, trid
195
SUFILTER - applies a zero-phase, sine-squared tapered filter
sufilter <stdin >stdout [optional parameters]
f=f1,f2,... array of filter frequencies(HZ)
amps=a1,a2,... array of filter amplitudes
verbose=0 =1 for advisory messages
Defaults:f=.10*(nyquist),.15*(nyquist),.45*(nyquist),.50*(nyquist)
(nyquist calculated internally)
amps=0.,1.,...,1.,0. trapezoid-like bandpass filter
Examples of filters:
Bandpass: sufilter <data f=10,20,40,50 | ...
Bandreject: sufilter <data f=10,20,30,40 amps=1.,0.,0.,1. | ..
Lowpass: sufilter <data f=10,20,40,50 amps=1.,1.,0.,0. | ...
Highpass: sufilter <data f=10,20,40,50 amps=0.,0.,1.,1. | ...
Notch: sufilter <data f=10,12.5,35,50,60 amps=1.,.5,0.,.5,1. |..
Credits:
CWP: John Stockwell, Jack Cohen
Possible optimization: Do assignments instead of crmuls where

filter is 0.0.
196
SUFLIP - flip a data set in various ways
suflip <data1 >data2 flip=1 verbose=0
none
flip=1 rotational sense of flip
+1 = flip 90 deg clockwise
-1 = flip 90 deg counter-clockwise
0 = transpose data
2 = flip right-to-left
3 = flip top-to-bottom
prefix for storing temporary files; else if
verbose=0 verbose = 1 echoes flip info
NOTE: tr.dt header field is lost if flip=-1,+1. It can be

reset using sushw.
EXAMPLE PROCESSING SEQUENCES:

1. suflip flip=-1 <data1 | sushw key=dt a=4000 >data2
2. suflip flip=2 <data1 | suflip flip=2 >data1_again
3. suflip tmpdir=/scratch <data1 | ...
Caveat: may fail on large files.
Credits:
CWP: Chris Liner, Jack K. Cohen, John Stockwell
Caveat:
right-left flip (flip = 2) and top-bottom flip (flip = 3)
don’t require the matrix approach. We sacrificed efficiency
for uniform coding.

Trace header fields modified: ns, dt, tracl
197
198
SUFNZERO - get Time of First Non-ZERO sample by trace
sufnzero <stdin >stdout [optional parameters]
none
mode=first Output time of first non-zero sample
=last for last non-zero sample
=both for both first & last non-zero samples
min=0 Threshold value for considering as zero

Any abs(sample)<min is considered to be zero
key=key1,... Keyword(s) to print
Credits:
Geocon : Garry Perratt
based on surms by the same, itself based on sugain & sumax by:
CWP : John Stockwell
199
SUFRAC -- take general (fractional) time derivative or integral of
data, plus a phase shift. Input is CAUSAL time-indexed
or depth-indexed data.
sufrac power= [optional parameters] <indata >outdata
power=0 exponent of (-i*omega)
=0 ==> phase shift only
>0 ==> differentiation
<0 ==> integration
sign=-1 sign in front of i * omega

dt=(from header) time sample interval (in seconds)
phasefac=0 phase shift by phase=phasefac*PI
Examples:
preprocess to correct 3D data for 2.5D migration
sufrac < sudata power=.5 sign=1 | ...
preprocess to correct susynlv, susynvxz, etc. (2D data) for 2D migration
sufrac < sudata phasefac=.25 | ...
The filter is applied in frequency domain.
Algorithm:
g(t) = Re[INVFTT{ ( (sign) iw)^power FFT(f)}]
Caveat:
Large amplitude errors will result if the data set has too few points.
Good numerical integration routine needed!

For example, see Gnu Scientific Library.
Credits:
CWP: Chris Liner, Jack K. Cohen, Dave Hale (pfas)
CWP: Zhenyue Liu and John Stockwell added phase shift option
Trace header fields accessed: ns, dt, trid, d1

/
200
SUFXDECON - random noise attenuation by FX-DECONvolution
sufxdecon <stdin >stdout [...]
taper=.1 length of taper
fmin=6. minimum frequency to process in Hz (accord to twlen)
fmax=.6/(2*dt) maximum frequency to process in Hz
twlen=entire trace time window length (minimum .3 for lower freqs)
ntrw=10 number of traces in window
ntrf=4 number of traces for filter (smaller than ntrw)
for storing temporary files; else, if the CWP_TMPDIR
environment variable is set, use its value for the path;
else use tmpfile()
Notes: Each trace is transformed to the frequency domain.

For each frequency, Wiener filtering, with unity prediction in
space, is used to predict the next sample.
At the end of the process, data is mapped back to t-x domain. ",
Credits:
CWP: Carlos E. Theodoro (10/07/97)
References:
Canales(1984):’Random noise reduction’ 54th. SEGM
Gulunay(1986):’FXDECON and complex Wiener Predicition
filter’ 56th. SEGM
Galbraith(1991):’Random noise attenuation by F-X
prediction: a tutorial’ 61th. SEGM
Algorithm:
- read data
- loop over time windows
- select data
- FFT (t -> f)
- loop over space windows
201
- select data
- loop over frequencies
- autocorelation
- matrix problem
- construct filter
- filter data
- loop along space window
- FFT (f -> t)
- reconstruct data
- output data

Trace header fields modified:
202
SUGABOR - Outputs a time-frequency representation of seismic data via
the Gabor transform-like multifilter analysis technique
presented by Dziewonski, Bloch and Landisman, 1969.
sugabor <stdin >stdout [optional parameters]
dt=(from header) time sampling interval (sec)
fmin=0 minimum frequency of filter array (hz)
fmax=NYQUIST maximum frequency of filter array (hz)
beta=3.0 ln[filter peak amp/filter endpoint amp]
band=.05*NYQUIST filter bandwidth (hz)
alpha=beta/band^2 filter width parameter
verbose=0 =1 supply additional info
holder=0 =1 output Holder regularity estimate
=2 output linear regularity estimate
Notes: This program produces a muiltifilter (as opposed to moving window)

representation of the instantaneous amplitude of seismic data in the
time-frequency domain. (With Gaussian filters, moving window and multi-
filter analysis can be shown to be equivalent.)
An input trace is passed through a collection of Gaussian filters

to produce a collection of traces, each representing a discrete frequency
range in the input data. For each of these narrow bandwidth traces, a
quadrature trace is computed via the Hilbert transform. Treating the narrow
bandwidth trace and its quadrature trace as the real and imaginary parts
of a "complex" trace permits the "instantaneous" amplitude of each
narrow bandwidth trace to be compute. The output is thus a representation
of instantaneous amplitude as a function of time and frequency.
Some experimentation with the "band" parameter may necessary to produce

the desired time-frequency resolution. A good rule of thumb is to run
sugabor with the default value for band and view the image. If band is
too big, then the t-f plot will consist of stripes parallel to the frequency
axis. Conversely, if band is too small, then the stripes will be parallel
to the time axis.
Caveat:
The Gabor transform is not a wavelet transform, but rather are sharp
203
frame basis. However, it is nearly a Morlet continuous wavelet transform
so the concept of Holder regularity may have some meaning. If you are
computing Holder regularity of, say, a migrated seismic section, then
set band to 1/3 of the frequency band of your data.
Examples:
suvibro | sugabor | suximage
suvibro | sugabor | suxmovie n1= n2= n3=
(because suxmovie scales it’s amplitudes off of the first panel,
may have to experiment with the wclip and bclip parameters
suvibro | sugabor | supsimage | ... ( your local PostScript utility)
Credits:
CWP: John Stockwell, Oct 1994

CWP: John Stockwell Oct 2004, added holder=1 option
Algorithm:
This programs takes an input seismic trace and passes it

through a collection of truncated Gaussian filters in the frequency
domain.
The bandwidth of each filter is given by the parameter "band". The

decay of these filters is given by "alpha", and the number of filters
is given by nfilt = (fmax - fmin)/band. The result, upon inverse
Fourier transforming, is that nfilt traces are created, with each
trace representing a different frequency band in the original data.
For each of the resulting bandlimited traces, a quadrature (i.e. pi/2

phase shifted) trace is computed via the Hilbert transform. The
bandlimited trace constitutes a "complex trace", with the bandlimited
trace being the "real part" and the quadrature trace being the
"imaginary part". The instantaneous amplitude of each bandlimited
trace is then computed by computing the modulus of each complex trace.
(See Taner, Koehler, and Sheriff, 1979, for a discussion of complex
trace analysis.
The final output for a given input trace is a map of instantaneous

amplitude as a function of time and frequency.
This is not a wavelet transform, but rather a redundant frame

representation.
204
References: Dziewonski, Bloch, and Landisman, 1969, A technique
for the analysis of transient seismic signals,
Bull. Seism. Soc. Am., 1969, vol. 59, no.1, pp.427-444.
Taner, M., T., Koehler, F., and Sheriff, R., E., 1979,
Complex seismic trace analysis, Geophysics, vol. 44,
pp.1041-1063.
Chui, C., K.,1992, Introduction to Wavelets, Academic

Press, New York.
Trace header fields accessed: ns, dt, trid, ntr

Trace header fields modified: tracl, tracr, d1, f2, d2, trid, ntr
205
SUGAIN - apply various types of gain to display traces
sugain <stdin >stdout [optional parameters]
none (no-op)
panel=0 =1 gain whole data set (vs. trace by trace)
tpow=0.0 multiply data by t^tpow
epow=0.0 multiply data by exp(epow*t)
gpow=1.0 take signed gpowth power of scaled data
agc=0 flag; 1 = do automatic gain control
gagc=0 flag; 1 = ... with gaussian taper
wagc=0.5 agc window in seconds (use if agc=1 or gagc=1)
trap=none zero any value whose magnitude exceeds trapval
clip=none clip any value whose magnitude exceeds clipval
pclip=none clip any value greater than clipval
nclip=none clip any value less than clipval
qclip=1.0 clip by quantile on absolute values on trace
qbal=0 flag; 1 = balance traces by qclip and scale
pbal=0 flag; 1 = bal traces by dividing by rms value
mbal=0 flag; 1 = bal traces by subtracting the mean
maxbal=0 flag; 1 = balance traces by subtracting the max
scale=1.0 multiply data by overall scale factor
norm=0.0 divide data by overall scale factor
bias=0.0 bias data by adding an overall bias value
jon=0 flag; 1 means tpow=2, gpow=.5, qclip=.95
verbose=0 verbose = 1 echoes info
mark=0 apply gain only to traces with tr.mark=0
=1 apply gain only to traces with tr.mark!=0

Operation order:
if (norm) scale/norm
out(t) = scale * BAL{CLIP[AGC{[t^tpow * exp(epow * t) * ( in(t)-bias )]^gpow}]}
206
Notes:
The jon flag selects the parameter choices discussed in
Claerbout’s Imaging the Earth, pp 233-236.
Extremely large/small values may be lost during agc. Windowing

these off and applying a scale in a preliminary pass through
sugain may help.
Sugain only applies gain to traces with tr.mark=0. Use sushw,

suchw, suedit, or suxedit to mark traces you do not want gained.
See the selfdocs of sushw, suchw, suedit, and suxedit for more
information about setting header fields. Use "sukeyword mark
for more information about the mark header field.
Credits:
SEP: Jon Claerbout
CWP: Jack K. Cohen, Brian Sumner, Dave Hale
Note: Have assumed tr.deltr >= 0 in tpow routine.
Jon’s second book, pages 233-236.
Trace header fields accessed: ns, dt, delrt, mark
207
SUGAUSSTAPER - Multiply traces with gaussian taper
sugausstaper < stdin > stdout [optional parameters]
<none>
key=offset keyword of header field to weight traces by
x0=300 centre of gaussian window
xw=50 width of gaussian window
Traces are multiplied with a symmetrical gaussian taper

w(t)=exp(-((key-x0)/xw)**2)
Credits:
Thomas Bohlen, 04.01.2002
208
SUGAZMIGQ - SU version of Jeno GAZDAG’s phase-shift migration
for zero-offset data, with attenuation Q.
sugazmig <infile >outfile vfile= [optional parameters]
dt=from header(dt) or .004 time sampling interval
dx=from header(d2) or 1.0 midpoint sampling interval
ft=0.0 first time sample
ntau=nt(from data) number of migrated time samples
dtau=dt(from header) migrated time sampling interval
ftau=ft first migrated time sample
vfile= name of file containing velocities
Q=1e6 quality factor
ceil=1e6 gain ceiling beyond which migration ceases

Note: ray bending effects not accounted for in this version.
The tmig and vmig arrays specify an interval velocity function of time.

containing one velocity for every time sample in the data that is to be
migrated. If vfile is specified, then the tmig and vmig arrays are ignored.
Caveat: Adding Q is a first attempt to address GPR issues.
Credits:
Constant Q attenuation correction by Chuck Oden 5 May 2004
CWP John Stockwell 12 Oct 1992
Based on a constant v version by Dave Hale.
209
Trace header fields accessed: ns, dt, delrt, d2
Trace header fields modified: ns, dt, delrt
210
SUGETHW - sugethw writes the values of the selected key words
sugethw key=key1,... [output=] <infile [>outfile]
key=key1,... At least one key word.
output=ascii output written as ascii for display
=binary for output as binary floats
=geom ascii output for geometry setting
verbose=0 quiet
=1 chatty
Output is written in the order of the keys on the command

line for each trace in the data set.
Example:
sugethw <stdin key=sx,gx
writes sx, gx values as ascii trace by trace to the terminal.
Comment:
Users wishing to edit one or more header field (as in geometry setting)
may do this via the following sequence:
sugethw < sudata output=geom key=key1,key2,... > hdrfile
Now edit the ASCII file hdrfile with any editor, setting the fields
appropriately. Convert hdrfile to a binary format via:
a2b < hdrfile n1=nfields > binary_file
Then set the header fields via:
sushw < sudata infile=binary_file key=key1,key2,... > sudata.edited
Credits:
SEP: Shuki Ronen

CWP: Jack K. Cohen
CWP: John Stockwell, added geom stuff, and getparstringarray
211
SUGET - Connect SU program to file descriptor for input stream.
suget fd=$1 | next_su_module
This program is for interfacing " outside processing systems

with SU. Typically, an outside system would execute the su command file
and a file descriptor would be passed by an outside system to
the su command file so that output data from the outside system
could be piped into the su programs executing inside the command file.
Example: suget fd=$1 | next_su_module
fd=-1 file_descriptor_for_input_stream
verbose=0 minimal listing
=1 asks for message with each trace processed.
Author: John Anderson (visiting scholar from Mobil) July 1994
212
SUGOUPILLAUDPO - calculate Primaries-Only impulse response of a lossless
GOUPILLAUD medium for plane waves at normal incidence
sugoupillaudpo < stdin > stdout [optional parameters]
none
l=1 source layer number; 1 <= l <= tr.ns
Source is located at the top of layer l.
k=1 receiver layer number; 1 <= k
Receiver is located at the top of layer k.
tmax number of output time-samples;
default: long enough to capture all primaries
pV=1 flag for vector field seismogram
(displacement, velocity, acceleration);
=-1 for pressure seismogram.
verbose=0 silent operation, =1 list warnings
Input: Reflection coefficient series:
impedance[i]-impedance[i+1]
r[i] = -----------------------------
impedance[i]+impedance[i+1]
r[0]= surface refl. coef. (as seen from above)

r[n]= refl. coef. of the deepest interface
Input file is to be in SU format, i.e., binary floats with a SU header.
Remarks:
1. For vector fields, a buried source produces a spike of amplitude 1
propagating downwards and a spike of amplitude -1 propagating upwards.
A buried pressure source produces spikes of amplitude 1 both in the up-
and downward directions.
A surface source induces only a downgoing spike of amplitude 1 at the
top of the first layer (both for vector and pressure fields).
2. The sampling interval dt in the header of the input reflectivity file
is interpreted as a two-way traveltime thicknes of the layers. The sampling
interval of the output seismogram is the same as that of the input file.
213
Credits:
CWP: Albena Mateeva, April 2001.
214
SUGOUPILLAUD - calculate 1D impulse response of
non-absorbing Goupillaud medium
sugoupillaud < stdin > stdout [optional parameters]
none
l=1 source layer number; 1 <= l <= tr.ns
Source is located at the top of layer l.
k=1 receiver layer number; 1 <= k
Receiver is located at the top of layer k.
tmax number of output time-samples; default:
tmax=NINT((2*tr.ns-(l-1)-(k-1))/2) if k < tr.ns
tmax=k if k >=tr.ns
pV=1 flag for vector field seismogram
(displacement, velocity, acceleration);
=-1 for pressure seismogram.
verbose=0 silent operation, =1 list warnings
Input: Reflection coefficient series:
impedance[i]-impedance[i+1]
r[i] = -----------------------------
impedance[i]+impedance[i+1]
r[0]= surface refl. coef. (as seen from above)

r[n]= refl. coef. of the deepest interface
Input file is to be in SU format, i.e., binary floats with a SU header.
Remarks:
1. For vector fields, a buried source produces a spike of amplitude 1
propagating downwards and a spike of amplitude -1 propagating upwards.
A buried pressure source produces spikes of amplitude 1 both in the up-
and downward directions.
A surface source induces only a downgoing spike of amplitude 1 at the
top of the first layer (both for vector and pressure fields).
2. The sampling interval dt in the header of the input reflectivity file
is interpreted as a two-way traveltime thicknes of the layers. The sampling
interval of the output seismogram is the same as that of the input file.
215
Credits:
CWP: Albena Mateeva, May 2000, a summer project at Western Geophysical
ANOTATION used in the code comments [arises from the use of z-transforms]:
Z-sampled: sampling interval equal to the TWO-way
traveltime of the layers;
z-sampled: sampling interval equal to the ONE-way
traveltime of the layers;
REFERENCES:
1. Ganley, D. C., 1981, A method for calculating synthetic seismograms

which include the effects of absorption and dispersion.
Geophysics, Vol.46, No. 8, p. 1100-1107.
The burial of the source is based on the Appendix of that article.
2. Robinson, E. A., Multichannel Time Series Analysis with Digital

Computer Programs: 1983 Goose Pond Press, 2nd edition.
The recursive polynomials Q, P used in this code are described

in Chapter 3 of the book: Wave Propagation in Layered Media.
My polynomial multiplication and division functions "prod" and

"pratio" are based on Robinson’s Fortran subroutines in Chapter 1.
4. Clearbout, J. F., Fundamentals of Geophysical Data Processing with

Applications to Petroleum Prospecting: 1985 Blackwell Scientific
Publications.
Chapter 8, Section 3: Introduces recursive polynomials F, G in a

more intuitive way than Robinson.
The connection between the Robinson’s P_k, Q_k and Clearbout’s

F_k, G_k is:
P_k(Z) = F_k(Z)
Q_k(Z) = - Z^(k) G_k(1/Z)
216
SUHARLAN - signal-noise separation by the invertible linear
transformation method of Harlan, 1984
susnlinsep <infile >outfile [optional parameters]
<none>
FLAGS:
niter=1 number of requested iterations
anenv=1 =1 for positive analytic envelopes
=0 for no analytic envelopes (not recommended)
scl=0 =1 to scale output traces (not recommended)
plot=3 =0 for no plots. =1 for 1-D plots only
=2 for 2-D plots only. =3 for all plots
norm=1 =0 not to normalize reliability values
verbose=1 =0 not to print processing information
rgt=2 =1 for uniform random generator
=2 for gaussian random generator
sts=1 =0 for no smoothing (not recommended)

General Parameters:
dx=20 offset sampling interval (m)
fx=0 offset on first trace (m)
dt=0.004 time sampling interval (s)
Tau-P Transform Parameters:

gopt=1 =1 for parabolic transform. =2 for Foster/Mosher
=3 for linear. =4 for absolute value of linear
pmin1=-400 minimum moveout at farthest offset for fwd transf(ms)
pmax1=400 maximum moveout at farthest offset for fwd transf(ms)
pmin2=pmin1 minimum moveout at farthest offset for inv transf(ms)
pmax2=pmax1 maximum moveout at farthest offset for inv transf(ms)
np=100 number of p-values for taup transform
prewhite=0.01 prewhitening value (suggested between 0.1 and 0,01)
offref=2000 reference offset for p-values (m)
depthref=500 reference depth for Foster/Mosher taup (if gopt=4)
217
pmula=pmax1 maximum p-value preserved in the data (ms)
pmulb=pmax1 minimum p-value muted on the data (ms)
ninterp=0 number of traces to interpolate in input data
Extraction Parameters:
nintlh=50 number of intervals (bins) in histograms
sditer=5 number of steepest descent iterations to compute ps
c=0.04 maximum noise allowed in a sample of signal(%)
rel1=0.5 reliability value for first pass of the extraction
rel2=0.75 reliability value for second pass of the extraction
Smoothing Parameters: ",

r1=10 number of points for damped lsq vertical smoothing
r2=2 number of points for damped lsq horizontal smoothing
Output Files:
signal=out_signal name of output file for extracted signal
noise=out_noise name of output file for extracted noise
Notes:
The signal-noise separation algorithm was developed by Dr. Bill Harlan
in 1984. It can be used to separate events that can be focused by a
linear transformation (signal) from events that can’t (noise). The
linear transform is whatever is well siuted for the application at
hand. Here, only the discrete Radon transform is used, so the program
is capable of separating events focused by that transform (linear,
parabolic or time-invariantly hyperbolic). Should other transform be
required, the changes to the program will be relatively
straightforward.
The reliability parameter is the most critical one to determine what

to extract as signal and what to reject as noise. It should be tested
for every dataset. The way to test it is to start with a small value,
say 0.1 or 0.01. If too much noise is present in the extracted noise,
it is too low. If too much signal was extracted, that is, part of the
signal was lost, it is too big. All other parameters have good default
values and should perhaps not be changed in a first encounter with the
program. The transform parameters are also critical. They should be
chosen such that no aliasing is present and such that the range of
interesting slopes is spanned by the transform but not much more. The
program suradon.c has more documentation on the transform paramters.
218
Credits:
Gabriel Alvarez CWP (1995)
Some subroutines are direct translations to C from Fortran versions
written by Dr. Bill Harlan (1984)
References:
Harlan, S., Claerbout, J., and Roca, F. (1984), Signal/noise

separation and velocity estimation, Geophysics, v. 49, no. 11,
p 1869-1880.
Harlan, S. (1988), Separation of signal and noise applied to

vertical seismic profiles, Geophysics, v. 53, no. 7,
p 932-946.
Alvarez, G. (1995), Comparison of moveout-based approaches to

ground roll and multiple suppression, MSc., Department of
Geophysics, Colorado School of Mines, (Chapter 3 deals
exclusively with this method).
219
SUHILB - Hilbert transform
suhilb <stdin >sdout
Note: the transform is computed in the direct (time) domain
Credits:
CWP: Jack Cohen
CWP: John Stockwell, modified to use Dave Hale’s hilbert() subroutine
220
char *sdoc[] = {
SUHISTOGRAM - create histogram of input amplitudes
suhistogram <in.su >out.dat
required parameters:
min= - minimum bin

max= - maximum bin
bins= - number of bins
optional parameters
trend=0 - =0 1-D histogram

=1 2-D histogram picks on cumulate
=2 2-D histogram in trace format
clip= threshold value to drop outliers
dt= - sample rate in feet or milliseconds. Defaults to

tr.dt*1e-3
datum= - header key to get datum shift if desired (e.g. to
hang from water bottom)
Notes:
trend=0 produces a two column ASCII output for use w/ gnuplot.
Extreme values are counted in the end bins.
trend=1 produces a 6 column ASCII output for use w/ gnuplot

The columns are time/depth and picks on the cumulate
at 2.28%, 15.87%, 50%, 84.13% & 97.72% of the total points
corresponding to the median and +- 1 or 2 standard deviations
for a Gaussian distribution.
trend=2 produces an SU trace panel w/ one trace per bin that

can be displayed w/ suximage, etc.
Example for plotting with xgraph:

suhistogram < data.su min=MIN max=MAX bins=BINS |
a2b n1=2 | xgraph n=BINS nplot=1
221
Author: Reginald H. Beardsley 2006 [email protected]
222
SUHROT - Horizontal ROTation of three-component data
suhrot <stdin >stdout [optional parameters]
none
angle=rad unit of angles, choose "rad", "deg", or "gon
inv=0 1 = inverse rotation (counter-clockwise)
verbose=0 1 = echo angle for each 3-C station
a=... array of user-supplied rotation angles

x=0.0,... array of corresponding header value(s)
key=tracf header word defining 3-C station ("x")
... or input angles from files:

n=0 number of x and a values in input files
xfile=... file containing the x values as specified by the
"key" parameter
afile=... file containing the a values
Notes:
dataset.
By default, the data will be rotated from the Z-North-East (Z,N,E)
coordinate system into Z-Radial-Transverse (Z,R,T).
If one of the parameters "a=" or "afile=" is set, the data

are rotated by these user-supplied angles. Specified x values
must be monotonically increasing or decreasing, and afile and
xfile are files of binary (C-style) floats.

Geophysics, Kiel University, Germany, 1999.
Trace header fields accessed: ns, sx, sy, gx, gy, key=keyword
Trace header fields modified: trid
223
SUHTMATH - do unary arithmetic operation on segy traces with
headers values
suhtmath <stdin >stdout
none
Optional parameter:
key=tracl header word to use
op=nop operation flag
nop : no operation
add : add header to trace
mult : multiply trace with header
div : divide trace by header
scale=1.0 scalar multiplier for header value

const=0.0 additive constant for header value
Operation order: ",
op=add: out(t) = in(t) + (scale * key + const)

op=mult: out(t) = in(t) * (scale * key + const)
op=div: out(t) = in(t) / (scale * key + const)
Credits:
Matthias Imhof, Virginia Tech, Fri Dec 27 09:17:29 EST 2002
224
SUIFFT - fft complex frequency traces to real time traces
suiftt <stdin >sdout sign=-1
none
Optional parameter:
sign=-1 sign in exponent of inverse fft
Output traces are normalized by 1/N where N is the fft size.
Note: sufft | suifft is not quite a no-op since the trace

length will usually be longer due to fft padding.
Credits:
CWP: Shuki, Chris, Jack

Trace header fields modified: ns, trid
225
SUILOG -- time axis inverse log-stretch of seismic traces
suilog nt= ntmin= <stdin >stdout
nt= nt output from sulog prog
ntmin= ntmin output from sulog prog
dt= dt output from sulog prog
none
NOTE: Parameters needed by suilog to reconstruct the

original data may be input via a parameter file.
EXAMPLE PROCESSING SEQUENCE:

sulog outpar=logpar <data1 >data2
suilog par=logpar <data2 >data3
where logpar is the parameter file
Credits:
CWP: Shuki Ronen, Chris Liner
Caveats:
amplitudes are not well preserved

Trace header fields modified: ns, dt
226
SUIMP2D - generate shot records for a line scatterer
embedded in three dimensions using the Born
integral equation ",
suimp2d [optional parameters] >stdout
Optional parameters
nrec=1 number of receivers
c=5000 speed
dt=.004 sampling rate
nt=256 number of samples
dsx=100 x-step in shot location
dsz=0 z-step in shot location
dgx=100 x-step in receiver location
dgz=0 z-step in receiver location
Example:
suimp2d nrec=32 | sufilter | supswigp | ...
Credits:
CWP: Norm Bleistein, Jack K. Cohen
Theory: Use the 3D Born integral equation (e.g., Geophysics,

v51, n8, p1554(7)). Use 2-D delta function for alpha and do
remaining y-integral by stationary phase.
Note: Setting a 2D offset in a single offset field beats the

hell out of us. We did _something_.
Trace header fields set: ns, dt, tracl, tracr, fldr, sx, selev,
gx, gelev, offset
227
SUIMP3D - generate inplane shot records for a point
scatterer embedded in three dimensions using
the Born integral equation ",
suimp3d [optional parameters] >stdout
Optional parameters
nrec=1 number of receivers
c=5000 speed
dt=.004 sampling rate
nt=256 number of samples
y0=0 point scatterer location
symin=0 first shot location
gymin=0 first receiver location
dsx=100 x-step in shot location
dsy=0 y-step in shot location
dsz=0 z-step in shot location
dgx=100 x-step in receiver location
dgy=0 y-step in receiver location
dgz=0 z-step in receiver location
Example:
suimp3d nrec=32 | sufilter | supswigp | ...
Credits:
CWP: Norm Bleistein, Jack K. Cohen
Theory: Use the 3D Born integral equation (e.g., Geophysics,

v51, n8, p1554(7)). Use 3-D delta function for alpha.
Note: Setting a 3D offset in a single offset field beats the

hell out of us. We did _something_.
228
Trace header fields set: ns, dt, tracl, tracr, fldr, tracf,
sx, sy, selev, gx, gy, gelev, offset
229
SUIMPEDANCE - Convert reflection coefficients to impedances.
suimpedance <stdin >stdout [optional parameters]
v0=1500. Velocity at first sample (m/sec)
rho0=1.0e6 Density at first sample (g/m^3)
Notes:
Implements recursion [1-R(k)]Z(k) = [1+R(k)]Z(k-1).
Credits:
SEP: Stew Levin
230
SUINTERP - interpolate traces using automatic event picking
suinterp < stdin > stdout
ninterp=1 number of traces to output between each pair of input traces

nxmax=500 maximum number of input traces
freq1=4. starting corner frequency of unaliased range
freq2=20. ending corner frequency of unaliased range
deriv=0 =1 means take vertical derivative on pick section
(useful if interpolating velocities instead of seismic)
linear=0 =0 means use 8 point sinc temporal interpolation
=1 means use linear temporal interpolation
(useful if interpolating velocities instead of seismic)
lent=5 number of time samples to smooth for dip estimate
lenx=1 number of traces to smooth for dip estimate
lagc=400 number of ms agc for dip estimate
xopt=0 0 compute spatial derivative via FFT
(assumes input traces regularly spaced and relatively
noise-free)
1 compute spatial derivative via differences
(will work on irregulary spaced data)
iopt=0 0 = interpolate
1 = output low-pass model: useful for QC if interpolator failing
2 = output dip picks in units of samples/trace

Notes:
This program outputs ’ninterp’ interpolated traces between each pair of
input traces. The values for lagc, freq1, and freq2 are only used for
event tracking. The output data will be full bandwidth with no agc. The
default parameters typically will do a satisfactory job of interpolation
for dips up to about 12 ms/trace. Using a larger value for freq2 causes
the algorithm to do a better job on the shallow dips, but to fail on the
steep dips. Only one dip is assumed at each time sample between each pair
of input traces.
The key assumption used here is that the low frequency data are unaliased
231
and can be used for event tracking. Those dip picks are used to interpolate
the original full-bandwidth data, giving some measure of interpolation
at higher frequencies which otherwise would be aliased. Using iopt equal
to 1 allows you to visually check whether the low-pass picking model is
aliased.
Trace headers for interpolated traces are not updated correctly.

The output header for an interpolated traces equals that for the preceding
trace in the original input data. The original input traces are passed
through this module without modification.
The place this code is most likely to fail is on the first breaks.
Example run: suplane | suinterp | suxwigb &
Credit: John Anderson (visiting scholar from Mobil) July 1994
232
SUINTVEL - convert stacking velocity model to interval velocity model
suintvel vs= t0= outpar=/dev/tty
vs= stacking velocities
t0= normal incidence times
outpar=/dev/tty output parameter file in the form:
h=layer thicknesses vector
v=interval velocities vector
Examples:
suintvel vs=5000,5523,6339,7264 t0=.4,.8,1.125,1.425 outpar=intpar
suintvel par=stkpar outpar=intpar
If the file, stkpar, contains:

vs=5000,5523,6339,7264
t0=.4,.8,1.125,1.425
then the two examples are equivalent.
Note: suintvel does not have standard su syntax since it does not
operate on seismic data. Hence stdin and stdout are not used.
Note: may go away in favor of par program, velconv, by Dave
Credits:
CWP: Jack
The Common Depth Point Stack
William A. Schneider
Proc. IEEE, v. 72, n. 10, p. 1238-1254
1984
Formulas:
Note: All sums on i are from 1 to k
From Schneider:
Let h[i] be the ith layer thickness measured at the cmp and
233
v[i] the ith interval velocity.
Set:
t[i] = h[i]/v[i]
Define:
t0by2[k] = 0.5 * t0[k] = Sum h[i]/v[i]
vh[k] = vs[k]*vs[k]*t0by2[k] = Sum v[i]*h[i]
Then:
dt[i] = h[i]/v[i] = t0by2[i] - t0by2[i-1]
dvh[i] = h[i]*v[i] = vh[i] - vh[i-1]
h[i] = sqrt(dvh[i] * dt[i])
v[i] = sqrt(dvh[i] / dt[i])
234
SUINVXZCO - Seismic INVersion of Common Offset data for a smooth
velocity function V(X,Z) plus a slowness perturbation vp(x,z)
suinvvxzco <infile >outfile [optional parameters]
vfile file containing velocity array v[nx][nz]
nx= number of x samples (2nd dimension) in velocity
nz= number of z samples (1st dimension) in velocity
nxm= number of midpoints of input traces
offs= or from header (offset) source-receiver offset
dxm= or from header (d2) sampling interval of midpoints
fxm=0 first midpoint in input trace
nxd=5 skipped number of midpoints (see note)
dx=50.0 x sampling interval of velocity
fx=0.0 first x sample of velocity
dz=50.0 z sampling interval of velocity
nxb=nx/2 band centered at midpoints (see note)
nxc=0 hozizontal range in which velocity is changed
nzc=0 vertical range in which velocity is changed
fxo=0.0 x-coordinate of first output trace
dxo=15.0 horizontal spacing of output trace
nxo=101 number of output traces ",
fzo=0.0 z-coordinate of first point in output trace
dzo=15.0 vertical spacing of output trace
nzo=101 number of points in output trace ",
fmax=0.25/dt Maximum frequency set for operator antialiasing
ang=180 Maximum dip angle allowed in the image
ls=0 =1 for line source; =0 for point source
pert=0 =1 calculate time correction from v_p[nx][nz]
vpfile file containing slowness perturbation array v_p[nx][nz]
Notes:
Traveltime and amplitude are calculated by finite difference which
is done only in one of every NXD midpoints; in the skipped midpoint,
interpolation is used to calculate traveltime and amplitude. ",
For each midpoint, traveltime and amplitude are calculated in the
horizontal range of (xm-nxb*dx, xm+nxb*dx). Velocity is changed by
constant extropolation in two upper trianglar corners whose width is
235
nxc*dx and height is nzc*dz.
Eikonal equation will fail to solve if there is a polar turned ray.

In this case, the program shows the related geometric information.
There are three ways to remove the turned ray: smoothing velocity,
reducing nxb, and increaing nxc and nzc (if the turned ray occurs
in the shallow areas). To prevent traveltime distortion from a over
smoothed velocity, traveltime is corrected based on the slowness
perturbation.
Offsets are signed - may be positive or negative.
Reference:
Bleistein, N., Cohen, J. K., and Hagin, F., 1987,
Two-and-one-half dimensional Born inversion with an arbitrary reference
Geophysics Vol. 52, no.1, 26-36.
236
SUINVZCO3D - Seismic INVersion of Common Offset data with V(Z) velocity
function in 3D
suinvzco3d <infile >outfile [optional parameters]
vfile file containing velocity array v[nz]
nxm= number of midpoints of input traces
ny= number of lines
offs= or from header (offset) source-receiver offset
fxm=0 first midpoint in input trace
nxd=5 skipped number of midpoints (see note)
dx=50.0 x sampling interval of velocity
fx=0.0 first x sample of velocity
dz=50.0 z sampling interval of velocity
nxb=nx/2 band centered at midpoints (see note)
fxo=0.0 x-coordinate of first output trace
dxo=15.0 horizontal spacing of output trace
nxo=101 number of output traces ",
fyo=0.0 y-coordinate of first output trace
dyo=15.0 y-coordinate spacing of output trace
nyo=101 number of output traces in y-direction
fzo=0.0 z-coordinate of first point in output trace
dzo=15.0 vertical spacing of output trace
nzo=101 number of points in output trace ",
fmax=0.25/dt Maximum frequency set for operator antialiasing
ang=180 Maximum dip angle allowed in the image
Notes:
This algorithm is based on formula (50) in Geophysics Vol. 51,

1552-1558, by Cohen, J., Hagin, F., and Bleistein, N.
Traveltime and amplitude are calculated by ray tracing.

Interpolation is used to calculate traveltime and amplitude. ",
For each midpoint, traveltime and amplitude are calculated in the
horizontal range of (xm-nxb*dx, xm+nxb*dx). Velocity is changed by
237
linear interpolation in two upper trianglar corners whose width is
nxc*dx and height is nzc*dz. ",

Offsets are signed - may be positive or negative. ",
238
SUK1K2FILTER - symmetric box-like K-domain filter defined by the
cartesian product of two sin^2-tapered polygonal
filters defined in k1 and k2
suk1k2filter <infile >outfile [optional parameters]
k1=val1,val2,... array of K1 filter wavenumbers
k2=val1,val2,... array of K2 filter wavenumbers
amps1=a1,a2,... array of K1 filter amplitudes
amps2=a1,a2,... array of K2 filter amplitudes
d1=tr.d1 or 1.0 sampling interval in first (fast) dimension
d2=tr.d1 or 1.0 sampling interval in second (slow) dimension
quad=0 =0 all four quandrants
=1 (quadrants 1 and 4)
=2 (quadrants 2 and 3)
Defaults:
k1=.10*(nyq1),.15*(nyq1),.45*(nyq1),.50*(nyq1)
k2=.10*(nyq2),.15*(nyq2),.45*(nyq2),.50*(nyq2)
amps1=0.,1.,...,1.,0. trapezoid-like bandpass filter
amps2=0.,1.,...,1.,0. trapezoid-like bandpass filter
The nyquist wavenumbers, nyq1 and nyq2, are computed internally.

Notes:
The filter is assumed to be symmetric, to yield real output
Because the data are assumed to be purely spatial (i.e. non-seismic),

the data are assumed to have trace id (30), corresponding to (z,x) data
The relation: w = 2 pi F is well known for frequency, but there

doesn’t seem to be a commonly used letter corresponding to F for the
spatial conjugate transform variables. We use K1 and K2 for this.
More specifically we assume a phase:
-i(k1 x1 + k2 x2) = -2 pi i(K1 x1 + K2 x2).
239
and K1, K2 define our respective wavenumbers.
Credits:
CWP: John Stockwell, November 1995.
Trace header fields accessed: ns, d1, d2
240
sukdmig2d infile= outfile= ttfile= [parameters]
... The following 9 parameters describe traveltime tables:

fxo=fxt x-coordinate of first output trace
dxo=0.5*dxt horizontal spacing of output trace
nxo=2*(nxt-1)+1 number of output traces ",
off0=0 first offest in output
doff=99999 offset increment in output
noff=1 number of offsets in output ",
absoff=0 flag for using absolute offsets of input traces
=0 means use offset=gx-sx
=1 means use abs(gx-sx)
limoff=0 flag for only using input traces that fall within the range
of defined output offset bins (off0,doff,noff)
=0 means use all input traces
=1 means limit traces used by offset
241
v0=1500(m/s) reference velocity value at surface ",
ls=1 flag for line source


npv=0 flag of computing quantities for velocity analysis
rscale=1000.0 scaling for roundoff error suppression
...if npv>0 specify the following three files:

tvfile=tvfile input file of traveltime variation tables
tv[ns][nxt][nzt]
csfile=csfile input file of cosine tables cs[ns][nxt][nzt]
outfile1=outfile1 file containning additional migration output
with extra amplitude
Notes:
migration process, traveltimes are interpolated into shot/gephone
2. Input seismic traces must be SU format and can be any type of
gathers (common shot, common offset, common CDP, and so on). ",
3. Migrated traces are output in CDP gathers if velocity analysis
is required, with dimension nxo*noff*nzo. ",
[ns*nxt*nzt+noff*nxo*nzo+4*nr*nzt+5*nxt*nzt+npa*(2*ns*nxt*nzt
+noff*nxo*nzo+4*nxt*nzt)]*4 bytes
where nr = 1+min(nxt*dxt,0.5*offmax+aperx)/dxo.
8. if limoff=0, input traces from outside the range defined by off0, doff,
noff, will get migrated into the extremal offset bins/planes. E.g. if
absoff=0 and limoff=0, all traces with gx<sx will get migrated into the
off0 bin.
242
Modifcations:
Gary Billings, Talisman Energy, Sept 2005: added absoff, limoff.

Trace header fields modified: sx, gx
243
sukdmig3d datain= dataout= [parameters]
ttfile file for input tttables
datain=stdin file for input seismic traces

dataout=stdout file for common offset migration output
crfile=NULL file for cos theta and ray paths
The following 17 parameters describe tttables: (from ttfile header)

fxgd= or from header (f1) first x-sample in tttable
nxt= or from header (ns) number of x-samples in tttable
dxgd= or from header (d1) x-interval in tttable
fygd= or from header (f2) first y-sample in tttable
nyt= or from header (ntr) number of y-samples in tttable
dygd= or from header (d2) y-interval in tttable
ixsf= or from header (sdel) x in dxgd of first source
nxs= or from header (nhs) number of sources in x
ixsr= or from header (swdep) ratio of source & gd spacing
iysf= or from header (gdel) y in dygd of first source
nys= or from header (nvs) number of sources in y
iysr= or from header (gwdep) ratio of source & gd spacing
fzs= or from header (sdepth/1000) first depth sample in tttable
nzs= or from header (duse) number of depth samples in tttable
dzs= or from header (ep/1000) depth interval in tttable
nxgd= or from header (selev) x size of the traveltime region
nygd= or from header (gelev) y size of the traveltime region
multit= or from header (scalel) number of multivalued traveltime
The following two parameters are from data header

dxm= or from header (d2) mid point spacing of input data
Default: output is 5 times finer in depth and 2 times finer in lateral

fzo=fzs z-coordinate of first point in output trace
dzo=0.2*dzs vertical spacing of output trace (5 times finer)
nzo=5*(nzs-1)+1 number of points in output trace (5 times finer) ",
fxo=fxgd x-coordinate of first output trace
244
dxo=0.5*dxgd horizontal spacing of output trace (2 times finer)
nxo=2*(nxgd-1)+1 number of output traces (2 times finer)
fyo=fygd y-coordinate of first output trace
dyo=0.5*dygd horizontal spacing of output trace (2 times finer)
nyo=2*(nygd-1)+1 number of output traces (2 times finer)
Default: poststack migration ",

fxoffset=0 first offest in output in x
fyoffset=0 first offest in output in y
dxoffset=99999 offset increment in output in x
dyoffset=99999 offset increment in output in y
nxoffset=1 number of offsets in output in x
nyoffset=1 number of offsets in output in y ",
xoffsetmax=99999 x-maximum absolute offset allowed in migration
yoffsetmax=99999 y-maximum absolute offset allowed in migration
xaper=nxt*dxgd/2.5 migration lateral aperature in x
yaper=nyt*dygd/2.5 migration lateral aperature in y
angmax=60 max angle to handle
fmax=0.25/dt max frequency in the data
pptr=100 print verbal information at every pptr traces
ntrmax=100000 maximum number of input traces to be migrated
ls=0 point =0 line source =1
Notes:
1. Traveltime tables were generated by program SUTETRARAY (or other
ones) on very sparse tetrahedral model, with dimension nys*nxs*nzs
*nyt*nxt.
is required, with dimension nyoffset*nxoffset*nyo*nxo*nzo. ",
[nys*nxs*nzs*nyt*nxt+nyoffset*nxoffset*nxo*nyo*nzo+
nys*nxo*nzo*nyt*nxt]
6. Input traces must specify source and receiver positions via header
fields tr.sx and tr.gx, as well as tr.sy and tr.gy. Offset is
computed automatically.
Disclaimer:
This is a research code that will take considerable work to get into
245
the form of a a production level 3D migration code. The code is
offered as is, along with tetramod and sutetraray, to provide a starting
point for researchers who wish to write their own 3D migration codes.
Author: Zhaobo Meng, 01/10/97, Colorado School of Mines

Trace header fields modified: sx, gx
246
SUKDSYN2D - Kirchhoff Depth SYNthesis of 2D seismic data from a
migrated seismic section
sukdsyn2d infile outfile [parameters]
infile=stdin input migrated section
outfile=stdout file for output seismic traces
ttfile file for input traveltime tables

fzt= first depth sample
nzt= number of depth samples
dzt= depth interval
fxt= first lateral sample
nxt= number of lateral samples
dxt= lateral interval
The following 6 parameters describe the migration section:

fz= first z-coordinate in migrated section
dz= vertical spacing of migrated section
nz= number of depth points in migrated section
fx= first x-coordinate of migrated section
dx= horizontal spacing of migrated section
nx= number of lateral points in migrated section
nt=501 number of time samples
dt=0.004 time sampling interval (sec)
ft=0.0 first time (sec)
nxo=1 number of source-receiver offsets
dxo=25 offset sampling interval
fxo=0.0 first offset
nxs=101 number of shotpoints
dxs=25 shotpoint sampling interval
fxs=0.0 first shotpoint
fmax=1/(4*dt) maximum frequency in migration section (Hz)
aperx=nxt*dxt/2 modeling lateral aperature
angmax=60 modeling angle aperature from vertical
247
Notes:
This program takes a migrated seismic section and a set of travel time
tables generated using rayt2d for a specific background velocity model
and generates synthetic seismic data in the form of common shot gathers.
(Common offset gathers may be generated by using nxo=1.) (Demigration.)
This program is a tool which may be used for the migration residual
statics estimation technique of Tjan, Audebert, and Larner 1994.
1. The traveltime tables are generated by program rayt2d (or other ones)
modeling process, traveltimes are interpolated into shot/geophone
positions and migration section grids.
2. The input migration section must be an array of binary floats (no SU
headers). ",
3. The synthesized traces are output in common-shot gathers in SU format.
4. The memory requirement for this program is about
(ns*nxt*nzt+nx*nz+4*nr*nzt+3*nxt*nzt)*4 bytes
where nr = 1+min(nxt*dxt,0.5*offmax+aperx)/dxo.
Author: CWP: Zhenyue Liu, 07/24/95, Colorado School of Mines
References:
Tjan, T., F. Audebert, and K. Larner, 1994,

Prestack migration for residual statics estimation in complex media
(Appeared in 1994 Project Review, CWP-153.)
Tjan, T., 1995, Residual statics estimation for data from structurally
complex areas using prestack depth migration: M.Sc. thesis, Colorado
School of Mines. (In progress.)
Larner, K., and Tjan, T., 1995, Simultaneous statics and velocity
estimation for data from structurally complex areas.
(Appeared in 1995 Project Review, CWP-185.)
248
Trace header fields set: ns, dt, delrt, tracl, tracr, fldr, tracf
offset, sx, gx, trid, counit
249
SUKEYCOUNT - sukeycount writes a count of a selected key
sukeycount key=keyword < infile [> outfile]
key=keyword One key word.
verbose=0 quiet
=1 chatty
Writes the key and the count to the terminal or a text

file when a change of key occurs. This does not provide
a unique key count (see SUCOUNTKEY for that).
Note that for key values 1 2 3 4 2 5
value 2 is counted once per occurrence since this program
only recognizes a change of key, not total occurrence.
Examples:
sukeycount < stdin key=fldr
sukeycount < stdin key=fldr > out.txt
Credits:
MTU: David Forel, Jan 2005
250
SUKFILTER - radially symmetric K-domain, sin^2-tapered, polygonal
filter
sukfilter <infile >outfile [optional parameters]
k=val1,val2,... array of K filter wavenumbers
amps=a1,a2,... array of K filter amplitudes
d1=tr.d1 or 1.0 sampling interval in first (fast) dimension
d2=tr.d1 or 1.0 sampling interval in second (slow) dimension
Defaults:
k=.10*(nyq),.15*(nyq),.45*(nyq),.50*(nyq)
amps=0.,1.,...,1.,0. trapezoid-like bandpass filter
The nyquist wavenumbers, nyq=sqrt(nyq1^2 + nyq2^2) is computed

internally.
Notes:
The filter is assumed to be symmetric, to yield real output.


-i(k1 x1 + k2 x2) = -2 pi i(K1 x1 + K2 x2).
Credits:
CWP: John Stockwell, June 1997.
251
SUKFRAC - apply FRACtional powers of i|k| to data, with phase shift
sukfrac <infile >outfile [optional parameters]
power=0 exponent of (i*sqrt(k1^2 + k2^2))^power
=0 ===> phase shift only
>0 ===> differentiation
<0 ===> integration
sign=1 sign on transform exponent
d1=1.0 x1 sampling interval
d2=1.0 x2 sampling interval
phasefac=0 phase shift by phase=phasefac*PI
Notes:
-i(k1 x1 + k2 x2) = -2 pi i(K1 x1 + K2 x2).
Algorithm:
g(x1,x2)=Re[2DINVFFT{ ( (sign) i |k|)^power 2DFFT(f)}e^i(phase)]
Caveat:
Large amplitude errors will result of the data set has too few points.
Examples:
Edge sharpening:
Laplacean :
sukfrac < image_data power=2 phasefac=-1 | ...
Image enhancement:
Derivative filter:
sukfrac < image_data power=1 phasefac=-.5 | ...
Image enhancement:
Half derivative (this one is the best for photographs):
sukfrac < image_data power=.5 phasefac=-.25 | ...
Credits:
CWP: John Stockwell, June 1997, based on sufrac.
252
253
SUKILL - zero out traces
sukill <stdin >stdout min= count=1
Required parameters
min= first trace to kill (one-based)
Optional parameters
count= 1 number of traces to kill
Credits:
CWP: Chris Liner, Jack K. Cohen
254
SULOG -- time axis log-stretch of seismic traces
sulog [optional parameters] <stdin >stdout
none
ntmin= .1*nt minimum time sample of interest
number of samples (nt=)
minimum time sample (ntmin=)
output number of samples (ntau=)
m=3 length of stretched data
is set according to
ntau = nextpow(m*nt)
ntau= pow of 2 override for length of stretched
data (useful for padding zeros
to avoid aliasing)
NOTES:
ntmin is required to avoid taking log of zero and to
keep number of outsamples (ntau) from becoming enormous.
Data above ntmin is zeroed out.
The output parameters will be needed by suilog to

reconstruct the original data.
EXAMPLE PROCESSING SEQUENCE:

sulog outpar=logpar <data1 >data2
suilog par=logpar <data2 >data3
Credits:
CWP: Shuki, Chris
Caveats:
Amplitudes are not well preserved.

255
char *sdoc[] = {
SUMAX - get trace by trace local/global maxima, minima, or absolute maximum
sumax <stdin >stdout [optional parameters]
none
output=ascii write ascii data to outpar
=binary for binary floats to stdout
=segy for SEGY traces to stdout
mode=maxmin output both minima and maxima

=max maxima only
=min minima only
=abs absolute maxima only
=rms RMS
=thd search first max above threshold
threshamp=0 threshold amplitude value

threshtime=0 tmin to start search for threshold
verbose=0 writes global quantities to outpar

=1 trace number, values, sample location
=2 key1 & key2 instead of trace number
key1=fldr key for verbose=2
key2=ep key for verbose=2
outpar=/dev/tty output parameter file; contains output

from verbose
Examples:
For global max and min values: sumax < segy_data
For local and global max and min values: sumax < segy_data verbose=1
To plot values specified by mode:
sumax < segy_data output=binary mode=modeval | xgraph n=npairs
To plot seismic data with the only values nonzero being those specified
by mode=modeval:
sumax < segy_data output=segy mode=modeval | suxwigb
Note: while traces are counted from 1, sample values are counted from 0.
256
Also, if multiple min, max, or abs max values exist on a trace,
only the first one is captured.
See also: suxmax, supsmax
Credits:
CWP : John Stockwell (total rewrite)
Geocon : Garry Perratt (all ASCII output changed from %e to %f)
(added mode=rms).
ESCI: Reginald Beardsley (added header key option)
based on an original program by:
SEP: Shuki Ronen
CWP: Jack K. Cohen
IFM-GEOMAR: Gerald Klein (added threshold option)
Trace header fields accessed: ns dt & user specified keys
257
SUMEAN - get the mean values of data traces ",
sumean < stdin > stdout [optional parameters]
power = 2.0 mean to the power
(e.g. = 1.0 mean amplitude, = 2.0 mean energy)
verbose = 0 writes mean value of section to outpar
= 1 writes mean value of each trace / section to
outpar
abs = 1 average absolute value
= 0 preserve sign if power=1.0
Notes:
Each sample is raised to the requested power, and the sum of all those
values is averaged for each trace (verbose=1) and the section.
The values power=1.0 and power=2.0 are physical, however other powers
represent other mathematical L-p norms and may be of use, as well.
Credits:
Bjoern E. Rommel, IKU, Petroleumsforskning / October 1997
[email protected]
258
SUMEDIAN - MEDIAN filter about a user-defined polygonal curve with
the distance along the curve specified by key header word
sumedian <stdin >stdout xshift= tshift= [optional parameters]
xshift= array of position values as specified by
the ‘key’ parameter
tshift= array of corresponding time values (sec)
... or input via files:
nshift= number of x,t values defining median times
xfile= file containing position values as specified by
tfile= file containing corresponding time values (sec)
key=tracl Key header word specifying trace number
=offset use trace offset instead
mix=.6,1,1,1,.6 array of weights for mix (weighted moving average)

median=0 =0 for mix
=1 for median filter
nmed=5 odd no. of traces to median filter
sign=-1 =-1 for upward shift
=+1 for downward shift
subtract=1 =1 subtract filtered data from input
=0 don’t subtract
verbose=0 =1 echoes information

Notes:
------
Median filtering is a process for suppressing a particular moveout on
seismic sections. Its advantage over traditional dip filtering is that
events with arbitrary moveout may be suppressed. Median filtering is
commonly used in up/down wavefield separation of VSP data.
The process generally consists of 3 steps:

1. A copy of the data panel is shifted such that the polygon in x,t
259
specifying moveout is flattened to horizontal. The x,t pairs are
specified either by the vector xshift,tshift or by the values in
the datafiles xfile,tfile. For fractional shift, the shifted data
is interpolated.
2. Then a mix (weighted moving average) is performed over the shifted
panel to emphasize events with the specified moveout and suppress
events with other moveouts.
3. The panel is then shifted back (and interpolated) to its original
moveout, and subtracted from the original data. Thus all events
with the user-specified moveout are removed.
For VSP data the following modifications are provided:

1. The moveout polygon in x,t is usually the first break times for
each trace. The parameter sign allows for downward shift in order
align upgoing events.
2. Alternative to a mix, a median filter can be applied by setting
the parameter median=1 and nmed= to the number of traces filtered.
3. By setting subtract=0 the filtered panel is only shifted back but
not subtracted from the original data.
The values of tshift are linearly interpolated for traces falling

between given xshift values. The tshift interpolant is extrapolated
to the left by the smallest time sample on the trace and to the right
by the last value given in the tshift array.
The files tfile and xfile are files of binary (C-style) floats.
The number of values defined by mix=val1,val2,... determines the

number of traces to be averaged, the values determine the weights.
Caveat:
The median filter may perform poorly on the edges of a section.
Choosing larger beginning and ending mix values may help, but may
also introduce additional artifacts.
Examples:
Credits:
CWP: John Stockwell, based in part on sumute, sureduce, sumix

CENPET: Werner M. Heigl - fixed various errors, added VSP functionality
260
ideas for improvement:
* a versatile median filter needs to do:
shift traces by fractional amounts -> needs sinc interpolation
positive and negative shifts similar to SUSTATIC
make subtraction of filtered events a user choice
provide a median stack as well as a weighted average stack
Trace header fields accessed: ns, dt, delrt, key=keyword
261
SUMIGFD - 45-90 degree Finite difference depth migration for
zero-offset data.
sumigfd <infile >outfile vfile= [optional parameters]
nz= number of depth sapmles
dz= depth sampling interval
(see Notes below concerning format of this file)
dip=45,65,79,80,87,89,90 Maximum angle of dip reflector

Notes: ",
The computation cost by dip angle is 45=65=79<80<87<89<90
The input velocity file \’vfile\’ consists of C-style binary floats. ",
The structure of this file is vfile[iz][ix]. Note that this means that
the x-direction is the fastest direction instead of z-direction! Such a
structure is more convenient for the downward continuation type
migration algorithm than using z as fastest dimension as in other SU
programs. (In C v[iz][ix] denotes a v(x,z) array, whereas v[ix][iz]
denotes a v(z,x) array, the opposite of what Matlab and Fortran
programmers may expect.) ",
Because most of the tools in the SU package (such as unif2, unisam2,

and makevel) produce output with the structure vfile[ix][iz], you will
need to transpose the velocity files created by these programs. You may
use the SU program \’transp\’ in SU to transpose such files into the
required vfile[iz][ix] structure.
Credits: CWP Baoniu Han, April 20th, 1998
262
263
SUMIGFFD - Fourier finite difference depth migration for
zero-offset data. This method is a hybrid migration which
combines the advantages of phase shift and finite difference",
migrations.
sumigffd <infile >outfile vfile= [optional parameters]
nz= number of depth sapmles ",
fz=0.0 first depth sample

migration algorithm than using z as fastest dimension as in other SU ",
programs. (In C v[iz][ix] denotes a v(x,z) array, whereas v[ix][iz]
Because most of the tools in the SU package (such as unif2, unisam2, ",
Credits: CWP Baoniu Han, July 21th, 1997
264
265
SUMIGGBZOAN - MIGration via Gaussian beams ANisotropic media (P-wave)
sumiggbzoan <infile >outfile vfile= nt= nx= nz= [optional parameters]
a3333file= name of file containing a3333(x,z)
nx= number of inline samples (traces)
nz= number of depth samples
dt=tr.dt time sampling interval
dx=tr.d2 inline sampling interval (trace spacing)
fmin=0.025/dt minimum frequency
fmax=10*fmin maximum frequency
amin=-amax minimum emergence angle; must be > -90 degrees
bwh=0.5*vavg/fmin beam half-width; vavg denotes average velocity
Files for general anisotropic parameters confined to a vertical plane:

For transversely isotropic media Thomsen’s parameters could be used:

deltafile= name of file containing delta(x,z)
epsilonfile= name of file containing epsilon(x,z)
if anisotropy parameters are not given the program considers ",

the medium to be isotropic.
Credits:
CWP: Tariq Alkhalifah, based on MIGGBZO by Dave Hale
CWP: repackaged as an SU program by John Stockwell, April 2006
Alkhailfah, T., 1993, Gaussian beam migration for

anisotropic media: submitted to Geophysics.
266
Cerveny, V., 1972, Seismic rays and ray intensities
in inhomogeneous anisotropic media:
Hale, D., 1992, Migration by the Kirchhoff,

slant stack, and Gaussian beam methods:
CWP,1992 Report 121, Colorado School of Mines.
Hale, D., 1992, Computational Aspects of Gaussian

Beam migration:
CWP,1992 Report 121, Colorado School of Mines.
267
SUMIGGBZO - MIGration via Gaussian Beams of Zero-Offset SU data
sumiggbzo <infile >outfile vfile= nz= [optional parameters]
vfile= name of file containing v(z,x)
nz= number of depth samples
dt=from header time sampling interval
dx=from header(d2) or 1.0 spatial sampling interval
fmin=0.025/dt minimum frequency
fmax=10*fmin maximum frequency
amin=-amax minimum emergence angle; must be > -90 degrees
bwh=0.5*vavg/fmin beam half-width; vavg denotes average velocity
verbose=0 =0 silent; =1 chatty
Note: spatial units of v(z,x) must be the same as those on dx.

v(z,x) is represented numerically in C-style binary floats v[x][z],
where the depth direction is the fast direction in the data. Such
models can be created with unif2 or makevel.
(In C v[iz][ix] denotes a v(x,z) array, whereas v[ix][iz]

Credits:
CWP: Dave Hale (algorithm), Jack K. Cohen, and John Stockwell

(reformatting for SU)
268
SUMIGPREFD --- The 2-D prestack common-shot 45-90 degree
finite-difference depth migration.
sumigprefd <indata >outfile [parameters] ",
Required Parameters: ",

nxo= number of total horizontal output samples
nxshot= number of shot gathers to be migrated
dx= horizontal sampling interval ",
vfile= velocity profile, it must be binary format (see Notes)
dip=79 the maximum dip to migrate, possible values are:
45,65,79,80,87,89,90 degrees
The computation cost is 45=65=79<80<87<89<90
fmax=25 peak frequency of Ricker wavelet used as source wavelet
f1=5,f2=10,f3=40,f4=50 frequencies to build a Hamming window
lpad=9999,rpad=9999 number of zero traces padded on both

sides of depth section to determine the
migration aperature, the default values
are using the full aperature.
verbose=0 silent, =1 additional runtime information
Notes:
The input velocity file \’vfile\’ consists of C-style binary floats.
programs.
269
Also, sx must be monotonically increasing throughout the dataset, and
and gx must be monotonically increasing within a shot. You may resort
your data with \’susort\’, accordingly.
The scalco header field is honored so this field must be set correctly.
See selfdocs of \’susort\’, \’suchw\’. Also: sukeyword scalco
Credits: CWP, Baoniu Han, [email protected], April 19th, 1998

Modified: Chris Stolk, 11 Dec 2005, - changed data input
to remove erroneous time delay.
Modified: CWP, John Stockwell 26 Sept 2006 - replaced Han’s
"goto-loop" in two places with "do { }while loops".
Fixed it so that sx, gx, and scalco are honored.
Trace header fields accessed: ns, dt, delrt, d2, sx, gx,
270
SUMIGPREFFD - The 2-D prestack common-shot Fourier finite-difference
depth migration.
sumigpreffd <indata >outfile [parameters] ",
Required Parameters: ",

dx= horizontal sampling interval
vfile= velocity profile, it must be binary format.
fmax=25 the peak frequency of Ricker wavelet used as source wavelet
migration aperature, the default
values are using the full aperature.
Notes:
The input velocity file consists of C-style binary floats. ",
programs.

271

"goto-loop" with "do { }while loops".

272
SUMIGPREPSPI --- The 2-D PREstack commom-shot Phase-Shift-Plus
interpolation depth MIGration.
sumigprepspi <indata >outfile [parameters] ",

dx= horizontal sampling interval ",
fmax=25 the peak frequency of Ricker wavelet used as source wavelet
Notes:
programs.


273


274
SUMIGPRESP - The 2-D prestack common-shot split-step Fourier ",
migration
sumigpresp <indata >outfile [parameters] ",
dx= horizontal sampling interval
fmax=25 The peak frequency of Ricker wavelet used as source wavelet
Notes:
The input velocity file consists of C-style binary floats.
programs.

275

Trace header fields accessed: ns, dt, delrt, d2, sx, gx, scalco
276
SUMIGPSPI - Gazdag’s phase-shift plus interpolation depth migration
for zero-offset data, which can handle the lateral
velocity variation.
sumigpspi <infile >outfile vfile= [optional parameters]
(Please see Notes below concerning the format of vfile)

Notes:
programs.
Credits: CWP, Baoniu Han, April 20th, 1998

277
SUMIGPS - MIGration by Phase Shift with turning rays
sumigps <stdin >stdout [optional parms]
None
dx=from header(d2) or 1.0 distance between sucessive cdp’s
ffil=0,0,0.5/dt,0.5/dt trapezoidal window of frequencies to migrate
vfile= binary (non-ascii) file containing velocities v(t)
nxpad=0 number of cdps to pad with zeros before FFT
ltaper=0 length of linear taper for left and right edges",

Notes:
Input traces must be sorted by either increasing or decreasing cdp.

then the tmig and vmig arrays are ignored.
The time of first sample is assumed to be zero, regardless of the value

of the trace header field delrt.
Credits:
CWP: Dave Hale (originally called supsmig.c)
278
279
SUMIGPSTI - MIGration by Phase Shift for TI media with turning rays
sumigpsti <stdin >stdout [optional parms]
None
dx=from header(d2) or 1.0 distance between sucessive cdp’s
ffil=0,0,0.5/dt,0.5/dt trapezoidal window of frequencies to migrate
vnmig=1500.0 interval NMO velocities corresponding to times in tmig
etamig=0.0 interval eta values corresponding to times in tmig
vnfile= binary (non-ascii) file containing NMO velocities vn(t)
etafile= binary (non-ascii) file containing eta values eta(t)
ltaper=0 length of linear taper for left and right edges ",
Notes:
The tmig, vnmig, vmig and etamig arrays specify an interval values
function of time. Linear interpolation and constant extrapolation is
used to determine interval velocities at times not specified. Values
specified in tmig must increase monotonically.
containing one velocity for every time sample. If vnfile is specified,
then the tmig and vnmig arrays are ignored.

Trace header fields accessed: ns and dt
280
SUMIGSPLIT - Split-step depth migration for zero-offset data.
sumigsplit <infile >outfile vfile= [optional parameters]
fz= first depth sample

Notes:
programs.
Because most of the tools in the SU package (such as unif2, unisam2,

Credits: CWP Baoniu Han, July 21th, 1997
Reference:
281
Stoffa, P. L. and Fokkema, J. T. and Freire, R. M. and Kessinger, W. P.,
1990, Split-step Fourier migration, Geophysics, 55, 410-421.

282
SUMIGTK - MIGration via T-K domain method for common-midpoint stacked data
sumigtk <stdin >stdout dxcdp= [optional parms]
dxcdp distance between successive cdps
fmax=Nyquist maximum frequency
ltaper=0 length of linear taper for left and right edges",

Notes:

then the tmig and vmig arrays are ignored.

Credits:
CWP: Dave Hale
Trace header fields accessed: ns and dt
283
SUMIGTOPO2D - Kirchhoff Depth Migration of 2D postack/prestack data
from the (variable topography) recording surface
sumigtopo2d infile= outfile= [parameters]
ttfile file for input traveltime tables
fzt first depth sample in traveltime table
nzt number of depth samples in traveltime table
dzt depth interval in traveltime table
fxt first lateral sample in traveltime table
nxt number of lateral samples in traveltime table
dxt lateral interval in traveltime table
fs x-coordinate of first source
ns number of sources
ds x-coordinate increment of sources
fxi x-coordinate of the first input trace

dxi horizontal spacing of input data
nxi number of input trace locations in surface
surf="0,0;99999,0" Recording surface "x1,z1;x2,z2;x3,z3;...
fxo=fxt x-coordinate of first output trace
dxo=0.5*dxt horizontal spacing of output trace
nxo=2*(nxt-1)+1 number of output traces ",
off0=0 first offest in output
doff=99999 offset increment in output
noff=1 number of offsets in output ",
v0=1500(m/s) reference velocity value at surface ",
284
Notes:
1. Traveltime tables were generated by program rayt2dtopo (or any
other one that considers topography )on relatively coarse grids,
with dimension ns*nxt*nzt. In the migration process, traveltimes
are interpolated into shot/gephone positions and output grids.
is required, with dimension nxo*noff*nzo. ",
Trino Salinas, 07/01/96, Colorado School of Mines,

extended the code to migrate data from a nonflat
recording surface.
References :
Bleistein, N., Cohen, J., and Hagin, F., 1987, Two and one-half
dimensional Born inversion with arbitrary reference: Geophysics,
52, 26-36.
Liu,Z., 1993, A Kirchhoff approach to seismic modeling and

pre-stack depth migration: CWP Annual Report, CWP, Colorado
School of Mines.
Wiggins, J. W., 1984, Kirchhoff integral extrapolation and

migration of nonplanar data: Geophysics, 49, 1239-
1248.
285
SUMIXGATHERS - mix two gathers
sumixgathers file1 file2 > stdout [optional parameters]
ntr=tr.ntr if ntr header field is not set, then ntr is mandatory
Optional Parameters: none
Notes: Both files have to be sorted by offset

Mixes two gathers keeping only the traces of the first file
if the offset is the same. The purpose is to substitute only
traces non existing in file1 by traces interpolated store in file2. ",
Example. If file1 is original data file and file 2 is obtained by

resampling with Radon transform, then the output contains original
traces with gaps filled
Credits:
Daniel Trad. UBC
Trace header fields accessed: ns, dt, ntr
Copyright (c) University of British Columbia, 1999.
286
SUMIX - compute weighted moving average (trace MIX) on a panel
of seismic data
sumix <stdin >sdout

mix=.6,1,1,1,.6 array of weights for weighted average
Note:
The number of values defined by mix=val1,val2,... determines the number
of traces to be averaged, the values determine the weights.
Examples:
sumix <stdin mix=.6,1,1,1,.6 >sdout (default) mix over 5 traces weights
sumix <stdin mix=1,1,1 >sdout simple 3 trace moving average
Author:
CWP: John Stockwell, Oct 1995
287
SUMUTE - mute above (or below) a user-defined polygonal curve with ",
the distance along the curve specified by key header word
sumute <stdin >stdout xmute= tmute= [optional parameters]
xmute= array of position values as specified by
tmute= array of corresponding time values (sec)
in case of air wave muting, correspond to
air blast duration
nmute= number of x,t values defining mute
xfile= file containing position values as specified by
tfile= file containing corresponding time values (sec)
key=offset Key header word specifying trace offset
=tracl use trace number instead
ntaper=0 number of points to taper before hard
mute (sine squared taper)
mode=0 =1 to zero BELOW the polygonal curve
=2 to mute below AND above a straight line
in that case xmute,tmute describes the total
time length of the muted zone
linvel=330 linear velocity
tm0=0 time shift of linear mute at \’key\’=0
Notes:
The tmute interpolant is extrapolated to the left by the smallest time

sample on the trace and to the right by the last value given in the
tmute array.
The files tfile and xfile are files of binary (C-style) floats.
In the context of this program "above" means earlier time and

"below" means later time (above and below as seen on a seismic section.
The below=2 option is intended for removing air waves. The mute is
is over a narrow window above and below the polygonal line specified
288
by the values in tmute, xmute or tfile and xfile.
If data are spatial, such as the (z-x) output of a migration, then

depth values are used in place of times in tmute and tfile. The value
of the depth sampling interval is given by the d1 header field
Caveat: if data are seismic time sections, then tr.dt must be set. If
data are seismic depth sections, then tr.trid must be set to 30, and
tr.d1 header field must be set.
Credits:
SEP: Shuki Ronen

CWP: Jack K. Cohen, Dave Hale, John Stockwell
DELPHI: Alexander Koek
Trace header fields accessed: ns, dt, delrt, key=keyword, trid, d1

Trace header fields modified: muts or mute
289
SUNAN - remove NaNs & Infs from the input stream
sunan <in.su >out.su
verbose=1 echo locations of NaNs or Infs to stderr
=0 silent
Notes:
A simple program to remove NaNs and Infs from an input stream.
The program sets NaNs and Infs to 0.0.
Author: Reginald H. Beardsley 2003 [email protected]
A simple program to remove NaNs & Infs from an input stream. They
shouldn’t be there, but it can be hard to find the cause and fix
the problem if you can’t look at the data.
290
SUNHMOSPIKE - generates SPIKE test data set with a choice of several
Non-Hyperbolic MOveouts
sunhmospike [optional parameters] > out_data_file
ntr=20 number of traces
dt=0.001 time sample rate in seconds
offref=2000 reference offset
gopt= 1 = parabolic transform model

2 = Foster/Mosher pseudo hyperbolic option model
3 = linear tau-p model
depthref=400 reference depth used when gopt=2
offinc=100 offset increment
nspk=4 number of events
p1 = 0 event moveout for event #1 in ms on reference offset

t1 = 100 intercept time ms event #1
a1 = 1.0 amplitude for event #1

t2 = 100 intercept time ms for spike #2
p3 = 0; event moveout for event #3 in ms on reference offset

t3 = 200 intercept time for spike #3
t4 = 200 intercept time s for spike #4

cdp = 1 output cdp number
Notes:
Creates a common cdp su data file with up to four spike events
for impulse response studies for suradon, and sutifowler
Credits:
CWP: Shuki Ronen, Chris Liner,
291
Modified: CWP by John Anderson, April, 1994, to have
appropriate trace header words and default values
for SUTIFOWLER tests
292
SUNMO - NMO for an arbitrary velocity function of time and CDP
sunmo <stdin >stdout [optional parameters]
tnmo=0 NMO times corresponding to velocities in vnmo
vnmo=1500 NMO velocities corresponding to times in tnmo
anis1=0 two anisotropy coefficients making up quartic term
anis2=0 in traveltime curve, corresponding to times in tnmo
cdp= CDPs for which vnmo & tnmo are specified (see Notes)
smute=1.5 samples with NMO stretch exceeding smute are zeroed
lmute=25 length (in samples) of linear ramp for stretch mute
sscale=1 =1 to divide output samples by NMO stretch factor
invert=0 =1 to perform (approximate) inverse NMO
ixoffset=0 do not consider cross-line offset
=1 read cross-line offset from trace header
upward=0 =1 to scan upward to find first sample to kill
Notes:
For constant-velocity NMO, specify only one vnmo=constant and omit tnmo.
The anisotropy coefficients anis1, anis2 permit non-hyperbolicity due

to layering, mode conversion, or anisotropy. Default is isotropic NMO.
For NMO with a velocity function of time only, specify the arrays
vnmo=v1,v2,... tnmo=t1,t2,...
where v1 is the velocity at time t1, v2 is the velocity at time t2, ...
The times specified in the tnmo array must be monotonically increasing.
Linear interpolation and constant extrapolation of the specified velocities
is used to compute the velocities at times not specified.
The same holds for the anisotropy coefficients as a function of time only.
For NMO with a velocity function of time and CDP, specify the array
cdp=cdp1,cdp2,...
and, for each CDP specified, specify the vnmo and tnmo arrays as described
above. The first (vnmo,tnmo) pair corresponds to the first cdp, and so on.
Linear interpolation and constant extrapolation of 1/velocity^2 is used
to compute velocities at CDPs not specified.
The same holds for the anisotropy coefficients as a function of time and
CDP.
Moveout is defined by
293
1 anis1
--- x^2 + ------------- x^4.
v^2 1 + anis2 x^2
Note: In general, the user should set the cdp parameter. The default is
to use tr.cdp from the first trace and assume only one cdp.
Caveat:
Nmo cannot handle negative moveout as in triplication caused by
anisotropy. But negative moveout happens necessarily for negative anis1 at
sufficiently large offsets. Then the error-negative moveout- is printed.
Check anis1. An error (anis2 too small) is also printed if the
denominator of the quartic term becomes negative. Check anis2. These errors
are prompted even if they occur in traces which would not survive the
NMO-stretch threshold. Chop off enough far-offset traces (e.g. with suwind)
if anis1, anis2 are fine for near-offset traces.
NMO interpolation error is less than 1% for frequencies less than 60% of
the Nyquist frequency.
Exact inverse NMO is impossible, particularly for early times at large

offsets and for frequencies near Nyquist with large interpolation errors.
Credits:
SEP: Shuki, Chuck Sword
CWP: Shuki, Jack, Dave Hale, Bjoern Rommel
Modified: 08/08/98 - Carlos E. Theodoro - option for lateral offset
Modified: 07/11/02 - Sang-yong Suh -
added "upward" option to handle decreasing velocity function.
Proc. IEEE, v. 72, n. 10, p. 1238-1254
1984
Trace header fields accessed: ns, dt, delrt, offset, cdp, sy
294
SUNORMALIZE - Trace balancing by rms, max, or median ",
sunormalize <stdin >stdout t0=0 t1=TMAX norm=rms
dt=tr.dt if not set in header, dt is mandatory
ns=tr.ns if not set in header, ns is mandatory
norm=rms type of norm rms, max, med
t0=0.0 startimg time for window
t1=TMAX ending time for window
Author: Ramone Carbonell,

Inst. Earth Sciences-CSIC Barcelona, Spain, April 1998.
Modifications: Nils Maercklin,
RISSC, University of Naples, Italy, September 2006
(fixed user input of ns, dt, if values are not set in header).

295
SUNULL - create null (all zeroes) traces
sunull nt= [optional parameters] >outdata
Required parameter
nt= number of samples per trace
Optional parameters
ntr=5 number of null traces to create
Rationale: It is sometimes useful to insert null traces

between "panels" in a shell loop.
See also: sukill, sumute, suzero
Credits:
CWP: Jack K. Cohen
Trace header fields set: ns, dt, tracl
296
SUOCEXT - smaller Offset EXTrapolation via Offset Continuation
method for common-offset gathers
suocext <stdin >stdout cdpmin= cdpmax= dxcdp= noffmix= offextr= [...]
cdpmin= minimum cdp (integer number) for which to apply DMO
cdpmax= maximum cdp (integer number) for which to apply DMO
dxcdp= distance between adjacent cdp bins (m)
noffmix= number of offsets to mix (see notes)
offextr= offset to extrapolate
else use tmpfile()
Notes:
headers changes. One common-offset gather usually is enough.
The times in tdmo must be monotonically increasing.
297
A term for better amplitude reconstruction was added to Hale’s formulation.
Credits: Carlos E. Theodoro (modification of Hale’s SUDMOFK program)
C. Theodoro & K. Larner, 1998
Extrapolation of seismic data to small offsets (CWP-276).

Dave Hale, 1988
Bleistein, Cohen & Jaramillo, 1997

True amplitude transformation to zero offset of data from
curved reflectors (CWP-262).

Trace header fields modified: offset.
298
SUOLDTONEW - convert existing su data to xdr format
suoldtonew <oldsu >newsu
none
none
Notes:
This program is used to convert native machine datasets to
xdr-based, system-independent format.
Author: Stewart A. Levin, Mobil, 1966
299
SUOP2 - do a binary operation on two data sets
suop2 data1 data2 op=diff [trid=111] >stdout
none
Optional parameter:
op=diff difference of two panels of seismic data
=sum sum of two panels of seismic data
=prod product of two panels of seismic data
=quo quotient of two panels of seismic data
=ptdiff differences of a panel and single trace
=ptsum sum of a panel and single trace
=ptprod product of a panel and single trace
=ptquo quotient of a panel and single trace
=zipper do "zipper" merge of two panels
trid=FUNPACKNYQ output trace identification code. (This option

is active only for op=zipper)
For SU version 39 FUNPACNYQ=111
(See: sukeyword trid)
Note1: Output = data1 "op" data2 with the header of data1
Note2: For convenience and backward compatibility, this

program may be called without an op code as:
For: panel "op" panel operations:

susum file1 file2 == suop2 file1 file2 op=sum
sudiff file1 file2 == suop2 file1 file2 op=diff
suprod file1 file2 == suop2 file1 file2 op=prod
suquo file1 file2 == suop2 file1 file2 op=quo
For: panel "op" trace operations:

suptsum file1 file2 == suop2 file1 file2 op=ptsum
suptdiff file1 file2 == suop2 file1 file2 op=ptdiff
suptprod file1 file2 == suop2 file1 file2 op=ptprod
suptquo file1 file2 == suop2 file1 file2 op=ptquo
Note3: If an explicit op code is used it must FOLLOW the

filenames.
300
Note4: With op=quo and op=ptquo, divide by 0 is trapped and 0 is returned.
Note5: Weighted operations can be specified by setting weighting

coefficients for the two datasets:
w1=1.0
w2=1.0
Note6: With op=zipper, it is possible to set the output trace id code

(See: sukeyword trid)
This option processes the traces from two files interleaving its samples.
Both files must have the same trace length and must not longer than
SU_NFLTS/2 (as in SU 39 SU_NFLTS=32768).
Being "tr1" a trace of data1 and "tr2" the corresponding trace of

data2, The merged trace will be :
tr[2*i]= tr1[i]
tr[2*i+1] = tr2[i]
The default value of output tr.trid is that used by sufft and suifft,
which is the trace id reserved for the complex traces obtained through
the application of sufft. See also, suamp.
For operations on non-SU binary files use:farith
Credits:
SEP: Shuki Ronen
CWP: Jack K. Cohen
CWP: John Stockwell, 1995, added panel op trace options.
: Fernando M. Roxo da Motta <[email protected]> - added zipper op
Notes:
and repeating operation code within the branches of the switch.
301
SUOP - do unary arithmetic operation on segys
suop <stdin >stdout op=abs
none
Optional parameter:
op=abs operation flag
abs : absolute value
avg : remove average value
ssqrt : signed square root
sqr : square
ssqr : signed square
sgn : signum function
exp : exponentiate
slog : signed natural log
slog10: signed common log
cos : cosine
sin : sine
tan : tangent
cosh : hyperbolic cosine
sinh : hyperbolic sine
tanh : hyperbolic tangent
norm : divide trace by Max. Value
db : 20 * slog10 (data)
neg : negate value
posonly : pass only positive values
negonly : pass only negative values
sum : running sum trace integration
diff : running diff trace differentiation
refl : (v[i+1] - v[i])/(v[i+1] + v[i])
mod2pi : modulo 2 pi
inv : inverse
rmsamp : rms amplitude
nop : no operation
Note: Binary ops are provided by suop2.

Operations inv, slog and slog10 are "punctuated", ",
meaning that if, the input contains 0 values,
0 values are returned. ",
For file operations on non-SU format binary files use: farith
302
Credits:
CWP: Shuki, Jack

Toralf Foerster: norm and db operations, 10/95.
Notes:
and repeating operation code within the branches of the switch.
Note on db option. The following are equivalent:

... | sufft | suamp | suop op=norm | suop op=slog10 |\
sugain scale=20| suxgraph style=normal
... | sufft | suamp | suop op=db | suxgraph style=normal
303
SUPACK1 - pack segy trace data into chars
supack1 <segy_file >packed_file gpow=0.5
none
Optional parameter:
gpow=0.5 exponent used to compress the dynamic
range of the traces
Credits:
CWP: Jack K. Cohen, Shuki Ronen, Brian Sumner
Caveats:
This program is for single site use. Use segywrite to make
a portable tape.
We are storing the local header words, ungpow and unscale,

required by suunpack1 as floats. Although not essential
(compare the handling of such fields as dt), it allows us
to demonstrate the convenience of using the natural data type.
In any case, the data itself is non-portable floats in general,
so we aren’t giving up any intrinsic portability.
Notes:
ungpow and unscale are defined in segy.h
trid = CHARPACK is defined in su.h and segy.h

Trace header fields modified: ungpow, unscale, trid
304
SUPACK2 - pack segy trace data into 2 byte shorts
supack2 <segy_file >packed_file gpow=0.5
none
Optional parameter:
gpow=0.5 exponent used to compress the dynamic
range of the traces
Credits:
Revised: 7/4/95 Stewart A. Levin Mobil

Changed encoding to ensure 2 byte length (short is
8 bytes on Cray).
Caveats:
This program is for single site use. Use segywrite to make
a portable tape.
We are storing the local header words, ungpow and unscale,

required by suunpack2 as floats.
Notes:
trid = SHORTPACK is defined in su.h and segy.h

Trace header fields modified: ungpow, unscale, trid
305
SUPASTE - paste existing SEGY headers on existing data
supaste <bare_data >segys ns= head=headers ftn=0
Required parameter:
ns=the number of samples per trace
head=headers file with segy headers
ftn=0 Fortran flag
0 = unformatted data from C
1 = ... from Fortran
verbose=0 1= echo number of traces pasted
Caution:
An incorrect ns field will munge subsequent processing.
Notes:
This program is used when the option head=headers is used in
sustrip. See: sudoc sustrip for more details
Credits:
CWP: Jack K. Cohen
306
SUPEF - Wiener predictive error filtering
supef <stdin >stdout [optional parameters]
dt is mandatory if not set in header
cdp= CDPs for which minlag, maxlag, pnoise, mincorr,
maxcorr are set (see Notes)
minlag=dt first lag of prediction filter (sec)
maxlag=last lag default is (tmax-tmin)/20
pnoise=0.001 relative additive noise level
mincorr=tmin start of autocorrelation window (sec)
maxcorr=tmax end of autocorrelation window (sec)
showwiener=0 =1 to show Wiener filter on each trace
mix=1,... array of weights (floats) for moving
average of the autocorrelations
outpar=/dev/null output parameter file, contains the Wiener filter
if showwiener=1 is set
method=linear for linear interpolation of cdp values
=mono for monotonic cubic interpolation of cdps
=akima for Akima’s cubic interpolation of cdps
=spline for cubic spline interpolation of cdps

Notes:
1) To apply spiking decon (Wiener filtering with no gap):
Run the following command
suacor < data.su | suximage perc=95
You will see horizontal strip running across the center of your plot.
This is the autocorrelation wavelet for each trace. The idea of spiking
decon is to apply a Wiener filter with no gap to the data to collapse
the waveform into a spike. The idea is to pick the width of the
autocorrelation waveform _from beginning to end_ (not trough to trough)
and use this time for MAXLAG_SPIKING:
307
supef < data.su maxlag=MAXLAG_SPIKING > dataspiked.su
2) Prediction Error Filter (i.e. gapped Wiener filtering)

The purpose of gapped decon is to suppress repetitions in the data
such as those caused by water bottom multiples.
To look for the period of the repetitions
suacor ntout=1000 < dataspiked.su | suximage perc=95
The value of ntout must be larger than the default 100. The idea is
to look for repetitions in the autocorrelation. These repetitions will
appear as a family of parallel stripes above and below the main
autocorrelation waveform. Set MAXLAG_PEF to the period of the repetitions
Set MINLAG_PEF to be slightly larger than the value of MAXLAG_SPIKING
that you used to spike the data.
supef < dataspiked.su minlag=MINLAG_PEF maxlag=MAXLAG_PEF > datapef.su
Some experimentation may be required to get a satisfactory result.
3) It may be effective to sort your data into cdp gathers with susort,
and perform sunmo correction to the water speed with sunmo, prior to
attempts to suppress water bottom multiples. After applying supef, the
user should apply inverse nmo to undo the nmo to water speed prior to
further processing.
For a filter expressed as a function of cdp, specify the array

cdp=cdp1,cdp2,...
and for each cdp specified, specify the minlag and maxlag arrays as
minlag=min1,min2,... maxlag=max1,max2,...
It is required that the number of minlag and maxlag values be equal to

the number of cdp’s specified. If the number of
values in these arrays does not equal the number of cdp’s, only the first
value will be used.
Caveat:
The showweiner=1 option writes out the wiener filter to outpar, and
the prediction error filter to stdout, which is ",
1,0,0,...,-wiener[0],...,-wiener[imaxlag-1]
where the sample value of -wiener[0], is iminlag in the pe-filter.
The pe-filter is output as a SU format datafile, one pe-filter for each
308
trace input.
Credits:
CWP: Shuki Ronen, Jack K. Cohen, Ken Larner
CWP: John Stockwell, added mixing feature (April 1998)
CSM: Tanya Slota (September 2005) added cdp feature
A. Ziolkowski, "Deconvolution", for value of maxlag default:
page 91: imaxlag < nt/10. I took nt/20.
Notes:
The prediction error filter is 1,0,0...,0,-wiener[0], ...,
so no point in explicitly forming it.
If imaxlag < 2*iminlag - 1, then we don’t need to compute the

autocorrelation for lags:
imaxlag-iminlag+1, ..., iminlag-1
It doesn’t seem worth the duplicated code to implement this.
309
SUPERMUTE - permute or transpose a 3d datacube
supermute <stdin >sdout
none
n1=ns from header number of samples in the fast direction
n2=ntr from header number of samples in the med direction ",
n3=1 number of samples in the slow direction
o1=1 new fast direction

o2=2 new med direction
o3=3 new slow direction
d1=1 output interval in new fast direction

d2=1 output interval in new med direction
d3=1 output interval in new slow direction
Credits:
VT: Matthias Imhof
Trace header fields accessed: ns, ntr

Trace header fields modified: d1=1, f1=1, d2=1, f2=1, ns, ntr
310
SUPGC - Programmed Gain Control--apply agc like function
but the same function to all traces preserving
relative amplitudes spatially.
Required parameter:
file= name of input file
ntrscan=200 number of traces to scan for gain function
lwindow=1.0 length of time window in seconds
Author: John Anderson (visitor to CWP from Mobil)
311
SUPICKAMP - pick amplitudes within user defined and resampled window
supickamp <stdin >stdout d2= [optional parameters]
d2= sampling interval for slow dimension
(required if key-parameter not specified)
key= Key header word specifying trace offset
(alternatively, specify d2,x2beg)
x_above= array of lateral position values

(upper window corner)
t_above= array of time values

t_xabove= file containing time and lateral position values
t_xbelow= file containing time and lateral position values
(lower window corner)
wl= window width if t_xbelow is not specified
(No windowing if not specified)
dt_resamp=dt resampling interval within pick window

(dt has to come from trace headers)
tmin=0 minimum time in input trace
x2beg=0 first lateral position
format=ascii write ascii data to stdout
=binary for binary floats to stdout
verbose=1 writes complete pick information into outpar
=2 writes complete pick information into outpar
in tab-delimited column format
outpar=/dev/tty output parameter file; contains output
from verbose
arg1=max output (first dimension) to stdout
arg2=i2 output (second dimension) to stdout
(see notes for other options)
Notes:
Window can be defined using

(1) vectors x_above, t_above, [wl]
(2) file t_xabove, [wl] or
312
(3) files t_xabove, t_xbelow
files t_xabove, t_xbelow can be generated using xwigb’s picking

algorithm. The lateral positions have to be monotonically increasing
or decreasing for both vector and file input.
verbose=1 or 2 writes min, max, abs[max], energy and associated times
tmin,tmax,tabs to outpar, together with global values. verbose=0
only outputs global values.
Acceptable arg-parameters for lateral positions are
(1) x2 (2) i2 = trace number
If key=keyword is set, then the values of x2 are taken from the header
field represented by the keyword (for example key=offset)
Type sukeyword -o to see the complete list of SU keywords.
Credits:
CWP: Andreas Rueger July 06, 1996

MTU: David Forel, Jan. 26, 2005 Add verbose=2 option
313
SUPLANE - create common offset data file with up to 3 planes
suplane [optional parameters] >stdout
npl=3 number of planes
taper=0 no end-of-plane taper
= 1 taper planes to zero at the end
offset=400 offset
dt=0.004 time sample interval in seconds
...plane 1 ...
dip1=0 dip of plane #1 (ms/trace)
len1= 3*ntr/4 HORIZONTAL extent of plane (traces)
ct1= nt/2 time sample for center pivot
cx1= ntr/2 trace for center pivot
...plane 2 ...
...plane 3 ...
liner=0 use parameters

= 1 parameters set for 64x64 data set
with separated dipping planes.
Credits:
CWP: Chris Liner
Trace header fields set: ns, dt, offset, tracl
314
SUPOFILT - POlarization FILTer for three-component data
supofilt <stdin >stdout [optional parameters]
dfile=polar.dir file containing the 3 components of the
direction of polarization
wfile=polar.rl file name of weighting polarization parameter
smooth=1 1 = smooth filter operators, 0 do not
sl=0.05 smoothing window length in seconds
wpow=1.0 raise weighting function to power wpow
dpow=1.0 raise directivity functions to power dpow
Notes:
dataset.
This program SUPOFILT is an extension to the polarization analysis

program supolar. The files wfile and dfile are SU files as written
by SUPOLAR.

GeoForschungsZentrum (GFZ) Potsdam, Germany, 1999-2000.
References:
Benhama, A., Cliet, C. and Dubesset, M., 1986: Study and
Application of spatial directional filtering in three
component recordings.
Geophysical Prospecting, vol. 36.
Kanasewich, E. R., 1981: Time Sequence Analysis in Geophysics,
The University of Alberta Press.
Kanasewich, E. R., 1990: Seismic Noise Attenuation,
Handbook of Geophysical Exploration, Pergamon Press, Oxford.
315
316
SUPOLAR - POLarization analysis of three-component data
supolar <stdin [optional parameters]
none
wl=0.1 correlation window length in seconds
win=boxcar correlation window shape, choose "boxcar",
"hanning", "bartlett", or "welsh
file=polar base of output file name(s)
rl=1 1 = rectilinearity evaluating 2 eigenvalues,
2, 3 = rectilinearity evaluating 3 eigenvalues
rlq=1.0 contrast parameter for rectilinearity
dir=1 1 = 3 components of direction of polarization
(the only three-component output file)
tau=0 1 = global polarization parameter
ellip=0 1 = principal, subprincipal, and transverse
ellipticities e21, e31, and e32
pln=0 1 = planarity measure
f1=0 1 = flatness or oblateness coefficient
l1=0 1 = linearity coefficient
amp=0 1 = amplitude parameters: instantaneous,
quadratic, and eigenresultant ir, qr, and er
theta=0 1, 2, 3 = incidence angle of principal axis
phi=0 1, 2, 3 = horizontal azimuth of principal axis
angle=rad unit of angles theta and phi, choose "rad",
"deg", or "gon
all=0 1, 2, 3 = set all output flags to that value
Notes:
dataset.
Correct calculation of angles theta and phi requires the first of
these traces to be the vertical component, followed by the two
horizontal components (e.g. Z, N, E, or Z, inline, crossline).
Significant signal energy on Z is necessary to resolve the 180 deg
ambiguity of phi (options phi=2,3 only).
317
Each calculated polarization attribute is written into its own
SU file. These files get the same base name (set with "file=")
and the parameter flag as an extension (e.g. polar.rl).
In case of a tapered correlation window, the window length wl may

have to be increased compared to the boxcar case, because of their
smaller effective widths (Bartlett, Hanning: 1/2, Welsh: 1/3).
Range of values:
parameter option interval
rl 1, 2 0.0 ... 1.0 (1.0: linear polarization)
rl 3 -1.0 ... 1.0
tau, l1 1 0.0 ... 1.0 (1.0: linear polarization)
pln, f1 1 0.0 ... 1.0 (1.0: planar polarization)
e21, e31, e32 1 0.0 ... 1.0 (0.0: linear polarization)
theta 1 -pi/2... pi/2 rad
theta 2, 3 0.0 ... pi/2 rad
phi 1 -pi/2... pi/2 rad
phi 2 -pi ... pi rad (see notes above)
phi 3 0.0 ... 2 pi rad (see notes above)

GeoForschungsZentrum (GFZ) Potsdam, Germany, 1998-2001.
References:
Jurkevics, A., 1988: Polarization analysis of three-component
array data. Bulletin of the Seismological Society of America,
vol. 78, no. 5.
Kanasewich, E. R., 1981: Time Sequence Analysis in Geophysics.
The University of Alberta Press.
Kanasewich, E. R., 1990: Seismic Noise Attenuation.
Handbook of Geophysical Exploration, Pergamon Press, Oxford.
Meyer, J. H. 1988: First Comparative Results of Integral and
Instantaneous Polarization Attributes for Multicomponent Seismic
Data. Institut Francais du Petrole.
Press, W. H., Teukolsky, S. A., Vetterling, W. T., and Flannery, B. P.
1996: Numerical Recipes in C - The Art of Scientific Computing.
Cambridge University Press, Cambridge.
318
Samson, J. C., 1973: Description of the Polarisation States of Vector
Processes: Application to ULF Electromagnetic Fields.
Geophysical Journal vol. 34, p. 403-419.
Sheriff, R. E., 1991: Encyclopedic Dictionary of Exploration
Geophysics. 3rd ed., Society of Exploration Geophysicists, Tulsa.

319
SUPUT - Connect SU program to file descriptor for output stream.
su_module | suput fp=$1
This program is for interfacing " outside processing systems

with SU. Typically, the outside system would execute the SU command file.
The outside system provides the file descriptor it would like to read
from to the command file to be an argument for suput.
Example: su_module | suput fp=$1
fd=-1 file_descriptor_for_output_stream_from_su
verbose=0 minimal listing
=1 asks for message with each trace processed.
Author: John Anderson (visiting scholar from Mobil) July 1994
320
SUPWS - Phase stack or phase-weighted stack (PWS) of adjacent traces
having the same key header word
supws <stdin >stdout [optional parameters]
none
key=cdp key header word to stack on
pwr=1.0 raise phase stack to power pwr
sl=0.0 window length in seconds used for smoothing
of the phase stack (weights)
ps=0 0 = output is PWS, 1 = output is phase stack
Note:
Phase weighted stacking is a tool for efficient incoherent noise
reduction. An amplitude-unbiased coherency measure is designed
based on the instantaneous phase, which is used to weight the
samples of an ordinary, linear stack. The result is called the
phase-weighted stack (PWS) and is cleaned from incoherent noise.
PWS thus permits detection of weak but coherent arrivals.
The phase-stack (coherency measure) has values between 0 and 1.
If the stacking is over cdp and the PWS option is set, then the
offset header field is set to zero. Otherwise, output traces get
their headers from the first trace of each data ensemble to stack,
including the offset field. Use "sushw" afterwards, if this is
not acceptable.

References:
B. L. N. Kennett, 2000: Stacking three-component seismograms.
Geophysical Journal International, vol. 141, p. 263-269.
M. Schimmel and H. Paulssen, 1997: Noise reduction and detection
321
of weak , coherent signals through phase-weighted stacks.
Geophysical Journal International, vol. 130, p. 497-505.
M. T. Taner, A. F. Koehler, and R. E. Sheriff, 1979: Complex
seismic trace analysis. Geophysics, vol. 44, p. 1041-1063.

Trace header fields modified: nhs, offset
322
SUQUANTILE - display some quantiles or ranks of a data set
suquantile <stdin >stdout [optional parameters]
none (no-op)
panel=1 flag; 0 = do trace by trace (vs. whole data set)
quantiles=1 flag; 0 = give ranks instead of quantiles

Credits:
CWP: Jack K. Cohen
Trace header fields accessed: ns, tracl, mark
323
SURADON - compute forward or reverse Radon transform or remove multiples
by using the parabolic Radon transform to estimate multiples
and subtract.
suradon <stdin >stdout [Optional Parameters]
choose=0 0 Forward Radon transform
1 Compute data minus multiples
2 Compute estimate of multiples
3 Compute forward and reverse transform
4 Compute inverse Radon transform
igopt=1 1 parabolic transform: g(x) = offset**2
2 Foster/Mosher psuedo hyperbolic transform
g(x) = sqrt(depth**2 + offset**2)
3 Linear tau-p: g(x) = offset
4 abs linear tau-p: g(x) = abs(offset)
offref=2000. reference maximum offset to which maximum and minimum
moveout times are associated
interoff=0. intercept offset to which tau-p times are associated
pmin=-200 minimum moveout in ms on reference offset
pmax=400 maximum moveout in ms on reference offset
dp=16 moveout increment in ms on reference offset
pmula=80 moveout in ms on reference offset where multiples begin
at maximum time
pmulb=200 moveout in ms on reference offset where multiples begin
at zero time
depthref=500. Reference depth for Foster/Mosher hyperbolic transform
nwin=1 number of windows to use through the mute zone
f1=60. High-end frequency before taper off
f2=80. High-end frequency
prewhite=0.1 Prewhitening factor in percent.
cdpkey=cdp name of header word for defining ensemble
offkey=offset name of header word with spatial information
nxmax=120 maximum number of input traces per ensemble
ltaper=7 taper (integer) for mute tapering function
Optimizing Parameters:
The following parameters are occasionally used to avoid spatial aliasing
problems on the linear tau-p transform. Not recommended for other
transforms...
ninterp=0 number of traces to interpolate between each input trace
prior to computing transform
324
freq1=3.0 low-end frequency in Hz for picking (good default: 3 Hz)
(Known bug: freq1 cannot be zero)
freq2=20.0 high-end frequency in Hz for picking (good default: 20 Hz)
lagc=400 length of AGC operator for picking (good default: 400 ms)
lent=5 length of time smoother in samples for picker
(good default: 5 samples)
lenx=1 length of space smoother in samples for picker
(good default: 1 sample)
xopt=1 1 = use differences for spatial derivative
(works with irregular spacing)
0 = use FFT derivative for spatial derivatives
(more accurate but requires regular spacing and
at least 16 input tracs--will switch to differences
automatically if have less than 16 input traces)
Credits:
CWP: John Anderson (visitor to CSM from Mobil) Spring 1993
Multiple removal notes:

Usually the input data are NMO corrected CMP gathers. The
first pass is to compute a parabolic Radon transform and
identify the multiples in the transform domain. Then, the
module is run on all the data using "choose=1" to estimate
and subtract the multiples. See the May, 1993 CWP Project
Review for more extensive documentation.
NWIN notes:
The parabolic transform runs with higher resolution if the
mute zone is honored. When "nwin" is specified larger than
one (say 6), then multiple windows are used through the mute
zone. It is assumed in this case that the input data are
sorted by the offkey header item from small offset to large
offset. This causes the code to run 6 times longer. The
mute time is taken from the "muts" header word.
You may have to manually set this header field yourself, if
it is not already set.
References:
Anderson, J. E., 1993, Parabolic and linear 2-D, tau-p transforms
using the generalized radon tranform, in May 11-14, 1993
Project Review, Consortium Project on Seismic Inverse methods
for Complex Structures, CWP-137, Center for Wave Phenomena
325
internal report.
Other References cited in above paper:
Beylkin, G,.1987, The discrete Radon transform: IEEE Transactions
of Acoustics, Speech, and Signal Processing, 35, 162-712.
Chapman, C.H.,1981, Generalized Radon transforms and slant stacks:
Geophysical Journal of the Royal Astronomical Society, 66,
445-453.
Foster, D. J. and Mosher, C. C., 1990, Multiple supression
using curvilinear Radon transforms: SEG Expanded Abstracts 1990,
1647-1650.
Foster, D. J. and Mosher, C. C., 1992, Suppression of multiples
using the Radon transform: Geophysics, 57, No. 3, 386-395.
Gulunay, N., 1990, F-X domain least-squares Tau-P and Tau-Q: SEG
Expanded Abstracts 1990, 1607-1610.
Hampson, D., 1986, Inverse velocity stacking for multiple elimination:
J. Can. Soc. Expl. Geophs., 22, 44-55.
Hampson, D., 1987, The discrete Radon transform: a new tool for image
enhancement and noise suppression: SEG Expanded Abstracts 1978,
141-143.
Johnston, D.E., 1990, Which multiple suppression method should I use?
SEG Expanded Abstracts 1990, 1750-1752.
Trace header words accessed: ns, dt, cdpkey, offkey, muts
326
SURAMP - Linearly taper the start and/or end of traces to zero.
suramp <stdin >stdout [optional parameters]
Optional parameters
tmin=tr.delrt/1000 end of starting ramp (sec)
tmax=(nt-1)*dt beginning of ending ramp (sec)
dt = (from header) sampling interval (sec)
The taper is a linear ramp from 0 to tmin and/or tmax to the

end of the trace. Default is a no-op!
Credits:
CWP: Jack K. Cohen, Ken Larner
327
SURANDSPIKE - make a small data set of RANDom SPIKEs
surandspike [optional parameters] > out_data_file
Creates a common offset su data file with random spikes
n1=500 number of time samples
n2=200 number of traces
nspk=20 number of spikes per trace
amax=0.2 abs(max) spike value
mode=1 different spikes on each trace
=2 same spikes on each trace
Credits:
ARAMCO: Chris Liner
Trace header fields set: ns, dt, offset
328
SURANDSTAT - Add RANDom time shifts STATIC errors to seismic traces
surandstat <stdin >stdout [optional parameters]
none
max=tr.dt maximum random time shift (ms)
scale=1.0 scale factor for shifts
Credits:
ARAMCO: Chris Liner 3/26/05
329
SURANGE - get max and min values for non-zero header entries
surange <stdin
key= Header key(s) to range (default=all)
Note: Gives partial results if interrupted
Credits:
Geocon: Garry Perratt (output one header per line;
option to specify headers to range;
added first & last values where min<max)
Based upon original by:
SEP: Stew Levin
CWP: Jack K. Cohen
Note: the use of "signal" is inherited from BSD days and may
break on some UNIXs. It is dicy in that the responsibility
for program termination is lateraled back to the main.
330
SURECIP - sum opposing offsets in prepared data (see below)
surecip <stdin >stdout
Sum traces with equal positive and negative offsets (i.e. assume
reciprocity holds).
Usage:
suabshw <data >absdata
susort cdp offset <absdata | surecip >sumdata
Note that this processing stream can be simply evoked by:
recip data sumdata
Credits:
SEP: Shuki Ronen
CWP: Jack Cohen
Caveat:
The assumption is that this operation is not a mainstay processing
item. Hence the recommended implemention via the ’recip’ shell
script. If it becomes a mainstay, then a much faster code can
quickly drummed up by incorporating portions of suabshw and
susort.

Trace header fields modified: nhs, tracl, sx, gx
331
SUREDUCE - convert traces to display in reduced time ",
sureduce <stdin >stdout rv=
dt=tr.dt if not set in header, dt is mandatory
rv=8.0 reducing velocity in km/sec ",
Note: Useful for plotting refraction seismic data.

To remove reduction, do:
suflip < reduceddata.su flip=3 | sureduce rv=RV > flip.su
suflip < flip.su flip=3 > unreduceddata.su
Trace header fields accessed: dt, ns, offset

Author: UC Davis: Mike Begnaud March 1995
Trace header fields accessed: ns, dt, offset
332
SUREFCON - Convolution of user-supplied Forward and Reverse
refraction shots using XY trace offset in reverse shot
surefcon <forshot sufile=revshot xy=trace offseted >stdout
sufile= file containing SU trace to use as reverse shot
xy= Number of traces offseted from the 1st trace in sufile
none

Notes:
This code implements the Refraction Convolution Section (RCS) method
of generalized reciprocal refraction traveltime analysis developed by
Derecke Palmer and Leoni Jones.
The time sampling interval on the output traces is half of that on the
traces in the input files.
Example:
surefcon <DATA sufile=DATA xy=1 | ...
Here, the su data file, "DATA", convolved the nth trace by

(n+xy)th trace in the same file
Credits: (based on suconv)

UNSW: D. Palmer, K.T. LEE
CAVEATS: no space-variable or time-variable capacity.
The more than one trace allowed in sufile is the
beginning of a hook to handle the spatially variant case.

Notes:
This code implements the refraction convolution
333
section (RCS) method
method described in:
Palmer, D, 2001a, Imaging refractors with the convolution section,

Geophysics 66, 1582-1589.
Palmer, D, 2001b, Resolving refractor ambiguities with amplitudes,
Geophysics 66, 1590-1593.
Exploration Geophysics (2005) 36, 1825

Butsuri-Tansa (Vol. 58, No.1)
Mulli-Tamsa (Vol. 8,
A simple approach to refraction statics with the
Generalized Main Reciprocal Method and the Refraction
Convolution Section Heading
by Derecke Palmer Leonie Jones
334
SURELANAN - REsiduaL-moveout semblance ANalysis for ANisotropic media
surelan refl= npicks= [optional parameters]
reflector file: reflec =
number of points in the reflector file =
nr1=51 number of r1-parameter samples
dr1=0.01 r1-parameter sampling interval
fr1=-0.25 first value of r1-parameter
nr2=51 number of r2-parameter samples
dr2=0.01 r2-parameter sampling interval
fr2=-0.25 first value of r2-parameter
dzratio=5 ratio of output to input depth sampling intervals
nsmooth=dzratio*2+1 length of semblance num and den smoothing window
verbose=0 =1 for diagnostic print on stderr
method=linear for linear interpolation of the interface
Note:
1. This program is part of Debashish Sarkar’s anisotropic model building
technique.
2. Input migrated traces should be sorted by cdp - surelan outputs a
group of semblance traces every time cdp changes. Therefore, the
output will be useful only if cdp gathers are input.
3. The residual-moveout semblance for cdp gathers is based
on z(h)*z(h) = z(0)*z(0) + r1*h^2 + r2*h^4/[h^2+z(0)^2] where z
depth and h is the half-offset.
335
SURELAN - compute residual-moveout semblance for cdp gathers based
on z(h)*z(h) = z(0)*z(0) + r*h*h where z depth and h offset.
surelan <stdin >stdout [optional parameters]
nr=51 number of r-parameter samples
dr=0.01 r-parameter sampling interval
fr=-0.25 first value of b-parameter
smute=1.5 samples with RMO stretch exceeding smute are zeroed
dzratio=5 ratio of output to input depth sampling intervals
nsmooth=dzratio*2+1 length of semblance num and den smoothing window
Note:
1. This program is part of Zhenyue Liu’s velocity analysis technique.
2. Input migrated traces should be sorted by cdp - surelan outputs a
group of semblance traces every time cdp changes. Therefore, the
output will be useful only if cdp gathers are input.
3. The parameter r may take negative values. The range of r can be
controlled by maximum of (z(h)*z(h)-z(0)*z(0))/(h*h)
336
SURESAMP - Resample in time
suresamp <stdin >stdout [optional parameters]
none
nt=tr.ns number of time samples on output
dt= time sampling interval on output
default is:
tr.dt/10^6 seismic data
tr.d1 non-seismic data
tmin= time of first sample in output
default is:
tr.delrt/10^3 seismic data
tr.f1 non-seismic data
rf= resampling factor;
if defined, set nt=nt_in*rf and dt=dt_in/rf
Example 1: (assume original data had dt=.004 nt=256)

sufilter <data f=40,50 amps=1.,0. |
suresamp nt=128 dt=.008 | ...
Using the resampling factor rf, this example translates to:
sufilter <data f=40,50 amps=1.,0. | suresamp rf=0.5 | ...
Note the typical anti-alias filtering before sub-sampling!
Example 2: (assume original data had dt=.004 nt=256)

suresamp <data nt=512 dt=.002 | ...
or use:
suresamp <data rf=2 | ...
Example 3: (assume original data had d1=.1524 nt=8192)

sufilter <data f=0,1,3,3.28 amps=1,1,1,0 |
Example 4: (assume original data had d1=.5 nt=4096)

337
Credits:
CWP: Dave (resamp algorithm), Jack (SU adaptation)
CENPET: Werner M. Heigl - modified for well log support
RISSC: Nils Maercklin 2006 - minor fixes, added rf option
Algorithm:
Resampling is done via 8-coefficient sinc-interpolation.
See "$CWPROOT/src/cwp/lib/intsinc8.c" for technical details.
Trace header fields accessed: ns, dt, delrt, d1, f1, trid
Trace header fields modified: ns, dt, delrt (only when set tmin)
d1, f1 (only when set tmin)
338
SURESSTAT - Surface consistent source and receiver statics calculation
suresstat <stdin [optional parameters]
ssol= output file source statics
rsol= output file receiver statics
ntraces= number of traces in input data set (must be correct!)
ntpick=50 maximum static shift (samples)
niter=5 number of iterations
nshot=240 largest shot number (fldr=1 to nshot)
nr=335 largest receiver number (tracf=1 to nr)
nc=574 maximum number of cmp’s (for array allocation)
sfold=96 maximum shot gather fold
rfold=96 maximum receiver gather fold
cfold=48 maximum cmp gather fold
sub=0 subtract super trace 1 from super trace 2 (=1)
sub=0 strongly biases static to a value of 0
mode=0 use global maximum in cross-correllation window
=1 choose the peak perc=percent smaller than the global max.
perc=10. percent of global max (used only for mode=1)
verbose=0 print diagnostic output (verbose=1)
Notes:
Estimates surface-consistent source and receiver statics, meaning that
there is one static correction value estimated for each shot and receiver
position.
The method employed here is based on the method of Ronen and Claerbout:
Geophysics 50, 2759-2767 (1985).
The input consists of moveout-corrected SU data sorted in shot gathers.

The output files are ascii files containing the source and receiver
statics, as a function of shot number (trace header fldr) and
receiver station number (trace header tracf).
The code builds a supertrace1 and supertrace2, which are subsequently

cross-correllated. The program then picks the time lag associated with
the largest peak in the cross-correllation according to two possible
criteria set by the parameter "mode". If mode=0, the maximum of the
cross-correllation window is chosen. If mode=1, the program will pick
339
a peak which is up to perc=percent smaller than the global maximum, but
closer to zero lag than the global maximum. (Choosing mode=0 is
recommended.)
The geometry can be irregular: the program simply computes a static

correction for each shot record (fldr=1 to fldr=nshot), with any missing
shots being assigned a static of 0. A static correction for each
receiver station (tracf=1 to tracf=nr) is calculated, with missing
receivers again assigned a static of 0. ",
to apply the static corrections, use sustatic with hdrs=3
Reference:
Ronen, J. and Claerbout, J., 1985, Surface-consistent residual statics

estimation by stack-power maximization: Geophysics, vol. 50,
2759-2767.
Credits:
CWP: Timo Tjan, 4 October 1994
rewritten by Thomas Pratt, USGS, Feb. 2000.
Trace header fields accessed: ns, dt, tracf, fldr, cdp
340
SUSHAPE - Wiener shaping filter
sushape <stdin >stdout [optional parameters]
w= vector of input wavelet to be shaped or ...
...or ...
wfile= ... file containing input wavelet in SU (SEGY trace) format
d= vector of desired output wavelet or ...
...or ...
dfile= ... file containing desired output wavelet in SU format
dt=tr.dt if tr.dt is not set in header, then dt is mandatory
nshape=trace length of shaping filter
pnoise=0.001 relative additive noise level
showshaper=0 =1 to show shaping filter
Notes:
Example of commandline input wavelets:

sushape < indata w=0,-.1,.1,... d=0,-.1,1,.1,... > shaped_data
sushape < indata wfile=inputwavelet.su dfile=desire.su > shaped_data
To get the shaping filters into an ascii file:

... | sushape ... showwshaper=1 2>file | ... (sh or ksh)
(... | sushape ... showshaper=1 | ...) >&file (csh)
Credits:
CWP: Jack Cohen
CWP: John Stockwell, added wfile and dfile options

341
SUSHIFT - shifted/windowed traces in time
sushift <stdin >stdout [tmin= ] [tmax= ]
tmin= min time to pass

tmax= max time to pass
dt= sample rate in microseconds
fill=0.0 value to place in padded samples
(defaults for tmin and tmax are calculated from the first trace.
verbose= 1 echos parameters to stdout
Background :
tmin and tmax must be given in seconds
In the high resolution single channel seismic profiling the sample

interval is short, the shot rate and the number of samples are high.
To reduce the file size the delrt time is changed during a profiling
trip. To process and display a seismic section a constant delrt is
needed. This program does this job.
The SEG-Y header variable delrt (delay in ms) is a short integer.

That’s why in the example shown below delrt is rounded to 123 !
... | sushift tmin=0.1234 tmax=0.2234 | ...
The dt= and fill= options are intended for manipulating velocity
volumes in trace format. In particular models which were hung
from the water bottom when created & which then need to have the
water layer added." ,
Author:
Toralf Foerster
Institut fuer Ostseeforschung Warnemuende
Sektion Marine Geologie
Seestrasse 15
D-18119 Rostock, Germany
Trace header fields accessed: ns, delrt

Trace header fields modified: ns, delrt
342
343
SUSHW - Set one or more Header Words using trace number, mod and
integer divide to compute the header word values or input
the header word values from a file
... compute header fields

sushw <stdin >stdout key=cdp,.. a=0,.. b=0,.. c=0,.. d=0,.. j=..,..
... or read headers from a binary file

sushw <stdin > stdout key=key1,.. infile=binary_file
Required Parameters for setting headers from infile:

key=key1,key2 ... is the list of header fields as they appear in infile
infile= binary file of values for field specified by
key1,key2,...
Optional parameters ():

key=cdp,... header key word(s) to set
a=0,... value(s) on first trace
b=0,... increment(s) within group
c=0,... group increment(s)
d=0,... trace number shift(s)
j=ULONG_MAX,ULONG_MAX,... number of elements in group
Notes:
Fields that are getparred must have the same number of entries as key
words being set. Any field that is not getparred is set to the default
value(s) above. Explicitly setting j=0 will set j to ULONG_MAX.
The value of each header word key is computed using the formula:
i = itr + d
val(key) = a + b * (i % j) + c * (i / j)
where itr is the trace number (first trace has itr=0, NOT 1)
Examples:
1. set every dt field to 4ms
sushw <indata key=dt a=4000 |...
2. set the sx field of the first 32 traces to 6400, the second 32 traces
to 6300, decrementing by -100 for each 32 trace groups
...| sushw key=sx a=6400 c=-100 j=32 |...
3. set the offset fields of each group of 32 traces to 200,400,...,6400
...| sushw key=offset a=200 b=200 j=32 |...
4. perform operations 1., 2., and 3. in one call
344
..| sushw key=dt,sx,offset a=4000,6400,200 b=0,0,200 c=0,-100,0 j=0,32,32 |
In this example, we set every dt field to 4ms. Then we set the first
32 shotpoint fields to 6400, the second 32 shotpoint fields to 6300 and
so forth. Next we set each group of 32 offset fields to 200, 400, ...,
6400.
Example of a typical processing sequence using suchw:

sushw <indata key=dt a=4000 |
sushw key=sx a=6400 c=-100 j=32 |
sushw key=offset a=200 b=200 j=32 |
suchw key1=gx key2=offset key3=sx b=1 c=1 |
suchw key1=cdp key2=gx key3=sx b=1 c=1 d=2 >outdata
Again, it is possible to eliminate the multiple calls to both sushw and

sushw, as in Example 4.
Reading header values from a binary file:

If the parameter infile=binary_file is set, then the values that are to
be set for the fields specified by key=key1,key2,... are read from that
file. The values are read sequentially from the file and assigned trace
by trace to the input SU data. The infile consists of C (unformated)
binary floats in the form of an array of size (nkeys)*(ntraces) where
nkeys is the number of floats in the first (fast) dimension and ntraces
is the number of traces.
Comment:
Users wishing to edit one or more header fields (as in geometry setting)
may do this via the following sequence:
sugethw < sudata output=geom key=key1,key2 ... > hdrfile
Now edit the ASCII file hdrfile with any editor, setting the fields
appropriately. Convert hdrfile to a binary format via:
a2b < hdrfile n1=nfields > binary_file
Then set the header fields via:
sushw < sudata infile=binary_file key1 key2 ... > sudata.edited
Caveat:
If the (number of traces)*(number of key words) exceeds the number of
values in the infile then the user may still set a single header field
on the remaining traces via the parameters key=keyword a,b,c,d,j.
Example:
sushw < sudata=key1,key2 ... infile=binary_file [Optional Parameters]
345
Credits:
SEP: Einar Kajartansson
CWP: Jack K. Cohen
CWP: John Stockwell, added multiple fields and infile= options
Caveat:
All constants are cast to doubles.
346
SUSORT - sort on any segy header keywords
susort <stdin >stdout [[+-]key1 [+-]key2 ...]
Susort supports any number of (secondary) keys with either

ascending (+, the default) or descending (-) directions for
each. The default sort key is cdp.
Note: Only the following types of input/output are supported

Disk input --> any output
Pipe input --> Disk output
Note: If the the CWP_TMPDIR environment variable is set use

Example:
To sort traces by cdp gather and within each gather
by offset with both sorts in ascending order:
susort <INDATA >OUTDATA cdp offset
Caveat: In the case of Pipe input a temporary file is made

to hold the ENTIRE data set. This temporary is
either an actual disk file (usually in /tmp) or in some
implementations, a memory buffer. It is left to the
user to be SENSIBLE about how big a file to pipe into
susort relative to the user’s computer.
Credits:
SEP: Einar, Stew
CWP: Shuki, Jack
Caveats:
Since the algorithm depends on sign reversal of the key value
to obtain a descending sort, the most significant figure may
be lost for unsigned data types. The old SEP support for tape
input was removed in version 1.16---version 1.15 is in the
Portability directory for those who may want to input SU data
stored on tape.
Trace header fields modified: tracl, tracr
347
SUSORTY - make a small 2-D common shot off-end
data set in which the data show geometry
values to help visualize data sorting.
susorty [optional parameters] > out_data_file
dshot=10 shot interval (m)
noff=20 number of offsets
doff=20 offset increment (m)
Notes:
Creates a common shot su data file for sort visualization
time samples quantity
---------------- ----------
first 25% shot coord
second 25% rec coord
third 25% offset
fourth 25% cmp coord
1. default is shot ordered (hsv2 cmap looks best to me)

susorty | suximage legend=1 units=meters cmap=hsv2
2. sort on cmp (note random order within a cmp)

susorty | susort cdp > junk.su
suximage < junk.su legend=1 units=meters cmap=hsv2
3. sort to cmp and subsort on offset

susorty | susort cdp offset > junk.su
suximage < junk.su legend=1 units=meters cmap=hsv2
Credits:
CWP: Chris Liner 10.09.01
Trace header fields set: ns, dt, sx, gx, offset, cdp, tracl
348
SUSPECFK - F-K Fourier SPECtrum of data set
suspecfk <infile >outfile [optional parameters]
dt=from header time sampling interval

dx=from header(d2) or 1.0 spatial sampling interval

Note: To facilitate further processing, the sampling intervals

in frequency and wavenumber as well as the first
frequency (0) and the first wavenumber are set in the
output header (as respectively d1, d2, f1, f2).
Note: The relation: w = 2 pi F is well known, but there

doesn’t seem to be a commonly used letter corresponding
to F for the spatial conjugate transform variable. We
use K for this. More specifically we assume a phase:
i(w t - k x) = 2 pi i(F t - K x).
and F, K define our notion of frequency, wavenumber.
Credits:
CWP: Dave (algorithm), Jack (reformatting for SU)

Trace header fields modified: tracl, ns, dt, trid, d1, f1, d2, f2
349
SUSPECFX - Fourier SPECtrum (T -> F) of traces
suspecfx <infile >outfile
Note: To facilitate further processing, the sampling interval

in frequency and first frequency (0) are set in the
output header.
Credits:
CWP: Dave (algorithm), Jack (reformatting for SU)

Trace header fields modified: ns, dt, trid, d1, f1
350
SUSPECK1K2 - 2D (K1,K2) Fourier SPECtrum of (x1,x2) data set
suspeck1k2 <infile >outfile [optional parameters]
d1=from header(d1) or 1.0 spatial sampling interval in first (fast)

dimension
d2=from header(d2) or 1.0 spatial sampling interval in second
(slow) dimension

Notes:
To facilitate further processing, the sampling intervals in wavenumber

as well as the first frequency (0) and the first wavenumber are set in
the output header (as respectively d1, d2, f1, f2).

-i(k1 x1 + k2 x2) = -2 pi i(K1 x1 + K2 x2).
Credits:
CWP: John Stockwell, 26 April 1995, based on original code by
Dave Hale and Jack Cohen
Trace header fields accessed: ns, d1, d2, trid

Trace header fields modified: tracl, ns, dt, trid, d1, f1, d2, f2
351
SUSPIKE - make a small spike data set
suspike [optional parameters] > out_data_file
Creates a common offset su data file with up to four spikes

for impulse response studies
offset=400 offset
nspk=4 number of spikes
ix1= ntr/4 trace number (from left) for spike #1
it1= nt/4 time sample to spike #1
ix2 = ntr/4 trace for spike #2
it2 = 3*nt/4 time for spike #2
ix3 = 3*ntr/4; trace for spike #3
it3 = nt/4; time for spike #3
ix4 = 3*ntr/4; trace for spike #4
it4 = 3*nt/4; time for spike #4
Credits:
CWP: Shuki Ronen, Chris Liner
Trace header fields set: ns, dt, offset
352
SUSPLIT - Split traces into different output files by keyword value
susplit <stdin >stdout [options]
none
key=cdp Key header word to split on (see segy.h)
stem=split_ Stem name for output files
middle=key middle of name of output files
suffix=.su Suffix for output files
numlength=3 Length of numeric part of filename
verbose=0 =1 to echo filenames, etc.
close=0 =1 to close files before opening new ones
Credits:
Geocon: Garry Perratt hacked together from various other codes
353
SUSTACK - stack adjacent traces having the same key header word
sustack <stdin >stdout [Optional parameters]
none
key=cdp header key word to stack on
normpow=1.0 each sample is divided by the
normpow’th number of non-zero values
stacked (normpow=0 selects no division)
repeat=0 =1 repeats the stack trace nrepeat times
nrepeat=10 repeats stack trace nrepeat times in
output file
verbose=0 verbose = 1 echos information
Notes:
------
The offset field is set to zero on the output traces, unless
the user is stacking with key=offset. In that case, the value
of the offset field is left unchanged.
Sushw can be used afterwards if this is not acceptable.
For VSP users:

The stack trace appears ten times in the output file when
setting repeat=1 and nrepeat=10. Corridor stacking can be
achieved by properly muting the upgoing data with SUMUTE
before stacking.
Credits:
CWP: Jack K. Cohen, Dave Hale
CENPET: Werner M. Heigl - added repeat trace functionality
Note:
The "valxxx" subroutines are in su/lib/valpkge.c. In particular,
"valcmp" shares the annoying attribute of "strcmp" that
if (valcmp(type, val, valnew) {
...
}
354
will be performed when val and valnew are different.

Trace header fields modified: nhs, tracl, offset
355
SUSTATICRRS - Elevation STATIC corrections, apply corrections from
headers or from a source and receiver statics file,
includes application of Residual Refraction Statics
sustaticrrs <stdin >stdout [optional parameters]
none
v0=v1 or user-defined or from header, weathering velocity
v1=user-defined or from header, subweathering velocity
hdrs=0 =1 to read statics from headers
=2 to read statics from files
sign=1 =-1 to subtract statics from traces(up shift)
Options when hdrs=2:
sou_file= input file for source statics (ms)
rec_file= input file for receiver statics (ms)
ns=240 number of sources
nr=335 number of receivers
no=96 number of offsets

blvl_file= base of the near-surface model file (sampled
at CMP locations)
refr_file= horizontal reference datum file (sampled at
CMP locations)
nsamp= number of midpoints on line
fx= first x location in velocity model
dx= midpoint interval
V_r= replacement velocity
mx= number of velocity model samples in
lateral direction
mz= number of velocity model samples in
vertical direction
dzv= velocity model depth interval
vfile= near-surface velocity model

nsamp= number of midpoints on line
fx= first x location in velocity model ",
dx= midpoint interval ",
356
none
Notes:
For hdrs=1, statics calculation is not performed, statics correction
is applied to the data by reading statics (in ms) from the header.
For hdrs=0, field statics are calculated, and

input field sut is assumed measured in ms.
output field sstat = 10^scalel*(sdel - selev + sdepth)/swevel
output field gstat = sstat - sut/1000.
output field tstat = sstat + gstat + 10^scalel*(selev - gelev)/wevel
For hdrs=2, statics are surface consistently obtained from the

statics files. The geometry should be regular.
The source- and receiver-statics files should be unformated C binary
floats and contain the statics (in ms) as a function of surface location.
For hdrs=3, residual refraction statics and average refraction statics

are computed. For hdrs=4, residual refraction statics are applied,
and for hdrs=5, average refraction statics are applied (Cox, 1999).
These three options are coupled in many data processing sequences:
before stack residual and average refraction statics are computed but
only residual refractions statics are applied, and after stack average
refraction statics are applied. Refraction statics are often split
like this to avoid biasing stacking velocities. The files blvl_file
and refr_file are the base of the velocity model defined in vfile and
the final reference datum, as described by Cox (1999), respectively.
Residual refraction statics are stored in the header field gstat, and
the average statics are stored in the header field tstat. V_r is the
replacement velocity as described by Cox (1999). The velocity file,
vfile, is designed to work with a horizontal upper surface defined in
refr_file. If the survey has irregular topography, the horizontal
upper surface should be above the highest topographic point on the
line, and the velocity between this horizontal surface and topography
should be some very large value, such as 999999999, so that the
traveltimes through that region are inconsequential.
Credits:
CWP: Jamie Burns
CWP: Modified by Mohammed Alfaraj, 11/10/1992, for reading

statics from headers and including sign (+-) option
357
CWP: Modified by Timo Tjan, 29 June 1995, to include input of
source and receiver statics from files.
CWP: Modified by Chris Robinson, 11/2000, to include the splitting

of refraction statics into residuals and averages
Trace header fields accessed: ns, dt, delrt, gelev, selev,

sdepth, gdel, sdel, swevel, sut, scalel
Trace header fields modified: sstat, gstat, tstat
References:
Cox, M., 1999, Static corrections for seismic reflection surveys:

Soc. Expl. Geophys.
358
SUSTATIC - Elevation static corrections, apply corrections from
headers or from a source and receiver statics file
sustatic <stdin >stdout [optional parameters]
none
v0=v1 or user-defined or from header, weathering velocity
v1=user-defined or from header, subweathering velocity
hdrs=0 =1 to read statics from headers
=2 to read statics from files
=3 to read from output files of suresstat
sign=1 =-1 to subtract statics from traces(up shift)
Options when hdrs=2 and hdrs=3:
sou_file= input file for source statics (ms)
rec_file= input file for receiver statics (ms)
ns=240 number of souces
nr=335 number of receivers
no=96 number of offsets
Notes:
For hdrs=1, statics calculation is not performed, statics correction
is applied to the data by reading statics (in ms) from the header.
For hdrs=0, field statics are calculated, and

input field sut is assumed measured in ms.
output field sstat = 10^scalel*(sdel - selev + sdepth)/swevel
output field gstat = sstat - sut/1000.
output field tstat = sstat + gstat + 10^scalel*(selev - gelev)/wevel
For hdrs=2, statics are surface consistently obtained from the

statics files. The geometry should be regular.
The source- and receiver-statics files should be unformated C binary
floats and contain the statics (in ms) as a function of surface location.
For hdrs=3, statics are read from the output files of suresstat, with
the same options as hdrs=2 (but use no=max traces per shot and assume
that ns=max shot number and nr=max receiver number).
For each shot number (trace header fldr) and each receiver number
(trace header tracf) the program will look up the appropriate static
correction. The geometry need not be regular as each trace is treated
independently.
359
Credits:
CWP: Jamie Burns
CWP: Modified by Mohammed Alfaraj, 11/10/1992, for reading

statics from headers and including sign (+-) option
CWP: Modified by Timo Tjan, 29 June 1995, to include input of

source and receiver statics from files.
modified by Thomas Pratt, USGS, Feb, 2000 to read statics from

the output files of suresstat
Trace header fields accessed: ns, dt, delrt, gelev, selev,

sdepth, gdel, sdel, swevel, sut, scalel, fldr, tracf
Trace header fields modified: sstat, gstat, tstat
360
SUSTKVEL - convert constant dip layer interval velocity model to the
stacking velocity model required by sunmo
sustkvel v= h= dip=0.0 outpar=/dev/tty
v= interval velocities
h= layer thicknesses at the cmp
dip=0.0 (constant) dip of the layers (degrees)
outpar=/dev/tty output parameter file in the form
required by sunmo:
tv=zero incidence time pick vector
v=stacking velocities vector
Examples:
sustkvel v=5000,6000,8000,10000 h=1000,1200,1300,1500 outpar=stkpar
sunmo <data.cdp par=stkpar >data.nmo
sustkvel par=intpar outpar=stkpar

sunmo <data.cdp par=stkpar >data.nmo
If the file, intpar, contains:

v=5000,6000,8000,10000
h=1000,1200,1300,1500
then the two examples are equivalent. The created parameter file,
stkpar, is in the form of the velocity model required by sunmo.
Note: sustkvel does not have standard su syntax since it does not
operate on seismic data. Hence stdin and stdout are not used.
Caveat: Does not accept a series of interval velocity models to

produce a variable velocity file for sunmo.
Credits:
CWP: Jack
Proc. IEEE, v. 72, n. 10, p. 1238-1254
361
1984
Formulas:
Note: All sums on i are from 1 to k
From Schneider:
Let h[i] be the ith layer thickness measured at the cmp and
v[i] the ith interval velocity.
Set:
t[i] = h[i]/v[i]
t0[k] = 2 Sum t[i] * cos(dip)
vs[k] = (1.0/cos(dip)) sqrt(Sum v[i]*v[i]*t[i] / Sum t[i])
Define:
t0by2[k] = Sum h[i]/v[i]
vh[k] = Sum v[i]*h[i]
Then:
t0[k] = 2 * t0by2[k] * cos(dip)
vs[k] = sqrt(vh[k] / t0by2[k]) / cos(dip)
362
SUSTOLT - Stolt migration for stacked data or common-offset gathers
sustolt <stdin >stdout cdpmin= cdpmax= dxcdp= noffmix= [...]
cdpmin minimum cdp (integer number) in dataset
cdpmax maximum cdp (integer number) in dataset
noffmix=1 number of offsets to mix (for unstacked data only)
tmig=0.0 times corresponding to rms velocities in vmig (s)
vmig=1500.0 rms velocities corresponding to times in tmig (m/s)
smig=1.0 stretch factor (0.6 typical if vrms increasing)
vscale=1.0 scale factor to apply to velocities
fmax=Nyquist maximum frequency in input data (Hz)
lstaper=0 length of side tapers (# of traces)
lbtaper=0 length of bottom taper (# of samples)
Notes:
If unstacked traces are input, they should be sorted into common-offset
gathers. One common-offset gather ends and another begins when the offset
field of the trace headers changes.
The number of offsets to mix (noffmix) should be specified for

unstacked data only. noffmix should typically equal the ratio of the
shotpoint spacing to the cdp spacing. This choice ensures that every
contribute through migration to other traces in adjacent cdps within
that mix.
The tmig and vmig arrays specify a velocity function of time that is
used to implement Stolt’s stretch for depth-variable velocity. The
stretch factor smig is often referred to as the "W" factor.
The times in tmig must be monotonically increasing.
363
Credits:
CWP: Dave Hale
Trace header fields accessed: ns, dt, delrt, offset, cdp
364
SUSTRIP - remove the SEGY headers from the traces
sustrip <stdin >stdout head=/dev/null outpar=/dev/tty ftn=0
none
head=/dev/null file to save headers in

number of samples (n1=)
number of traces (n2=)
sample rate in seconds (d1=)
ftn=0 Fortran flag

0 = write unformatted for C
1 = ... for Fortran
Notes:
Invoking head=filename will write the trace headers into filename.
You may paste the headers back onto the traces with supaste
See: sudoc supaste for more information
Credits:
CWP: Jack K. Cohen
365
SUSWAPBYTES - SWAP the BYTES in SU data to convert data from big endian
to little endian byte order, and vice versa
suswapbytes < stdin [optional parameter] > sdtout
format=0 foreign to native

=1 native to foreign
ns=from header if ns not set in header, must be set by hand
Notes:
The ’native’ endian is the endian (byte order) of the machine you are
running this program on. The ’foreign’ endian is the opposite byte order.
Examples of big endian machines are: IBM RS6000, SUN, NeXT

Examples of little endian machines are: PCs, DEC
Caveat: this code has not been tested on DEC
Credits:
CWP: adapted for SU by John Stockwell
based on a code supplied by:
Institute fur Geophysik, Hamburg: Jens Hartmann (June 1993)
366
SUSYNCZ - SYNthetic seismograms for piecewise constant V(Z) function
True amplitude (primaries only) modeling for 2.5D
susyncz > outfile [parameters]
none
ninf=4 number of interfaces (not including upper surface)
dip=5*i dips of interfaces in degrees (i=1,2,3,4)
zint=100*i z-intercepts of interfaces at x=0 (i=1,2,3,4)
v=1500+ 500*i velocities below surface & interfaces (i=0,1,2,3,4)
rho=1,1,1,1,1 densities below surface & interfaces (i=0,1,2,3,4)
nline=1 number of (identical) lines
dx=10 trace interval
tdelay=0 delay in recording time after source initiation
dt=0.004 time interval
Notes:
The original purpose of this code was to create some nontrivial
data for Brian Sumner’s CZ suite.
The program produces zero-offset data over dipping reflectors.
In the original fortran code, some arrays had the index

interval 1:ninf, as a natural way to index over the subsurface
reflectors. This indexing was preserved in this C translation.
Consequently, some arrays in the code do not use the 0 "slot".
Example:
susyncz | sufilter | sugain tpow=1 | display_program
Trace header fields set: tracl, ns, dt, delrt, ntr, sx, gx
Credits:
CWP: Brian Sumner, 1983, 1985, Fortran design and code
CWP: Stockwell & Cohen, 1995, translation to C
367
368
SUSYNLVCW - SYNthetic seismograms for Linear Velocity function
for mode Converted Waves
susynlvcw >outfile [optional parameters]
dxo=0.05 offset sampling interval (km)
fxo=0.0 first offset (km, see notes below)
xo=fxo,fxo+dxo,... array of offsets (use only for non-uniform offsets)
nxm=101 number of midpoints (see notes below)
dxm=0.05 midpoint sampling interval (km)
fxm=0.0 first midpoint (km)
nxs=101 number of shotpoints (see notes below)
dxs=0.05 shotpoint sampling interval (km)
fxs=0.0 first shotpoint (km)
x0=0.0 distance x at which v00 is specified
z0=0.0 depth z at which v00 is specified
v00=2.0 velocity at x0,z0 (km/sec)
gamma=1.0 velocity ratio, upgoing/downgoing
dvdx=0.0 derivative of velocity with distance x (dv/dx)
dvdz=0.0 derivative of velocity with depth z (dv/dz)
fpeak=0.2/dt peak frequency of symmetric Ricker wavelet (Hz)
ref="1:1,2;4,2" reflector(s): "amplitude:x1,z1;x2,z2;x3,z3;...
smooth=0 =1 for smooth (piecewise cubic spline) reflectors
er=0 =1 for exploding reflector amplitudes
ls=0 =1 for line source; default is point source
ob=1 =1 to include obliquity factors
sp=1 =1 to account for amplitude spreading
=0 for constant amplitudes throught out
tmin=10.0*dt minimum time of interest (sec)
ndpfz=5 number of diffractors per Fresnel zone
Notes:
Offsets are signed - may be positive or negative. Receiver locations

are computed by adding the signed offset to the source location.
Specify either midpoint sampling or shotpoint sampling, but not both.
369
If neither is specified, the default is the midpoint sampling above.
More than one ref (reflector) may be specified. When obliquity factors
are included, then only the left side of each reflector (as the x,z
reflector coordinates are traversed) is reflecting. For example, if x
coordinates increase, then the top side of a reflector is reflecting.
Note that reflectors are encoded as quoted strings, with an optional
reflector amplitude: preceding the x,z coordinates of each reflector.
Default amplitude is 1.0 if amplitude: part of the string is omitted.
Note that gamma<1 implies P-SV mode conversion, gamma>1 implies SV-P,
and gamma=1 implies no mode conversion.
based on Dave Hale’s code susynlv, but modified

by Mohammed Alfaraj to handle mode conversion
Date of modification: 01/07/92
Trace header fields set: trid, counit, ns, dt, delrt,

tracl. tracr, fldr, tracf,
cdp, cdpt, d2, f2, offset, sx, gx
370
SUSYNLVFTI - SYNthetic seismograms for Linear Velocity function in a
Factorized Transversely Isotropic medium
susynlvfti >outfile [optional parameters]
kilounits=1 input length units are km or kilo-feet
=0 for m or ft
Note: Output (sx,gx,offset) are always m or ft
dxo=0.05 offset sampling interval (kilounits)
fxo=0.0 first offset (kilounits, see notes below)
dxm=0.05 midpoint sampling interval (kilounits)
fxm=0.0 first midpoint (kilounits)
dxs=0.05 shotpoint sampling interval (kilounits)
fxs=0.0 first shotpoint (kilounits)
v00=2.0 velocity at x0,z0 (kilounits/sec)
ref=1:1,2;4,2 reflector(s): "amplitude:x1,z1;x2,z2;x3,z3;...
For transversely isotropic media:

angxs=0.0 angle of symmetry axis with the vertical (degrees)
define the media using either
a=1.0 corresponding to the ratio of elastic coef.(c1111/c3333)
f=0.4 corresponding to the ratio of elastic coef. (c1133/c3333)
l=0.3 corresponding to the ratio of elastic coef. (c1313/c3333)
371
Alternately use Tompson\’s parameters:
delta=0 Thomsen’s 1986 defined parameter
epsilon=0 Thomsen’s 1986 defined parameter
ntries=40 number of iterations in Snell’s law and offset searches
epsx=.001 lateral offset tolerance
epst=.0001 reflection time tolerance
nitmax=12 max number of iterations in travel time integrations
Notes:


More than one ref (reflector) may be specified. When obliquity factors
Concerning the choice of delta and epsilon. The difference between delta",
and epsilon should not exceed one. A possible break down of the program
is the result. This is caused primarly by the break down in the two point",
ray-tracing. Also keep the values of delta and epsilon between 2 and -2.
372
SUSYNLV - SYNthetic seismograms for Linear Velocity function
susynlv >outfile [optional parameters]
kilounits=1 input length units are km or kilo-feet
=0 for m or ft
Note: Output (sx,gx,offset) are always m or ft
dxo=0.05 offset sampling interval (kilounits)
fxo=0.0 first offset (kilounits, see notes below)
dxm=0.05 midpoint sampling interval (kilounits)
fxm=0.0 first midpoint (kilounits)
dxs=0.05 shotpoint sampling interval (kilounits)
fxs=0.0 first shotpoint (kilounits)
v00=2.0 velocity at x0,z0 (kilounits/sec)
Notes:

373
More than one ref (reflector) may be specified. Do this by putting
additional ref= entries on the commandline. When obliquity factors
Credits: CWP Dave Hale, 09/17/91, Colorado School of Mines

UTulsa Chris Liner 5/22/03 added kilounits flag

374
SUSYNVXZCS - SYNthetic seismograms of common shot in V(X,Z) media via
Kirchhoff-style modeling
susynvxzcs<vfile >outfile nx= nz= [optional parameters]
>outfile file containing seismograms of common ofset
nx= number of x samples (2nd dimension) in velocity
nxg= number of receivers of input traces
dxg=15 receiver sampling interval (m)
fxg=0.0 first receiver (m)
nxd=5 skipped number of receivers
nxs=1 number of offsets
dxs=50 shot sampling interval (m)
fxs=0.0 first shot (m)
dx=50 x sampling interval (m)
fx=0. first x sample (m)
dz=50 z sampling interval (m)
nxb=nx/2 band width centered at midpoint (see note)
nxc=0 hozizontal range in which velocity is changed
nzc=0 vertical range in which velocity is changed
pert=0 =1 calculate time correction from v_p[nx][nz]
vpfile file containing slowness perturbation array v_p[nx][nz]
ls=0 =1 for line source; =0 for point source
Notes:
This algorithm is based on formula (58) in Geo. Pros. 34, 686-703,
by N. Bleistein.
Traveltime and amplitude are calculated by finite difference which
375
is done only in one of every NXD receivers; in skipped receivers,
interpolation is used to calculate traveltime and amplitude. ",
For each receiver, traveltime and amplitude are calculated in the
horizontal range of (xg-nxb*dx, xg+nxb*dx). Velocity is changed by
constant extropolation in two upper trianglar corners whose width is
nxc*dx and height is nzc*dz.

There are three ways to remove the turned rays: smoothing velocity,
reducing nxb, and increaing nxc and nzc (if the turned ray occurs
in shallow areas). To prevent traveltime distortion from an over-
smoothed velocity, traveltime is corrected based on the slowness
perturbation.
More than one ref (reflector) may be specified.

Author: Zhenyue Liu, 07/20/92, Center for Wave Phenomena

Many subroutines borrowed from Dave Hale’s program: SUSYNLV
Trino Salinas, 07/30/96, fixed a bug in the geometry

setting to allow the spread move with the shots.

sx, gx
376
SUSYNVXZ - SYNthetic seismograms of common offset V(X,Z) media via
Kirchhoff-style modeling
susynvxz >outfile [optional parameters]
nxb=nx band centered at midpoint
nxd=1 skipped number of midponits
dx=100 x sampling interval (m)
dz=100 z sampling interval (m)
nxo=1 number of offsets
dxo=50 offset sampling interval (m)
fxo=0.0 first offset (m)
nxm=101 number of midpoints
dxm=50 midpoint sampling interval (m)
fxm=0.0 first midpoint (m)
Notes:
This algorithm is based on formula (58) in Geo. Pros. 34, 686-703,
by N. Bleistein.
Offsets are signed - may be positive or negative. ",

Traveltime and amplitude are calculated by finite differences which
is done only in part of midpoints; in the skiped midpoint, interpolation
is used to calculate traveltime and amplitude. ",
More than one ref (reflector) may be specified.

377
CWP: Zhenyue Liu, 07/20/92

Many subroutines borrowed from Dave Hale’s program: SUSYNLV

tracl. tracr,
378
SUTAB - print non zero header values and data for non-graphic terminals
sutab <stdin itmin=0 itmax=last_sample count=all
none
itmin=0 first time sample (zero-based) to plot
itmax= (last sample) last time sample (zero-based) to plot
count= (all traces) number of traces to plot
Example:
sutab <DATA itmin=32 itmax=63 count=10
Requests tab plot of samples 32 to 63 on the first 10 traces of DATA.
Credits:
CWP: Shuki Ronen, Jack K. Cohen
379
SUTAPER - Taper the edge traces of a data panel to zero.
sutaper <stdin >stdout [optional parameters]
ntr=tr.ntr number of traces. If tr.ntr is not set, then
ntr is mandatory
tr1=0 number of traces to be tapered at beginning
tr2=tr1 number of traces to be tapered at end
min=0. minimum amplitude factor of taper
tbeg=0 length of taper (ms) at trace start
tend=0 length of taper (ms) at trace end
taper=1 taper type
=1 linear (default)
=2 sine
=3 cosine
=4 gaussian (+/-3.8)
Notes:
To eliminate the taper, choose tbeg=0. and tend=0. and tr1=0
Credits:
Trace header fields accessed: ns, ntr
Rewrite: Tagir Galikeev, October 2002
380
SUTAUP - forwared and inverse T-X and F-K global slant stacks
sutaup <infile >outfile [optional parameters]
option=1 =1 for forward F-K domian computation
=2 for forward T-X domain computation
=3 for inverse F-K domain computation
=4 for inverse T-X domain computation
dt=tr.dt (from header) time sampling interval (secs)
nx=ntr (counted from data) number of horizontal samples (traces)
dx=1 horizontal sampling interval (m)
npoints=71 number of points for rho filter
pmin=0.0 minimum slope for Tau-P transform (s/m)
pmax=1/500 maximum slope for Tau-P transform (s/m)
np=nx number of slopes for Tau-P transform
ntau=nt number of time samples in Tau-P domain
fmin=3 minimum frequency of interest

Notes:
The cascade of a forward and inverse tau-p transform preserves the
relative amplitudes in a data panel, but not the absolute amplitudes
meaning that a scale factor must be applied to data output by such a
a cascade before the output may be compared to the original data.
This is a characteristic of the algorithm employed in this program.
(Suradon does not have this problem.)
Credits: CWP: Gabriel Alvarez, 1995.
Reference:
Levin, F., editor, 1991, Slant-Stack Processing, Geophysics Reprint
Series #14, SEG Press, Tulsa.
381
Trace header fields modified: dt,d2,f2
382
SUTIFOWLER VTI constant velocity prestack time migration
velocity analysis via Fowler’s method
sutifowler ncdps=250 vmin=1500 vmax=6000 dx=12.5
Required Parameter:
ncdps= number of input cdp’s
choose=1 1 do full prestack time migration
2 do only DMO
3 do only post-stack migrations
4 do only stacking velocity analysis
getcvstacks=0 flag to set to 1 if inputting precomputed cvstacks
(vmin, nvstack, and ncdps must match SUCVS4FOWLER job)
vminstack=vmin minimum velocity panel in m/s in input cvstacks
etamin=0. minimum eta (see paper by Tariq Alkhalifah)
etamax=0.5 maximum eta (see paper by Tariq Alkhalifah)
neta=1 number of eta values to image
d=0. Thomsen’s delta
vpvs=0.5 assumed vp/vs ratio (not critical -- default almost always ok)
dx=25. cdp x increment
vmin=1500. minimum velocity panel in m/s to output
vmax=8000. maximum velocity panel in m/s to output
nv=75 number of velocity panels to output
nvstack=180 number of stacking velocity panels to compute
( Let offmax be the maximum offset, fmax be
the maximum freq to preserve, and tmute be
the starting mute time in sec on offmax, then
the recommended value for nvstack would be
nvstack = 4 +(offmax*offmax*fmax)/(0.6*vmin*vmin*tmute)
---you may want to make do with less---)
nxpad=0 number of traces to padd for spatial fft
Ideally nxpad = (0.5*tmax*vmax+0.5*offmax)/dx
lmute=24 length of mute taper in ms
nonhyp=1 1 if do mute at 2*offset/vmin to avoid non-hyperbolic
moveout, 0 otherwise
lbtaper=0 length of bottom taper in ms
lstaper=0 length of side taper in traces
dtout=1.5*dt output sample rate in s, note: typically
fmax=salias*0.5/dtout
mxfold=120 maximum number of offsets/input cmp
salias=0.8 fraction of output frequencies to force within sloth
antialias limit. This controls muting by offset of
383
the input data prior to computing the cv stacks
for values of choose=1 or choose=2.
file=sutifowler root name for temporary files
p=not Enter a path name where to put temporary files.
specified Can enter multiple times to use multiple disk
systems.
The default uses information from the .VND file
in the current directory if it exists, or puts
unique temporary files in the current directory.
ngroup=20 Number of cmps per velocity analysis group.
printfile=stderr The output file for printout from this program.
Required trace header words on input are ns, dt, cdp, offset.
On output, trace headers are rebuilt from scratch with
ns - number of samples
dt - sample rate in usec
cdp - the output cmp number (0 based)
offset - the output velocity
tracf - the output velocity index (0 based)
fldr - index for velocity analysis group (0 based, groups of ngroup cdps)
ep - central cmp for velocity analysis group
igc - index for choice of eta (0 based)
igi - eta*100
sx=gx - x coordinate as icmp*dx
tracl=tracr -sequential trace count (1 based)
Note: Due to aliasing considerations, the small offset-to-depth

ratio assumption inherent in the TI DMO derivation, and the
poor stacking of some large-offset events associated with TI non-hyperbolic
moveout, a fairly stiff initial mute is recommended for the
long offsets. As a result, this method may not work well
where you have multiple reflections to remove via stacking.
Note: The temporary files can be split over multiple disks by building
a .VND file in your working directory. The .VND file is ascii text
with the first line giving the number of directories followed by
successive lines with one line per directory name.
Note: The output data order has primary key equal to cdp, secondary
key equal to eta, and tertiary key equal to velocity.
Credits:
CWP: John Anderson (visitor to CSM from Mobil) Spring 1993
384
385
SUTIHALEDMO - TI Hale Dip MoveOut (based on Hale’s PhD thesis)
sutihaledmo <infile >outfile [optional parameters]
nxmax maximum number of midpoints in common offset gather
option=1 1 = traditional Hale DMO (from PhD thesis)
2 = Bleistein’s true amplitude DMO
3 = Bleistein’s cos*cos weighted DMO
4 = Zhang’s DMO
5 = Tsvankin’s anisotropic DMO
6 = Tsvankin’s VTI DMO weak anisotropy approximation
dx=50. midpoint sampling interval between traces
in a common offset gather. (usually shot
interval in meters)
v=1500.0 velocity (in meters/sec)
(must enter a positive value for option=3)
(for excluding evanescent energy)
h=200.0 source-receiver half-offset (in meters)
ntpad=0 number of time samples to pad
nxpad=h/dx number of midpoints to pad
file=vnmo name of file with vnmo as a function of p
used for option=5--otherwise not used
(Generate this file by running program
sutivel with appropriate list of Thomsen’s
parameters.)
e=0. Thompsen’s epsilon
d=0. Thompsen’s delta
Note:
This module assumes a single common offset gather after NMO is

to be input, DMO corrected, and output. It is useful for computing
theoretical DMO impulse responses. The Hale algorithm is
computationally intensive and not commonly used for bulk processing
of all of the offsets on a 2-D line as there are cheaper alternative
algorithms. The Hale algorithm is commonly used in theoretical studies.
Bulk processing for multiple common offset gathers is typically done
using other modules.
386
Test run: suspike | sutihaledmo nxmax=32 option=1 v=1500 | suxwigb &
Author: (Visitor to CSM from Mobil) John E. Anderson Spring 1994

References: Anderson, J.E., and Tsvankin, I., 1994, Dip-moveout by
Fourier transform in anisotropic media, CWP-146
387
SUTIVEL - SU Transversely Isotropic velocity table builder
computes vnmo or vphase as a function of Thomsen’s parameters and
theta and optionally interpolate to constant increments in slowness
a=2500. alpha (vertical p velocity)
b=1250. beta (vertical sv velocity)
e=.20 epsilon (horiz p-wave anisotropy)
d=.10 delta (strange parameter)
maxangle=90.0 max angle in degrees
nangle=9001 number of angles to compute
verbose=0 set to 1 to see full listing
np=8001 number of slowness values to output
option=1 1=output vnmo(p) (result used for TI DMO)
2=output vnmo(theta) in degrees
3=output vnmo(theta) in radians
4=output vphase(p)
5=output vphase(theta) in degrees
6=output vphase(theta) in radians
7=output first derivative vphase(p)
8=output first derivative vphase(theta) in degrees
9=output first derivative vphase(theta) in radians
10=output second derivative vphase(p)
11=output second derivative vphase(theta) in degrees
12=output second derivative vphase(theta) in radians
13=( 1/vnmo(0)^2 -1/vnmo(theta)^2 )/p^2 test vs theta
(result should be zero for all theta for d=e)
14=return vnmo(p) for weak anisotropy
normalize=0 =1 means scale vnmo by cosine and scale vphase by
1/sqrt(1+2*e*sin(theta)*sin(theta)
(only useful for vphase when d=e for constant
result)
=0 means output vnmo or vphase unnormalized
Output on standard output is ascii text with:

line 1: number of values
line 2: abscissa increment (p or theta increment, always starts at zero)
line 3-n: one value per line
Author: (visitor to CSM form Mobil) John E. Anderson, Spring 1994
388
SUTSQ -- time axis time-squared stretch of seismic traces
sutsq [optional parameters] <stdin >stdout
none
tmin= .1*nt*dt minimum time sample of interest
(only needed for forward transform)
dt= .004 output sample rate
(only needed for inverse transform)
flag= 1 1=forward transform: time to time squared
-1=inverse transform: time squared to time
Note: The output of the forward transform always starts with

time squared equal to zero. ’tmin’ is used to avoid aliasing
the early times.
Caveats:
Amplitudes are not well preserved.

389
SUTTOZ - resample from time to depth
suttoz <stdin >stdout [optional parms]
nz=1+(nt-1)*dt*vmax/(2.0*dz) number of depth samples in output
dz=vmin*dt/2 depth sampling interval (defaults avoids aliasing)
fz=v(ft)*ft/2 first depth sample
t=0.0,... times corresponding to interval velocities in v
v=1500.0,... interval velocities corresponding to times in v
verbose=0 >0 to print depth sampling information
Notes:
Default value of nz set to avoid aliasing
The t and v arrays specify an interval velocity function of time.
Note that t and v are given as arrays of floats separated by commas,

for example:
t=0.0,0.01,.2,... v=1500.0,1720.0,1833.5,... with the number of t values
equaling the number of v values.

interval velocities at times not specified. Values specified in t

then the t and v arrays are ignored.
see selfdoc of suztot for depth to time conversion
Trace header fields accessed: ns, dt, and delrt

Trace header fields modified: trid, ns, d1, and f1
Credits:
CWP: Dave Hale c. 1992
390
SUTVBAND - time-variant bandpass filter (sine-squared taper)
sutvband <stdin >stdout tf= f=
tf= times for which f-vector is specified
f=f1,f2,f3,f4 Corner frequencies corresponding to the
times in tf. Specify as many f= as
there are entries in tf.
The filters are applied in frequency domain.
Example:
sutvband <data tf=.2,1.5 f=10,12.5,40,50 f=10,12.5,30,40 | ...
Credits:
CWP: Jack, Ken
391
SUTXTAPER - TAPER in (X,T) the edges of a data panel to zero.
sutxtaper <stdin >stdout [optional parameters]
low=0. minimum amplitude factor of taper
tbeg=0 length of taper (ms) at trace start
tend=0 length of taper (ms) at trace end
taper=1 taper type
=1 linear (default)
=2 sine
=3 cosine
key=tr set key to compute x-domain taper weights
default is using internal tracecount (tr)
tr1=0 number of traces to be tapered at beg (key=tr)
tr2=tr1 number of traces to be tapered at end (key=tr)
min=0. minimum value of key where taper starts (amp=1.)

max=0. maximum value of key where taper starts (amp=1.)
dx=1. length of taper (in key units)
if key=tr (unset) length is tr1 and (ntr-tr2)
Notes:
Taper type is used for trace (x-domain) tapering as well
as for time domain tapering.
The taper is applied to all traces <tr1 (or key<min) and
>tr2 (or key >max) and all time samples <tbeg and >tend.
Taper weights are amp*1 for traces n tr1<n<tr2 and computed
for all other traces corresponding to the taper typ.
If key is given the taper length is defined by dx, otherwise
the length of taper is tr1 and (ntr-tr2) respectively.
To eliminate the taper, choose tbeg=0. and tend=0. and tr1=0
If key is set, min,max values take precedence over tr1,tr2.
Credits: (based on sutaper)
392
Rewrite: Tagir Galikeev, October 2002
Rewrite: Gerald Klein, IFM-GEOMAR, April 2004
393
SUUNPACK1 - unpack segy trace data from chars to floats
suunpack1 <packed_file >unpacked_file
suunpack1 is the approximate inverse of supack1
Credits:
Caveats:
This program is for single site use with supack1. See the
supack1 header comments.
Notes:
trid = CHARPACK is defined in su.h and segy.h
Trace header fields accessed: ns, trid, ungpow, unscale

Trace header fields modified: trid, ungpow, unscale
394
SUUNPACK2 - unpack segy trace data from shorts to floats
suunpack2 <packed_file >unpacked_file
suunpack2 is the approximate inverse of supack2
Credits:

Changed decoding to parallel 2 byte encoding of supack2
Caveats:
This program is for single site use with supack2. See the
supack2 header comments.
Notes:
trid = SHORTPACK is defined in su.h and segy.h
Trace header fields accessed: ns, trid, ungpow, unscale

Trace header fields modified: trid, ungpow, unscale
395
SUUTM - UTM projection of longitude and latitude in SU trace headers
suutm <stdin >stdout [optional parameters]
counit=(from header) input coordinate units code:
=1: length (meters or feet; no UTM projection)
=2: seconds of arc
=3: decimal degrees
=4: degrees, minutes, seconds
idx=23 reference ellipsoid index (default is WGS 1984)
a=(from idx) user-specified semimajor axis of ellipsoid
f=(from idx) user-specified flattening of ellipsoid
zkey= if set, header key to store UTM zone number
verbose=0 =1: echo ellipsoid parameters
lon0= central meridian for TM projection in degrees

(default uses the 60 standard UTM longitude zones)
xoff=500000 false Easting (default: UTM)
ysoff=10000000 false Northing, southern hemisphere (default: UTM)
ynoff=0 false Northing, northern hemisphere (default: UTM)
Notes:
Universal Transverse Mercator (UTM) coordinates are defined between
latitudes 80S (-80) and 84N (84). Longitude values must be between
-180 degrees (west) and 179.999... degrees (east).
Latitudes are read from sy and gy (N positive), and longitudes

are read from sx and gx (E positive).
The UTM zone is determined from the receiver coordinates gy and gx.
Use suazimuth to calculate shot-receiver azimuths and offsets.
Reference ellipsoids:
An ellipsoid may be specified by its semimajor axis a and its
flattening f, or one of the following ellipsoids may be selected
by its index idx (semimajor axes in meters):
0 Sphere with radius of 6371000 m
1 Airy 1830
2 Australian National 1965
3 Bessel 1841 (Ethiopia, Indonesia, Japan, Korea)
4 Bessel 1841 (Namibia)
5 Clarke 1866
396
6 Clarke 1880
7 Everest (Brunei, E. Malaysia)
10 Everest (Pakistan)
11 Everest (W. Malaysia, Singapore 1948)
12 Everest (W. Malaysia 1969)
13 Geodetic Reference System 1980 (GRS 1980)
14 Helmert 1906
15 Hough 1960
16 Indonesian 1974
17 International 1924 / Hayford 1909
18 Krassovsky 1940
19 Modified Airy
20 Modified Fischer 1960
21 South American 1969
22 World Geodetic System 1972 (WGS 1972)
23 World Geodetic System 1984 (WGS 1984) / NAD 1983
UTM grid:
The Universal Transverse Mercator (UTM) system is a world wide
coordinate system defined between 80S and 84N. It divides the
Earth into 60 six-degree zones. Zone number 1 has its central
meridian at 177W (-177 degrees), and numbers increase eastward.
Within each zone, an Easting of 500,000 m is assigned to its

central meridian to avoid negative coordinates. On the northern
hemisphere, Northings start at 0 m at the equator and increase
northward. On the southern hemisphere a false Northing of
10,000,000 m is applied, i.e. Northings start at 10,000,000 m at
the equator and decrease southward.
Coordinate encoding (sx,sy,gx,gy):

counit=1 units of length (coordinates are not converted)
counit=2 seconds of arc
counit=3 decimal degrees
counit=4 degrees, minutes and seconds encoded as integer DDDMMSS
with scalco=1 or DDDMMSS.ss with scalco=-100 (see segy.h)
Units of length are also assumed, if counit <= 0 or counit >= 5.
Author:
397
Nils Maercklin, RISSC, University of Naples, Italy, March 2007
References:
NIMA (2000). Department of Defense World Geodetic System 1984 -
its definition and relationships with local geodetic systems.
Technical Report TR8350.2. National Imagery and Mapping Agency,
Geodesy and Geophysics Department, St. Louis, MO. 3rd edition.
J. P. Snyder (1987). Map Projections - A Working Manual.
U.S. Geological Survey Professional Paper 1395, 383 pages.
U.S. Government Printing Office.
Trace header fields accessed: sx, sy, gx, gy, scalco, counit
Trace header fields modified: sx, sy, gx, gy, scalco, counit
398
SUVCAT - append one data set to another, with or without an ",
overlapping region. Data in the overlap may be
determined by one of several methods.
suvcat data1 data2 >stdout
none
Optional parameters for overlapping sections:
taplen=0 Length of overlap in integer number of

samples.(Default is 0.)
taptype=0 Type of taper or combination method in the

overlap region. 0 - average
1 - maximum magnitude
2 - cosine scaled
3 - summation
Computational Notes:
This program vertically concatenates traces from data2 onto

the end of the corresponding traces in data1, with a region
of overlap, defined by taplen. Data in the overlapping ",
region is combined by the method specified by taptype. The
currently available methods are:
taptype=0 output is assigned the unweighted average of

each point in the overlap
taptype=1 output is assigned the value of the maximum
absolute value of each point in the overlap
taptype=2 output is assigned the weighted average of
each point in the overlap, where the output
is the sum of cos(x) times the values on the
first section, and 1-cos(x) times the values
on the second section, where x is factor that
goes from 0 to pi/2 across the overlap. This
favors the upper section in the upper part of
the overlap, and favors the lower section in
the lower part of the overlap.
taptype=3 output is assigned the sum of the amplitudes
at each sample in the overlap
399
Credits:
CWP: Jack K. Cohen, Michel Dietrich (Original SUVCAT)
Steven D. Sheaffer (modifed to include overlap)
IfG Kiel: Thies Beilecke (added taptype=3)

400
SUVEL2DF - compute stacking VELocity semblance for a single time in
over Vnmo and eta in 2-D
suvel2df <stdin >stdout [optional parameters]
tn zero-offset time of reflection
offsetm Maximum offset considerd
nv=50 number of velocities
dv=50.0 velocity sampling interval
fv=1500.0 first velocity
nvh=50 number of horizotal velocities
dvh=50.0 horizontal velocity sampling interval
fvh=1500.0 first horizontal velocity
xod=1.5 maximum offset-to-depth ratio to resolve
dtratio=5 ratio of output to input time sampling intervals
nsmooth=dtratio*2+1 length of semblance num and den smoothing window
vavg=fv+0.5*(nv-1)*dv average velocity used in the search
Notes:
Semblance is defined by the following quotient:
n-1
[ sum q(t,j) ]^2
j=0
s(t) = ------------------
n-1
n sum [q(t,j)]^2
j=0
where n is the number of non-zero samples after muting.

Smoothing (nsmooth) is applied separately to the numerator and denominator
before computing this semblance quotient.
Input traces should be sorted by cdp - suvel2df outputs a group of

semblance traces every time cdp changes. Therefore, the output will
be useful only if cdp gathers are input.
Credits:
401
CWP: Tariq Alkhalifah, February 1997
Trace header fields modified: ns, dt, offset.
402
SUVELAN_NCCS - compute stacking VELocity panel for cdp gathers
using Normalized CrossCorrelation Sum
suvelan_uccs <stdin >stdout [optional parameters]
nx=tr.cdpt number of traces in cdp
nsmooth=dtratio*2+1 length of smoothing window
pwr=1.0 semblance value to the power
Notes:
Normalized CrossCorrelation sum: sum all possible crosscorrelation
trace pairs in a CMP gather for each trial velocity and zero-offset
two-way travel time inside a time window. This coherence measure is
normalized by dividing each crosscorrelation trace pair by the geometric
mean of the energy, inside the chosen time window, of each trace pair
involved in each crosscorrelation. Then, to achieve a maximum amplitude
of unity, the result is multiplied by 2/(M(M-1)), which is the inverse
of the total number of crosscorrelation. The normalization allows to
bring out weak reflection as long as these reflections have moveouts close
to a hyperbola.
Credits:
CWP: Valmore Celis, Sept 2002
Based on the original code: suvelan.c

Colorado School of Mines: Dave Hale, c. 1989
Trace header fields accessed: ns, dt, delrt, offset, cdp, cdpt
Trace header fields modified: ns, dt, offset, cdp
Reference: Neidell, N.S., and Taner, M.T., 1971, Semblance and

other coherency measures for multichannel data:
Geophysics, 36, 498-509.
403
404
SUVELAN_NSEL - compute stacking VELocity panel for cdp gathers
using the Normalized Selective CrossCorrelation sum
suvelan_usel <stdin >stdout [optional parameters]
dx=tr.d2 offset increment
tau=0.5 threshold for significance values
Notes:
Normalized Selective CrossCorrelation Sum: is based on the coherence
measure known as crosscorrelation sum. The difference is that the selective
approach sum only crosscorrelation pairs with relatively large differential
moveout, thus increasing the resolving power in the velocity spectra
compared to that achieved by conventional methods. The normalization is
achieved in much the same way of normalizing the conventional
crosscorrelation sum.
Each crosscorrelation is divided by the geometric mean

of the energy of the traces involved, and the multiplying by a constant to
achieve maximum amplitude of unity. The constant is just the inverse of the
total number of crosscorrelations included in the sum. The selection is
made using a parabolic approximation of the differential moveout and
imposing a threshold for those differential moveouts.
That threshold is the parameter tau in this program, which varies between 0
to 1. A value of tau=0, means conventional crosscorrelation sum is applied
implying that all crosscorrelations are included in the sum. In contrast,
a value of tau=1 (not recomended) means that only the crosscorrelation
formed by the trace pair involving the shortest and longest offset is
included in the sum. Intermediate values will produce percentages of the
crosscorrelations included in the sum that will be shown in the screen
before computing the velocity spectra. Typical values for tau are between
0.2 and 0.6, producing approximated percentages of crosscorrelations summed
405
between 60% and 20%. The higher the value of tau the lower the percentage
and higher the increase in the resolving power of velocity spectra.
Keeping the percentage of crosscorrelations included in the sum between 20%

and 60% will increase resolution and avoid the precense of artifacts in
the results. In data contaminated by random noise or statics distortions
is recomended to mantaing the percentage of crosscorrelations included in
the sum above 25%. After computing the velocity spectra one might want to
adjust the level and number of contours before velocity picking.
Credits: CWP: Valmore Celis, Sept 2002

Colorado School of Mines: Dave Hale c. 1989
References:
Neidell, N.S., and Taner, M.T., 1971, Semblance and other
coherency measures for multichannel data: Geophysics, 36, 498-509.
Celis, V. T., 2002, Selective-correlation velocity analysis: CSM thesis.

406
SUVELAN - compute stacking velocity semblance for cdp gathers
suvelan <stdin >stdout [optional parameters]
anis1=0.0 quartic term, numerator of an extended quartic term
anis2=0.0 in denominator of an extended quartic term
nsmooth=dtratio*2+1 length of semblance num and den smoothing window
Notes:
Velocity analysis is usually a two-dimensional screen for optimal values of
the vertical two-way traveltime and stacking velocity. But if the travel-
time curve is no longer close to a hyperbola, the quartic term of the
traveltime series should be considered. In its easiest form (with anis2=0)
the optimizion of all parameters requires a three-dimensional screen. This
is done by a repetition of the conventional two-dimensional screen with a
variation of the quartic term. The extended quartic term is more accurate,
though the function is no more a polynomial. When screening for optimal
values the theoretical dependencies between these paramters can be taken
into account. The traveltime function is defined by
1 anis1
t^2 = t_0^2 + --- x^2 + ------------- x^4
v^2 1 + anis2 x^2
The coefficients anis1, anis2 are assumed to be small, that means the non-
hyperbolicity is assumed to be small. Triplications cannot be handled.
Semblance is defined by the following quotient:
n-1
[ sum q(t,j) ]^2
j=0
s(t) = ------------------
n-1
n sum [q(t,j)]^2
407
j=0
where n is the number of non-zero samples after muting.

Smoothing (nsmooth) is applied separately to the numerator and denominator
before computing this semblance quotient.
Then, the semblance is set to the power of the parameter pwr. With pwr > 1
the difference between semblance values is stretched in the upper half of
the range of semblance values [0,1], but compressed in the lower half of
it; thus, the few large semblance values are enhanced. With pwr < 1 the
many small values are enhanced, thus more discernible against background
noise. Of course, always at the expanse of the respective other feature.
Input traces should be sorted by cdp - suvelan outputs a group of

semblance traces every time cdp changes. Therefore, the output will
be useful only if cdp gathers are input.
Credits:
CWP, Colorado School of Mines:
Dave Hale (everything except ...)
Bjoern Rommel (... the quartic term)
SINTEF, IKU Petroleumsforskning
Bjoern Rommel (... the power-of-semblance function)

408
SUVELAN_UCCS - compute stacking VELocity panel for cdp gathers
using UnNormalized CrossCorrelation Sum
suvelan_uccs <stdin >stdout [optional parameters]
Notes:
Unnormalized crosscorrelation sum: sum all possible crosscorrelation trace
pairs in a CMP gather for each trial velocity and zero-offset two-way
travel time inside a time window. This unnormalized coherency measure
produces large spectral amplitudes for strong reflections and small
spectral amplitudes for weaker ones. If M is the number of traces in the
CMP gather M(M-1)/2 is the total number of crosscorrelations for each trial
velocity and zero-offset two-way traveltime.

Reference: Neidell, N.S., and Taner, M.T., 1971, Semblance and other
coherency measures for multichannel data: Geophysics, 36, 498-509.

409
SUVELAN_USEL - compute stacking velocity panel for cdp gathers
using the UnNormalized Selective CrossCorrelation Sum
suvelan_usel <stdin >stdout [optional parameters]
dx=tr.d2 offset increment
tau=0.5 threshold for significance values
Notes:
UnNormalized Selective CrossCorrelation sum: is based on the coherence
measure known as crosscorrelation sum. The difference is that the selective
approach sum only crosscorrelation pairs with relatively large differential
moveout, thus increasing the resolving power in the velocity spectra
compared to that achieved by conventional methods. The selection is made
using a parabolic approximation of the differential moveout and imposing a
threshold for those differential moveouts.
That threshold is the parameter tau in this program, which varies between
0 to 1. A value of tau=0, means conventional crosscorrelation sum is
applied implying that all crosscorrelations are included in the sum. In
contrast, a value of tau=1 (not recomended) means that only the
crosscorrelation formed by the trace pair involving the shortest and longest
offset is included in the sum. Intermediate values will produce percentages
of the crosscorrelations included in the sum that will be shown in the
screen before computing the velocity spectra. Typical values for tau are
between 0.2 and 0.6, producing approximated percentages of crosscorrelations
summed between 60% and 20%. The higher the value of tau the lower the
percentage and higher the increase in the resolving power of velocity
spectra.
Keeping the percentage of crosscorrelations included in the sum between 20%

and 60% will increase resolution and avoid the precense of artifacts in the
results. In data contaminated by random noise or statics distortions is
410
recomended to mantaing the percentage of crosscorrelations included in the
sum above 25%. After computing the velocity spectra one might want to
adjust the level and number of contours before velocity picking.
Based on the original code: suvelan_.c

References:
Neidell, N.S., and Taner, M.T., 1971, Semblance and other coherency
measures for multichannel data: Geophysics, 36, 498-509.
Celis, V. T., 2002, Selective-correlation velocity analysis: CSM thesis.

411
SUVIBRO - Generates a Vibroseis sweep (linear, linear-segment,
dB per Octave, dB per Hertz, T-power)
suvibro [optional parameters] > out_data_file
sweep=1 linear sweep
=2 linear-segment
=3 decibel per octave
=4 decibel per hertz
=5 t-power
swconst=0.0 sweep constant (see note)
f1=10.0 sweep frequency at start
f2=60.0 sweep frequency at end
tv=10.0 sweep length
phz=0.0 initial phase (degrees)
fseg=10.0,60.0 frequency segments (see notes)
tseg=0.0,10.0 time segments (see notes)
t1=1.0 length of taper at start (see notes)
t2=1.0 length of taper at end (see notes)
taper=1 linear
=2 sine
=3 cosine
Notes:
The default tapers are linear envelopes. To eliminate the
taper, choose t1=t2=0.0.
"swconst" is active only with nonlinear sweeps, i.e. when

sweep=3,4,5. ",
"tseg" and "fseg" arrays are used when only sweep=2
Author: CWP: Michel Dietrich

Rewrite: Tagir Galikeev, CWP, 7 October 1994
Trace header fields set: ns, dt, tracl, sfs, sfe, slen, styp
412
SUVLENGTH - Adjust variable length traces to common length
suvlength <vdata >stdout
none
ns output number of samples (default: 1st trace ns)
413
SUWAVEFORM - generate a seismic wavelet
suwaveform <stdin >stdout [optional parameters]
one of the optional parameters listed below
type=akb wavelet type
akb: AKB wavelet defined by max frequency fpeak
berlage: Berlage wavelet
gauss: Gaussian wavelet defined by frequency fpeak
gaussd: Gaussian first derivative wavelet
ricker1: Ricker wavelet defined by frequency fpeak
ricker2: Ricker wavelet defined by half and period
spike: spike wavelet, shifted by time tspike
unit: unit wavelet, i.e. amplitude = 1 = const.
dt=0.004 time sampling interval in seconds

ns= if set, number of samples in output trace
fpeak=20.0 peak frequency of a Berlage, Ricker, or Gaussian,

and maximum frequency of an AKB wavelet in Hz
half=1/fpeak Ricker wavelet "ricker2": half-length

period=c*half Ricker wavelet "ricker2": period (c=sqrt(6)/pi)
distort=0.0 Ricker wavelet "ricker2": distortion factor
decay=4*fpeak Berlage wavelet: exponential decay factor in 1/sec
tn=2 Berlage wavelet: time exponent
ipa=-90 Berlage wavelet: initial phase angle in degrees
tspike=0.0 Spike wavelet: time at spike in seconds
verbose=0 1: echo output wavelet length
Notes:
If ns is not defined, the program determines the trace length
depending on the dominant signal period.
The Ricker wavelet "ricker1" and the Gaussian wavelet "gauss

are zero-phase. For these two wavelets, the trace header word
delrt is set such that the peak amplitude is at t=0 seconds.
If this is not acceptable, use "sushw key=delrt a=0".
414
The Ricker wavelets can be defined either by the peak frequency
fpeak ("ricker1") or by its half-length, the period, and a
distortion factor ("ricker2"). "ricker" is an acceptable
alias for "ricker1".
The Berlage wavelet is defined by the peak frequency fpeak, a time

time exponent tn describing the wavelet shape at its beginning,
and an exponential decay factor describing the amplitude decay
towards later times. The parameters tn and decay are non-negative,
real numbers; tn is typically a small integer number and decay a
multiple of the dominant signal period 1/fpeak. Additionally, an
initial phase angle can be given; use -90 or 90 degrees for
zero-amplitude at the beginning.
For an AKB wavelet, fpeak is the maximum frequency; the peak

frequency is about 1/3 of the fpeak value.
The output wavelet can be normalized or scaled with "sugain".

Use "suvibro" to generate a Vibroseis sweep.
Example:
A normalized, zero-phase Ricker wavelet with a peak frequency
of 15 Hz is generated and convolved with a spike dataset:
suwaveform type=ricker1 fpeak=15 | sugain pbal=1 > wavelet.su
suplane npl=1 | suconv sufile=wavelet.su | suxwigb
Caveat:
This program does not check for aliasing.
Author:
Nils Maercklin, RISSC, University of Napoli, Italy, 2006
References:
Aldridge, D. F. (1990). The Berlage wavelet.
Geophysics, vol. 55(11), p. 1508-1511.
Alford, R., Kelly, K., and Boore, D. (1947). Accuracy
of finite-difference modeling of the acoustic wave
equation. Geophysics, vol. 39, p. 834-842. (AKB wavelet)
Sheriff, R. E. (2002). Encyclopedic dictionary of
applied geophysics. Society of Exploration Geophysicists,
Tulsa. (Ricker wavelet, page 301)
415
Notes:
For more information on the wavelets type "sudoc waveforms"
or have a look at "$CWPROOT/src/cwp/lib/waveforms.c".
Credits:
CWP, the authors of the subroutines in "waveforms.c".
Trace header fields set: ns, dt, trid, delrt
416
SUWEIGHT - weight traces by header parameter, such as offset
suweight < stdin > stdout [optional parameters]
<none>
key=offset keyword of header field to weight traces by
a=1.0 constant weighting parameter (see notes below)
b=.0005 variable weighting parameter (see notes below)
... or use values of a header field for the weighting ...
key2= keyword of header field to draw weights from

scale=.0001 scale factor to apply to header field values
inv=0 weight by header value

=1 weight by inverse of header value
Notes:
This code is initially written with offset weighting in mind, but may
be used for other, user-specified schemes.
The rationale for this program is to correct for unwanted linear

amplitude trends with offset prior to either CMP stacking or AVO work.
The code has to be edited should other functions of a keyword be required.
The default form of the weighting is to multiply the amplitudes of the

traces by a factor of: ( a + b*keyword).
If key2= header field is set then this program uses the weighting
values read from that header field, instead. Note, that because most
header fields are integers, the scale=.0001 permits 10001 in the header
to represent 1.0001.
To see the list of available keywords, type: sukeyword -o <CR>
Credits:
Author: CWP: John Stockwell February 1999.
Written for Chris Walker of UniqueStep Ltd., Bedford, U.K.
inv option added by Garry Perratt (Geocon).
417
header fields accessed: ns, keyword
418
SUWELLRF - convert WELL log depth, velocity, density data into a
uniformly sampled normal incidence Reflectivity Function of time
suwellrf [required parameters] [optional parameters] > [stdout]
dvrfile= file containing depth, velocity, and density values
...or...
dvfile= file containing depth and velocity values
drfile= file containing depth and density values
...or...
dfile= file containing depth values
vfile= file containing velocity log values
rhofile= file containing density log values
nval= number of triplets of d,v,r values if dvrfile is set,
number of pairs of d,v and d,r values dvfile and drfile
are set, or number of values if dfile, vfile, and rhofile
are set.
dtout=.004 desired time sampling interval (sec) in output
ntr=1 number of traces to output
Notes:
The format of the input file(s) is C-style binary float. These files
may be constructed from ascii file via:
a2b n1=3 < dvrfile.ascii > dvrfile.bin

...or...
a2b n1=2 < dvfile.ascii > dvfile.bin
a2b n1=2 < drfile.ascii > drfile.bin
...or...
a2b n1=1 < dfile.ascii > dfile.bin
a2b n1=1 < vfile.ascii > dfile.bin
a2b n1=1 < rhofile.ascii > rhofile.bin
A raw normal-incidence impedence reflectivity as a function of time is

is generated using the smallest two-way traveltime implied by the
input velocities as the time sampling interval. This raw reflectivity
trace is then resampled to the desired output time sampling interval
via 8 point sinc interpolation. If the number of samples on the output
exceeds SU_NFLTS the output trace will be truncated to that value.
419
Caveat:
This program is really only a first rough attempt at creating a well
log utility. User input and modifications are welcome.
See also: suresamp
Author: CWP: John Stockwell, Summer 2001, updated Summer 2002.

inspired by a project by GP grad student Leo Brown
420
SUWIND - window traces by key word
suwind <stdin >stdout [options]
none
verbose=0 =1 for verbose
key=tracl Key header word to window on (see segy.h)
min=LONG_MIN min value of key header word to pass
max=LONG_MAX max value of key header word to pass
abs=0 =1 to take absolute value of key header word

j=1 Pass every j-th trace ...
s=0 ... based at s (if ((key - s)%j) == 0)
skip=0 skip the initial N traces
count=ULONG_MAX ... up to count traces
reject=none Skip traces with specified key values
accept=none Pass traces with specified key values(see notes)
processing, but do no window the data
ordered=0 =1 if traces sorted in increasing keyword value
=-1 if traces are sorted in a decreasing order
Options for vertical windowing (time gating):

dt=tr.dt (from header) time sampling interval (sec) (seismic data)
=tr.d1 (nonseismic)
f1=tr.delrt (from header) first sample (seismic data)
=tr.f1 (nonseismic)
tmin=0.0 min time to pass

tmax=(from header) max time to pass
itmin=0 min time sample to pass
itmax=(from header) max time sample to pass
nt=itmax-itmin+1 number of time samples to pass
Notes:
On large data sets, the count parameter should be set if
possible. Otherwise, every trace in the data set will be
examined. However, the count parameter overrides the accept
parameter, so you can’t specify count if you want true
unconditional acceptance.
421
The skip= option allows the user to skip over traces, which helps
for selecting traces far from the beginning of the dataset.
Caveat: skip only works with disk input.
The ordered= option will speed up the process if the data are
sorted in according to the key.
The accept option is a bit strange--it does NOT mean accept ONLY
the traces on the accept list! It means accept these traces,
even if they would otherwise be rejected (except as noted in the
previous paragraph). To implement accept-only, you can use the
max=0 option (rejecting everything). For example, to accept
only the tracl values 4, 5 and 6:
... | suwind max=0 accept=4,5,6 | ...
Another example is the case of suppressing nonseismic traces in

a seismic data set. By the SEGY standard header field trace id,
trid=1 designates traces as being seismic traces. Other traces,
such as calibration traces may be designated by another value.
Example: trid=1 seismic and trid=0 is nonseismic. To reject

nonseismic traces
... | suwind key=trid reject=0 | ...
On most 32 bit machines, LONG_MIN, LONG_MAX and ULONG_MAX are

about -2E9,+2E9 and 4E9, they are defined in limits.h.
Selecting times beyond the maximum in the data induces

zero padding (up to SU_NFLTS).
The time gating here is to the nearest neighboring sample or

time value. Gating to the exact temporal value requires
resampling if the selected times fall between samples on the
trace. Use suresamp to perform the time gating in this case.
It doesn’t really make sense to specify both itmin and tmin,

but specifying itmin takes precedence over specifying tmin.
Similarly, itmax takes precedence over tmax and tmax over nt.
If dt in header is not set, then dt is mandatory
Credits:
422
CWP: Shuki Ronen, Jack Cohen, Chris Liner
Warnemuende: Toralf Foerster
CENPET: Werner M. Heigl (modified to include well log data)
Trace header fields accessed: ns, dt, delrt, keyword

Trace header fields modified: ns, delrt, ntr
423
SUXCOR - correlation with user-supplied filter
suxcor <stdin >stdout filter= [optional parameters]
Required parameters: ONE of

sufile= file containing SU traces to use as filter
filter= user-supplied correlation filter (ascii)
vibroseis=0 =nsout for correlating vibroseis data
first=1 supplied trace is default first element of
correlation. =0 for it to be second.
panel=0 use only the first trace of sufile as filter
=1 xcor trace by trace an entire gather
ftwin=0 first sample on the first trace of the window
(only with panel=1)
ltwin=0 first sample on the last trace of the window
(only with panel=1)
ntwin=nt number of samples in the correlation window
(only with panel=1)
ntrc=48 number of traces on a gather
(only with panel=1)

Notes: It is quietly assumed that the time sampling interval on the

single trace and the output traces is the same as that on the traces
in the input file. The sufile may actually have more than one trace,
but only the first trace is used when panel=0. When panel=1 the number
of traces in the sufile MUST be the same as the number of traces in
the input.
Examples:
suplane | suwind min=12 max=12 >TRACE
suxcor<DATA sufile=TRACE |...
Here, the su data file, "DATA", is correlated trace by trace with the
the single su trace, "TRACE".
suxcor<DATA filter=1,2,1 | ...

Here, the su data file, "DATA", is correlated trace by trace with the
the filter shown.
424
Correlating vibroseis data with a vibroseis sweep:
suxcor < data sufile=sweep vibroseis=nsout |...
is equivalent to, but more efficient than:
suxcor < data sufile=sweep |

suwind itmin=nsweep itmax=nsweep+nsout | sushw key=delrt a=0.0 |...
sweep=vibroseis sweep in SU format, nsweep=number of samples on

the vibroseis sweep, nsout=desired number of samples on output
or
suxcor < data sufile=sweep |
suwind itmin=nsweep itmax=nsweep+nsout | sushw key=delrt a=0.0 |...
tsweep=sweep length in seconds, tout=desired output trace length in seconds
In the spatially variant case (panel=1), a window with linear slope

can be defined:
ftwin is the first sample of the first trace in the gather,
ltwin is the first sample of the last trace in the gather,
ntwin is the lengthe of the window,
ntrc is the the number of traces in a gather.
If the data consists of a number gathers which need to be

correlated with the same number gathers in the sufile, ntrc
assures that the correlating window re-starts for each gather.
The default window is non-sloping and takes the entire trace

into account (ftwin=ltwin=0, ntwin=nt).
Credits:
CWP: modified by Ttjan to include cross correlation of panels
permitting spatially and temporally varying cross correlation.
UTK: modified by Rick Williams for vibroseis correlation option.
CAVEATS:
In the option, panel=1 the number of traces in the sufile must be
the same as the number of traces on the input.
425
426
SUXEDIT - examine segy diskfiles and edit headers
suxedit diskfile (open for possible header modification if writable)

suxedit <diskfile (open read only)
The following commands are recognized:

number read in that trace and print nonzero header words
<CR> go to trace one step away (step is initially -1)
+ read in next trace (step is set to +1)
- read in previous trace (step is set to -1)
dN advance N traces (step is set to N)
% print some percentiles of the trace data
r print some ranks (rank[j] = jth smallest datum)
p [n1 [n2]] tab plot sample n1 to n2 on current trace
g [tr1 tr2] ["opts"] wiggle plot (graph) the trace
[traces tr1 to tr2]
f wiggle plot the Fourier transform of the trace
! key=val change a value in a field (e.g. ! tracr=101)
? print help file
q quit
NOTE: sample numbers are 1-based (first sample is 1).
Credits:
SEP: Einar Kjartansson, Shuki Ronen, Stew Levin
CWP: Jack K. Cohen

Trace header fields modified: ntr (only for internal plotting)
427
SUZERO -- zero-out (or set constant) data within a time window
suzero itmax= < indata > outdata
Required parameters
itmax= last time sample to zero out
Optional parameters
itmin=0 first time sample to zero out
value=0 value to set
See also: sukill, sumute
Credits:
CWP: Chris
Geocon: Garry Perratt (added value= option)
428
SUZTOT - resample from depth to time
suztot <stdin >stdout [optional parms]
nt=1+(nz-1)*2.0*dz/(vmax*dt) number of time samples in output
dt=2*dz/vmin time sampling interval (defaults avoids aliasing)
ft=2*fz/v(fz) first time sample
z=0.0,... depths corresponding to interval velocities in v
v=1500.0,... interval velocities corresponding to depths in v
vfile= binary (non-ascii) file containing velocities v(z)
verbose=0 >0 to print depth sampling information
Notes:
Default value of nt set to avoid aliasing
The z and v arrays specify an interval velocity function of depth.
Note that z and v are given as arrays of floats separated by commas,

for example:
z=0.0,100,200,... v=1500.0,1720.0,1833.5,... with the number of z values
equaling the number of v values.

interval velocities at times not specified. Values specified in z

then the z and v arrays are ignored.
see the selfdoc of suttoz for time to depth conversion

Trace header fields accessed: ns, dt, and delrt
Trace header fields modified: trid, ns, d1, and f1
Credits:
CWP: John Stockwell, 2005,
based on suttoz.c written by Dave Hale c. 1992
429
supscontour <stdin [optional parameters] | ...
n2=tr.ntr or number of traces in the data set (ntr is an alias for n2)
d1=tr.d1 or tr.dt/10^6 sampling interval in the fast dimension

=.004 for seismic (if not set)
=1.0 for nonseismic (if not set)
d2=tr.d2 sampling interval in the slow dimension

=1.0 (if not set)
f1=tr.f1 or tr.delrt/10^3 or 0.0 first sample in the fast dimension
f2=tr.f2 or tr.tracr or tr.tracl first sample in the slow dimension

=1.0 for seismic (if not set)
=d2 for nonseismic (if not set)

Note that for seismic time domain data, the "fast dimension" is
time and the "slow dimension" is usually trace number or range.
Also note that "foreign" data tapes may have something unexpected
in the d2,f2 fields, use segyclean to clear these if you can afford
the processing time or use d2= f2= to override the header values if
not.
See the pscontour selfdoc for the remaining parameters.
On NeXT: supscontour < infile [optional parameters] | open
Trace header fields accessed: ns, ntr, tracr, tracl, delrt, trid,
dt, d1, d2, f1, f2
Credits:
430
CWP: Dave Hale and Zhiming Li (pscontour, etc.)
Jack Cohen and John Stockwell (supscontour, etc.)
Delphi: Alexander Koek, added support for irregularly spaced traces
Aarhus University: Morten W. Pedersen copied everything from the xwigb
source and just replaced all occurencies of the word
Notes:
When the number of traces isn’t known, we need to count
the traces for pscontour. You can make this value "known"
either by getparring n2 or by having the ntr field set
in the trace header. A getparred value takes precedence
over the value in the trace header.
When we do have to count the traces, we use the "tmpfile"

routine because on many machines it is implemented
as a memory area instead of a disk file.
If your system does make a disk file, consider altering

the code to remove the file on interrupt. This could be
done either by trapping the interrupt with "signal"
or by using the "tmpnam" routine followed by an immediate
"remove" (aka "unlink" in old unix).
When we must compute ntr, we don’t allocate a 2-d array,

but just content ourselves with copying trace by trace from
the data "file" to the pipe into the plotting program.
Although we could use tr.data, we allocate a trace buffer
for code clarity.
431
supscubecontour <stdin [optional parameters] | ...
n2 is the number of traces per frame. If not getparred then it

is the total number of traces in the data set.
n3 is the number of frames. If not getparred then it

is the total number of frames in the data set measured by ntr/n2


=1.0 (if not set)


the processing time or use d2= f2= to over-ride the header values if
not.
See the pscubecontour selfdoc for the remaining parameters.
example: supscubecontour < infile [optional parameters] | gv -
432
Credits:
CWP: Dave Hale and Zhiming Li (pscube)

Jack K. Cohen (suxmovie)
John Stockwell (supscubecontour)
Notes:
When n2 isn’t getparred, we need to count the traces
for pscube. Although we compute ntr, we don’t allocate a 2-d array
and content ourselves with copying trace by trace from
for code clarity.
433
supscube <stdin [optional parameters] | ...




=1.0 (if not set)


not.
See the pscube selfdoc for the remaining parameters.
On NeXT: supscube < infile [optional parameters] | open
434
Credits:

John Stockwell (supscube)
Notes:
for code clarity.
435
supsgraph <stdin [optional parameters] | ...
style=seismic seismic is default here, =normal is alternate
(see psgraph selfdoc for style definitions)
nplot is the number of traces (ntr is an acceptable alias for nplot)


=1.0 (if not set)


not.
See the psgraph selfdoc for the remaining parameters.
On NeXT: supsgraph < infile [optional parameters] | open
Credits:
436
CWP: Dave Hale and Zhiming Li (pswigp, etc.)
Jack Cohen and John Stockwell (supswigp, etc.)
Notes:
the traces for pswigp. You can make this value "known"
either by getparring nplot or by having the ntr field set

for code clarity.
437
supsimage <stdin [optional parameters] | ...


=1.0 (if not set)


not.
See the psimage selfdoc for the remaining parameters.
On NeXT: supsimage < infile [optional parameters] | open
dt, d1, d2, f1, f2
Credits:
438
CWP: Dave Hale and Zhiming Li (psimage, etc.)
Jack Cohen and John Stockwell (supsimage, etc.)
Notes:
the traces for psimage. You can make this value "known"


for code clarity.
439
of a SEGY (SU) data set
supsmax <stdin >postscript file [optional parameters]
mode=max max value
=min min value
=abs absolute max value


=1.0 (if not set)


not.
See the sumax selfdoc for additional parameter.

440
Credits:
CWP: John Stockwell, based on Jack Cohen’s SU JACKet
Notes:
the traces for psgraph. You can make this value "known"

441
supsmovie <stdin [optional parameters] | ...




=1.0 (if not set)


not.
See the psmovie selfdoc for the remaining parameters.
On NeXT: supsmovie < infile [optional parameters] | open
442
Credits:
CWP: Dave Hale and Zhiming Li (psmovie)

John Stockwell (supsmovie)
Notes:
for psmovie. In this case:
we are using tmpfile because on many machines it is
implemented as a memory area instead of a disk file.
Although we compute ntr, we don’t allocate a 2-d array
for code clarity.
443
supswigb <stdin [optional parameters] | ...
key=(keyword) if set, the values of x2 are set from header field
specified by keyword
=1.0 (if not set)
style=seismic normal (axis 1 horizontal, axis 2 vertical) or

vsp (same as normal with axis 2 reversed)
Note: vsp requires use of a keyword

not.
field represented by the keyword (for example key=offset, will show
traces in true offset). This permit unequally spaced traces to be plotted.
This program is really just a wrapper for the plotting program: pswigb
See the pswigb selfdoc for the remaining parameters.
444
dt, d1, d2, f1, f2, keyword (if set)
Credits:
CWP: Dave Hale and Zhiming Li (pswigb, etc.)

Jack Cohen and John Stockwell (supswigb, etc.)
Modified by Brian Zook, Southwest Research Institute, to honor

scale factors, added vsp style
Notes:
the traces for pswigb. You can make this value "known"

for code clarity.
445
supswigp <stdin [optional parameters] | ...
key=(keyword) if set, values of x2 are set from header field
=1.0 (if not set)


not.
This program is really just a wrapper for the plotting program: pswigp
See the pswigp selfdoc for the remaining parameters.
446
On NeXT: supswigp < infile [optional parameters] | open
dt, d1, d2, f1, f2, key specified by key
Credits:


Notes:



for code clarity.
447
suxwigb <stdin [optional parameters] | ...
=1.0 (if not set)

not.
This program is really just a wrapper for the plotting program: xcontour
See the xcontour selfdoc for the remaining parameters.
Credits:
448
CWP: Dave Hale and Zhiming Li (xwigb, etc.)
Jack Cohen and John Stockwell (suxwigb, etc.)
xwigb with xcountour ;-)
Notes:
the traces for xcontour. You can make this value "known"

for code clarity.
449
suxgraph <stdin [optional parameters] | ...
(see xgraph selfdoc for style definitions)
nplot= number of traces (ntr is an acceptable alias for nplot)


=1.0 (if not set)


not.
See the xgraph selfdoc for the remaining parameters.
On NeXT: suxgraph < infile [optional parameters] | open
Credits:
450
Notes:


for code clarity.
451
suximage infile= [optional parameters] | ... (direct I/O)

or
suximage <stdin [optional parameters] | ... (sequential I/O)
infile=NULL SU data to be ploted, default stdin with sequential access

if ’infile’ provided, su data read by (fast) direct access
with ftr,dtr and n2 suximage will pass a subset of data

to the plotting program-ximage:
ftr=1 First trace to be plotted
dtr=1 Trace increment to be plotted
n2=tr.ntr (Max) number of traces to be plotted (ntr is an alias for n2)
Priority: first try to read from parameter line;
if not provided, check trace header tr.ntr;
if still not provided, figure it out using ftello


=1.0 (if not set or was set to 0)
key= key for annotating d2 (slow dimension)

If annotation is not at proper increment, try
setting d2; only first trace’s key value is read


452
not.
See the ximage selfdoc for the remaining parameters.
Credits:
CWP: Dave Hale and Zhiming Li (ximage, etc.)

Jack Cohen and John Stockwell (suximage, etc.)
MTU: David Forel, June 2004, added key for annotating d2
ConocoPhillips: Zhaobo Meng, Dec 2004, added direct I/O
Notes:
When provide ftr and dtr and infile, suximage can be used to plot
multi-dimensional volumes efficiently. For example, for a Offset-CDP
dataset with 32 offsets, the command line
suximage infile=volume3d.su ftr=1 dtr=32 ... &
will display the zero-offset common offset data with ranrom access.
It is highly recommend to use infile= to view large datasets, since
using stdin only allows sequential access, which is very slow for
large datasets.

the traces for ximage. You can make this value "known"

for code clarity.
453
each trace of a SEGY (SU) data set
suxmax <stdin [optional parameters]
mode=max max value
=min min value


=1.0 (if not set)


not.

454
Credits:
Notes:
the traces for xgraph. You can make this value "known"

455
suxmovie <stdin [optional parameters]
n1=tr.ns number of samples per trace

ntr=tr.ntr number of traces in the data set
n2=tr.shortpad or tr.ntr number of traces in inline direction
n3=ntr/n2 number of traces in crossline direction


=1.0 (if not set)
d3=1.0 sampling interval in the slowest dimension
f1=tr.f1 or 0.0 first sample in the z dimension

f2=tr.f2 or 1.0 first sample in the x dimension
f3=1.0
mode=0 0= x,z slice movie through y dimension (in line)

1= y,z slice movie through x dimension (cross line)
2= x,y slice movie through z dimension (time slice)

Notes:
For seismic data, the "fast dimension" is either time or
depth and the "slow dimension" is usually trace number.
The 3D data set is expected to have n3 sets of n2 traces representing
the horizontal coverage of n2*d2 in x and n3*d3 in y direction.
The data is read to memory with and piped to xmovie with the
respective sampling parameters.
456
See the xmovie selfdoc for the remaining parameters and X functions.
457
suxpicker <stdin [optional parameters] | ...



=1.0 (if not set)


not.
458
See the xpicker selfdoc for the remaining parameters.
Credits:
CWP: Dave Hale and Zhiming Li (xpicker, etc.)

Jack Cohen and John Stockwell (suxpicker, etc.)
Notes:
the traces for xpicker. You can make this value "known"



for code clarity.
459
This is a modified suxwigb that uses the depth or coordinate scaling
when such values are used as keys.
=1.0 (if not set)


not.
460
This program is really just a wrapper for the plotting program: xwigb
See the xwigb selfdoc for the remaining parameters.
Credits:


Notes:
the traces for xwigb. You can make this value "known"

for code clarity.
461
PSBBOX - change BoundingBOX of existing PostScript file
psbbox < PostScriptfile [optional parameters] > PostScriptfile
llx= new llx
lly= new lly
urx= new urx
ury= new ury
verbose=1 =1 for info printed on stderr (0 for no info)
462
PSCONTOUR - PostScript CONTOURing of a two-dimensional function f(x1,x2)
pscontour n1= [optional parameters] <binaryfile >postscriptfile
n1 number of samples in 1st (fast) dimension
x1=f1,f1+d1,... array of monotonic sampled values in 1st dimension
x2=f2,f2+d2,... array of monotonic sampled values in 2nd dimension
nc=5 number of contour values
dc=(zmax-zmin)/nc contour interval
fc=min+dc first contour
c=fc,fc+dc,... array of contour values
cwidth=1.0,... array of contour line widths
cgray=0.0,... array of contour grays (0.0=black to 1.0=white)
ccolor=none,... array of contour colors; none means use cgray
cdash=0.0,... array of dash spacings (0.0 for solid)
labelcf=1 first labeled contour (1,2,3,...)
labelcper=1 label every labelcper-th contour
nlabelc=nc number of labeled contours (0 no contour label)
nplaces=6 number of decimal places in contour label
xbox=1.5 offset in inches of left side of axes box
ybox=1.5 offset in inches of bottom side of axes box
wbox=6.0 width in inches of axes box
hbox=8.0 height in inches of axes box
x1beg=x1min value at which axis 1 begins
x1end=x1max value at which axis 1 ends
d1num=0.0 numbered tic interval on axis 1 (0.0 for automatic)
f1num=x1min first numbered tic on axis 1 (used if d1num not 0.0)
n1tic=1 number of tics per numbered tic on axis 1
grid1=none grid lines on axis 1 - none, dot, dash, or solid
label1= label on axis 1
463
labelfont=Helvetica font name for axes labels
labelsize=18 font size for axes labels
title= title of plot
titlefont=Helvetica-Bold font name for title
titlesize=24 font size for title
labelcfont=Helvetica-Bold font name for contour labels
labelcsize=6 font size of contour labels
labelccolor=black color of contour labels
titlecolor=black color of title
axescolor=black color of axes
gridcolor=black color of grid
axeswidth=1 width (in points) of axes
ticwidth=axeswidth width (in points) of tic marks
gridwidth=axeswidth width (in points) of grid lines
seismic (axis 1 vertical, axis 2 horizontal)
Note.
The line width of unlabeled contours is designed as a quarter of that
of labeled contours.
All color specifications may also be made in X Window style Hex format
example: axescolor=#255
type: sudoc pscontour for more information
Notes:
For nice even-numbered contours, use the parameters fc and dc
Example: if the range of the z values of a data set range between

approximately -1000 and +1000, then use fc=-1000 nc=10 and dc=100
to get contours spaced by even 100’s.

MODIFIED: Craig Artley, Colorado School of Mines, 08/30/91
BoundingBox moved to top of PostScript output
MODIFIED: Zhenyue Liu, Colorado School of Mines, 08/26/93
464
Values are labeled on contours
Added color options (Courtesy of Dave Hale, Advance Geophysical).
Modified: Morten Wendell Pedersen, Aarhus University, 23/3-97
Added ticwidth,axeswidth, gridwidth parameters
465
PSCCONTOUR - PostScript Contour plot of a data CUBE
pscubecontour n1= n2= n3= [optional parameters] <binaryfile >postscriptfile

or
pscubecontour n1= n2= n3= front= side= top= [optional parameters] >postscriptfile
Data formats supported:

1. Entire cube read from stdin (n1*n2*n3 floats) [default format]
2. Faces read from stdin (n1*n2 floats for front, followed by n1*n3
floats for side, and n2*n3 floats for top) [specify faces=1]
3. Faces read from separate data files [specify filenames]
n1 number of samples in 1st (fastest) dimension
n2 number of samples in 2nd dimension
n3 number of samples in 3rd (slowest) dimension
front name of file containing front panel
side name of file containing side panel
top name of file containing top panel
faces=0 =1 to read faces from stdin (data format 2)
f1=0.0 first sample in 1st dimension
f2=0.0 first sample in 2nd dimension
d3=1.0 sampling interval in 3rd dimension
f3=0.0 first sample in 3rd dimension
d1s=1.0 factor by which to scale d1 before imaging
cgray=0.0,... array of contour grays (0.0=black to 1.0=white)
nplaces=6 number of decimal places in contour label
466
size1=4.0 size in inches of 1st axes (vertical)
size2=4.0 size in inches of 2nd axes (horizontal)
size3=3.0 size in inches of 3rd axes (projected)
angle=45 projection angle of cube in degrees (0<angle<90)
(angle between 2nd axis and projected 3rd axis)
labelcfont=Helvetica-Bold font name for contour labels
labelcsize=6 font size of contour labels
labelccolor=black color of contour labels
Note: The values of x1beg=x1min, x2end=x2max and x3beg=x3min cannot

be changed.
467
(Original codes pscontour and pscube)
AUTHOR: Craig Artley, Colorado School of Mines, 03/12/93

NOTE: Original written by Zhiming Li & Dave Hale, CSM, 07/01/90
Completely rewritten, the code now bears more similarity to
psimage than the previous pscube. Faces of cube now rendered
as three separate images, rather than as a single image. The
output no longer suffers from stretching artifacts, and the
code is simpler. -Craig
Added color options.
PSCCONTOUR: mashed together from pscube and pscontour

to generate 3d contour plots by Claudia Vanelle, Institute of Geophysics,
University of Hamburg, Germany somewhen in 2000
PSCUBE was "merged" with PSCONTOUR to create PSCUBECONTOUR

by Claudia Vanelle, Applied Geophysics Group Hamburg
somewhen in 2000
468
PSCUBE - PostScript image plot with Legend of a data CUBE
pscube n1= n2= n3= [optional parameters] <binaryfile >postscriptfile

or
pscube n1= n2= n3= front= side= top= [optional parameters] >postscriptfile

1. Entire cube read from stdin (n1*n2*n3 floats) [default format]
2. Faces read from stdin (n1*n2 floats for front, followed by n1*n3
floats for side, and n2*n3 floats for top) [specify faces=1]
3. Faces read from separate data files [specify filenames]
n1 number of samples in 1st (fastest) dimension
n2 number of samples in 2nd dimension
n3 number of samples in 3rd (slowest) dimension
front name of file containing front panel
side name of file containing side panel
top name of file containing top panel
faces=0 =1 to read faces from stdin (data format 2)
f3=0.0 first sample in 3rd dimension
perc=100.0 percentile used to determine clip
clip=(perc percentile) clip used to determine bclip and wclip
bperc=perc percentile for determining black clip value
wperc=100.0-perc percentile for determining white clip value
bclip=clip data values outside of [bclip,wclip] are clipped
wclip=-clip data values outside of [bclip,wclip] are clipped
brgb=0.0,0.0,0.0 red, green, blue values corresponding to black
wrgb=1.0,1.0,1.0 red, green, blue values corresponding to white
bhls=0.0,0.0,0.0 hue, lightness, saturation corresponding to black
whls=0.0,1.0,0.0 hue, lightness, saturation corresponding to white
bps=12 bits per sample for color plots, either 12 or 24
469
size1=4.0 size in inches of 1st axes (vertical)
size2=4.0 size in inches of 2nd axes (horizontal)
size3=3.0 size in inches of 3rd axes (projected)
angle=45 projection angle of cube in degrees (0<angle<90)
(angle between 2nd axis and projected 3rd axis)
legend=0 =1 display the color scale
if ==1, resize xbox,ybox,width,height
lstyle=vertleft Vertical, axis label on left side
=vertright (Vertical, axis label on right side)
=horibottom (Horizontal, axis label on bottom)
units= unit label for legend
legendfont=times_roman10 font name for title
following are defaults for lstyle=0. They are changed for other lstyles
lwidth=1.2 colorscale (legend) width in inches
lheight=height/3 colorscale (legend) height in inches
470
lx=1.0 colorscale (legend) x-position in inches
ly=(height-lheight)/2+xybox colorscale (legend) y-position in pixels
lbeg= lmin or wclip-5*perc value at which legend axis begins
lend= lmax or bclip+5*perc value at which legend axis ends
ldnum=0.0 numbered tic interval on legend axis (0.0 for automatic)
lfnum=lmin first numbered tic on legend axis (used if d1num not 0.0)
lntic=1 number of tics per numbered tic on legend axis
lgrid=none grid lines on legend axis - none, dot, dash, or solid
Note: The values of x1beg=x1min, x2end=x2max and x3beg=x3min cannot

be changed.
471
PSEPSI - add an EPSI formatted preview bitmap to an EPS file
psepsi <epsfile >epsifile
Note:
This application requires
(1) that gs (the Ghostscript interpreter) exist, and
(2) that the input EPS file contain a BoundingBox and EndComments.
Ghostscript is used to build the preview bitmap, which is then
merged with the input EPS file to make the output EPSI file.
472
PSGRAPH - PostScript GRAPHer
Graphs n[i] pairs of (x,y) coordinates, for i = 1 to nplot.
psgraph n= [optional parameters] <binaryfile >postscriptfile
n array containing number of points per plot

1.a. x1,y1,x2,y2,...,xn,yn
b. x1,x2,...,xn,y1,y2,...,yn (must set pairs=0)
2. y1,y2,...,yn (must give non-zero d1[]=)
3. x1,x2,...,xn (must give non-zero d2[]=)
4. nil (must give non-zero d1[]= and non-zero d2[]=)
The formats may be repeated and mixed in any order, but if
formats 2-4 are used, the d1 and d2 arrays must be specified including
d1[]=0.0 d2[]=0.0 entries for any internal occurences of format 1.
Similarly, the pairs array must contain place-keeping entries for
plots of formats 2-4 if they are mixed with both formats 1.a and 1.b.
Also, if formats 2-4 are used with non-zero f1[] or f2[] entries, then
the corresponding array(s) must be fully specified including f1[]=0.0
and/or f2[]=0.0 entries for any internal occurences of format 1 or
formats 2-4 where the zero entries are desired.
Available colors are all the common ones and many more. The complete
list of 68 colors is in the file $CWPROOT/src/psplot/basic.c.
nplot=number of n’s number of plots
d1=0.0,... x sampling intervals (0.0 if x coordinates input)
f1=0.0,... first x values (not used if x coordinates input)
d2=0.0,... y sampling intervals (0.0 if y coordinates input)
f2=0.0,... first y values (not used if y coordinates input)
pairs=1,... =1 for data pairs in format 1.a, =0 for format 1.b
linewidth=1.0,... line widths (in points) (0.0 for no lines)
linegray=0.0,... line gray levels (black=0.0 to white=1.0)
linecolor=none,... line colors; none means use linegray
Typical use: linecolor=red,yellow,blue,...
lineon=1.0,... length of line segments for dashed lines (in points)
lineoff=0.0,... spacing between dashes (0.0 for solid line)
mark=0,1,2,3,... indices of marks used to represent plotted points
marksize=0.0,0.0,... size of marks (0.0 for no marks)
473
style=normal normal (axis 1 horizontal, axis 2 vertical) or
reverse=0 =1 to reverse sequence of plotting curves ",
Note: n1 and n2 are acceptable aliases for n and nplot, respectively.
mark index:
1. asterisk
2. x-cross
3. open triangle
4. open square
5. open circle
6. solid triangle
7. solid square
8. solid circle
474
Example:
psgraph n=50,100,20 d1=2.5,1,0.33 <datafile >psfile
plots three curves with equally spaced x values in one plot frame
x1-coordinates are x1(i) = f1+i*d1 for i = 1 to n (f1=0 by default)
number of x2’s and then x2-coordinates for each curve are read
sequentially from datafile.
475
PSIMAGE - PostScript IMAGE plot of a uniformly-sampled function f(x1,x2)
psimage n1= [optional parameters] <binaryfile >postscriptfile
bclip and wclip will be set to be inside
[lbeg,lend] if lbeg and/or lend are supplied
threecolor=1 supply 3 color values instead of only two,
using not only black and white, but f.e. red,
green and blue
brgb=0.0,0.0,0.0 red, green, blue values corresponding to black
grgb=1.0,1.0,1.0 red, green, blue values corresponding to grey
wrgb=1.0,1.0,1.0 red, green, blue values corresponding to white
bhls=0.0,0.0,0.0 hue, lightness, saturation corresponding to black
ghls=0.0,1.0,0.0 hue, lightness, saturation corresponding to grey
whls=0.0,1.0,0.0 hue, lightness, saturation corresponding to white
bps=12 bits per sample for color plots, either 12 or 24
width=6.0 width in inches of axes box
height=8.0 height in inches of axes box
476
lnice=0 =1 nice legend arrangement
(overrides ybox,lx,width,height parameters)
lstyle=vertleft Vertical, axis label on left side
=vertright (Vertical, axis label on right side)
=horibottom (Horizontal, axis label on bottom)
following are defaults for lstyle=0. They are changed for other lstyles
lwidth=1.2 colorscale (legend) width in inches
lheight=height/3 colorscale (legend) height in inches
lx=1.0 colorscale (legend) x-position in inches
ly=(height-lheight)/2+xybox colorscale (legend) y-position in pixels
lbeg= lmin or wclip-5*perc value at which legend axis begins
lend= lmax or bclip+5*perc value at which legend axis ends
ldnum=0.0 numbered tic interval on legend axis (0.0 for automatic)
lfnum=lmin first numbered tic on legend axis (used if d1num not 0.0)
lntic=1 number of tics per numbered tic on legend axis
lgrid=none grid lines on legend axis - none, dot, dash, or solid
curve=curve1,curve2,... file(s) containing points to draw curve(s)
477
npair=n1,n2,n2,... number(s) of pairs in each file
curvecolor=black,.. color of curve(s)
curvewidth=axeswidth width (in points) of curve(s)
NOTES:
The curve file is an ascii file with the points specified as x1 x2
pairs, one pair to a line. A "vector" of curve files and curve
colors may be specified as curvefile=file1,file2,etc.
and curvecolor=color1,color2,etc, and the number of pairs of values
in each file as npair=npair1,npair2,... .
You may eliminate the blocky appearance of psimages by adjusting the

d1s= and d2s= parameters, so that psimages appear similar to ximages.
Some example colormap settings:

red white blue: wrgb=1.0,0,0 grgb=1.0,1.0,1.0 brgb=0,0,1.0
white red blue: wrgb=1.0,1.0,1.0 grgb=1.0,0.0,0.0 brgb=0,0,1.0
blue red white: wrgb=0.0,0.0,1.0 grgb=1.0,0.0,0.0 brgb=1.0,1.0,1.0
red green blue: wrgb=1.0,0,0 grgb=0,1.0,0 brgb=0,0,1.0
orange light-blue green: wrgb=1.0,.5,0 grgb=0,.7,1.0 brgb=0,1.0,0
red light-blue dark blue: wrgb=0.0,0,1.0 grgb=0,1.0,1.0 brgb=0,0,1.0

BoundingBox moved to top of PostScript output
Added color options (Courtesy of Dave Hale, Advance Geophysical).
MODIFIED: Torsten Schoenfelder, Koeln, Germany 006/07/97
colorbar (legend) (as in ximage (by Berend Scheffers, Delft))
MODIFIED: Brian K. Macy, Phillips Petroleum, 01/14/99
Added curve plotting option
MODIFIED: Torsten Schoenfelder, Koeln, Germany 02/10/99
color scale with interpolation of three colors
478
PSLABEL - output PostScript file consisting of a single TEXT string
on a specified background. (Use with psmerge to label plots.)
pslabel t= [t=] [optional parameters] > epsfile
Required Parameters (can have multiple specifications to mix fonts):

t= text string to write to output
f=Times-Bold font for text string
(multiple specifications for each t)
size=30 size of characters in points (72 points/inch)
tcolor=black color of text string
bcolor=white color of background box
nsub=0 number of characters to subtract when
computing size of background box (not all
characters are the same size so the
background box may be too big at times.)
Example:
pslabel t="(a) " f=Times-Bold t="h" f=Symbol t="=0.04" nsub=3 > epsfile
This example yields the PostScript equivalent of the string

(written here in LaTeX notation) $ (a)\\; \\eta=0.04 $
Notes:
This program produces a (color if desired) PostScript text string that
can be positioned and pasted on a PostScript plot using psmerge
see selfdoc of psmerge for further information.
Possible fonts: Helvetica, Helvetica-Oblique, Helvetica-Bold,

Helvetica-BoldOblique,Times-Roman,Times-Italic,Times-Bold,
Times-BoldItalic,Courier,Courier-Bold,Courier-Oblique,
Courier-BoldOblique,Symbol
Possible colors: greenyellow,yellow,goldenrod,dandelion,apricot,

peach,melon,yelloworange,orange,burntorange,bittersweet,redorange,
mahogany,maroon,brickred,red,orangered,rubinered,wildstrawberry,
salmon,carnationpink,magenta,violetred,rhodamine,mulberry,redviolet,
fuchsia,lavender,thistle,orchid,darkorchid,purple,plum,violet,royalpurple,
blueviolet,periwinkle,cadetblue,cornflowerblue,midnightblue,naveblue,
royalblue,blue,cerulean,cyan,processblue,skyblue,turquoise,tealblue,
aquamarine,bluegreen,emerald,junglegreen,seagreen,green,forestgreen,
479
pinegreen,limegreen,yellowgreen,springgreen,olivegreen,rawsienna,sepia,
brown,tan,white,black,gray
example: tcolor=#255
AUTHOR: John E. Anderson, Visiting Scientist from Mobil, 1994
480
PSMANAGER - printer MANAGER for HP 4MV and HP 5Si Mx Laserjet
PostScript printing
psmanager < stdin [optional parameters] > stdout
none
papersize=0 paper size (US Letter default)
=1 US Legal
=2 A4
=3 11x17
orient=0 paper orientation (Portrait default)

=1 Landscape
tray=3 printing tray (Bottom tray default)

=1 tray 1 (multipurpose slot)
=2 tray 2
manual=0 no manual feed

=1 (Manual Feed)
media=0 regular paper

=1 Transparency
=2 Letterhead
=3 Card Stock
=4 Bond
=5 Labels
=6 Prepunched
=7 Recyled
=8 Preprinted
=9 Color (printing on colored paper)
Notes:
The option manual=1 implies tray=1. The media options apply
only to the HP LaserJet 5Si MX model printer.
Examples:
overheads:
psmanager < postscript_file manual=1 media=1 | lpr
labels:
psmanager < postscript_file manual=1 media=5 | lpr
481
Notes: This code was reverse engineered using output from
the NeXTStep printer manager.
Author: John Stockwell, June 1995, October 1997
Reference:
PostScript Printer Description File Format Specification,
version 4.2, Adobe Systems Incorporated
482
PSMERGE - MERGE PostScript files
psmerge in= [optional parameters] >postscriptfile
in= postscript file to merge
origin=0.0,0.0 x,y origin in inches
scale=1.0,1.0 x,y scale factors
rotate=0.0 rotation angle in degrees
translate=0.0,0.0 x,y translation in inches
Notes:
More than one set of in, origin, scale, rotate, and translate
parameters may be specified. Output x and y coordinates are
determined by:
x = tx + (x-ox)*sx*cos(d) - (y-oy)*sy*sin(d)
y = ty + (x-ox)*sx*sin(d) + (y-oy)*sy*cos(d)
where tx,ty are translate coordinates, ox,oy are origin coordinates,
sx,sy are scale factors, and d is the rotation angle. Note that the
order of operations is shift (origin), scale, rotate, and translate.
If the number of occurrences of a given parameter is less than the number

of input files, then the last occurrence of that parameter will apply to
all subsequent files.
483
PSMOVIE - PostScript MOVIE plot of a uniformly-sampled function f(x1,x2,x3)
psmovie n1= [optional parameters] <binaryfile >postscriptfile
484
n3=1 number of samples in third dimension
title2= second title to annotate different frames
loopdsp=3 display loop type (1=loop over n1; 2=loop over n2;
3=loop over n3)
f3=d3 first sample in 3rd dimension
NeXT: view movie via: psmovie < infile n1= [optional params...] | open
Note: currently only the Preview Application can handle the multipage
PostScript output by this program.
485
PSWIGB - PostScript WIGgle-trace plot of f(x1,x2) via Bitmap
Best for many traces. Use PSWIGP (Polygon version) for few traces.
pswigb n1= [optional parameters] <binaryfile >postscriptfile
x2=f2,f2+d2,... array of sampled values in 2nd dimension
bias=0.0 data value corresponding to location along axis 2
perc=100.0 percentile for determining clip
clip=(perc percentile) data values < bias+clip and > bias-clip are clipped
xcur=1.0 wiggle excursion in traces corresponding to clip
wt=1 =0 for no wiggle-trace; =1 for wiggle-trace
va=1 =0 for no variable-area; =1 for variable-area fill
=2 for variable area, solid/grey fill
SHADING: 2<= va <=5 va=2 lightgrey, va=5 black",
nbpi=72 number of bits per inch at which to rasterize
486
interp=0 no display interpolation
=1 use 8 point sinc interpolation
Notes:
The interp option may be useful for high nbpi values, however, it
tacitly assumes that the data are purely oscillatory. Non-oscillatory
data will not be represented correctly when this option is set.

pairs, one pair to a line. A "vector" of curve files and curve
colors may be specified as curvefile=file1,file2,etc. and
curvecolor=color1,color2,etc, and the number of pairs of values in each
file as npair=npair1,npair2,... .
487
PSWIGP - PostScript WIGgle-trace plot of f(x1,x2) via Polygons
Best for few traces. Use PSWIGB (Bitmap version) for many traces.
pswigp n1= [optional parameters] <binaryfile >postscriptfile
fill=1 =0 for no fill;
>0 for pos. fill;
<0 for neg. fill
=2 for pos. fill solid, neg. fill grey
=-2for neg. fill solid, pos. fill grey
SHADING: 2<=abs(fill)<=5 2=lightgrey 5=black
linewidth=1.0 linewidth in points (0.0 for thinest visible line)
tracecolor=black color of traces; should contrast with background
backcolor=none color of background; none means no background
488

Note: linewidth=0.0 produces the thinest possible line on the output.

device. Thus the result is device-dependent, put generally looks the
best for seismic traces.
The curve file is an ascii file with the points specified as x1 x2 pairs,
one pair to a line. A "vector" of curve files and curve colors may
be specified as curvefile=file1,file2,etc. and similarly
curvecolor=color1,color2,etc, and the number of pairs of values
489
LCMAP - List Color Map of root window of default screen
Usage: lcmap
490
LPROP - List PROPerties of root window of default screen of display
Usage: lprop
491
SCMAP - set default standard color map (RGB_DEFAULT_MAP)
Usage: scmap
492
XCONTOUR - X CONTOUR plot of f(x1,x2) via vector plot call
xcontour n1= [optional parameters] <binaryfile [>psplotfile]
X Functionality:
Button 1 Zoom with rubberband box
Button 2 Show mouse (x1,x2) coordinates while pressed
q or Q key Quit
s key Save current mouse (x1,x2) location to file
p or P key Plot current window with pswigb (only from disk files)
mpicks=/dev/tty file to save mouse picks in
nplaces=6 number of decimal places in contour labeling
xbox=50 x in pixels of upper left corner of window
ybox=50 y in pixels of upper left corner of window
wbox=550 width in pixels of window
hbox=700 height in pixels of window
493
labelfont=Erg14 font name for axes labels
titlefont=Rom22 font name for title
windowtitle=xwigb title on window
labelcolor=blue color for axes labels
titlecolor=red color for title
gridcolor=blue color for grid lines
labelccolor=black color of contour labels ",
labelcfont=fixed font name for contour labels
Notes:
For some reason the contour might slight differ from ones generated
by pscontour (propably due to the pixel nature of the plot
coordinates)
The line width of unlabeled contours is designed as a quarter of that

of labeled contours.
AUTHOR: Morten Wendell Pedersen, Aarhus University
All the coding is based on snippets taken from xwigb, ximage and pscontour
All I have done is put the parts together and put in some bugs ;-)
So credits should go to the authors of these packages...
Caveats and Notes:

The code has been developed under Linux 1.3.20/Xfree 3.1.2E (X11 6.1)
with gcc-2.7.0 But hopefully it should work on other platforms as well
Since all the contours are drawn by Vector plot call’s everytime the
494
Window is exposed, the exposing can be darn slow
OOPS This should be history... Now I keep my window content with backing
store so I won’t have to redraw my window unless I really have to...
Portability Question: I guess I should check if the display supports

backingstore and redraw if it doesn’t (see DoesBackingStore(3) )
I have to be able to use CWBackingStore==Always (other values can be
NonUseful and WhenMapped
Since I put the contour labels everytime I draw one contour level the area
that contains the label will be crossed by the the next contour lines...
(this bug also seems to be present in pscontour)
To fix this I have to redraw all the labels after been through all
the contour calls
Right now I can’t see a way to fix this without actually to through
the entire label positioning again....Overkill I would say
The relative short length of the contour segments will propably mask the
cdash settings
which means it is disposable (but I know how to draw dashed lines :)
A way of fixing this could be to get all connected point and then use
XDrawlines or XDrawSegments... just an idea...No idea if it’ll work.
I think there is a bug in xContour since my plot coordinates increase

North and west ward instead of south and eastward
I need to check the Self Doc so if the right parameters are described
(I have been through it a couple of times but....)
All functions need a heavy cleanup for unused variables

I suppose there is a couple of memory leaks due to missing free’ing of
numerous pointers (especially fonts,GC’s & colors could be a problem...
I have to browse through the internal pscontour call... basically what

I have done is just putting pscontour instead of pswigb... Instead of
repositioning the input file pointer (which doesnt work with pipes) one
should consider the use of temporary file
or write your zoombox to pscontour (...how one does that?)
Wish List:
The use of cgray’s unused until now... I guess I’ll need to allocate
a gray Colormap -> meaning that the code not will run at other display
495
than 8 bit Pseudocolor :( (with the use of present version of the colormap
library (code in $CWPROOT/src/xplot/lib ) )
The format of contour label should be open for the user..
It could be nice if one could choose to have a pixmap (like ximage )

underlying the contours... this should be defined either by the input
data or by a seperate file
eg useful for viewing traveltime contours on top a plot of the velocity
field
496
XESPB - X windows display of Encapsulated PostScript as a single Bitmap
Usage: xepsb < stdin
Caveat: your system must have Display PostScript Graphics
NOTE: This program is included as a demo of EPS -> X programming.

See: xepsp and xpsp these are more advanced versions
497
XEPSP - X windows display of Encapsulated PostScript
Usage: xepsp < stdin

Note:
this program is a more advanced version of xepsb. See also: xepsp
498
XIMAGE - X IMAGE plot of a uniformly-sampled function f(x1,x2)
ximage n1= [optional parameters] <binaryfile
X Functionality:
q or Q key Quit
a or page up keys enhance clipping by 10%
c or page down keys reduce clipping by 10%
up,down,left,right keys move zoom window by half width/height
i or +(keypad) zoom in by factor 2
o or -(keypad) zoom out by factor 2
... change colormap interactively

r install next RGB - colormap
R install previous RGB - colormap
h install next HSV - colormap
H install previous HSV - colormap
H install previous HSV - colormap
(Move mouse cursor out and back into window for r,R,h,H to take effect)
balance=0 bclip & wclip individually
=1 set them to the same abs value
if specified via perc (avoids colorbar skew)
499
cmap=hsv\’n\’ or rgb\’m\’ \’n\’ is a number from 0 to 13
\’m\’ is a number from 0 to 11
cmap=rgb0 is equal to cmap=gray
cmap=hsv1 is equal to cmap=hue
(compatibility to older versions)
lwidth=16 colorscale (legend) width in pixels
lheight=hbox/3 colorscale (legend) height in pixels
lx=3 colorscale (legend) x-position in pixels
ly=(hbox-lheight)/3 colorscale (legend) y-position in pixels
windowtitle=ximage title on window
blank=0 This indicates what portion of the lower range
to blank out (make the background color). The
value should range from 0 to 1.
500
plotfile=plotfile.ps filename for interactive ploting (P)
curvecolor=color1,color2,... color(s) for curve(s)
blockinterp=0 whether to use block interpolation (0=no, 1=yes)
NOTES:
pairs separated by a space, one pair to a line. A "vector" of curve
files and curve colors may be specified as curvefile=file1,file2,etc.
Stewart A. Levin, Mobil - Added ps print option
Brian Zook, Southwest Research Institute, 6/27/96, added blank option
Toralf Foerster, Baltic Sea Research Institute, 9/15/96, new colormaps
Berend Scheffers, Delft, colorbar (legend)
Brian K. Macy, Phillips Petroleum, 11/27/98, added curve plotting option
G.Klein, GEOMAR Kiel, 2004-03-12, added cursor scrolling and

interactive change of zoom and clipping.
Zhaobo Meng, ConocoPhillips, 12/02/04, added amplitude display
Garry Perratt, Geocon, 08/04/05, modified perc handling to center colorbar if balan
INTL2B_block - blocky interpolation of a 2-D array of bytes
intl2b_block blocky interpolation of a 2-D array of bytes
Function Prototype:
void intl2b_block(int nxin, float dxin, float fxin,
int nyin, float dyin, float fyin, unsigned char *zin,
int nxout, float dxout, float fxout,
int nyout, float dyout, float fyout, unsigned char *zout);
501
Input:
nxin number of x samples input (fast dimension of zin)
dxin x sampling interval input
fxin first x sample input
nyin number of y samples input (slow dimension of zin)
dyin y sampling interval input
fyin first y sample input
zin array[nyin][nxin] of input samples (see notes)
nxout number of x samples output (fast dimension of zout)
dxout x sampling interval output
fxout first x sample output
nyout number of y samples output (slow dimension of zout)
dyout y sampling interval output
fyout first y sample output
Output:
zout array[nyout][nxout] of output samples (see notes)
Notes:
The arrays zin and zout must passed as pointers to the first element of
a two-dimensional contiguous array of unsigned char values.
Constant extrapolation of zin is used to compute zout for

output x and y outside the range of input x and y.
Author: James Gunning, CSIRO Petroleum 1999. Hacked from

intl2b() by Dave Hale, Colorado School of Mines, c. 1989-1991
502
XPICKER - X wiggle-trace plot of f(x1,x2) via Bitmap with PICKing
xpicker n1= [optional parameters] <binaryfile
X Menu functionality:
Pick Filename Window default is pick_file
Load load an existing Pick Filename
Save save to Pick Filename
View only/Pick default is View, click to enable Picking
Add/Delete default is Add, click to delete picks
Cross off/on default is Cross off, click to enable Crosshairs
In View mode:
l lock the zoom while moving the coursor
u unlock the zoom
Notes:
Menu selections and toggles ("clicks") are made with button 1
Pick selections are made with button 3
Edit a pick selection by dragging it with button 3 down or
by making a new pick on that trace
Reaching the window limits while moving within changes the zoom
factor in this direction. The use of zoom locking(l) disables it
Other X Mouse functionality:

Mouse Button 1 Zoom with rubberbox
Mouse Button 2 Show mouse (x1,x2) coordinates while pressed
The following keys are active in View Only mode:
n1= number of samples in 1st (fast) dimension
mpicks=pick_file name of output (input) pick file
503
SHADING: 2<=va<=5 va=2 light grey, va=5 black
endian= =0 little endian, =1 big endian
interp=0 no sinc interpolation
=1 perform sinc interpolation
x2file= file of "acceptable" x2 values
x1x2=1 save picks in the order (x1,x2)
504
=0 save picks in the order (x2,x1)
Notes:
Xpicker will try to detect the endian value of the X-display and will
set it to the right value. If it gets obviously wrong information the
endian value will be set to the endian value of the machine that is
given at compile time as the value of CWPENDIAN defined in cwp.h
and set via the compile time flag ENDIANFLAG in Makefile.config.
The only time that you might want to change the value of the endian
variable is if you are viewing traces on a machine with a different
byte order than the machine you are creating the traces on AND if for
some reason the automaic detection of the display byte order fails.
Set endian to that of the machine you are viewing the traces on.
The interp flag is useful for making better quality wiggle trace for
making plots from screen dumps. However, this flag assumes that the
data are purely oscillatory. This option may not be appropriate for all
data sets.
If the x2file= option is set, then the values from the specified file
will define the set of "acceptable" values of x2 for xpicker to
output. The format is a single column of ASCII values. The number of
specified values is arbitrary.
Such a file can be built from an SU data set via:

sugethw < sudata key=offset output=geom > x2example
If the value of x2file= is not set, then

xpicker will use the values specified via: x2=.,.,.,. or those that are",
computed from the values of f2= and d2= as being the "acceptible
values.
See the selfdoc of suxpicker for information on using key fields from
the SU trace headers directly.

with picking by Wenying Cai of University of Utah.
Endian stuff by Morten Pedersen and John Stockwell of CWP.
Interp stuff by Tony Kocurko of Memorial University of Newfoundland
Modified to include acceptable values by Bill Lutter of the
505
Department of Geology, University of Wisconsin 10/96
MODIFIED: P. Michaels, Boise State Univeristy 29 December 2000
Added solid/grey color scheme for peaks/troughs
G.Klein, IFG Kiel University, 2003-09-29, added cursor scrolling and

NOTES:
Interactive picker improved to allow x-axis of picks to be
coordinated with "key=header" parameter set in driver routine
suxpicker. Multiple picks per trace are now allowed.
Input:
The command line of suxpicker is unchanged. The parameter"key=header"
set in suxpicker controls a) trace x-axis displayed via xpicker and
b) the header values in the first column of a pick file either read in
or written out from xpicker c) header values expected in optional file
or written out from xpicker c) header values expected in optional file
x2file= which reads into xpicker allowable trace x-axis values.
a) example command line: suxpicker key=offset < shot10.plotpik
b) pick file format:

x-axis_value_1 time_1
etc.
x-axis_value_n time_n
pick file example:

1000.000000 0.500000
2000.000000 1.000000
3000.000000 1.500000
4000.000000 2.000000
5000.000000 2.500000
c) format of optional file x2file=:

header_value_1
header_value_2
etc.
header_val_m
If file "x2file=" exists in directory from which suxpicker is
506
invoked, then these trace header x-axis values are the only allowable
x-axis pick values used in the pick "add" or "delete" menu operation.
Header values do not need to be sorted or 1 to 1 with input traces.
Further, pick file x-axis values can be read into xpicker via load
operation without having to match key_pickx1_val x-axis values and
can also be rewritten out an output pickfile. As indicated, only
the "add" and "delete" pick operations are influenced by existence
of this file.
Offset header values for "x2file=" can be generated by the

command line:
sugethw < su_segyfile key=offset output=geom > x2examplefile=
Output: Only change is in format of pick_file (format described above).

If x2file= file exists then x-axis value of added picks
will be forced to nearest allowable trace x-axis value (input values
of x2file= file). If x2file= is not set, then the values of x2
that are either assigned uniformly to the traces via f2 and d2,
or by the vector of values of x2=.,.,.,. will be the "acceptable"
values.
Strategy:
a) malloc() and realloc() used to dynamically allocate memory
for array of x-axis value read in from optional file
x2file=. This is done in function read_keyval().
b) The pick file dimensions are set in main program via malloc()
and then initialized (*apick)[i].picked = FALSE) in function
init_picks(). The pick file is declared as pick_t **apick, in
order to use realloc() as needed in functions load_picks where the
pick file is read in and edit_picks where picks are added. The
call to realloc() and further initializing is performed in
function realloc_picks().
c) If x2file= file exists the mouse derived x-axis value

for a pick to be added is checked against allowable x-axis values
to find the closest match via function add_pick called from
edit_picks. If the pick is to be deleted, first a search is done
to find the closest x-axis value, then the existing pick values
are searched to find the closest radial value (x**2 + t**2) via
function del_pick() invoked from edit_picks.
507
d) Code modifications are limited to above mentioned functions,
except for additional parameters passed to functions edit_picks,
load_picks, save_picks, and check_buttons.
508
XPSP - Display conforming PostScript in an X-window
xpsp < stdin
Note: this is the most advanced version of xepsb and xepsp.

509
XWIGB - X WIGgle-trace plot of f(x1,x2) via Bitmap
xwigb n1= [optional parameters] <binaryfile
X Functionality:
q or Q key Quit
l lock the zoom while moving the coursor
u unlock the zoom
1,2,...,9 Zoom/Move factor of the window size
Notes:
Reaching the window limits while moving within changes the zoom
factor in this direction. The use of zoom locking(l) disables it
SHADING: 2<=va<=5 va=2 light grey, va=5 black
510
windowtitle=xwigb title on window
endian= =0 little endian =1 big endian
interp=0 no interpolation in display
=1 use 8 point sinc interpolation
wigclip=0 If 0, the plot box is expanded to accommodate
the larger wiggles created by xcur>1. If this
flag is non-zero, the extra-large wiggles are
are clipped at the boundary of the plot box.
plotfile=plotfile.ps filename for interactive ploting (P)
curvecolor=color1,color2,... color(s) for curve(s)
Notes:
Xwigb will try to detect the endian value of the X-display and will
set it to the right value. If it gets obviously wrong information the
endian value will be set to the endian value of the machine that is
given at compile time as the value of CWPENDIAN defined in cwp.h
511
and set via the compile time flag ENDIANFLAG in Makefile.config.
The only time that you might want to change the value of the endian
variable is if you are viewing traces on a machine with a different
byte order than the machine you are creating the traces on AND if for
some reason the automaic detection of the display byte order fails.
Set endian to that of the machine you are viewing the traces on.
The interp flag is useful for making better quality wiggle trace for
making plots from screen dumps. However, this flag assumes that the
data are purely oscillatory. This option may not be appropriate for all
data sets.

pairs, separated by a space, one pair to a line. A "vector" of curve
files and curve colors may be specified as curvefile=file1,file2,etc.
Endian stuff by:

Morten Wendell Pedersen, Aarhus University (visiting CSM, June 1995)
& John Stockwell, Colorado School of Mines, 5 June 1995
Stewart A. Levin, Mobil - Added ps print option

John Stockwell - Added optional sinc interpolation
Stewart A. Levin, Mobil - protect title, labels in pswigb call
Brian J. Zook, SwRI - Added style=normal and wigclip flag
Brian K. Macy, Phillips Petroleum, 11/27/98, added curve plotting option

Curve plotting notes:
MODIFIED: P. Michaels, Boise State Univeristy 29 December 2000
Added solid/grey color scheme for peaks/troughs
G.Klein, IFG Kiel University, 2002-09-29, added cursor scrolling and

IFM-GEOMAR Kiel, 2004-03-12, added zoom locking
IFM-GEOMAR Kiel, 2004-03-25, interactive plotting fixed
512
XGRAPH - X GRAPHer
Graphs n[i] pairs of (x,y) coordinates, for i = 1 to nplot.
xgraph n= [optional parameters] <binaryfile
X Functionality:
q or Q key Quit
n array containing number of points per plot
nplot=number of n’s number of plots
d1=0.0,... x sampling intervals (0.0 if x coordinates input)
f1=0.0,... first x values (not used if x coordinates input)
d2=0.0,... y sampling intervals (0.0 if y coordinates input)
f2=0.0,... first y values (not used if y coordinates input)
pairs=1,... =1 for data pairs in format 1.a, =0 for format 1.b
linewidth=1,1,... line widths in pixels (0 for no lines)
linecolor=2,3,... line colors (black=0, white=1, 2,3,4 = RGB, ...)
mark=0,1,2,3,... indices of marks used to represent plotted points
marksize=0,0,... size of marks in pixels (0 for no marks)
reverse=0 =1 to reverse sequence of plotting curves ",
Optional resource parameters (defaults taken from resource database):

windowtitle= title on window
width= width in pixels of window
height= height in pixels of window
nTic1= number of tics per numbered tic on axis 1
grid1= grid lines on axis 1 - none, dot, dash, or solid
labelFont= font name for axes labels
titleFont= font name for title
titleColor= color for title
axesColor= color for axes
513
gridColor= color for grid lines
style= normal (axis 1 horizontal, axis 2 vertical) or

1.a. x1,y1,x2,y2,...,xn,yn
b. x1,x2,...,xn,y1,y2,...,yn
2. y1,y2,...,yn (must give non-zero d1[]=)
3. x1,x2,...,xn (must give non-zero d2[]=)
4. nil (must give non-zero d1[]= and non-zero d2[]=)
The formats may be repeated and mixed in any order, but if
formats 2-4 are used, the d1 and d2 arrays must be specified including
d1[]=0.0 d2[]=0.0 entries for any internal occurences of format 1.
Similarly, the pairs array must contain place-keeping entries for
plots of formats 2-4 if they are mixed with both formats 1.a and 1.b.
Also, if formats 2-4 are used with non-zero f1[] or f2[] entries, then
the corresponding array(s) must be fully specified including f1[]=0.0
and/or f2[]=0.0 entries for any internal occurences of format 1 or
formats 2-4 where the zero entries are desired.
mark index:
1. asterisk
2. x-cross
3. open triangle
4. open square
5. open circle
6. solid triangle
7. solid square
8. solid circle
Note: n1 and n2 are acceptable aliases for n and nplot, respectively.
Example:
xgraph n=50,100,20 d1=2.5,1,0.33 <datafile
plots three curves with equally spaced x values in one plot frame
x1-coordinates are x1(i) = f1+i*d1 for i = 1 to n (f1=0 by default)
number of x2’s and then x2-coordinates for each curve are read
sequentially from datafile.
514
XMOVIE - image one or more frames of a uniformly sampled function f(x1,x2)
xmovie n1= n2= [optional parameters] <fileoffloats
X Functionality:
Button 2 reverse the direction of the movie.
Button 3 stop and start the movie.
q or Q key Quit
s or S key stop display and switch to Step mode
b or B key set frame direction to Backward
f or F key set frame direction to Forward
n or N key same as ’f’
c or C key set display mode to Continuous mode
n1= number of samples in 1st (fast) dimension
n2= number of samples in 2nd (slow) dimension
fframe=1 value corresponding to first frame
dframe=1 frame sampling interval
loop=0 =1 to loop over frames after last frame is input
=2 to run movie back and forth
interp=1 =0 for a non-interpolated, blocky image
idm=0 =1 to set initial display mode to stepmode
515
windowtitle= title on window and icon
sleep= delay between frames in microseconds
Color options:
cmap=gray gray, hue, saturation, or default colormaps may be specified
bhue=0 hue mapped to bclip (hue and saturation maps)
whue=240 hue mapped to wclip (hue and saturation maps)
sat=1 saturation (hue map only)
bright=1 brightness (hue and saturation maps)
white=(bclip+wclip)/2 data value mapped to white (saturation map only)
Notes:
Colors are specified using the HSV color wheel model:
Hue: 0=360=red, 60=yellow, 120=green, 180=cyan, 240=blue, 300=magenta
Saturation: 0=white, 1=pure color
Value (brightness): 0=black, 1=maximum intensity
For the saturation mapping (cmap=sat), data values between white and bclip
are mapped to bhue, with saturation varying from white to the pure color.
Values between wclip and white are similarly mapped to whue.
For the hue mapping (cmap=hue), data values between wclip and bclip are
mapped to hues between whue and bhue. Intermediate hues are found by
moving counterclockwise around the circle from bhue to whue. To reverse
the polarity of the image, exchange bhue and whue. Equivalently,
exchange bclip and wclip (setting perc=0 is an easy way to do this).
Hues in excess of 360 degrees can be specified in order to reach the
opposite side of the color circle, or to wrap around the circle more
than once.
516
The title string may contain a C printf format string containing a
conversion character for the frame number. The frame number is
computed from dframe and fframe. E.g., try setting title="Frame %g".
517
XRECTS - plot rectangles on a two-dimensional grid
xrects x1min= x1max= x2min= x2max= [optional parameters] <rectangles
x1min minimum x1 coordinate
x1max maximum x1 coordinate
x2min minimum x2 coordinate
x2max maximum x2 coordinate
color=red color used for rectangules

518
FFTLAB - Motif-X based graphical 1D Fourier Transform
Usage: fftlab
Caveat: you must have the Motif Developer’s package to install this code
519
supscontour <stdin [optional parameters] | ...


=1.0 (if not set)


not.
See the pscontour selfdoc for the remaining parameters.
On NeXT: supscontour < infile [optional parameters] | open
dt, d1, d2, f1, f2
Credits:
520
CWP: Dave Hale and Zhiming Li (pscontour, etc.)
Jack Cohen and John Stockwell (supscontour, etc.)
Notes:
the traces for pscontour. You can make this value "known"



for code clarity.
521
supscubecontour <stdin [optional parameters] | ...




=1.0 (if not set)


not.
See the pscubecontour selfdoc for the remaining parameters.
example: supscubecontour < infile [optional parameters] | gv -
522
Credits:

John Stockwell (supscubecontour)
Notes:
for code clarity.
523
supscube <stdin [optional parameters] | ...




=1.0 (if not set)


not.
See the pscube selfdoc for the remaining parameters.
On NeXT: supscube < infile [optional parameters] | open
524
Credits:

John Stockwell (supscube)
Notes:
for code clarity.
525
supsgraph <stdin [optional parameters] | ...
(see psgraph selfdoc for style definitions)
nplot is the number of traces (ntr is an acceptable alias for nplot)


=1.0 (if not set)


not.
On NeXT: supsgraph < infile [optional parameters] | open
Credits:
526
Notes:

for code clarity.
527
supsimage <stdin [optional parameters] | ...


=1.0 (if not set)


not.
See the psimage selfdoc for the remaining parameters.
On NeXT: supsimage < infile [optional parameters] | open
dt, d1, d2, f1, f2
Credits:
528
CWP: Dave Hale and Zhiming Li (psimage, etc.)
Jack Cohen and John Stockwell (supsimage, etc.)
Notes:
the traces for psimage. You can make this value "known"


for code clarity.
529
of a SEGY (SU) data set
supsmax <stdin >postscript file [optional parameters]
mode=max max value
=min min value


=1.0 (if not set)


not.

530
Credits:
Notes:
the traces for psgraph. You can make this value "known"

531
supsmovie <stdin [optional parameters] | ...




=1.0 (if not set)


not.
See the psmovie selfdoc for the remaining parameters.
On NeXT: supsmovie < infile [optional parameters] | open
532
Credits:
CWP: Dave Hale and Zhiming Li (psmovie)

John Stockwell (supsmovie)
Notes:
for psmovie. In this case:
we are using tmpfile because on many machines it is
implemented as a memory area instead of a disk file.
Although we compute ntr, we don’t allocate a 2-d array
for code clarity.
533
supswigb <stdin [optional parameters] | ...
=1.0 (if not set)


not.
This program is really just a wrapper for the plotting program: pswigb
See the pswigb selfdoc for the remaining parameters.
534
dt, d1, d2, f1, f2, keyword (if set)
Credits:
CWP: Dave Hale and Zhiming Li (pswigb, etc.)

Jack Cohen and John Stockwell (supswigb, etc.)

Notes:
the traces for pswigb. You can make this value "known"

for code clarity.
535
supswigp <stdin [optional parameters] | ...
key=(keyword) if set, values of x2 are set from header field
=1.0 (if not set)


not.
This program is really just a wrapper for the plotting program: pswigp
See the pswigp selfdoc for the remaining parameters.
536
On NeXT: supswigp < infile [optional parameters] | open
dt, d1, d2, f1, f2, key specified by key
Credits:


Notes:



for code clarity.
537
=1.0 (if not set)

not.
This program is really just a wrapper for the plotting program: xcontour
See the xcontour selfdoc for the remaining parameters.
Credits:
538
xwigb with xcountour ;-)
Notes:
the traces for xcontour. You can make this value "known"

for code clarity.
539
suxgraph <stdin [optional parameters] | ...
(see xgraph selfdoc for style definitions)
nplot= number of traces (ntr is an acceptable alias for nplot)


=1.0 (if not set)


not.
On NeXT: suxgraph < infile [optional parameters] | open
Credits:
540
Notes:


for code clarity.
541
suximage infile= [optional parameters] | ... (direct I/O)

or
suximage <stdin [optional parameters] | ... (sequential I/O)
infile=NULL SU data to be ploted, default stdin with sequential access

if ’infile’ provided, su data read by (fast) direct access
with ftr,dtr and n2 suximage will pass a subset of data

to the plotting program-ximage:
ftr=1 First trace to be plotted
dtr=1 Trace increment to be plotted
n2=tr.ntr (Max) number of traces to be plotted (ntr is an alias for n2)
Priority: first try to read from parameter line;
if not provided, check trace header tr.ntr;
if still not provided, figure it out using ftello


=1.0 (if not set or was set to 0)
key= key for annotating d2 (slow dimension)

If annotation is not at proper increment, try
setting d2; only first trace’s key value is read


542
not.
See the ximage selfdoc for the remaining parameters.
Credits:
CWP: Dave Hale and Zhiming Li (ximage, etc.)

Jack Cohen and John Stockwell (suximage, etc.)
MTU: David Forel, June 2004, added key for annotating d2
ConocoPhillips: Zhaobo Meng, Dec 2004, added direct I/O
Notes:
When provide ftr and dtr and infile, suximage can be used to plot
multi-dimensional volumes efficiently. For example, for a Offset-CDP
dataset with 32 offsets, the command line
suximage infile=volume3d.su ftr=1 dtr=32 ... &
will display the zero-offset common offset data with ranrom access.
It is highly recommend to use infile= to view large datasets, since
using stdin only allows sequential access, which is very slow for
large datasets.

the traces for ximage. You can make this value "known"

for code clarity.
543
each trace of a SEGY (SU) data set
suxmax <stdin [optional parameters]
mode=max max value
=min min value


=1.0 (if not set)


not.

544
Credits:
Notes:
the traces for xgraph. You can make this value "known"

545
suxmovie <stdin [optional parameters]
n1=tr.ns number of samples per trace

ntr=tr.ntr number of traces in the data set
n2=tr.shortpad or tr.ntr number of traces in inline direction
n3=ntr/n2 number of traces in crossline direction


=1.0 (if not set)
d3=1.0 sampling interval in the slowest dimension
f1=tr.f1 or 0.0 first sample in the z dimension

f2=tr.f2 or 1.0 first sample in the x dimension
f3=1.0
mode=0 0= x,z slice movie through y dimension (in line)

1= y,z slice movie through x dimension (cross line)
2= x,y slice movie through z dimension (time slice)

Notes:
For seismic data, the "fast dimension" is either time or
depth and the "slow dimension" is usually trace number.
The 3D data set is expected to have n3 sets of n2 traces representing
the horizontal coverage of n2*d2 in x and n3*d3 in y direction.
The data is read to memory with and piped to xmovie with the
respective sampling parameters.
546
See the xmovie selfdoc for the remaining parameters and X functions.
547
suxpicker <stdin [optional parameters] | ...



=1.0 (if not set)


not.
548
See the xpicker selfdoc for the remaining parameters.
Credits:
CWP: Dave Hale and Zhiming Li (xpicker, etc.)

Jack Cohen and John Stockwell (suxpicker, etc.)
Notes:
the traces for xpicker. You can make this value "known"



for code clarity.
549
This is a modified suxwigb that uses the depth or coordinate scaling
when such values are used as keys.
=1.0 (if not set)


not.
550
This program is really just a wrapper for the plotting program: xwigb
See the xwigb selfdoc for the remaining parameters.
Credits:


Notes:
the traces for xwigb. You can make this value "known"

for code clarity.
551
SXPLOT - X Window plot a triangulated sloth function s(x1,x2)
sxplot <modelfile [optional parameters]
edgecolor=cyan color to draw fixed edges
tricolor=yellow color to draw non-fixed edges of triangles
=none non-fixed edges of triangles are not shown
bclip=minimum sloth sloth value corresponding to black
wclip=maximum sloth sloth value corresponding to white
x1beg=x1min value at which x1 axis begins
x1end=x1max value at which x1 axis ends
x2beg=x2min value at which x2 axis begins
x2end=x2max value at which x2 axis ends
cmap=gray gray, hue, or default colormaps may be specified

552
GBBEAM - Gaussian beam synthetic seismograms for a sloth model
gbbeam <rayends >syntraces xg= zg= [optional parameters]
xg= x coordinates of receiver surface
zg= z coordinates of receiver surface
ng=101 number of receivers (uniform distributed along surface)
krecord=1 integer index of receiver surface (see notes below)
ang=0.0 array of angles corresponding to amplitudes in amp
amp=1.0 array of amplitudes corresponding to angles in ang
bw=0 beamwidth at peak frequency
reftrans=0 =1 complex refl/transm. coefficients considered
prim =1, only single-reflected rays are considered ",
=0, only direct hits are considered
atten=0 =1 add noncausal attenuation
=2 add causal attenuation
lscale= if defined restricts range of extrapolation
aperture= maximum angle of receiver aperture
fpeak=0.1/dt peak frequency of ricker wavelet
infofile ASCII-file to store useful information
NOTES:
Only rays that terminate with index krecord will contribute to the
synthetic seismograms at the receiver (xg,zg) locations. The
receiver locations are determined by cubic spline interpolation
of the specified (xg,zg) coordinates.

MODIFIED: Andreas Rueger, Colorado School of Mines, 08/18/93
Modifications include: 2.5-D amplitudes, computation of reflection/
transmission losses, attenuation,
timewindow, lscale, aperture, beam width, etc.
553
NORMRAY - dynamic ray tracing for normal incidence rays in a sloth model
normray <modelfile >rayends [optional parameters]
caustic 0: show all rays 1: show only caustic rays
nonsurface 0: show rays which reach surface 1: show all rays
surface 0: shot ray from subsurface 1: from surface
nrays number of location to shoot rays
dangle increment of ray angle for one location
nangle number of rays shot from one location
ashift shift first taking off angle
xs1 x of shooting location
zs1 z of shooting location
nangle=101 number of takeoff angles
fangle=-45 first takeoff angle (in degrees)
rayfile file of ray x,z coordinates of ray-edge intersections
nxz=101 number of (x,z) in optional rayfile (see notes below)
wavefile file of ray x,z coordinates uniformly sampled in time
nt=101 number of (x,z) in optional wavefile (see notes below)
infofile ASCII-file to store useful information
fresnelfile used if you want to plot the fresnel volumes.
default is <fresnelfile.bin>
outparfile contains parameters for the plotting software.
default is <outpar>
krecord if specified, only rays incident at interface with index
krecord are displayed and stored
prim =1, only single-reflected rays are plotted ",
=0, only direct hits are displayed
ffreq=-1 FresnelVolume frequency
refseq=1,0,0 index of reflector followed by sequence of reflection (1)
transmission(0) or ray stops(-1).
The default rayend is at the model boundary.
NOTE:refseq must be defined for each reflector
NOTES:
The rayends file contains ray parameters for the locations at which
the rays terminate.
The rayfile is useful for making plots of ray paths.

nxz should be larger than twice the number of triangles intersected
by the rays.
The wavefile is useful for making plots of wavefronts.
554
The time sampling interval in the wavefile is tmax/(nt-1),
where tmax is the maximum time for all rays.
The infofile is useful for collecting information along the

individual rays. The fresnelfile contains data used to plot
the Fresnel volumes. The outparfile stores information used
for the plotting software.

Modifications include: functions writeFresnel, checkIfSourceIsOnEdge;
options refseq=, krecord=, prim=, infofile=;
computation of reflection/transmission losses, attenuation.
MODIFIED: Boyi Ou, Colorado School of Mines, 4/14/95
Notes:
This code can shoot rays from specified interface by users, normally you
need to use gbmodel2 to generate interface parameters for this code, both
code have a parameter named nrays, it should be same. If you just want to
shoot rays from one specified location, you need to specify xs1,zs1,
otherwise, leave them alone. If you want to shoot rays from surface, you need
to define surface equal to 1. The rays from one location will be
approximately symetric with direction Normal_direction - ashift.(if nangle is
odd, it is symetric, even, almost symetric. The formula for the first take
off angle is: angle=normal_direction-nangle/2*dangle-ashift. If you only want to
see caustics, you specify caustic=1, if you want to see rays which does not
reach surface, you specify nonsurface=1.
/
555
TRI2UNI - convert a TRIangulated model to UNIformly sampled model
tri2uni <triangfile >uniformfile n2= n1= [optional parameters]
n1= number of samples in the first (fast) dimension
n2= number of samples in the second dimension
d1=1.0 sampling interval in first (fast) dimension
d2=1.0 sampling interval in second dimension
f1=0.0 first value in dimension 1 sampled
f2=0.0 first value in dimension 2 sampled
Note:
The triangulated/uniformly-sampled quantity is assumed to be sloth=1/v^2
556
TRIMODEL - make a triangulated sloth (1/velocity^2) model
trimodel >modelfile [optional parameters]
xmin=0.0 minimum horizontal coordinate (x)
xmax=1.0 maximum horizontal coordinate (x)
zmin=0.0 minimum vertical coordinate (z)
zmax=1.0 maximum vertical coordinate (z)
xedge= x coordinates of an edge
zedge= z coordinates of an edge
sedge= sloth along an edge
kedge= array of indices used to identify edges
normray 0:do not generate parameters 1: does it
normface specify which interface to shoot rays
nrays number of locations to shoot rays
sfill= x, z, x0, z0, s00, dsdx, dsdz to fill a region
densfill= x, z, dens to fill a region
qfill= x, z, Q-factor to fill a region
maxangle=5.0 maximum angle (deg) between adjacent edge segments
Notes:
More than set of xedge, zedge, and sedge parameters may be
specified, but the numbers of these parameters must be equal.
Within each set, vertices will be connected by fixed edges.
Edge indices in the k array are used to identify edges

specified by the x and z parameters. The first k index
corresponds to the first set of x and z parameters, the
second k index corresponds to the second set, and so on.
After all vertices and their corresponding sloth values have

been inserted into the model, the optional sfill parameters
are used to fill closed regions bounded by fixed edges.
Let (x,z) denote any point inside a closed region. Sloth inside
this region is determined by s(x,z) = s00+(x-x0)*dsdx+(z-z0)*dsdz.
The (x,z) component of the sfill parameter is used to identify a
closed region.
557
Fill regions with attenuation Q-factors and density values.
Corrected bug in computing s00 in makeSlothForTri() function.
MODIFIED: Boyi Ou, Colorado School of Mines, 4/14/95
Make code to generate interface parameters for shooting rays
from specified interface
NOTE:
When you use normface to specify interface, the number of interface might
not be the number of interface in the picture, for example, you build a one
interface model, this interface is very long, so in the shell, you use three
part of xedge,zedge,sedge to make this interface, so when you use normface to
specify interface, this interface is just part of whole interface. If you
want see the normal rays from entire interface, you need to maek normal ray
picture few times, and merge them together.
558
TRIRAY - dynamic RAY tracing for a TRIangulated sloth model
triray <modelfile >rayends [optional parameters]
xs=(max-min)/2 x coordinate of source (default is halfway across model)
zs=min z coordinate of source (default is at top of model)
nangle=101 number of takeoff angles
fangle=-45 first takeoff angle (in degrees)
langle=45 last takeoff angle (in degrees)
rayfile= file of ray x,z coordinates of ray-edge intersections
nxz=101 number of (x,z) in optional rayfile (see notes below)
wavefile= file of ray x,z coordinates uniformly sampled in time
nt=101 number of (x,z) in optional wavefile (see notes below)
infofile= ASCII-file to store useful information
fresnelfile= used if you want to plot the fresnel volumes.
default is <fresnelfile.bin>
outparfile= contains parameters for the plotting software.
default is <outpar>
krecord= if specified, only rays incident at interface with index
krecord are displayed and stored
prim= =1, only single-reflected rays are plotted
=0, only direct hits are displayed
ffreq=-1 FresnelVolume frequency
refseq=1,0,0 index of reflector followed by sequence of reflection (1)
transmission(0) or ray stops(-1).
The default rayend is at the model boundary.
NOTE:refseq must be defined for each reflector
NOTES:
The rayends file contains ray parameters for the locations at which
the rays terminate.
The rayfile is useful for making plots of ray paths.

nxz should be larger than twice the number of triangles intersected
by the rays.
The wavefile is useful for making plots of wavefronts.

The time sampling interval in the wavefile is tmax/(nt-1),
where tmax is the maximum time for all rays.
The infofile is useful for collecting information along the

individual rays. The fresnelfile contains data used to plot
the Fresnel volumes. The outparfile stores information used
559
for the plotting software.

Modifications include: functions writeFresnel, checkIfSourceIsOnEdge;
options refseq=, krecord=, prim=, infofile=;
computation of reflection/transmission losses, attenuation.
560
TRISEIS - Gaussian beam synthetic seismograms for a sloth model
triseis <modelfile >seisfile xs= zs= xg= zg= [optional parameters]
xs= x coordinates of source surface
zs= z coordinates of source surface
xg= x coordinates of receiver surface
zg= z coordinates of receiver surface
ns=1 number of sources uniformly distributed along s surface
ds= increment between source locations (see notes below)
fs=0.0 first source location (relative to start of s surface)
ng=101 number of receivers uniformly distributed along g surface
dg= increment between receiver locations (see notes below)
fg=0.0 first receiver location (relative to start of g surface)
dgds=0.0 change in receiver location with respect to source location
krecord=1 integer index of receiver surface (see notes below)
kreflect=-1 integer index of reflecting surface (see notes below)
prim =1, only single-reflected rays are considered ",
=0, only direct hits are considered
bw=0 beamwidth at peak frequency
nangle=101 number of ray takeoff angles
fangle=-45 first ray takeoff angle (in degrees)
langle=45 last ray takeoff angle (in degrees)
reftrans=0 =1 complex refl/transm. coefficients considered
atten=0 =1 add noncausal attenuation
=2 add causal attenuation
lscale= if defined restricts range of extrapolation
fpeak=0.1/dt peak frequency of ricker wavelet
aperture= maximum angle of receiver aperture
NOTES:
Only rays that terminate with index krecord will contribute to the
synthetic seismograms at the receiver (xg,zg) locations. The
source and receiver locations are determined by cubic spline
interpolation of the specified (xs,zs) and (xg,zg) coordinates.
The default source location increment (ds) is determined to span
the source surface defined by (xs,zs). Likewise for dg.
561
Modifications include: 2.5-D amplitudes, correction for ref/transm,
timewindow, lscale, aperture, beam width, etc.
562
UNI2TRI - convert UNIformly sampled model to a TRIangulated model
uni2tri <slothfile >modelfile n2= n1= [optional parameters]
n1= number of samples in first (fast) dimension
n2= number of samples in second dimension
d1=1.0 sampling interval in dimension 1
d2=1.0 sampling interval in dimension 2
f1=0.0 first value in dimension 1
f2=0.0 first value in dimesion 2
ifile= triangulated model file used as initial model
errmax= maximum sloth error (see notes below)
verbose=1 =0 for silence
=1 to report maximum error at each stage to stderr
=2 to also write the normalized error to efile
efile=emax.dat filename for error file (for verbose=2)
mm=0 output every mm-th intermediate model (0 for none)
mfile=intmodel intermediate models written to intmodel%d
method=3 =1 add 1 vertex at maximum error
=2 add vertex to every triangle that exceeds errmax
=3 method 2, but avoid closely spaced vertices
tol=10 closeness criterion for (in samples)
sfill= x, z, x0, z0, s00, dsdx, dsdz to fill a region
Notes:
Triangles are constructed until the maximum error is
not greater than the user-specified errmax. The default errmax
is 1% of the maximum value in the sampled input file.
After the uniform values have been triangulated, the optional sfill
parameters are used to fill closed regions bounded by fixed edges.
Let (x,z) denote any point inside a closed region. Values inside
this region is determined by s(x,z) = s00+(x-x0)*dsdx+(z-z0)*dsdz.
The (x,z) component of the sfill parameter is used to identify a
closed region.
The uniformly sampled quantity is assumed to be sloth=1/v^2.
563
AUTHOR: Craig Artley, Colorado School of Mines, 03/31/94
NOTE: After a program outlined by Dave Hale, 12/27/90.
564
SPSPLOT - plot a triangulated sloth function s(x,z) via PostScript
spsplot <modelfile >postscriptfile [optional parameters]
gedge=0.0 gray to draw fixed edges (in interval [0.0,1.0])
gtri=1.0 gray to draw non-fixed edges of triangles
gmin=0.0 min gray to shade triangles (in interval [0.0,1.0])
gmax=1.0 max gray to shade triangles (in interval [0.0,1.0])
sgmin=minimum s(x,z) s(x,y) corresponding to gmin
sgmax=maximum s(x,z) s(x,y) corresponding to gmax
xbeg=xmin value at which x axis begins
xend=xmax value at which x axis ends
dxnum=0.0 numbered tic interval on x axis (0.0 for automatic)
fxnum=xmin first numbered tic on x axis (used if dxnum not 0.0)
nxtic=1 number of tics per numbered tic on x axis
gridx=none grid lines on x axis - none, dot, dash, or solid
labelx= label on x axis
zbeg=zmin value at which z axis begins
zend=zmax value at which z axis ends
dznum=0.0 numbered tic interval on z axis (0.0 for automatic)
fznum=zmin first numbered tic on z axis (used if dynum not 0.0)
nztic=1 number of tics per numbered tic on z axis
gridz=none grid lines on z axis - none, dot, dash, or solid
labelz= label on z axis
style=seismic normal (z axis horizontal, x axis vertical) or
seismic (z axis vertical, x axis horizontal)
Note: A value of gedge or gtri outside the interval [0.0,1.0]

results in that class of edge not being drawn.
565
Tweaks to improve PostScript header, add basic color support.
NOTE: Have observed errors in output when compiled with optimization

under NEXTSTEP 3.1. Caveat Emptor.

566
Shells:
______
ARGV - give examples of dereferencing char **argv
Usage: argv
______
COPYRIGHT - insert CSM COPYRIGHT lines at top of files in working directory
Usage: copyright file(s)
______
CPALL , RCPALL - for local and remote directory tree/file transfer
Usage: cpall sourcedir destinationdir

Caveat: destinationdir must exist and be writeable by the user
Usage: rcpall sourcedir remotemachine destinationdir
If user name is different on the remote machine, then second"

entry is "remotemachine -l remoteusername"
Caveats: rsh, copy, and write permissions required
You must be on the source machine,
destinationdir must exist and be writeable by the user.
Notes: Both of these shell scripts use tar to do the transfer.
______
CWPFIND - look for files with patterns in CWPROOT/src/cwp/lib
Usage: cwpfind pattern_fragment

cwpfind -e exact_pattern
______
DIRTREE - show DIRectory TREE
Usage: dirtree
567
______
FILETYPE - list all files of given type
Usage: filetype string_from_file_output
Examples:
filetype text - list printable files
filetype stripped - list unstripped files
______
Grep - recursively call egrep in pwd
Usage: Grep [-egrep_options] pattern
Caution: Do NOT redirect into file in pwd, either use something

like >../Grep.out or perhaps pipe output into mail to yourself.
Author: Jack, 7/95
______
NEWCASE - Changes the case of all the filenames in a directory, dir
Usage: newcase -l dir change all filenames to lower case

-u dir change all filenames to upper case
Notes: Useful for files downloaded from VAX.
______
OVERWRITE - copy stdin to stdout after EOF
This shell is called from the shell script: replace
______
PRECEDENCE - give table of C precedences from Kernighan and Ritchie
Usage: precedence
______
REPLACE - REPLACE string1 with string2 in files
Usage: replace string1 string2 files
568
______
THIS_YEAR - print the current year
Usage: this_year
NOTES - useful for building dated filenames, etc.
______
TIME_NOW - prints time in ZULU format with no spaces
Usage: time_now
Note: Useful for building dated filenames
______
TODAYS_DATE - prints today’s date in ZULU format with no spaces
Usage: todays_date
______
USERNAMES - get list of all login names
Usage: usernames
______
VARLIST - list variables used in a Fortran program
Usage: varlist file.f ...
Output is in the file: vars.file
______
WEEKDAY - prints today’s WEEKDAY designation
Usage: weekday
569
______
ZAP - kill processes by name
Typical usages:
zap ximage
zap ’xmovie|xgraph’
Zap accepts full pattern matching for the process names
Caveat: zap assumes that the FIRST field produced by the

Unix "ps" command is the pid (process identifier) number.
If not, change the number in the awk print statement to
the appropriate field.
Author: Jack, 6/95 -- after Kernighan and Pike’s zap
570
______
GENDOCS - generate complete list of selfdocs in latex form
Usage: gendocs -o output filename is: selfdocs.tex
______
STRIPTOTXT - put files from $CWPROOT/src/doc/Stripped into a new
directory in the form $CWPROOT/src/TXT/NAME.txt
Usage: striptotxt
Author: John Stockwell, Sept 2001
______
UPDATEDOCALL - put self-docs in ../doc/Stripped
Usage: updatedocall
Note: this shell uses updatedoc to update the database used by

suname and gendocs
______
UPDATEDOC - put self-docs in ../doc/Stripped and ../doc/Headers
Usage: updatedoc path
Notes:
Paths include: cwp/main cwp/lib cwp/shell par/main par/lib par/shell
xplot/main xplot/lib psplot/main psplot/lib psplot/shell
Xtcwp/main Xtcwp/lib su/main su/lib su/shell su/graphics/psplot
su/graphics/xplot tri/main tri/lib xtri tri/graphics/psplot
tetra/lib tetra/main
comp/dct/lib comp/dct/main comp/dct/libutil comp/dwpt/1d/lib
comp/dwpt/1d/main comp/dwpt/2d/lib comp/dwpt/2d/main
Use: updatedocall to update full directory, use updatehead to

to update the master header file.
This shell builds the database used by suname and gendocs
______
UPDATEHEAD - update ../doc/Headers/Headers.all
571
Usage: updatehead
Notes:
This file builds the database used by suname
572
______
LOOKPAR - show getpar lines in SU code with defines evaluated
Usage: lookpar filename ...
______
MAXDIFF - find absolute maximum difference in two segy data sets
Usage: maxdiff file1 file2
______
RECIP - sum opposing (reciprocal) offsets in cdp sorted data
Usage: recip <stdin >stdout
______
RMAXDIFF - find percentage maximum difference in two segy data sets
Usage: rmaxdiff file1 file2
______
SUAGC - perform agc on SU data
Note: this is an interface to sugain for backward compatibility

See selfdoc of: sugain for more information
______
SUBAND - Trapezoid-like Sin squared tapered Bandpass filter via SUFILTER
Usage: suband < stdin > stdout
Note: this shell mimmics the old program SUBAND, supersceded by SUFILTER
See selfdoc of: sufilter for more information
______
SUDIFF, SUSUM, SUPROD, SUQUO, SUPTDIFF, SUPTSUM,
SUPTPROD, SUPTQUO - difference, sum, product, quotient of two SU data
sets via suop2
Usage:
573
sudiff file1 file2 > stdout
susum file1 file2 > stdout
...etc
Note: uses suop2 to perform the computation
______
SUDOC - get DOC listing for code
Usage: sudoc name
Note: Use this shell script to get selfdoc information for

codes labeled with and asterisk (*) or pound sign (#) in suname list
______
SUENV - Instantaneous amplitude, frequency, and phase via: SUATTRIBUTES
Usage: suenv < stdin > stdout
Note: this shell mimmics the old program SUENV, supersceded by SUATTRIBUTES
See selfdoc of: suattributes for more information
______
SUFIND - get info from self-docs
Usage: sufind [-v -n] string
sufind string gives a brief synopsis

sufind -v string is a verbose hunt for relevant items
sufind -n name_fragment searches for command name
______
SUFIND - get info from self-docs
Usage: sufind [-v -n] string
sufind string gives a brief synopsis

sufind -v string is a verbose hunt for relevant items
sufind -n name_fragment searches for command name
Author: CWP: Jack K. Cohen, 1992

Modified by: CWP: S. Narahara, 04/11/1998.
574
______
SUGENDOCS - generate complete list of selfdocs in latex form
Usage: sugendocs -o output filename is: selfdocs.tex

Note: this shell simply calls gendocs
______
SUHELP - list the CWP/SU programs and shells
Usage: suhelp
______
SUKEYWORD -- guide to SU keywords in segy.h
Usage: sukeyword -o to begin at the top of segy.h

sukeyword [string] to find [string]
Note: keyword= occurs in many SU programs.
______
SUNAME - get name line from self-docs
Usage: suname [name]
Note: dummy selfdocs have been included in all cwp and shell programs
that don’t have automatic selfdocs.
______
UNGLITCH - zonk outliers in data
Usage: unglitch < stdin
Note: this shell just invokes: sugain < stdin qclip=.99 > stdout
See selfdoc of: sugain for further information
575
______
MERGE2 - put 2 standard size PostScript figures on one page
Usage: merge2 fig1 fig2
Notes: Translation values are hard-coded numbers that work well for
standard size (8.5 x 11) figures.
See selfdoc of: psmerge for details
______
MERGE4 - put 4 standard size PostScript plots on one page
Usage: merge4 ulfig urfig llfig lrfig
Note: Translation values are hard-coded numbers that work well for
standard size (8.5 x 11) figures.
See selfdoc of: psmerge for further information
576
Libs:
BASIC - Basic C function interface to PostScript
beginps write PostScript prolog (including %%Pages comment)

endps write PostScript trailer (including %%Pages comment)
begineps write encapsulated PostScript prolog (no %%Pages comment)
endeps write encapsulated PostScript trailer (no %%Pages comment)
boundingbox set BoundingBox to llx lly urx ury
newpage print "%%%%Page: label ordinal" to stdout
showpage print "showpage" to stdout
gsave print "GS" to stdout
grestore print "GR" to stdout
newpath print "NP" to stdout
closepath print "CP" to stdout
clip print "clip" to stdout
translate print "tx ty TR" to stdout, tx,ty = translation in x,y
scale print "sx sy SC" to stdout, sx,sy = scaling in x,y
rotate print "angle RO" to stdout, angle = rotation angle
concat print "m[0] m[1] m[2] m[3] m[4] m[5] CAT" to stdout
setgray print "gray setgray" to stdout, gray is 0-255 gray level
setrgbcolor print "red green blue setrgbcolor" to stdout
red,green,blue = 0-255 red,green,blue levels
setcolor set color by name based on definition in color structure
setlinewidth print "width SLW" to stdout, width = desired line width
setlinejoin print "code setlinejoin"
setdash print "[ dash ] offset setdash" to stdout
dash = array defining dash, offset = dash offset
moveto print "x y M" to stdout, move to x,y
rmoveto print "x y RM" to stdout, move to x,y
lineto print "x y L" to stdout, draw a line to x,y
rlineto print "x y RL" to stdout, draw a line to x,y
arc print "x y r ang1 ang2 arc" to stdout, draw an arc
x,y = vertex r = radius from ang1 to ang2
stroke print "S" to stdout
fill print "F" to stdout
show print "str SH" to stdout, show a string
justshow justify and show a string
image write a sampled gray-scale image
rgbimage write sampled color (rgb) image
setfont execute findfont, scalefont, and setfont for specified font
and size
fontbbox determine font bounding box for specified font and size
fontheight return maximum height for specified font and size
577
fontwidth return maximum width for specified font and size
fontcapheight return maximum capheight for specified font and size
fontxheight return maximum xheight for specified font and size
fontdescender return maximum descender for specified font and size
polyline draw a segmented line
markto draw a mark at specified location
rectclip set a rectangular clipping path
rectfill draw a filled rectangle
strokerect stroke a rectangle
Function Prototypes:
void beginps (void);
void endps (void);
void begineps (void);
void endeps (void);
void newpage (const char *label, int ordinal);
void boundingbox (int llx, int lly, int urx, int ury);
void showpage (void);
void gsave (void);
void grestore (void);
void newpath (void);
void closepath (void);
void clip(void);
void translate (float tx, float ty);
void scale (float sx, float sy);
void rotate (float angle);
void concat (float m[]);
void setgray (float gray);
void setrgbcolor (float red, float green, float blue);
void setcolor (const char *name);
void setlinewidth (float width);
void setlinejoin (int code);
void setdash (float dash[], int ndash, float offset);
void moveto (float x, float y);
void rmoveto (float x, float y);
void lineto (float x, float y);
void rlineto (float x, float y);
void arc (float x, float y, float r, float ang1, float ang2);
void stroke (void);
void fill (void);
void show (const char *str);
void justshow (float just, const char *str);
void image (int w, int h, int bps, float m[], unsigned char *samples);
578
void rgbimage (int w, int h, int bpc, float m[], unsigned char *samples);
void setfont (const char *fontname, float fontsize);
void fontbbox (const char *fontname, float fontsize, float bbox[]);
float fontheight (const char *fontname, float fontsize);
float fontwidth (const char *fontname, float fontsize);
float fontcapheight (const char *fontname, float fontsize);
float fontxheight (const char *fontname, float fontsize);
float fontdescender (const char *fontname, float fontsize);
float fontascender (const char *fontname, float fontsize);
void polyline (const float *x, const float *y, int n);
void markto (float x, float y, int index, float size);
void rectclip (float x, float y, float width, float height);
void rectfill (float x, float y, float width, float height);
void rectstroke (float x, float y, float width, float height);
justshow:
Input:
just justification factor
str string
image:
Input:
w width of image (in samples)
h height of image (in samples)
bps number of bits per sample
m array[6] containing image matrix
samples array[w*h] of sample values
rgbimage:
Input:
w width of image (in samples)
h height of image (in samples)
bpc number of bits per component
m array[6] containing image matrix
samples array[3*w*h] of sample values
polyline:
Input:
x array[n] of x-coordinates
y array[n] of y-coordinates
n number of points
markto:
579
Input:
x x-coordinate of mark
y y-coordinate of mark
index type of mark to draw
size size of mark
rectclip:
Input:
x x-coordinate of clipping path origin
y y-coordinate of clipping path origin
width width of clipping path
height height of clipping path
rectfill:
Input:
x x-coordinate of rectangle origin
y y-coordinate of rectangle origin
width width of rectangle
height height of rectangle
strokerect:
Input:
x x-coordinate of rectangle origin
y y-coordinate of rectangle origin
width width of rectangle
height height of rectangle
Notes:
The majority of these routines are self explanatory. They are just
C wrappers that echo PostScript graphics commands.
justshow:
The justification factor positions the string relative to the current point.
just" may assume any value, but the common uses are:
-1.0 right-justify the string
-0.5 center the string on the current point
0.0 left-justify the string (like using "show")
image:
Level 1 PostScript implementations support 1, 2, 4, and 8 bits per
sample. Level 2 adds support for 12 bits per sample.
Samples are hex-encoded, and output lines are limited to 78 characters.
580
rgbimage:
In general, Level 1 PostScript implementations do not support rgbimage.
Level 2 supports 1, 2, 4, 8, and 12 bits per color component. The
samples array should contain three color components (in R,G,B... order)
for each sample value. Samples are hex-encoded, and output lines are
limited to 78 characters.
polyline:
The path is stroked every 200 points.
References:
Author: Dave Hale, Colorado School of Mines, 1989

with modifications by Craig Artley, Colorado School of Mines, 1991, and
additions by Dave Hale, Advance Geophysical, 1992.
581
PSAXESBOX3 - Functions draw an axes box via PostScript, estimate bounding box
these are versions of psAxesBox and psAxesBox3 for psmovie.
psAxesBox3 draw an axes box via PostScript

psAxesBBox3 estimate bounding box for an axes box drawn via psAxesBox3
void psAxesBox3(
float x, float y, float width, float height,
float x1Beg, float x1End, float p1Beg, float p1End,
float d1Num, float f1Num, int n1Tic, int grid1, char *label1,
char *labelFont, float labelSize,
char *title, char *titleFont, float titleSize,
int style, char *title2);
void psAxesBBox3(
char *titleFont, float titleSize,
int style, int bbox[]);
Input:
x x coordinate of lower left corner of box
y y coordinate of lower left corner of box
width width of box
height height of box
x1Beg axis value at beginning of axis 1
x1End axis value at end of axis 1
p1Beg pad value at beginning of axis 1
p1End pad value at end of axis 1
d1Num numbered tic increment for axis 1 (0.0 for automatic)
f1Num first numbered tic for axis 1
n1Tic number of horizontal tics per numbered tic for axis 1
grid1 grid code for axis 1: NONE, DOT, DASH, or SOLID
label1 label for axis 1
d2Num vertical numbered tic increment (0.0 for automatic)
f2Num first numbered vertical tic
n2Tic number of vertical tics per numbered tic
582
grid2 grid code for vertical axis: NONE, DOT, DASH, or SOLID
label2 vertical axis label
labelFont name of font to use for axes labels
labelSize size of font to use for axes labels
title axes box title
titleFont name of font to use for title
titleSize size of font to use for title
style NORMAL (axis 1 on bottom, axis 2 on left)
SEISMIC (axis 1 on left, axis 2 on top)
title2 second title
psAxesBBox3:
Input:
width width of box
Output:
bbox bounding box (bbox[0:3] = llx, lly, ulx, uly)
Notes:
psAxesBox3:
psAxesBox will determine the numbered tic increment and first
numbered tic automatically, if the specified increment is zero.
Pad values must be specified in the same units as the corresponding

axes values. These pads are useful when the contents of the axes box
requires more space than implied by the axes values. For example,
the first and last seismic wiggle traces plotted inside an axes box
will typically extend beyond the axes values corresponding to the
first and last traces. However, all tics will lie with the limits
specified in the axes values (x1Beg, x1End, x2Beg, x2End).
psAxesBBox3:
psAxesBBox uses font sizes to estimate the bounding box for
an axes box drawn with psAxesBox. To be on the safe side,
psAxesBBox overestimates.
583
psAxesBBox assumes that the axes labels and titles do not extend
beyond the corresponding edges of the axes box.
References:
(see references in basic.c)
Author: Dave Hale, Colorado School of Mines, 06/27/89
modified by Zhiming Li, CSM, 7/1/90
584
PSAXESBOX - Functions to draw PostScript axes and estimate bounding box
psAxesBox Draw an axes box via PostScript

psAxesBBox estimate bounding box for an axes box drawn via psAxesBox
void psAxesBox(
char *titleColor, char *axesColor, char *gridColor,
float ticwidth, float axeswidth, float gridwidth,
int style);
void psAxesBBox(
char *titleFont, float titleSize,
psAxesBox:
Input:
width width of box
585
d2Num vertical numbered tic increment (0.0 for automatic)
f2Num first numbered vertical tic
n2Tic number of vertical tics per numbered tic
grid2 grid code for vertical axis: NONE, DOT, DASH, or SOLID
label2 vertical axis label
titleColor color to use for title
axesColor color to use for axes and axes labels
gridColor color to use for grid lines
axeswidth width (in points) of axes
ticwidth width (in points) of tic marks
gridwidth width (in points) of grid lines
psAxesBBox:
Input:
width width of box
Output:
Notes:
psAxesBox:
psAxesBox will determine the numbered tic increment and first numbered
tic automatically, if the specified increment is zero. Axis numbering
is in scientific notation, if necessary and is plotted to four
significant digits.

586
requires more space than implied by the axes values. For example, the
first and last seismic wiggle traces plotted inside an axes box will
typically extend beyond the axes values corresponding to the first and
last traces. However, all tics will lie with the limits specified in
the axes values (x1Beg, x1End, x2Beg, x2End).
psAxesBBox:
psAxesBBox uses font sizes to estimate the bounding box for
an axes box drawn with psAxesBox. To be on the safe side,
psAxesBBox overestimates.
psAxesBBox assumes that the axes labels and titles do not extend
beyond the corresponding edges of the axes box.
References:
(see References for basic.c)
Modified: Ken Larner, Colorado School of Mines, 08/30/90
Modified: Dave Hale, Advance Geophysical, 10/18/92
Added color parameters for title, axes, and grid.
587
PSCAXESBOX - Draw an axes box for cube via PostScript
psCubeAxesBox Draw an axes box for cube via PostScript
Function Prototype:
void psCubeAxesBox(
float x, float y, float size1, float size2, float size3, float angle,
char *titleColor, char *axesColor, char *gridColor);
Input:
size1 size of 1st dimension of the input cube
size2 size of 2nd dimension of the input cube
size3 size of 3rd dimension of the input cube
angle projection angle of the cube
n1Tic number of tics for axis 1
588
titleColor color to use for title
Authors: Zhiming Li & Dave Hale, Colorado School of Mines, 6/90
Modified: Craig Artley, Colorado School of Mines, 3/12/93
Changed name to psCubeAxesBox (from psAxes3), fixed minor bugs.
589
PSCONTOUR - draw contour of a two-dimensional array via PostScript
psContour draw contour of a two-dimensional array via PostScript
Function Prototype:
void psContour (float c, int nx, float x[], int ny, float y[], float z[],
float lcs, char *lcf, char *lcc, float *w, int nplaces);
Input:
c contour value
nx number of x-coordinates
x array of x-coordinates (see notes below)
ny number of y-coordinates
y array of y-coordinates (see notes below)
lcs font size of contour label
lcf font name of contour label
lcc color of contour label
LSB flag arrays (see Notes):
z array of nx*ny z(x,y) values (see notes below)
w array of nx*ny z(x,y) values (see notes below)
Notes:
The two-dimensional array z is actually passed as a one-dimensional
array containing nx*ny values, stored with nx fast and ny slow.
The x and y arrays define a grid that is not necessarily

uniformly-sampled. Linear interpolation of z values on the
grid defined by the x and y arrays is used to determine z values
between the gridpoints.
The two least significant bits of z are used to mark intersections

of the contour with the x,y grid; therefore, the z values will
almost always be altered (slightly) by contour.
pscontour isolates the use of PostScript to four internal functions:
void coninit(void) - called before any contour drawing

void conmove(float x, float y) - moves current position to x,y
void condraw(float x, float y) - draws from current position to x,y
void condone(void) - called when contour drawing is done
These functions can usually be replaced with equivalent functions in

other graphics environments.
590
The w array is used to restrict the range of contour labeling that
occurs only if w<1.
As suggested in the reference, the following scheme is used to refer

to a cell of the two-dimensional array z:
north (0)
(ix,iy+1) --------- (ix+1,iy+1)
| cell |
west (3) | ix,iy | east (1)
| |
(ix,iy) --------- (ix+1,iy)
south (2)
Reference:
Cottafava, G. and Le Moli, G., 1969, Automatic contour map:
Commuincations of the ACM, v. 12, n. 7, July, 1969.

contour labeling added by: Zhenyue Liu, August 1993
591
PSDRAWCURVE - Functions to draw a curve from a set of points
psDrawCurve Draw a curve from a set of points via PostScript

void psDrawCurve(
float *x1curve, float *x2curve, int ncurve,
char *curveColor, float curvewidth, int style);
psDrawCurve:
Input:
width width of box
x1curve vector of x1 coordinates for points along curve
ncurve number of points along curve
curveColor color to use for curve
curvewidth width (in points) of curve
Author: Brian Macy, Phillips Petroleum Co., 11/20/98

(Adapted after Dave Hale and other’s psAxesBox routine)
592
PSLEGENDBOX - Functions to draw PostScript axes and estimate bounding box
psLegendBox Draw an legend box via PostScript

psLegendBBox estimate bounding box for an legend box drawn via psLegendBox
void psLegendBox(
char *axesColor, char *gridColor,
int style);
void psLegendBBox(
psLegendBox:
Input:
width width of box
style VERTLEFT (Vertical, axis label on left side)
VERTRIGHT (Vertical, axis label on right side)
HORIBOTTOM (Horizontal, axis label on bottom)
593
psLegendBBox:
Input:
width width of box
style VERTLEFT (Vertical, axis label on left side)
VERTRIGHT (Vertical, axis label on right side)
HORIBOTTOM (Horizontal, axis label on bottom)
Output:
Notes:
psLegendBox:
psLegendBox will determine the numbered tic increment and first numbered
tic automatically, if the specified increment is zero. Axis numbering
is in scientific notation, if necessary and is plotted to four
significant digits.

Legend values. These pads are useful when the contents of the Legend box
requires more space than implied by the Legend values. For example, the
first and last seismic wiggle traces plotted inside an Legend box will
typically extend beyond the Legend values corresponding to the first and
last traces. However, all tics will lie with the limits specified in
the Legend values (x1Beg, x1End, x2Beg, x2End).
psLegendBBox:
psLegendBBox uses font sizes to estimate the bounding box for
an Legend box drawn with psLegendBox. To be on the safe side,
psLegendBBox overestimates.
psLegendBBox assumes that the Legend labels and titles do not extend
beyond the corresponding edges of the Legend box.
References:
(see References for basic.c)
Modified: Ken Larner, Colorado School of Mines, 08/30/90
Modified: Dave Hale, Advance Geophysical, 10/18/92
594
Modified: Torsten Schoenfelder, Koeln, Germany, 07/06/97
Display a legend for ps file, move axis from left to right
Modified: Torsten Schoenfelder, Koeln, Germany, 10/02/98
Corrected width of bbox to include legend title
595
PSWIGGLE - draw wiggle-trace with (optional) area-fill via PostScript
psWiggle draw wiggle-trace with (optional) area-fill via PostScript
Function Prototype:
void psWiggle (int n, float z[], float zmin, float zmax, float zbase,
float yzmin, float yzmax, float xfirst, float xlast, int fill);
Inputs:
n number of samples to draw
z array to draw
zmin z values below zmin will be clipped
zmax z values above zmax will be clipped
zbase z values between zbase and either zmin or zmax will be filled
yzmin y-coordinate corresponding to zmin
yzmax y-coordinate corresponding to zmax
xfirst x-coordinate corresponding to z[0]
xlast x-coordinate corresponding to z[n-1]
fill = 0 for no fill
> 0 for fill between zbase and zmax
< 0 for fill between zbase and zmin
+2 for fill solid between zbase and zmax grey between zbase and zmin
-2 for fill solid between zbase and zmin grey between zbase and zmax
SHADING: 2<= abs(fill) <=5 abs(fill)=2 light grey abs(fill)=5 black
NOTES:
psWiggle reduces PostScript output by eliminating linetos when
z values are essentially constant. The tolerance for detecting
constant" z values is ZEPS*(zmax-zmin), where ZEPS is a fraction
defined below.
A more complete optimization would eliminate all connected line

segments that are essentially colinear.
psWiggle breaks up the wiggle into segments that can be drawn

without exceeding the PostScript pathlimit.

Corrected dead trace bug. Now the last point of each segment
is guaranteed to be drawn.
MODIFIED: Paul Michaels, Boise State University, 29 December 2000
added fill=+/-2 option of solid/grey color scheme
596
AXESBOX - Functions to draw axes in X-windows graphics
xDrawAxesBox draw a labeled axes box

xSizeAxesBox determine optimal origin and size for a labeled axes box
void xDrawAxesBox (Display *dpy, Window win,
int x, int y, int width, int height,
float x1beg, float x1end, float p1beg, float p1end,
float d1num, float f1num, int n1tic, int grid1, char *label1,
float d2num, float f2num, int n2tic, int grid2, char *label2,
char *labelfont, char *title, char *titlefont,
char *axescolor, char *titlecolor, char *gridcolor,
int style);
void xSizeAxesBox (Display *dpy, Window win,
char *labelfont, char *titlefont, int style,
int *x, int *y, int *width, int *height);
xDrawAxesBox:
Input:
dpy display pointer
win window
x x coordinate of upper left corner of box
y y coordinate of upper left corner of box
width width of box
x1beg axis value at beginning of axis 1
x1end axis value at end of axis 1
p1beg pad value at beginning of axis 1
p1end pad value at end of axis 1
d1num numbered tic increment for axis 1 (0.0 for automatic)
f1num first numbered tic for axis 1
n1tic number of tics per numbered tic for axis 1
d2num numbered tic increment for axis 2 (0.0 for automatic)
f2num first numbered tic for axis 2
n2tic number of tics per numbered tic for axis 2
597
labelfont name of font to use for axes labels
titlefont name of font to use for title
axescolor name of color to use for axes
titlecolor name of color to use for title
gridcolor name of color to use for grid
int style NORMAL (axis 1 on bottom, axis 2 on left)
xSizeAxesBox:
Input:
dpy display pointer
win window
Output:
width width of box
{
XFontStruct *fa,*ft;
Notes:
xDrawAxesBox:
will determine the numbered tic incremenet and first

first and last traces. However, all tics will lie within the limits
specified in the axes values (x1beg, x1end, x2beg, x2end).
xSizeAxesBox:
is intended to be used prior to xDrawAxesBox.
598
An "optimal" axes box is one that more or less fills the window,
with little wasted space around the edges of the window.
599
xCreateRGBDefaultMap create XA_RGB_DEFAULT_MAP property of root window if

it does not already exist
xGetFirstPixel return first pixel in range of contiguous pixels in
XA_RGB_DEFAULT_MAP
xGetLastPixel return last pixel in range of contiguous pixels in
XA_RGB_DEFAULT_MAP
xCreateHSVColormap create a 2 ramp colormap (HSV - Model)
xCreateRGBColormap create a 2 ramp colormap (RGB - Model)
Status xCreateRGBDefaultMap (Display *dpy, XStandardColormap *scmap);
unsigned long xGetFirstPixel (Display *dpy);
unsigned long xGetLastPixel (Display *dpy);
Colormap xCreateRGBColormap (Display *dpy, Window win,
char *str_cmap, int verbose)
Colormap xCreateHSVColormap (Display *dpy, Window win,
char *str_cmap, int verbose)
xCreateRGBDefaultMap:
Input:
dpy display
Output:
scmap the standard colormap structure
xGetFirstPixel, xGetLastPixel:
Input:
dpy display
Notes:
PROBLEM
-------
Most mid-range display devices today support what X calls

the "PseudoColor visual". Typically, only 256 colors (or gray
levels) may be displayed simultaneously. Although these 256 colors
may be chosen from a much larger (4096 or more) set of available
colors, only 256 colors can appear on a display at one time.
These 256 colors are indexed by pixel values in a table called

the colormap. Each window can have its own colormap, but only
600
one colormap can be installed in the display hardware at a time.
(Again, only 256 colors may be displayed at one time.) The window
manager is responsible for installing a window’s colormap when that
window becomes the key window.
Many of the applications we are likely to write require a large,

contiguous range of pixels (entries in the colormap). In this
range, we must be able to:
(1) given a color (or gray), determine the corresponding pixel.
(2) given a pixel, determine the corresponding color (or gray).
An example would be an imaging application that uses a gray scale
to display images in shades of gray between black and white.
Such applications are also likely to require a few additional colors
for drawing axes, text, etc.
The problem is to coordinate the use of the limited number of

256 simultaneous colors so that windows for different applications
appear reasonable, even when their particular colormaps are not
installed in the display hardware. For example, we might expect
an analog xclock’s hands to be visible even when xclock’s window
is not the key window, when its colormap is not installed.
We should ensure that the range of contiguous pixels used by one

application (perhaps for imaging) does not conflict with the pixels
used by other applications to draw text, clock hands, etc.
SOLUTION
--------
Applications that do not require special colormaps should simply

use the default colormap inherited from the root window when new
top-level windows are created.
Applications that do require a special colormap MUST create their

own colormap. They must not assume that space will be available
in the default colormap for a contiguous range of read/write pixels,
because the server or window manager may have already allocated
these pixels as read-only. Even if sufficient pixels are available
in the default colormap, they should not be allocated by a single
application. The default colormap should be used only for windows
requiring a limited number of typical colors, such as red, yellow, etc.
601
Applications that require a contiguous range of read/write pixels
should allocate these pixels in their window’s private colormaps.
They should determine which contiguous pixels to allocate from
parameters in the standard colormap XA_RGB_DEFAULT_MAP. In particular,
the first pixel in the range of contiguous pixels should be
base_pixel
and the last pixel in the range should be
base_pixel+red_max*red_mult+green_max*green_mult+blue_max*blue_mult,
where base_pixel, red_max, etc. are members in the XStandardColormap
structure. On an 8-bit display, this range will typically provide 216
contiguous pixels, which may be set to a gray scale, color scale, or
whatever. This leaves 40 colors for drawing text, axes, etc.
If the XA_RGB_DEFAULT_MAP does not exist, it should be created to

consist of various colors composed of an equal number of reds,
greens, and blues. For example, if 216 colors are to be allocated,
then red_max=green_max=blue_max=5, red_mult=36, green_mult=6, and
blue_mult=1. Because of the difficulty in forcing a particular
pixel to correspond to a particular color in read-only color cells,
these 216 colors will likely be read/write color cells unless
created by the X server. In any case, these 216 colors should not
be modified by any application. In creating custom colormaps, the
only use of XA_RGB_DEFAULT_MAP should be in determining which 216
pixels to allocate for contiguous pixels.
In creating a custom colormap for a window, the application should

initialize this colormap to the colors already contained in the
window’s colormap, which was inherited initially from its parent.
This will ensure that typical colors already allocated by other
applications will be consistent with pixels used by the application
requiring the custom colormap. Ideally, windows might have
different colormaps, but the only differences would be in the
range of contiguous colors used for imaging, rendering, etc.
Ideally, the pixels corresponding to colors used to draw text,
axes, etc. would be consistent for all windows.
Unfortunately, it is impractical to maintain complete consistency

among various private colormaps. For example, suppose a custom
colormap is created for a window before other applications have
had the opportunity to allocate their colors from the default
colormap. Then, when the window with the custom colormap becomes
the key window, the windows of the other applications may be
displayed with false colors, since the colormap of the key window
602
may not contain the true colors. The colors used by the other
applications did not exist when the custom colormap was created.
One solution to this problem might be to initially allocate a
set of "common" colors in the default colormap before launching
any applications. This will increase the likelihood that typical
colors will be consistent among various colormaps.
Functions are provided below to

(1) create the standard colormap XA_RGB_DEFAULT_MAP, if it does not exist,
(2) determine the first and last pixels in the contiguous range of pixels,
(3) create some common private colormaps
xCreateRGBDefaultMap:
This function returns 0 if the XA_RGB_DEFAULT_MAP property does not exist
and cannot be created. At least 8 contiguous color cells must be free
in the default colormap to create the XA_RGB_DEFAULT_MAP. If created, the
red_max, green_max, and blue_max values returned in scmap will be equal.
xGetFirstPixel, xGetLastPixel:
If it does not already exist, XA_RGB_DEFAULT_MAP will be created.
If XA_RGB_DEFAULT_MAP does not exist and cannot be created, then
this function returns 0.
xCreateRGBColormap, xCreateHSVColormap:
The returned colormap is only created; the window’s colormap attribute
is not changed, and the colormap is not installed by this function.
The returned colormap is a copy of the window’s current colormap, but
with an RGB color scale allocated in the range of contiguous cells
determined by XA_RGB_DEFAULT_MAP. If it does not already exist,
XA_RGB_DEFAULT_MAP will be created.
603
DRAWCURVE - Functions to draw a curve from a set of points
xDrawCurve draw a curve from a set of points

void xDrawCurve(Display *dpy, Window win,
float *x1curve, float *x2curve, int ncurve,
char *curvecolor, int style);
xDrawCurve:
Input:
dpy display pointer
win window
width width of box
ncurve number of points along curve
curvecolor name of color to use for axes
Author: Brian Macy, Phillips Petroleum Co., 11/14/98

(Adapted after Dave Hale’s xDrawAxesBox routine)
604
IMAGE - Function for making the image in an X-windows image plot
xNewImage make a new image of pixels from bytes
Function Prototype:
XImage *xNewImage (Display *dpy, unsigned long pmin, unsigned long pmax,
int width, int height, float blank, unsigned char *bytes);
Input:
dpy display pointer
pmin minimum pixel value (corresponding to byte=0)
pmax maximum pixel value (corresponding to byte=255)
width number of bytes in x dimension
height number of bytes in y dimension
blank portion for blanking (0 to 1)
bytes unsigned bytes to be mapped to an image

Revision: Brian Zook, Southwest Research Institute, 6/27/96 added blank option
This allows replacing the low end by the background.
605
LEGENDBOX - draw a labeled axes box for a legend (i.e. colorscale)
Function Prototype:
void xDrawLegendBox (Display *dpy, Window win,
float bclip, float wclip, char *units, char *legendfont,
char *labelfont, char *title, char *titlefont,
char *axescolor, char *titlecolor, char *gridcolor,
int style);
Input:
dpy display pointer
win window
width width of box
units label for legend
legendfont name of font to use for legend labels
axescolor name of color to use for axes
titlecolor name of color to use for title
gridcolor name of color to use for grid
Notes:
xDrawLegendBox will determine the numbered tic incremenet and first

specified in the axes values (x1beg, x1end, x2beg, x2end).
Author: Berend Scheffers , TNO Delft, 06/11/92
606
xRubberBox Track pointer with rubberband box
Function Prototype:
void xRubberBox (Display *dpy, Window win, XEvent event,
Input:
dpy display pointer
win window ID
event event of type ButtonPress
Output:
x x of upper left hand corner of box in pixels
y y of upper left hand corner of box in pixels
width width of box in pixels
height height of box in pixels
Notes:
xRubberBox assumes that event is a ButtonPress event for the 1st button;
i.e., it tracks motion of the pointer while the 1st button is down, and
it sets x, y, w, and h and returns after a ButtonRelease event for the
1st button.
Before calling xRubberBox, both ButtonRelease and Button1Motion events

must be enabled.
This is the same rubberbox.c as in Xtcwp/lib, only difference is

that xRubberBox here is XtcwpRubberBox there, and a shift has been
added to make the rubberbox more visible.
607
WINDOW - Function to create a window in X-windows graphics
xNewWindow Create a new window and return the window ID
Function Prototype:
Window xNewWindow (Display *dpy, int x, int y, int width, int height,
int border, int background, char *name);
Input:
dpy display pointer
x x in pixels of upper left corner
y y in pixels of upper left corner
width width in pixels
height height in pixels
border border pixel
background background pixel
name name of window (also used for icon)
Notes:
The parent window is the root window.
The border_width is 4 pixels.
608
XCONTOUR - draw contour of a two-dimensional array via X vectorplot calls
xContour draw contour of a two-dimensional array via X vector plot calls
Function Prototype:
void xContour(Display *dpy, Window win,GC gcc, GC gcl,
float *cp,int nx, float x[], int ny, float y[], float z[],
char lcflag,char *lcf,char *lcc, float *w, int nplaces)
Input:
c contour value
nx number of x-coordinates
x array of x-coordinates (see notes below)
ny number of y-coordinates
y array of y-coordinates (see notes below)
lcf font name of contour label
lcc color of contour label
Least Significat Bits:
z array of nx*ny z(x,y) values (see notes below)
w array of nx*ny z(x,y) values (see notes below)
Notes:
The two-dimensional array z is actually passed as a one-dimensional
array containing nx*ny values, stored with nx fast and ny slow.
The x and y arrays define a grid that is not necessarily

uniformly-sampled. Linear interpolation of z values on the
grid defined by the x and y arrays is used to determine z values
between the gridpoints.
The two least significant bits of z are used to mark intersections

of the contour with the x,y grid; therefore, the z values will
almost always be altered (slightly) by contour.
xContour is a modified version of psContour where the use of conmove

and condraw call have been changed to match X-Windows.
Since XDrawLine requires a start and end point, the use of a manually update
of the position variables x0 and y0 is used instead of conmove.
The w array is used to restrict the range of contour labeling that

occurs only if w<1.
As suggested in the reference, the following scheme is used to refer

to a cell of the two-dimensional array z:
609
north (0)
(ix,iy+1) --------- (ix+1,iy+1)
| cell |
west (3) | ix,iy | east (1)
| |
(ix,iy) --------- (ix+1,iy)
south (2)
Reference:
Cottafava, G. and Le Moli, G., 1969, Automatic contour map:
Commuincations of the ACM, v. 12, n. 7, July, 1969.
Author: Morten Wendell Pedersen Aarhus University 07/20/96

Heavily based on psContour by
Dave Hale, Colorado School of Mines, 06/28/89
and with contour labeling added by: Zhenyue Liu, June 1993
(actually most of the credit should go to these two guys)
610
AXES - the Axes Widget
XtcwpPointInAxesRectangle returns TRUE if point is inside axes

rectangle, otherwise FALSE
XtcwpSetAxesValues set axes values
XtcwpSetAxesPads set axes pads
Function Prototype:
Boolean XtcwpPointInAxesRectangle (Widget w, Position x, Position y);
void XtcwpSetAxesValues (Widget w, float x1beg, float x1end, float x2beg,
float x2end);
void XtcwpSetAxesPads (Widget w, float p1beg, float p1end, float p2beg,

float p2end);
XtcwpPointInAxesRectangle:
Input:
w axes widget
x x coordinate of point
y y coordinate of point
XtcwpSetAxesValues:
Input:
w axes widget
XtcwpSetAxesPads:
Input:
w axes widget
p1beg axis pad at beginning of axis 1
p1end axis pad at end of axis 1
p2beg axis pad at beginning of axis 2
p2end axis pad at end of axis 2
Notes:
XtcwpPointInAxesRectangle:
This function is useful for determining whether or not input events
occured with the pointer inside the axes rectangle. I.e., the input
callback function will typically call this function.
XtcwpSetAxesPads:
611
require more space than implied by the axes values. For example, the
first and last seismic wiggle traces plotted inside an axes box
specified in the axes values (x1beg, x1end, x2beg, and x2end).

Modified: Craig Artley, Colorado School of Mines, 06/03/93, Rotate label for
vertical axis (Courtesy Dave Hale, Advance Geophysical).
612
XtcwpCreateRGBDefaultMap create XA_RGB_DEFAULT_MAP property of root

window if it does not already exist
XtcwpGetFirstPixel return first pixel in range of contiguous
pixels in XA_RGB_DEFAULT_MAP
XtcwpGetLastPixel return last pixel in range of contiguous
pixels in XA_RGB_DEFAULT_MAP
XtcwpCreateRGBColormap create a colormap with an RGB color scale in
contiguous cells
XtcwpCreateGrayColormap create a colormap with a gray scale in contiguous cells
XtcwpCreateHueColormap create a colormap with varying hues (blue to red)
in contiguous cells
XtcwpCreateSatColormap create a colormap with varying saturations in
contiguous cells
Status XtcwpCreateRGBDefaultMap (Display *dpy, XStandardColormap *scmap);
unsigned long XtcwpGetFirstPixel (Display *dpy);
unsigned long XtcwpGetLastPixel (Display *dpy);
Colormap XtcwpCreateRGBColormap (Display *dpy, Window win);
Colormap XtcwpCreateGrayColormap (Display *dpy, Window win);
Colormap XtcwpCreateHueColormap (Display *dpy, Window win);
Colormap XtcwpCreateSatColormap (Display *dpy, Window win,
float fhue, float lhue, float wfrac, float bright)
XtcwpCreateRGBDefaultMap:
Input:
dpy display
Output:
scmap the standard colormap structure
XtcwpGetFirstPixel, XtcwpGetLastPixel:
Input:
dpy display
XtcwpCreateRGBColormap, XtcwpCreateGrayColormap, XtcwpCreateHueColormap:

Input:
dpy display
win window
XtcwpCreateSatColormap:
613
Input:
dpy display
win window
fhue first hue in colormap (saturation=1)
lhue last hue in colormap (saturation=1)
wfrac fractional position of white within the colormap (saturation=0)
bright brightness
Notes:
PROBLEM
-------
Most mid-range display devices today support what X calls

the "PseudoColor visual". Typically, only 256 colors (or gray
levels) may be displayed simultaneously. Although these 256 colors
may be chosen from a much larger (4096 or more) set of available
colors, only 256 colors can appear on a display at one time.
These 256 colors are indexed by pixel values in a table called

the colormap. Each window can have its own colormap, but only
one colormap can be installed in the display hardware at a time.
(Again, only 256 colors may be displayed at one time.) The window
manager is responsible for installing a window’s colormap when that
window becomes the key window.
Many of the applications we are likely to write require a large,

contiguous range of pixels (entries in the colormap). In this
range, we must be able to:
(1) given a color (or gray), determine the corresponding pixel.
(2) given a pixel, determine the corresponding color (or gray).
An example would be an imaging application that uses a gray scale
to display images in shades of gray between black and white.
Such applications are also likely to require a few additional colors
for drawing axes, text, etc.
The problem is to coordinate the use of the limited number of

256 simultaneous colors so that windows for different applications
appear reasonable, even when their particular colormaps are not
installed in the display hardware. For example, we might expect
an analog xclock’s hands to be visible even when xclock’s window
is not the key window, when its colormap is not installed.
We should ensure that the range of contiguous pixels used by one
614
application (perhaps for imaging) does not conflict with the pixels
used by other applications to draw text, clock hands, etc.
SOLUTION
--------
Applications that do not require special colormaps should simply

use the default colormap inherited from the root window when new
top-level windows are created.
Applications that do require a special colormap MUST create their

own colormap. They must not assume that space will be available
in the default colormap for a contiguous range of read/write pixels,
because the server or window manager may have already allocated
these pixels as read-only. Even if sufficient pixels are available
in the default colormap, they should not be allocated by a single
application. The default colormap should be used only for windows
requiring a limited number of typical colors, such as red, yellow, etc.
Applications that require a contiguous range of read/write pixels

should allocate these pixels in their window’s private colormaps.
They should determine which contiguous pixels to allocate from
parameters in the standard colormap XA_RGB_DEFAULT_MAP. In particular,
the first pixel in the range of contiguous pixels should be
base_pixel
and the last pixel in the range should be
base_pixel+red_max*red_mult+green_max*green_mult+blue_max*blue_mult,
where base_pixel, red_max, etc. are members in the XStandardColormap
structure. On an 8-bit display, this range will typically provide 216
contiguous pixels, which may be set to a gray scale, color scale, or
whatever. This leaves 40 colors for drawing text, axes, etc.
If the XA_RGB_DEFAULT_MAP does not exist, it should be created to

consist of various colors composed of an equal number of reds,
greens, and blues. For example, if 216 colors are to be allocated,
then red_max=green_max=blue_max=5, red_mult=36, green_mult=6, and
blue_mult=1. Because of the difficulty in forcing a particular
pixel to correspond to a particular color in read-only color cells,
these 216 colors will likely be read/write color cells unless
created by the X server. In any case, these 216 colors should not
be modified by any application. In creating custom colormaps, the
only use of XA_RGB_DEFAULT_MAP should be in determining which 216
615
pixels to allocate for contiguous pixels.
In creating a custom colormap for a window, the application should

initialize this colormap to the colors already contained in the
window’s colormap, which was inherited initially from its parent.
This will ensure that typical colors already allocated by other
applications will be consistent with pixels used by the application
requiring the custom colormap. Ideally, windows might have
different colormaps, but the only differences would be in the
range of contiguous colors used for imaging, rendering, etc.
Ideally, the pixels corresponding to colors used to draw text,
axes, etc. would be consistent for all windows.
Unfortunately, it is impractical to maintain complete consistency

among various private colormaps. For example, suppose a custom
colormap is created for a window before other applications have
had the opportunity to allocate their colors from the default
colormap. Then, when the window with the custom colormap becomes
the key window, the windows of the other applications may be
displayed with false colors, since the colormap of the key window
may not contain the true colors. The colors used by the other
applications did not exist when the custom colormap was created.
One solution to this problem might be to initially allocate a
set of "common" colors in the default colormap before launching
any applications. This will increase the likelihood that typical
colors will be consistent among various colormaps.
Functions are provided below to

(1) create the standard colormap XA_RGB_DEFAULT_MAP, if it does not exist,
(2) determine the first and last pixels in the contiguous range of pixels,
(3) create some common private colormaps - gray scale, hue scale, etc.
XtcwpCreateRGBDefaultMap:
This function returns 0 if the XA_RGB_DEFAULT_MAP property does not exist
and cannot be created. At least 8 contiguous color cells must be free
in the default colormap to create the XA_RGB_DEFAULT_MAP. If created, the
red_max, green_max, and blue_max values returned in scmap will be equal.
XtcwpGetFirstPixel, XtcwpGetLastPixel:
If it does not already exist, XA_RGB_DEFAULT_MAP will be created.
If XA_RGB_DEFAULT_MAP does not exist and cannot be created, then
this function returns 0.
616
XtcwpCreateRGBColormap, XtcwpCreateGrayColormap, XtcwpCreateHueColormap,
XtcwpCreateSatColormap:
The returned colormap is only created; the window’s colormap attribute
is not changed, and the colormap is not installed by this function.
The returned colormap is a copy of the window’s current colormap, but
with an RGB color scale allocated in the range of contiguous cells
determined by XA_RGB_DEFAULT_MAP. If it does not already exist,
XA_RGB_DEFAULT_MAP will be created.
617
FX - Functions to support floating point coordinates in X
FMapFX map float x to x

FMapFY map float y to y
FMapFWidth map float width to width
FMapFHeight map float height to height
FMapFAngle map float angle to angle
FMapFPoint map float x,y to x,y
FMapFPoints map float points to points
FMapX inverse map x to float x
FMapY inverse map y to float y
FMapWidth inverse map width to float width
FMapHeight inverse map height to float height
FMapAngle inverse map angle to float angle
FMapPoint map x,y to float x,y
FMapPoints map points to float points
FSetGC set graphics context
FSetMap set map (scales and shifts)
FSetClipRectangle set clip rectangle
FClipOn turn clip on
FClipOff turn clip off
FClipPoint clip point
FClipLine clip line
FClipRectangle clip rectangle
FXCreateFGC create float graphics context
FXFreeFGC free float graphic context
FXDrawPoint draw point at float x,y (with clipping)
FXDrawPoints draw float points (with clipping)
FXDrawLine draw line from float x1,y1 to float x2,y2
(with clipping)
FXDrawLines draw lines between float points (with clipping)
FXDrawRectangle draw rectangle with float x,y,width,height
(with clipping)
FXDrawArc draw arc with float x,y,width,height,angle1,angle2
FXDrawString draw string at float x,y
FXFillRectangle fill rectangle with float x,y,width,height (with
clipping)
int FMapFX (FGC fgc, float fx);
int FMapFY (FGC fgc, float fy);
int FMapFWidth (FGC fgc, float fwidth);
int FMapFHeight (FGC fgc, float fheight);
618
int FMapFAngle (FGC fgc, float fangle);
void FMapFPoint (FGC fgc, float fx, float fy, int *x_return, int *y_return);
void FMapFPoints (FGC fgc, FXPoint fpoints[], int npoints,
XPoint points_return[]);
float FMapX (FGC fgc, int x);
float FMapY (FGC fgc, int y);
float FMapWidth (FGC fgc, int width);
float FMapHeight (FGC fgc, int height);
float FMapAngle (FGC fgc, int angle);
void FMapPoint (FGC fgc, int x, int y, float *fx_return, float *fy_return);
void FMapPoints (FGC fgc, XPoint points[], int npoints,
FXPoint fpoints_return[]);
void FSetGC (FGC fgc, GC gc);
void FSetMap (FGC fgc, int x, int y, int width, int height,
float fx, float fy, float fwidth, float fheight);
void FSetClipRectangle(FGC fgc, float fxa, float fya, float fxb, float fyb);
void FClipOn (FGC fgc);
void FClipOff (FGC fgc);
int FClipPoint (FGC fgc, float fx, float fy);
int FClipLine (FGC fgc, float fx1, float fy1, float fx2, float fy2,
float *fx1c, float *fy1c, float *fx2c, float *fy2c);
int FClipRectangle (FGC fgc, float fx, float fy, float fwidth, float fheight,
float *fxc, float *fyc, float *fwidthc, float *fheightc);
FGC FXCreateFGC (GC gc, int x, int y, int width, int height,
void FXFreeFGC (FGC fgc);
void FXDrawPoint (Display *display, Drawable d, FGC fgc, float fx, float fy);
void FXDrawPoints (Display *display, Drawable d, FGC fgc,
FXPoint fpoints[], int npoints, int mode);
void FXDrawLine (Display *display, Drawable d, FGC fgc,
float fx1, float fy1, float fx2, float fy2);
void FXDrawLines (Display *display, Drawable d, FGC fgc,
FXPoint fpoints[], int npoints, int mode);
void FXDrawRectangle (Display *display, Drawable d, FGC fgc,
void FXDrawArc (Display *display, Drawable d, FGC fgc,
float fx, float fy, float fwidth, float fheight,
float fangle1, float fangle2);
void FXDrawString (Display *display, Drawable d, FGC fgc,
float fx, float fy, char *string, int length);
void FXFillRectangle (Display *display, Drawable d, FGC fgc,
619
Notes:
The functions defined below are designed to resemble the equivalent
X functions. For example, FXDrawLine() is analogous to XDrawLine.
Each of the FXDraw<xxx>() functions requires an FGC instead of a GC
(graphics context). An FGC contains a GC, along with the information
required to transform floating point coordinates to integer (pixel)
coordinates.
Additional functions are provided to transform floating point coordinates

to integer coordinates and vice versa. Where feasible, macros are also
provided to perform these coordinate transformations.
Clipping of floating point coordinates is supported, because clipping

after mapping to integer coordinates is not valid when the mapped
integer coordinates overflow the range of short integers. By clipping
the floating point coordinates before mapping to integers, this overflow
can be avoided. By default, clipping is turned off until a clip rectangle
is specified or until clipping is explicitly turned on. Clipping is
not currently supported for all FXDraw functions.

Modified: Dave Hale, Colorado School of Mines, 05/18/91
Added floating point clipping capability to some FXDraw functions.
620
MISC - Miscellaneous X-Toolkit functions
XtcwpDrawString90 Draw a string rotated 90 degrees counter-clockwise
Function Prototype:
void XtcwpDrawString90 (Display *dpy, Drawable d, GC gc,
int x, int y, char *string, int count);
Input:
dpy X display
d X drawable
gc X graphics context
x,y coordinates of baseline starting position of the string
string array[count] of characters to be drawn
count number of characters in string
Author: Dave Hale, Advance Geophysical, 06/03/93
621
RESCONV - general purpose resource type converters
XtcwpStringToFloat convert string to float in resource
Function Prototype:
void XtcwpStringToFloat (XrmValue *args, int *nargs,
XrmValue *fromVal, XrmValue *toVal);
622
XtcwpRubberbox Track pointer with rubberband box
Function Prototype:
void XtcwpRubberbox (Display *dpy, Window win, XEvent event,
Input:
dpy display pointer
win window ID
event event of type ButtonPress
Output:
x x of upper left hand corner of box in pixels
y y of upper left hand corner of box in pixels
width width of box in pixels
height height of box in pixels
Notes:
XtcwpRubberbox assumes that event is a ButtonPress event for the 1st button;
i.e., it tracks motion of the pointer while the 1st button is down, and
it sets x, y, w, and h and returns after a ButtonRelease event for the
1st button.
Before calling XtcwpRubberbox, both ButtonRelease and Button1Motion events

must be enabled.
623
RADIOBUTTONS - convenience functions creating and using radio buttons
XtcwpCreateStringRadioButtons create an XmFrame containing radio

buttons labeled with strings
Widget XtcwpCreateStringRadioButtons (Widget parent, char *label,
int nstrings, char **strings, int first,
void (*callback)(int selected, void *clientdata), void *clientdata);
Input:
parent parent widget
label label for this collection of radio bottons
nstrings number of strings
strings array[nstrings] of character strings, one per button
first index of button to be initially selected
callback function called when radio buttons change state
clientdata pointer to client data to be passed to callback
Notes:
This code depends on the Motif Developer’s Package.
An integer index of the selected button (along with the clientdata pointer)
is passed to the callback function.
The returned XmFrame is not managed.
624
SAMPLES - Motif-based Graphics Functions
samplesCreate
samplesDraw
samplesSetN
samplesSetData
samplesSetPlotValue
samplesSetEditMode
samplesSetOrigin
Samples *samplesCreate (Widget parent, char *title,
void (*editDone)(Samples *s));
void samplesDraw (Samples *s);
void samplesSetN (Samples *s, int n);
void samplesSetData (Samples *s, float *d);
void samplesSetPlotValue (Samples *s, float pv);
void samplesSetEditMode (Samples *s, EditMode m);
void samplesSetOrigin (Samples *s, int i);
Notes:
Watch this space.
Author: Dave Hale, Colorado School of Mines
625
fgethdr - get segy tape identification headers from the file by file pointer
Input:
fp file pointer
Output:
chdr 3200 bytes of segy character header
bhdr 400 bytes of segy binary header
Authors: zhiming li and j. dulac , unocal
modified for CWP/SU: R. Beardsley
626
FGETTR - Routines to get an SU trace from a file
fgettr get a fixed-length segy trace from a file by file pointer

fvgettr get a variable-length segy trace from a file by file pointer
fgettra get a fixed-length trace from disk file by trace number
gettr macro using fgettr to get a trace from stdin
vgettr macro using vfgettr to get a trace from stdin
gettra macro using fgettra to get a trace from stdin by trace number
Function Prototype:
int fgettr(FILE *fp, segy *tp);
int fvgettr(FILE *fp, segy *tp);
int fgettra(FILE *fp, segy *tp, int itr);
Returns:
fgettr, fvgettr:
int: number of bytes read on current trace (0 after last trace)
fgettra:
int: number of traces in disk file
Macros defined in segy.h

define gettr(x) fgettr(stdin, (x))
define vgettr(x) fgettr(stdin, (x))
Usage example:
segy tr;
...
tr.offset = abs(tr.offset);
puttr(&tr);
}
...
Authors: SEP: Einar Kjartansson, Stew Levin CWP: Shuki Ronen, Jack Cohen

Major rewrite: Use xdr library for portable su output file
format. Merge fgettr and fgettra into same source file.
Make input from multiple streams work (at long last!).
Always set ntr for DISK input. This fixes susort failure.
Revised: 1/9/96 jkc CWP
627
Set lastfp on nread <=0 return, too.
Revised: 28 Mar, 2006 Stewart A. Levin Landmark Graphics
Reworked XDR to support random seeks on > 2GB files
and to read big endian SHORTPACK data on little endian machines.
628
FPUTTR - Routines to put an SU trace to a file
fputtr put a segy trace to a file by file pointer

fvputtr put a segy trace to a file by file pointer (variable ns)
puttr macro using fputtr to put a trace to stdin
vputtr macro using fputtr to put a trace to stdin (variable ns)
Function Prototype:
void fputtr(FILE *fp, segy *tp);
void fvputtr(FILE *fp, segy *tp);
Returns:
void
Notes:
The functions puttr(x) vputtr(x) are macros defined in segy.h

define puttr(x) fputtr(stdin, (x))
define vputtr(x) fvputtr(stdin, (x))
Usage example:
segy tr;
...
tr.offset = abs(tr.offset);
puttr(&tr);
}
...
Authors: SEP: Einar Kjartansson, Stew Levin CWP: Shuki Ronen, Jack Cohen
629
HDRPKGE - routines to access the SEGY header via the hdr structure.
gethval get a trace header word by index

puthval put a trace header word by index
getbhval get a binary header word by index
putbhval put a binary header word by index
gethdval get a trace header word by name
puthdval put a trace header word by name
hdtype get the data type of a trace header word by name

getkey get the name of a trace header word from its index
getindex get the index of a trace header word from the name
swaphval swap the trace header words by index

swapbhval swap the binary header words by index
gettapehval get a tape trace header word by index
puttapehval put a tape trace header word by index
gettapebhval get a tape binary header word by index
puttapebhval put a tape binary header word by index
printheader display non-null header field values
void gethval(const segy *tr, int index, Value *valp);
void puthval(segy *tr, int index, Value *valp);
void putbhval(bhed *bh, int index, Value *valp);
void getbhval(const bhed *bh, int index, Value *valp);
void gethdval(const segy *tr, char *key, Value *valp);
void puthdval(segy *tr, char *key, Value *valp);
char *hdtype(const char *key);
char *getkey(const int index);
int getindex(const char *key);
void swaphval(segy *tr, int index);
void swapbhval(bhed *bh, int index);
void gettapehval(tapesegy *tapetr, int index, Value *valp);
void puttapehval(tapesegy *tapetr, int index, Value *valp);
void gettapebhval(tapebhed *tapetr, int index, Value *valp);
void puttapebhval(tapebhed *tapetr, int index, Value *valp);
void printheader(const segy *tp);
Notes:
This package includes only those routines that directly access
630
the "hdr" or "bhdr" structures. It does not include routines
such as printfval, printftype, printfhead that use the routines
in this package to indirectly access these structures.
Note that while gethdval and puthdval are more convenient to use
than gethval and puthval, they incur an inefficiency in the
common case of iterating code over a set of traces with a fixed
key or keys. In such cases, it is advisable to set the index
or indices outside the loop using getindex.
swaphval:
Byte-swapping is needed for converting SU data from big-endian to little-
endian formats, and vice versa. The swap_.... subroutines are based
on subroutines provided by Jens Hartmann of the Institut fur Geophysik
in Hamburg. These are found in .../cwp/lib/swapbyte.c.
Authors: SEP: Einar Kjartansson CWP: Jack Cohen, Shuki Ronen
swaphval: CWP: John Stockwell
631
TABPLOT - TABPLOT selected sample points on selected trace
tabplot tabplot selected sample points on selected trace
Function Prototype:
void tabplot(segy *tp, int itmin, int itmax);
Input:
tp pointer to a segy
itmin minimum time sample printed
itmax maximum time sample printed
Authors: CWP: Brian Sumner, Jack K. Cohen
632
VALPKGE - routines to handle variables of type Value
vtoi cast Value variable as an int

vtol cast Value variable as a long
vtof cast Value variable as a float
vtod cast Value variable as a double
atoval convert ascii to Value
valtoabs take absolute value of a Value variable
valcmp compare Value variables
printfval printf a Value variable
fprintfval fprintf a Value variable
scanfval scanf a Value variable
printftype printf for the type of a segy header word
int vtoi(register cwp_String type, Value val);
long vtol(register cwp_String type, Value val);
float vtof(register cwp_String type, Value val);
double vtod(register cwp_String type, Value val);
void atoval(cwp_String type, cwp_String keyval, Value *valp);
Value valtoabs(cwp_String type, Value val);
int valcmp(register cwp_String type, Value val1, Value val2);
void printfval(register cwp_String type, Value val);
void fprintfval(FILE *stream, register cwp_String type, Value val);
void scanfval(register cwp_String type, Value *valp);
Notes:
A Value is defined by the following in .../su/include/su.h:
typedef union { * storage for arbitrary type *

char s[8];
short h;
unsigned short u;
long l;
unsigned long v;
int i;
unsigned int p;
float f;
double d;
unsigned int U:16;
unsigned int P:32;
} Value;
633
The use of the valpkge routines, as well as the hdrpkge routines,
permits the user to change the definition of the types of the
various fields of the segy data type, without breaking codes that
look at part or all of these fields.
Authors: CWP: Jack K. Cohen, Shuki Ronen
634
ABEL - Functions to compute the discrete ABEL transform:
abelalloc allocate and return a pointer to an Abel transformer

abelfree free an Abel transformer
abel compute the Abel transform
Function prototypes:
void *abelalloc (int n);
void abelfree (void *at);
void abel (void *at, float f[], float g[]);
Input:
ns number of samples in the data to be transformed
f[] array of floats, the function being transformed
Output:
at pointer to Abel transformer returned by abelalloc(int n)
g[] array of floats, the transformed data returned by
abel(*at,f[],g[])
Notes:
The Abel transform is defined by:
Infinity
g(y) = 2 Integral dx f(x)/sqrt(1-(y/x)^2)
|y|
Linear interpolation is used to define the continuous function f(x)

corresponding to the samples in f[]. The first sample f[0] corresponds
to f(x=0) and the sampling interval is assumed to be 1. Therefore, the
input samples correspond to 0 <= x <= n-1. Samples of f(x) for x > n-1
are assumed to be zero. These conventions imply that
g[0] = f[0] + 2*f[1] + 2*f[2] + ... + 2*f[n-1]
References:
Hansen, E. W., 1985, Fast Hankel transform algorithm: IEEE Trans. on
Acoustics, Speech and Signal Processing, v. ASSP-33, n. 3, p. 666-671.
(Beware of several errors in the equations in this paper!)
Authors: Dave Hale and Lydia Deng, Colorado School of Mines, 06/01/90
635
AIRY - Approximate the Airy functions Ai(x), Bi(x) and their respective
derivatives Ai’(x), Bi’(x)
airya return approximation of Ai(x)

airypa return approximation of Ai’(x)
airyb return approximation of Bi(x)
airybp return approximation of Bi’(x)
float airya (float x);
float airyap (float x);
float airyb (float x);
float airybp (float x);
Input:
x value at which to evaluate Ai(x)
Returned:
airya Ai(x)
airypa Ai’(x)
airyb Bi(x)
airybp Bi’(x)
Reference:
The approximation is derived from tables and formulas in Abramowitz
and Stegun, p. 475-477.
636
ALLOC - Allocate and free multi-dimensional arrays
alloc1 allocate a 1-d array

realloc1 re-allocate a 1-d array
free1 free a 1-d array
free6 free a 6-d arrayalloc1int
allocate a 1-d array of ints
realloc1int re-allocate a 1-d array of ints
free1int free a 1-d array of ints
alloc2int allocate a 2-d array of ints
alloc1float allocate a 1-d array of floats
realloc1float re-allocate a 1-d array of floats
free1float free a 1-d array of floats
alloc5uchar allocate a 5-d array of unsigned chars
free5uchar free a 5-d array of unsiged chars
alloc5ushort allocate a 5-d array of unsigned shorts
free5ushort free a 5-d array of unsiged shorts
637
alloc6ushort allocate a 6-d array of unsigned shorts
free6ushort free a 6-d array of unsiged shorts
alloc1double allocate a 1-d array of doubles
realloc1double re-allocate a 1-d array of doubles
free1double free a 1-d array of doubles
alloc1complex allocate a 1-d array of complexs
realloc1complex re-allocate a 1-d array of complexs
free1complex free a 1-d array of complexs
alloc1dcomplex allocate a 1-d array of complexs

realloc1dcomplex re-allocate a 1-d array of complexs
free1dcomplex free a 1-d array of complexs
void *alloc1 (size_t n1, size_t size);
void *realloc1 (void *v, size_t n1, size_t size);
void free1 (void *p);
void **alloc2 (size_t n1, size_t n2, size_t size);
void free2 (void **p);
void ***alloc3 (size_t n1, size_t n2, size_t n3, size_t size);
void free3 (void ***p);
void ****alloc4 (size_t n1, size_t n2, size_t n3, size_t n4, size_t size);
void free4 (void ****p);
size_t size);
int *alloc1int (size_t n1);
int *realloc1int (int *v, size_t n1);
void free1int (int *p);
int **alloc2int (size_t n1, size_t n2);
void free2int (int **p);
int ***alloc3int (size_t n1, size_t n2, size_t n3);
void free3int (int ***p);
638
float *alloc1float (size_t n1);
float *realloc1float (float *v, size_t n1);
void free1float (float *p);
float **alloc2float (size_t n1, size_t n2);
void free2float (float **p);
float ***alloc3float (size_t n1, size_t n2, size_t n3);
void free3float (float ***p);
float ****alloc4float (size_t n1, size_t n2, size_t n3, size_t n4);
void free4float (float ****p);
size_t n6);
int ****alloc4int (size_t n1, size_t n2, size_t n3, size_t n4);
void free4int (int ****p);
size_t n5);
size_t n5);
size_t n5,size_t n6);
double *alloc1double (size_t n1);
double *realloc1double (double *v, size_t n1);
void free1double (double *p);
double **alloc2double (size_t n1, size_t n2);
void free2double (double **p);
double ***alloc3double (size_t n1, size_t n2, size_t n3);
void free3double (double ***p);
complex *alloc1complex (size_t n1);
complex *realloc1complex (complex *v, size_t n1);
void free1complex (complex *p);
complex **alloc2complex (size_t n1, size_t n2);
void free2complex (complex **p);
complex ***alloc3complex (size_t n1, size_t n2, size_t n3);
void free3complex (complex ***p);
complex *alloc1dcomplex (size_t n1);

complex *realloc1dcomplex (dcomplex *v, size_t n1);
void free1dcomplex (dcomplex *p);
complex **alloc2dcomplex (size_t n1, size_t n2);
void free2dcomplex (dcomplex **p);
complex ***alloc3dcomplex (size_t n1, size_t n2, size_t n3);
void free3dcomplex (dcomplex ***p);
Notes:
The functions defined below are intended to simplify manipulation
of multi-dimensional arrays in scientific programming in C. These
functions are useful only because true multi-dimensional arrays
639
in C cannot have variable dimensions (as in FORTRAN). For example,
the following function IS NOT valid in C:
void badFunc(a,n1,n2)
float a[n2][n1];
{
a[n2-1][n1-1] = 1.0;
}
However, the following function IS valid in C:
void goodFunc(a,n1,n2)
float **a;
{
a[n2-1][n1-1] = 1.0;
}
Therefore, the functions defined below do not allocate true
multi-dimensional arrays, as described in the C specification.
Instead, they allocate and initialize pointers (and pointers to
pointers) so that, for example, a[i2][i1] behaves like a 2-D array.
The array dimensions are numbered, which makes it easy to add

functions for arrays of higher dimensions. In particular,
the 1st dimension of length n1 is always the fastest dimension,
the 2nd dimension of length n2 is the next fastest dimension,
and so on. Note that the 1st (fastest) dimension n1 is the
first argument to the allocation functions defined below, but
that the 1st dimension is the last subscript in a[i2][i1].
(This is another important difference between C and Fortran.)
The allocation of pointers to pointers implies that more storage

is required than is necessary to hold a true multi-dimensional array.
The fraction of the total storage allocated that is used to hold
pointers is approximately 1/(n1+1). This extra storage is unlikely
to represent a significant waste for large n1.
The functions defined below are significantly different from similar

functions described by Press et al, 1988, NR in C.
In particular, the functions defined below:
(1) Allocate arrays of arbitrary size elements.
(2) Allocate contiguous storage for arrays.
(3) Return NULL if allocation fails (just like malloc).
(4) Do not provide arbitrary lower and upper bounds for arrays.
Contiguous storage enables an allocated multi-dimensional array to

be passed to a C function that expects a one-dimensional array.
640
For example, to allocate and zero an n1 by n2 two-dimensional array
of floats, one could use
a = alloc2(n1,n2,sizeof(float));
zeroFloatArray(n1*n2,a[0]);
where zeroFloatArray is a function defined as
void zeroFloatArray(int n, float *a)
{
int i;
for (i=0; i<n; i++)
a[i] = 0.0;
}
Internal error handling and arbitrary array bounds, if desired,

should be implemented in functions that call the functions defined
below, with the understanding that these enhancements may limit
portability.

Zhaobo Meng, added 4D, 5D and 6D functions, 1996
641
ANTIALIAS - Butterworth anti-aliasing filter
antialias use before increasing the sampling interval of data

i.e. subsampling
Function Prototype:
void antialias (float frac, int phase, int n, float p[], float q[]);
Input:
frac current sampling interval / future interval (should be <= 1)
phase =0 for zero-phase filter; =1 for minimum-phase filter
n number of samples
p array[n] of input samples
Output:
q array[n] of output (anti-alias filtered) samples
Notes:
The anti-alias filter is a recursive (Butterworth) filter. For zero-phase
anti-alias filtering, the recursive filter is applied forwards and backwards.
642
AXB - Functions to solve a linear system of equations Ax=b by LU
decomposition, invert a square matrix or directly multiply an
inverse matrix by another matrix (without explicitely computing
the inverse).
LU_decomposition Decompose a matrix (A) into a lower triangular (L)

and an upper triangular (U) such that A=LU
backward_substitution Apply backward substitution to an LU decomposed

matrix to solve the linear system of equations Ax=b
inverse_matrix compute the inverse of a square non-singular matrix
inverse_matrix_multiply computes the product A^(-1)*B without explicitely

computing the inverse matrix
Function prototypes:
void LU_decomposition (int nrows, float **matrix, int *idx, float *d);
void backward_substitution (int nrows, float **matrix, int *idx, float *b);
void inverse_matrix (int nrows, float **matrix);
void inverse_matrix_multiply (int nrows1, float **matrix1, int ncols2,
int nrows2, float **matrix2, float **out_matrix);
LU_decomposition:
Input:
nrows number of rows of matrix to invert
matrix matrix of coefficients in linear system Ax=b
Output:
matrix matrix containing LU decomposition (original matrix destroyed)
idx vector recording the row permutations effected by partial
pivoting
d +/- 1 depending on whether the number of row interchanges
was even or odd
backward_substitution
Input:
nrows number of rows (and columns) of input matrix
matrix matrix of coefficients (after LU decomposition)
idx permutation vector obtained from routine LU_decomposition
b right hand side vector in equation Ax=b
Output:
b vector with the solution
643
inverse_matrix
Input:
nrows number of rows (and columns) of input matrix
matrix matrix to invert
Output:
matrix inverse of input matrix
inverse_matrix_multiply
nrows1 number of rows (and columns) of matrix to invert
matrix1 square matrix to invert
ncols2 number of coulmns of second matrix
nrows2 number of rows of second matrix
matrix second matrix (multiplicator)
Output Parameters:
out_matrix matrix containing the product of the inverse of the first
matrix by the second one.
Note:
matrix1 and matrix2 are not destroyed, (not clobbered)
Notes:
To solve the set of linear equations Ax=b, first do the LU decomposition of
A (which will clobber A with its LU decomposition) and then do the backward
substitution with this new matrix and the right-hand side vector b. The vector
b will be clobbered with the solution. Both, the original matrix and vector B,
will have been destroyed.
The LU decomposition is carried out with the Crout’s method with implicit
partial pivoting that guaratees that the maximum pivot is used in every
step of the algorithm.
The operation count to solve a linear system of equations via LU decomposition

is 1/3N^3 and is a factor of 3 better than the standard Gauss-Jordan algorithm
To invert a matrix the count is the same with both algorithms: N^3.
Once a linear system Ax=b has been solved, to solve another linear system
with the same matrix A but with different vetor b, ONLY the back substitution
has to be repeated with the new b (remember that the matrix in backsubstitution
is not the original matrix but its LU decomposition)
If you want to compute A^(-1)*B from matrices A and B, it is better to

use the subroutine inverse_matrix_multiply rather than explicitely computing
the inverse. This saves a whole martix multiplication and is also more accurate.
644
Refferences:
Press, Teukolsky, Vettering and Flannery, Numerical Recipes in C:
The art of scientific computing. Cambridge University Press.
second edition. (1992).
Golub and Van Loan, Matrix Computations. John Hopkins University Press.
Second Edition. (1989).
Horn and Johnson, Matrix Analysis. Cambridge University Press. (1985).
Credits:
Adapted from discussions in Numerical Recipes, by Gabriel Alvarez (1995)
645
BIGMATRIX - Functions to manipulate 2-dimensional matrices that are too big
to fit in real memory, but that are small enough to fit in
virtual memory:
bmalloc allocate a big matrix

bmfree free a big matrix
bmread read a vector from a big matrix
bmwrite write a vector to a big matrix
void *bmalloc (int nbpe, int n1, int n2);
void bmfree (void *bm);
void bmread (void *bm, int dir, int k1, int k2, int n, void *v);
void bmwrite (void *bm, int dir, int k1, int k2, int n, void *v);
bmalloc:
Input:
nbpe number of bytes per matrix element
n1 number of elements in 1st (fastest) dimension
n2 number of elements in 2nd (slowest) dimension
Returned:
bm pointer to big matrix
bmfree:
Input:
bm pointer to big matrix state (returned by bmalloc)
bmread:
Input:
d = 1 or 2: direction in which to read matrix elements
k1 1st dimension index of first matrix element to read
k2 2nd dimension index of first matrix element to read
n number of elements to read
Output:
v array[n] to contain matrix elements read
bmwrite:
Input:
d = 1 or 2: direction in which to write matrix elements
646
k1 1st dimension index of first matrix element to write
k2 2nd dimension index of first matrix element to write
n number of elements to write
v array[n] containing matrix elements to write
Notes:
The bm functions provide access to a big 2-dimensional matrix along
either the 1st or 2nd dimensions. Although, the matrix must be small
enough to fit in virtual memory, it may be too large to fit in real memory.
These functions provide equally efficient (or equally inefficient) access
to vectors in a big matrix along either the 1st or 2nd dimensions.
For example, the following algorithm will efficiently transpose an

n1 by n2 array of (n1*n2) floats stored in a file:
void *bm;
float *v;
bm = bmalloc(sizeof(float),n1,n2);
for (i2=0; i2<n2; i2++) {
(read n1 floats from input file into array v);
bmwrite(bm,1,0,i2,n1,(char*)v);
}
for (i1=0; i1<n1; i1++) {
bmread(bm,2,i1,0,n2,(char*)v);
(write n2 floats in array v to output file);
}
bmfree(bm);
647
BUTTERWORTH - Functions to design and apply Butterworth filters:
bfdesign design a Butterworth filter

bfhighpass apply a high-pass Butterworth filter
bflowpass apply a low-pass Butterworth filter
void bfhighpass (int npoles, float f3db, int n, float p[], float q[]);
void bflowpass (int npoles, float f3db, int n, float p[], float q[]);
void bfdesign (float fpass, float apass, float fstop, float astop,
int *npoles, float *f3db);
bfdesign:
Input:
fpass frequency in pass band at which amplitude is >= apass
apass amplitude in pass band corresponding to frequency fpass
fstop frequency in stop band at which amplitude is <= astop
astop amplitude in stop band corresponding to frequency fstop
Output:
npoles number of poles
f3db frequency at which amplitude is sqrt(0.5) (-3 db)
bfhighpass and bflowpass:

Input:
npoles number of poles (and zeros); npoles>=0 is required
f3db 3 db frequency; nyquist = 0.5; 0.0<=f3db<=0.5 is required
n length of p and q
p array[n] to be filtered
Output:
q filtered array[n] (may be equivalent to p)
Notes:
(1) Nyquist frequency equals 0.5
(2) The following conditions must be true:

(0.0<fpass && fpass<0.5) &&
(0.0<fstop && fstop<0.5) &&
(fpass!=fstop) &&
(0.0<astop && astop<apass && apass<1.0)
(3) if (fpass<fstop)
648
bfdesign:
Butterworth filter: compute number of poles and -3 db frequency
for a low-pass or high-pass filter, given a frequency response
constrained at two frequencies.
649
COMPLEX - Functions to manipulate complex numbers
cadd add two complex numbers

csub subtract two complex numbers
cmul multiply two complex numbers
cdiv divide two complex numbers
cmplx make a complex number from two real numbers
conjg complex conjugate of a complex number
cneg negate a complex number
cinv invert a complex number
csqrt complex square root of a complex number
cexp complex exponential of a complex number
crmul multiply a complex number by a real number
rcabs real magnitude of a complex number
Structure:
typedef struct _complexStruct { complex number
float r,i;
} complex;
complex cadd (complex a, complex b);
complex csub (complex a, complex b);
complex cmul (complex a, complex b);
complex cdiv (complex a, complex b);
float rcabs (complex z);
complex cmplx (float re, float im);
complex conjg (complex z);
complex cneg (complex z);
complex cinv (complex z);
complex csqrt (complex z);
complex cexp (complex z);
complex crmul (complex a, float x);
Notes:
The function "rcabs" was originally called "fcabs". This produced
a collision on some systems so a new name was chosen.
Reference:
Adapted from Press et al, 1988, Numerical Recipes in C (Appendix E).

650
Added function cinv().
651
COMPLEXD - Functions to manipulate double-precision complex numbers
dcadd add two dcomplex numbers

dcsub subtract two dcomplex numbers
dcmul multiply two dcomplex numbers
dcdiv divide two dcomplex numbers
dcmplx make a dcomplex number from two real numbers
dconjg dcomplex conjugate of a dcomplex number
dcneg negate a dcomplex number
dcinv invert a dcomplex number
dcsqrt dcomplex square root of a dcomplex number
dcexp dcomplex exponential of a dcomplex number
dcrmul multiply a dcomplex number by a real number
drcabs real magnitude of a dcomplex number
Structure:
typedef struct _dcomplexStruct { dcomplex number
double r,i;
} dcomplex;
dcomplex dcadd (dcomplex a, dcomplex b);
dcomplex dcsub (dcomplex a, dcomplex b);
dcomplex dcmul (dcomplex a, dcomplex b);
dcomplex dcdiv (dcomplex a, dcomplex b);
double drcabs (dcomplex z);
dcomplex dcmplx (double re, double im);
dcomplex dconjg (dcomplex z);
dcomplex dcneg (dcomplex z);
dcomplex dcinv (dcomplex z);
dcomplex dcsqrt (dcomplex z);
dcomplex dcexp (dcomplex z);
dcomplex dcrmul (dcomplex a, double x);
Notes:
The function "drcabs" was originally called "fcabs". This produced
a collision on some systems so a new name was chosen.
Reference:
Adapted from Press et al, 1988, Numerical Recipes in C (Appendix E).

652
Added function dcinv().
653
COMPLEXF - Subroutines to perform operations on complex numbers.
This set of functions complement the one in complex.c
of the CWP library
cipow raise a complex number to an integer power

crpow raise a complex number to a real power
rcpow raise a real number to a complex power
ccpow raise a complex number to a complex power
ccos compute the complex cosine of a complex angle
csin compute the complex sine of a complex angle
ccosh compute the complex hyperbolic cosine of a complex angle
csinh compute the complex hyperbolic sine of a complex angle
cexp1 compute the complex exponential of a complex number
clog compute the complex logarithm of a complex number
complex cipow(complex a, int p);
complex crpow(complex a, float p);
complex rcpow(float a, complex p);
complex ccpow (complex a, complex p)
complex ccos(complex a);
complex csin(complex a);
complex ccosh(complex a);
complex csinh(complex a);
complex cexp1(complex a);
complex clog(complex a);
Credits:
Dave Hale, original version in C++
Gabriel Alvarez, translation to C
654
COMPLEXFD - Subroutines to perform operations on double complex numbers.
This set of functions complement the one in complexd.c
of the CWP library
dcipow raise a double complex number to an integer power

dcrpow raise a double complex number to a real power
rdcpow raise a real number to a double complex power
dcdcpow raise a double complex number to a double complex power
dccos compute the double complex cosine of a double complex angle
dcsin compute the double complex sine of a double complex angle
dccosh compute the double complex hyperbolic cosine of a double complex angle
dcsinh compute the double complex hyperbolic sine of a double complex angle
dcexp1 compute the double complex exponential of a double complex number
dclog compute the double complex logarithm of a double complex number
dcomplex dcipow(dcomplex a, int p);
dcomplex dcrpow(dcomplex a, float p);
dcomplex rdcpow(float a, dcomplex p);
dcomplex dcdcpow (dcomplex a, dcomplex p)
dcomplex dccos(dcomplex a);
dcomplex dcsin(dcomplex a);
dcomplex dccosh(dcomplex a);
dcomplex dcsinh(dcomplex a);
dcomplex dcexp1(dcomplex a);
dcomplex dclog(dcomplex a);
Credits:
Dave Hale, original version in C++
Gabriel Alvarez, translation to C
655
CONVOLUTION - Compute z = x convolved with y
conv compute the convolution of two input vector arrays
Input:
lx length of x array
ifx sample index of first x
x array[lx] to be convolved with y
ly length of y array
ify sample index of first y
y array[ly] with which x is to be convolved
lz length of z array
ifz sample index of first z
Output:
z array[lz] containing x convolved with y
Function Prototype:
void conv (int lx, int ifx, float *x, int ly, int ify, float *y,
int lz, int ifz, float *z);
Notes:
The operation z = x convolved with y is defined to be
ifx+lx-1
z[i] = sum x[j]*y[i-j] ; i = ifz,...,ifz+lz-1
j=ifx
The x samples are contained in x[0], x[1], ..., x[lx-1]; likewise for
the y and z samples. The sample indices of the first x, y, and z values
determine the location of the origin for each array. For example, if
z is to be a weighted average of the nearest 5 samples of y, one might
use
...
x[0] = x[1] = x[2] = x[3] = x[4] = 1.0/5.0;
conv(5,-2,x,lx,0,y,ly,0,z);
...
In this example, the filter x is symmetric, with index of first sample = -2.
This function is optimized for architectures that can simultaneously perform

a multiply, add, and one load from memory; e.g., the IBM RISC System/6000.
Because, for each value of i, it accumulates the convolution sum z[i] in a
scalar, this function is not likely to be optimal for vector architectures.
656
CUBICSPLINE - Functions to compute CUBIC SPLINE interpolation coefficients
cakima compute cubic spline coefficients via Akima’s method

(continuous 1st derivatives, only)
cmonot compute cubic spline coefficients via the Fritsch-Carlson method
(continuous 1st derivatives, only)
csplin compute cubic spline coefficients for interpolation
(continuous 1st and 2nd derivatives)
chermite compute cubic spline coefficients via Hermite Polynomial

(continuous 1st derivatives only)
void cakima (int n, float x[], float y[], float yd[][4]);
void cmonot (int n, float x[], float y[], float yd[][4]);
void csplin (int n, float x[], float y[], float yd[][4]);
void chermite (int n, float x[], float y[], float yd[][4]);
Input:
n number of samples
x array[n] of monotonically increasing or decreasing abscissae
y array[n] of ordinates
Output:
yd array[n][4] of cubic interpolation coefficients (see notes)
Notes:
The computed cubic spline coefficients are as follows:
yd[i][0] = y(x[i]) (the value of y at x = x[i])
yd[i][1] = y’(x[i]) (the 1st derivative of y at x = x[i])
yd[i][2] = y’’(x[i]) (the 2nd derivative of y at x = x[i])
yd[i][3] = y’’’(x[i]) (the 3rd derivative of y at x = x[i])
To evaluate y(x) for x between x[i] and x[i+1] and h = x-x[i],

use the computed coefficients as follows:
y(x) = yd[i][0]+h*(yd[i][1]+h*(yd[i][2]/2.0+h*yd[i][3]/6.0))
Akima’s method provides continuous 1st derivatives, but 2nd and

3rd derivatives are discontinuous. Akima’s method is not linear,
in that the interpolation of the sum of two functions is not the
same as the sum of the interpolations.
The Fritsch-Carlson method yields continuous 1st derivatives, but 2nd

and 3rd derivatives are discontinuous. The method will yield a
657
monotonic interpolant for monotonic data. 1st derivatives are set
to zero wherever first divided differences change sign.
The method used by "csplin" yields continuous 1st and 2nd derivatives.
References:
See Akima, H., 1970, A new method for
interpolation and smooth curve fitting based on local procedures,
Journal of the ACM, v. 17, n. 4, p. 589-602.
For more information, see Fritsch, F. N., and Carlson, R. E., 1980,
Monotone piecewise cubic interpolation: SIAM J. Numer. Anal., v. 17,
n. 2, p. 238-246.
Also, see the book by Kahaner, D., Moler, C., and Nash, S., 1989,
Numerical Methods and Software, Prentice Hall. This function was
derived from SUBROUTINE PCHEZ contained on the diskette that comes
with the book.
For more general information on spline functions of all types see the book by:
Greville, T.N.E, 1969, Theory and Applications of Spline Functions,
Academic Press.
Author: Dave Hale, Colorado School of Mines c. 1989, 1990, 1991
658
DBLAS - Double precision Basic Linear Algebra subroutines
(adapted from LINPACK FORTRAN):
idamax return index of element with maximum absolute value

dasum return sum of absolute values
daxpy compute y[i] = a*x[i]+y[i]
dcopy copy x[i] to y[i] (i.e., set y[i] = x[i])
ddot return sum of x[i]*y[i] (i.e., return the dot product of x and y)
dnrm2 return square root of sum of squares of x[i]
dscal compute x[i] = a*x[i]
dswap swap x[i] and y[i]
int idamax (int n, double *sx, int incx);
double dasum (int n, double *sx, int incx);
void daxpy (int n, double sa, double *sx, int incx, double *sy, int incy);
void dcopy (int n, double *sx, int incx, double *sy, int incy);
double ddot (int n, double *sx, int incx, double *sy, int incy);
double dnrm2 (int n, double *sx, int incx);
void dscal (int n, double sa, double *sx, int incx);
void dswap (int n, double *sx, int incx, double *sy, int incy);
idmax:
Input:
n number of elements in array
sx array[n] of elements
incx increment between elements
Returned:
index of element with maximum absolute value (idamax)
dasum:
Input:
Returned:
sum of absolute values (dasum)
daxpy:
Input:
n number of elements in arrays
659
sa the scalar multiplier
sx array[n] of elements to be scaled and added
incx increment between elements of sx
sy array[n] of elements to be added
incy increment between elements of sy
Output:
sy array[n] of accumulated elements
dcopy:
Input:
sx array[n] of elements to be copied
Output:
sy array[n] of copied elements
ddot:
Input:
sy array[n] of elements
Returned: dot product of the two arrays
dnrm2:
Input:
Returned: square root of sum of squares of x[i]
dscal:
Input:
660
Output:
sx array[n] of scaled elements
661
DGE - Double precision Gaussian Elimination matrix subroutines adapted
from LINPACK FORTRAN:
dgefa Gaussian elimination to obtain the LU factorization of a matrix.

dgeco Gaussian elimination to obtain the LU factorization and condition
number of a matrix.
dgesl Solve linear system Ax = b or A’x = b after LU factorization.
void dgefa (double **a, int n, int *ipvt, int *info);
void dgeco (double **a, int n, int *ipvt, double *rcond, double *z);
void dgesl (double **a, int n, int *ipvt, double *b, int job);
dgfa:
Input:
a matrix[n][n] to be factored (see notes below)
n dimension of a
Output:
a matrix[n][n] factored (see notes below)
ipvt indices of pivot permutations (see notes below)
info index of last zero pivot (or -1 if no zero pivots)
dgeco:
Input:
n dimension of a
Output:
rcond reciprocal of condition number (see notes below)
Workspace:
z array[n]
dgesl:
Input:
a matrix[n][n] that has been LU factored (see notes below)
n dimension of a
b right-hand-side vector[n]
job =0 to solve Ax = b
662
=1 to solve A’x = b
Output:
b solution vector[n]
Notes:
These functions were adapted from LINPACK FORTRAN. Because two-dimensional
arrays cannot be declared with variable dimensions in C, the matrix a
is actually a pointer to an array of pointers to floats, as declared
above and used below.
Elements of a are stored as follows:

a[0][0] a[1][0] a[2][0] ... a[n-1][0]
a[0][1] a[1][1] a[2][1] ... a[n-1][1]
a[0][2] a[1][2] a[2][2] ... a[n-1][2]
. .
. . .
. . .
. .
a[0][n-1] a[1][n-1] a[2][n-1] ... a[n-1][n-1]
Both the factored matrix a and the pivot indices ipvt are required
to solve linear systems of equations via dgesl.
dgeco:
Given the reciprocal of the condition number, rcond, and the double
epsilon, DBL_EPSILON, the number of significant decimal digits, nsdd,
in the solution of a linear system of equations may be estimated by:
nsdd = (int)log10(rcond/DBL_EPSILON)
Author: Dave Hale, Colorado School of Mines, 1989
663
npfa_d return valid n for complex-to-complex PFA

npfar_d return valid n for real-to-complex/complex-to-real PFA
npfao_d return optimal n for complex-to-complex PFA
npfaro_d return optimal n for real-to-complex/complex-to-real PFA
pfacc_d 1D PFA complex to complex
pfacr_d 1D PFA complex to real
pfarc_d 1D PFA real to complex
pfamcc_d multiple PFA complex to real
pfa2cc_d 2D PFA complex to complex
pfa2cr_d 2D PFA complex to real
pfa2rc_d 2D PFA real to complex
int npfa_d (int nmin);
int npfao_d (int nmin, int nmax);
int npfar_d (int nmin);
int npfaro_d (int nmin, int nmax);
void pfacc_d (int isign, int n, real_complex z[]);
void pfacr_d (int isign, int n, real_complex cz[], real rz[]);
void pfarc_d (int isign, int n, real rz[], real_complex cz[]);
void pfamcc_d (int isign, int n, int nt, int k, int kt, real_complex z[]);
void pfa2cc_d (int isign, int idim, int n1, int n2, real_complex z[]);
void pfa2cr_d (int isign, int idim, int n1, int n2, real_complex cz[], real rz[]);
void pfa2rc_d (int isign, int idim, int n1, int n2, real rz[], real_complex cz[]);
npfa_d:
Input:
nmin lower bound on returned value (see notes below)
Returned: valid n for prime factor fft
npfao_d
Input:
nmax desired (but not guaranteed) upper bound on returned value
npfar_d
Input:
nmin lower bound on returned value
664
Returned: valid n for real-to-real_complex/real_complex-to-real prime factor fft
npfaro_d:
Input:
Returned: valid n for real-to-real_complex/real_complex-to-real prime factor fft
pfacc_d:
Input:
isign sign of isign is the sign of exponent in fourier kernel
n length of transform (see notes below)
z array[n] of real_complex numbers to be transformed in place
Output:
z array[n] of real_complex numbers transformed
pfacr_d:
Input:
cz array[n/2+1] of real_complex values (may be equivalenced to rz)
Output:
rz array[n] of real values (may be equivalenced to cz)
pfarc_d:
Input:
n length of transform; must be even (see notes below)
Output:
cz array[n/2+1] of real_complex values (may be equivalenced to rz)
pfamcc_d:
Input:
n number of real_complex elements per transform (see notes below)
nt number of transforms
k stride in real_complex elements within transforms
665
kt stride in real_complex elements between transforms
z array of real_complex elements to be transformed in place
Output:
z array of real_complex elements transformed
pfa2cc_d:
Input:
idim dimension to transform, either 1 or 2 (see notes)
n1 1st (fast) dimension of array to be transformed (see notes)
n2 2nd (slow) dimension of array to be transformed (see notes)
z array[n2][n1] of real_complex elements to be transformed in place
Output:
z array[n2][n1] of real_complex elements transformed
pfa2cr_d:
Input:
idim dimension to transform, which must be either 1 or 2 (see notes)
cz array of real_complex values (may be equivalenced to rz)
Output:
rz array of real values (may be equivalenced to cz)
pfa2rc_d:
Input:
Output:
cz array of real_complex values (may be equivalenced to rz)
Notes:
Table of valid n and cost for prime factor fft. For each n, cost
was estimated to be the inverse of the number of ffts done in 1 sec
on an IBM RISC System/6000 Model 320H, by Dave Hale, 08/04/91.
666
(Redone by Jack Cohen for 15 sec to rebuild NTAB table on advice of
David and Gregory Chudnovsky, 05/03/94).
Cost estimates are least accurate for very small n. An alternative method
for estimating cost would be to count multiplies and adds, but this
method fails to account for the overlapping of multiplies and adds
that is possible on some computers, such as the IBM RS/6000 family.
npfa_d:
The returned n will be composed of mutually prime factors from
the set {2,3,4,5,7,8,9,11,13,16}. Because n cannot exceed
720720 = 5*7*9*11*13*16, 720720 is returned if nmin exceeds 720720.
npfao_d:
If nmin does not exceed 720720, then the returned n will not be
less than nmin. The optimal n is chosen to minimize the estimated
cost of performing the fft, while satisfying the constraint, if
possible, that n not exceed nmax.
npfar and npfaro:

Current implemenations of real-to-real_complex and real_complex-to-real prime
factor ffts require that the transform length n be even and that n/2
be a valid length for a real_complex-to-real_complex prime factor fft. The
value returned by npfar satisfies these conditions. Also, see notes
for npfa.
pfacc:
n must be factorable into mutually prime factors taken
from the set {2,3,4,5,7,8,9,11,13,16}. in other words,
n = 2**p * 3**q * 5**r * 7**s * 11**t * 13**u
where
0 <= p <= 4, 0 <= q <= 2, 0 <= r,s,t,u <= 1
is required for pfa to yield meaningful results. this
restriction implies that n is restricted to the range
1 <= n <= 720720 (= 5*7*9*11*13*16)
pfacr:
Because pfacr uses pfacc to do most of the work, n must be even
and n/2 must be a valid length for pfacc. The simplest way to
obtain a valid n is via n = npfar(nmin). A more optimal n can be
obtained with npfaro.
667
pfarc:
Because pfarc uses pfacc to do most of the work, n must be even
pfamcc:
To perform a two-dimensional transform of an n1 by n2 real_complex array
(assuming that both n1 and n2 are valid "n"), stored with n1 fast
and n2 slow:
pfamcc(isign,n1,n2,1,n1,z); (to transform 1st dimension)
pfamcc(isign,n2,n1,n1,1,z); (to transform 2nd dimension)
pfa2cc:
Only one (either the 1st or 2nd) dimension of the 2-D array is transformed.
If idim equals 1, then n2 transforms of n1 real_complex elements are performed;

else, if idim equals 2, then n1 transforms of n2 real_complex elements are
performed.
Although z appears in the argument list as a one-dimensional array,

z may be viewed as an n1 by n2 two-dimensional array: z[n2][n1].
Valid n is computed via the "np" subroutines.
To perform a two-dimensional transform of an n1 by n2 real_complex array

and n2 slow: pfa2cc(isign,1,n1,n2,z); pfa2cc(isign,2,n1,n2,z);
pfa2cr:
If idim equals 1, then n2 transforms of n1/2+1 real_complex elements to n1 real
elements are performed; else, if idim equals 2, then n1 transforms of n2/2+1
real_complex elements to n2 real elements are performed.
Although rz appears in the argument list as a one-dimensional array,

rz may be viewed as an n1 by n2 two-dimensional array: rz[n2][n1].
Likewise, depending on idim, cz may be viewed as either an n1/2+1 by
n2 or an n1 by n2/2+1 two-dimensional array of real_complex elements.
Let n denote the transform length, either n1 or n2, depending on idim.

Because pfa2rc uses pfa2cc to do most of the work, n must be even
and n/2 must be a valid length for pfa2cc. The simplest way to
668
pfa2rc:
If idim equals 1, then n2 transforms of n1 real elements to n1/2+1 real_complex
elements are performed; else, if idim equals 2, then n1 transforms of n2
real elements to n2/2+1 real_complex elements are performed.

n2 or an n1 by n2/2+1 two-dimensional array of real_complex elements.

References:
Temperton, C., 1985, Implementation of a self-sorting
in-place prime factor fft algorithm: Journal of
Computational Physics, v. 58, p. 283-299.
Temperton, C., 1988, A new set of minimum-add rotated

rotated dft modules: Journal of Computational Physics,
v. 75, p. 190-198.
Press et al, 1988, Numerical Recipes in C, p. 417.

Revised by Baoniu Han to handle double precision. 12/14/98
669
CWP_Exit - exit subroutine for CWP/SU codes
670
FRANNOR - functions to generate a pseudo-random float normally distributed
with N(0,1); i.e., with zero mean and unit variance.
frannor return a normally distributed random float

srannor seed random number generator for normal distribution
float frannor (void);
void srannor (int seed);
frannor:
Input: (none)
Returned: normally distributed random float
srannor:
Input:
seed different seeds yield different sequences of random numbers.
Notes:
Adapted from subroutine rnor in Kahaner, Moler and Nash (1988)
which in turn was based on an algorithm by
Marsaglia and Tsang (1984).
References:
Numerical Methods and Software", D.
Kahaner, C. Moler, S. Nash, Prentice Hall, 1988.
Marsaglia G. and Tsang, W. W., 1984,

A fast, easily implemented method for sampling from decreasing or symmetric
unimodal density functions: SIAM J. Sci. Stat. Comput., v. 5, no. 2,
p. 349-359.
671
FRANUNI - Functions to generate a pseudo-random float uniformly distributed
on [0,1); i.e., between 0.0 (inclusive) and 1.0 (exclusive)
franuni return a random float

sranuni seed random number generator
float franuni (void);
void sranuni (int seed);
franuni:
Input: (none)
Returned: pseudo-random float
sranuni:
seed different seeds yield different sequences of random numbers.
Notes:
Adapted from subroutine uni in Kahaner, Moler, and Nash (1988).
This book references a set of unpublished notes by
Marsaglia.
According to the reference, this random

number generator "passes all known tests and has a period that is ...
approximately 10^19".
References:
Numerical Methods and Software", D. Kahaner, C. Moler, S. Nash,
Prentice Hall, 1988.
Marsaglia G., "Comments on the perfect uniform random number generator

Unpublished notes, Wash S. U.
672
HANKEL - Functions to compute discrete Hankel transforms
hankelalloc allocate and return a pointer to a Hankel transformer

hankelfree free a Hankel transformer
hankel0 compute the zeroth-order Hankel transform
hankel1 compute the first-order Hankel transform
void *hankelalloc (int nfft);
void hankelfree (void *ht);
void hankel0 (void *ht, float f[], float h[]);
void hankel1 (void *ht, float f[], float h[]);
hankelalloc:
Input:
nfft valid length for real to complex fft (see notes below)
Returned:
pointer to Hankel transformer
hankelfree:
Input:
ht pointer to Hankel transformer (as returned by hankelalloc)
hankel0:
Input:
f array[nfft/2+1] to be transformed
Output:
h array[nfft/2+1] transformed
hankel1:
Input:
f array[nfft/2+1] to be transformed
Output:
h array[nfft/2+1] transformed
Notes:
The zeroth-order Hankel transform is defined by:
673
Infinity
h0(k) = Integral dr r j0(k*r) f(r)
0
where j0 denotes the zeroth-order Bessel function.
The first-order Hankel transform is defined by:
Infinity
h1(k) = Integral dr r j1(k*r) f(r)
0
where j1 denotes the first-order Bessel function.
The Hankel transform is its own inverse.
The Hankel transform is computed by an Abel transform followed by

a Fourier transform.
References:
Hansen, E. W., 1985, Fast Hankel transform algorithm: IEEE Trans. on
Acoustics, Speech and Signal Processing, v. ASSP-33, n. 3, p. 666-671.
(Beware of several errors in the equations in this paper!)
Authors: Dave Hale, Colorado School of Mines, 06/04/90
674
HILBERT - Compute Hilbert transform y of x
hilbert compute the Hilbert transform
Function Prototype:
void hilbert (int n, float x[], float y[]);
Input:
n length of x and y
x array[n] to be Hilbert transformed
Output:
y array[n] containing Hilbert transform of x
Notes:
The Hilbert transform is computed by convolving x with a
windowed (approximate) version of the ideal Hilbert transformer.
675
HOLBERG1D - Compute coefficients of Holberg’s 1st derivative filter
holberg1d comput the coefficients of Holberg’s 1st derivative filter
Function Prototype:
void holbergd1 (float e, int n, float d[]);
Input:
e maximum relative error in group velocity
n number of coefficients in filter (must be 2, 4, 6, or 8)
Output:
d array[n] of coefficients
Notes:
Coefficients are output in a form suitable for convolution. The
derivative is centered halfway between coefficients d[n/2-1] and d[n/2].
Coefficients are computed via the power series method of Kindelan et al.,
1990, On the construction and efficiency of staggered numerical
differentiators for the wave equation: Geophysics 55, 107-110.
See also, Holberg, 1987, Computational aspects of the choice of
operator and sampling interval for numerical differentiation in
large-scale simulation of wave phenomena: Geophys. Prosp., 35, 629-655
Reference:
Kindelan et al., 1990,
On the construction and efficiency of staggered numerical
differentiators for the wave equation: Geophysics 55, 107-110.
See also, Holberg, 1987, Computational aspects of the choice of

operator and sampling interval for numerical differentiation in
large-scale simulation of wave phenomena: Geophys. Prosp., 35, 629-655
676
INTCUB - evaluate y(x), y’(x), y’’(x), ... via piecewise cubic interpolation
intcub evaluate y(x), y’(x), y’’(x), ... via piecewise spline

interpolation
Function Prototype:
void intcub (int ideriv, int nin, float xin[], float ydin[][4],
Input:
ideriv =0 if y(x) desired; =1 if y’(x) desired, ...
nin length of xin and ydin arrays
xin array[nin] of monotonically increasing or decreasing x values
ydin array[nin][4] of y(x), y’(x), y’’(x), and y’’’(x)
nout length of xout and yout arrays
xout array[nout] of x values at which to evaluate y(x), y’(x), ...
Output:
yout array[nout] of y(x), y’(x), ... values
Notes:
xin values must be monotonically increasing or decreasing.
Extrapolation of the function y(x) for xout values outside the range
spanned by the xin values is performed using the derivatives in
ydin[0][0:3] or ydin[nin-1][0:3], depending on whether xout is closest
to xin[0] or xin[nin-1], respectively.
Reference:
See: Greville, T. N., Theory and Application of Spline Functions, Academic Press
for a general discussion of spline functions.
677
INTL2B - bilinear interpolation of a 2-D array of bytes
intl2b bilinear interpolation of a 2-D array of bytes
Function Prototype:
void intl2b (int nxin, float dxin, float fxin,
int nyin, float dyin, float fyin, unsigned char *zin,
int nxout, float dxout, float fxout,
int nyout, float dyout, float fyout, unsigned char *zout);
Input:
nxin number of x samples input (fast dimension of zin)
dxin x sampling interval input
fxin first x sample input
nyin number of y samples input (slow dimension of zin)
dyin y sampling interval input
fyin first y sample input
zin array[nyin][nxin] of input samples (see notes)
nxout number of x samples output (fast dimension of zout)
dxout x sampling interval output
fxout first x sample output
nyout number of y samples output (slow dimension of zout)
dyout y sampling interval output
fyout first y sample output
Output:
zout array[nyout][nxout] of output samples (see notes)
Notes:
The arrays zin and zout must passed as pointers to the first element of
a two-dimensional contiguous array of unsigned char values.
Constant extrapolation of zin is used to compute zout for

output x and y outside the range of input x and y.
For efficiency, this function builds a table of interpolation

coefficents pre-multiplied by byte values. To keep the table
reasonably small, the interpolation does not distinguish
between x and y values that differ by less than dxin/ICMAX
and dyin/ICMAX, respectively, where ICMAX is a parameter
defined above.
Author: Dave Hale, Colorado School of Mines, c. 1989-1991.
678
INTLINC - evaluate complex y(x) via linear interpolation of y(x[0]), y(x[1]), ...
Function Prototype:
void intlinc (int nin, float xin[], complex yin[], complex yinl, complex yinr,
int nout, float xout[], complex yout[]);
Input:
nin length of xin and yin arrays
yin array[nin] of input y(x) values
yinl value used to extraplate y(x) to left of input yin values
yinr value used to extraplate y(x) to right of input yin values
xout array[nout] of x values at which to evaluate y(x)
Output:
yout array[nout] of linearly interpolated y(x) values
Notes:
spanned by the xin values in performed as follows:
For monotonically increasing xin values,

yout=yinl if xout<xin[0], and yout=yinr if xout>xin[nin-1].
For monotonically decreasing xin values,

yout=yinl if xout>xin[0], and yout=yinr if xout<xin[nin-1].
If nin==1, then the monotonically increasing case is used.

/
679
INTLIN - evaluate y(x) via linear interpolation of y(x[0]), y(x[1]), ...
intlin evaluate y(x) via linear interpolation of y(x[0]), y(x[1]), ...
Function Prototype:
void intlin (int nin, float xin[], float yin[], float yinl, float yinr,
int nout, float xout[], float yout[]);
Input:
nin length of xin and yin arrays
yin array[nin] of input y(x) values
yinl value used to extraplate y(x) to left of input yin values
yinr value used to extraplate y(x) to right of input yin values
xout array[nout] of x values at which to evaluate y(x)
Output:
yout array[nout] of linearly interpolated y(x) values
Notes:
spanned by the xin values in performed as follows:
For monotonically increasing xin values,

yout=yinl if xout<xin[0], and yout=yinr if xout>xin[nin-1].
For monotonically decreasing xin values,

yout=yinl if xout>xin[0], and yout=yinr if xout<xin[nin-1].
If nin==1, then the monotonically increasing case is used.
680
INTSINC8 - Functions to interpolate uniformly-sampled data via 8-coeff. sinc
approximations:
ints8c interpolation of a uniformly-sampled complex function y(x) via an

8-coefficient sinc approximation.
ints8r Interpolation of a uniformly-sampled real function y(x) via a
table of 8-coefficient sinc approximations
void ints8c (int nxin, float dxin, float fxin, complex yin[],
complex yinl, complex yinr, int nxout, float xout[], complex yout[]);
void ints8r (int nxin, float dxin, float fxin, float yin[],
float yinl, float yinr, int nxout, float xout[], float yout[]);
Input:
nxin number of x values at which y(x) is input
dxin x sampling interval for input y(x)
fxin x value of first sample input
yin array[nxin] of input y(x) values: yin[0] = y(fxin), etc.
yinl value used to extrapolate yin values to left of yin[0]
yinr value used to extrapolate yin values to right of yin[nxin-1]
nxout number of x values a which y(x) is output
xout array[nxout] of x values at which y(x) is output
Output:
yout array[nxout] of output y(x): yout[0] = y(xout[0]), etc.
Notes:
Because extrapolation of the input function y(x) is defined by the
left and right values yinl and yinr, the xout values are not restricted
to lie within the range of sample locations defined by nxin, dxin, and
fxin.
The maximum error for frequiencies less than 0.6 nyquist is less than
one percent.
681
INTTABLE8 - Interpolation of a uniformly-sampled complex function y(x)
via a table of 8-coefficient interpolators
intt8c interpolation of a uniformly-sampled complex function y(x)

via a table of 8-coefficient interpolators
intt8r interpolation of a uniformly-sampled real function y(x) via a
table of 8-coefficient interpolators
Function Prototype:
void intt8c (int ntable, float table[][8],
int nxin, float dxin, float fxin, complex yin[],
complex yinl, complex yinr, int nxout, float xout[], complex yout[]);
void intt8r (int ntable, float table[][8],
int nxin, float dxin, float fxin, float yin[],
float yinl, float yinr, int nxout, float xout[], float yout[]);
Input:
ntable number of tabulated interpolation operators; ntable>=2
table array of tabulated 8-point interpolation operators
yin array of input y(x) values: yin[0] = y(fxin), etc.
xout array of x values at which y(x) is output
Output:
yout array of output y(x) values: yout[0] = y(xout[0]), etc.
NOTES:
ntable must not be less than 2.
The table of interpolation operators must be as follows:
Let d be the distance, expressed as a fraction of dxin, from a particular

xout value to the sampled location xin just to the left of xout. Then,
for d = 0.0,
table[0][0:7] = 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0, 0.0
are the weights applied to the 8 input samples nearest xout.
682
Likewise, for d = 1.0,
table[ntable-1][0:7] = 0.0, 0.0, 0.0, 0.0, 1.0, 0.0, 0.0, 0.0
are the weights applied to the 8 input samples nearest xout. In general,
for d = (float)itable/(float)(ntable-1), table[itable][0:7] are the
weights applied to the 8 input samples nearest xout. If the actual sample
distance d does not exactly equal one of the values for which interpolators
are tabulated, then the interpolator corresponding to the nearest value of
d is used.
Because extrapolation of the input function y(x) is defined by the left

and right values yinl and yinr, the xout values are not restricted to lie
within the range of sample locations defined by nxin, dxin, and fxin.
683
LINEAR_REGRESSION - Compute linear regression of (y1,y2,...,yn) against
(x1,x2,...,xn)
Function Prototype:
void linear_regression(float *y, float *x, int n, float coeff[4]);
Input:
y array of y values
x array of x values
n length of y and x arrays
Output:
coeff[4] where:
coeff[0] slope of best fit line

coeff[1] intercept of best fit line
coeff[2] correlation coefficient of fit (1 = perfect) [dimensionless]
coeff[3] standard error of fit (0 = perfect) [dimensions of y]
Notes:
y(x)
| * . fit is y(x) = a x + b
| .
| . *
| * .
| . *
------------------- x
n Sum[x*y] - Sum[x]*Sum[y]
a = --------------------------
n Sum[x*x] - Sum[x]*Sum[x]
Sum[y] - a*Sum[x]
b = -----------------
n
cc = std definition
se = std definition
Author: Chris Liner, UTulsa, 11/16/03
684
MKDIFF - Make an n-th order DIFFerentiator via Taylor’s series method.
mkdiff make discrete Taylor series approximation to n’th derivative.
Function Prototype:
void mkdiff (int n, float a, float h, int l, int m, float d[]);
Input:
n order of desired derivative (n>=0 && n<=m-l is required)
a fractional distance from integer sampling index (see notes)
h sampling interval
l sampling index of first coefficient (see notes below)
m sampling index of last coefficient (see notes below)
Output:
d array[m-l+1] of coefficients for n’th order differentiator
Notes:
The abscissae x of a sampled function f(x) can always be expressed as
x = (j+a)*h, where j is an integer, a is a fraction, and h is the
sampling interval. To approximate the n’th order derivative fn(x)
of the sampled function f(x) at x = (j+a)*h, use the m-l+1 coefficients
in the output array d[] as follows:
fn(x) = d[0]*f(j-l) + d[1]*f(j-l-1) +...+ d[m-l]*f(j-m)
i.e., convolve the coefficients in d with the samples in f.
m-l+1 (the number of coefficients) must not be greater than the

NCMAX parameter specified below.
For best approximations,

when n is even, use a = 0.0, l = -m
when n is odd, use a = 0.5, l = -m-1
685
MKHDIFF - Compute filter approximating the bandlimited HalF-DIFFerentiator.
mkhdiff - Compute filter approximating the bandlimited half-differentiator.
Function Prototype:
void mkhdiff (float h, int l, float d[]);
Input:
h sampling interval
l half-length of half-differentiator (length = 1+2*l is odd)
Output:
d array[1+2*l] of coefficients for half-differentiator
Notes:
The half-differentiator is defined by
pi
d[l+j] = sqrt(1/h)/(2pi) * integral dw sqrt(-iw)*exp(-iwj)
-pi
pi
= sqrt(2/h)/(2pi) * integral dw sqrt(w)*(cos(wj)-sin(wj))
0
for j = -l, -l+1, ... , l.
An alternative definition is that f’(j) = d(j)*d(j)*f(j), where

f’(j) denotes the derivative of a sampled function f(j) and *
denotes a convolution sum.
The half-derivative g(j) of f(j) may be computed by the following sum:
g(j) = d[0]*f(j+l) + d[1]*f(j+l-1) + ... + d[2*l]*f(j-l)
The integral over frequency is evaluated numerically using Simpson’s

method. Although the Filon method of numerical integration is more
appropriate for this integral, the truncation of d[l+j] for |j| > l
is probably the greatest source of error. In any case, d[l+j] is
cosine-tapered to reduce these truncation errors.
686
MKSINC - Compute least-squares optimal sinc interpolation coefficients.
mksinc Compute least-squares optimal sinc interpolation coefficients.
Function Prototype:
void mksinc (float d, int lsinc, float sinc[]);
Input:
d fractional distance to interpolation point; 0.0<=d<=1.0
lsinc length of sinc approximation; lsinc%2==0 and lsinc<=20
Output:
sinc array[lsinc] containing interpolation coefficients
Notes:
The coefficients are a least-squares-best approximation to the ideal
sinc function for frequencies from zero up to a computed maximum
frequency. For a given interpolator length, lsinc, mksinc computes
the maximum frequency, fmax (expressed as a fraction of the nyquist
frequency), using the following empirically derived relation (from
a Western Geophysical Technical Memorandum by Ken Larner):
fmax = min(0.066+0.265*log(lsinc),1.0)
Note that fmax increases as lsinc increases, up to a maximum of 1.0.

Use the coefficients to interpolate a uniformly-sampled function y(i)
as follows:
lsinc-1
y(i+d) = sum sinc[j]*y(i+j+1-lsinc/2)
j=0
Interpolation error is greatest for d=0.5, but for frequencies less

than fmax, the error should be less than 1.0 percent.
687
MNEWT - Solve non-linear system of equations f(x) = 0 via Newton’s method
mnewt Solve non-linear system of equations f(x) = 0 via Newton’s method
Function Prototype:
int mnewt (int maxiter, float ftol, float dxtol, int n, float *x, void *aux,
void (*fdfdx)(int n, float *x, float *f, float **dfdx, void *aux));
Input:
maxiter maximum number of iterations
ftol converged when sum of absolute values of f less than ftol
dxtol converged when sum of absolute values of dx less than dxtol
n number of equations
x array[n] containing initial guess of solution
aux pointer to auxiliary parameters to be passed to fdfdx
fdfdx pointer to function to evaluate f(x) and f’(x)
Output:
x array[n] containing solution
Returned: number of iterations; -1 if failed to converge in maxiter
Input to the user-supplied function fdfdx:

n number of equations
x array[n] of x0, x1, ...
aux pointer to auxiliary variables required by fdfdx.
Output from the user-supplied function fdfdx:

f array[n] of f0(x), f1(x), ...
dfdx array[n][n] of f’(x); dfdx[j][i] = dfi/dxj
688
npfa return valid n for complex-to-complex PFA

npfar return valid n for real-to-complex/complex-to-real PFA
npfao return optimal n for complex-to-complex PFA
npfaro return optimal n for real-to-complex/complex-to-real PFA
pfacc 1D PFA complex to complex
pfacr 1D PFA complex to real
pfarc 1D PFA real to complex
pfamcc multiple PFA complex to real
pfa2cc 2D PFA complex to complex
pfa2cr 2D PFA complex to real
pfa2rc 2D PFA real to complex
int npfa (int nmin);
int npfao (int nmin, int nmax);
int npfar (int nmin);
int npfaro (int nmin, int nmax);
void pfacc (int isign, int n, complex z[]);
void pfacr (int isign, int n, complex cz[], float rz[]);
void pfarc (int isign, int n, float rz[], complex cz[]);
void pfamcc (int isign, int n, int nt, int k, int kt, complex z[]);
void pfa2cc (int isign, int idim, int n1, int n2, complex z[]);
void pfa2cr (int isign, int idim, int n1, int n2, complex cz[], float rz[]);
void pfa2rc (int isign, int idim, int n1, int n2, float rz[], complex cz[]);
npfa:
Input:
npfao
Input:
npfar
Input:
689
Returned: valid n for real-to-complex/complex-to-real prime factor fft
npfaro:
Input:
Returned: valid n for real-to-complex/complex-to-real prime factor fft
pfacc:
Input:
z array[n] of complex numbers to be transformed in place
Output:
z array[n] of complex numbers transformed
pfacr:
Input:
cz array[n/2+1] of complex values (may be equivalenced to rz)
Output:
pfarc:
Input:
n length of transform; must be even (see notes below)
Output:
cz array[n/2+1] of complex values (may be equivalenced to rz)
pfamcc:
Input:
n number of complex elements per transform (see notes below)
nt number of transforms
k stride in complex elements within transforms
690
kt stride in complex elements between transforms
z array of complex elements to be transformed in place
Output:
z array of complex elements transformed
pfa2cc:
Input:
idim dimension to transform, either 1 or 2 (see notes)
z array[n2][n1] of complex elements to be transformed in place
Output:
z array[n2][n1] of complex elements transformed
pfa2cr:
Input:
cz array of complex values (may be equivalenced to rz)
Output:
pfa2rc:
Input:
Output:
cz array of complex values (may be equivalenced to rz)
Notes:
Table of valid n and cost for prime factor fft. For each n, cost
was estimated to be the inverse of the number of ffts done in 1 sec
on an IBM RISC System/6000 Model 320H, by Dave Hale, 08/04/91.
691
(Redone by Jack Cohen for 15 sec to rebuild NTAB table on advice of
David and Gregory Chudnovsky, 05/03/94).
Cost estimates are least accurate for very small n. An alternative method
for estimating cost would be to count multiplies and adds, but this
method fails to account for the overlapping of multiplies and adds
that is possible on some computers, such as the IBM RS/6000 family.
npfa:
npfao:
If nmin does not exceed 720720, then the returned n will not be
less than nmin. The optimal n is chosen to minimize the estimated
cost of performing the fft, while satisfying the constraint, if
possible, that n not exceed nmax.
npfar and npfaro:

Current implemenations of real-to-complex and complex-to-real prime
factor ffts require that the transform length n be even and that n/2
be a valid length for a complex-to-complex prime factor fft. The
value returned by npfar satisfies these conditions. Also, see notes
for npfa.
pfacc:
n must be factorable into mutually prime factors taken
from the set {2,3,4,5,7,8,9,11,13,16}. in other words,
n = 2**p * 3**q * 5**r * 7**s * 11**t * 13**u
where
0 <= p <= 4, 0 <= q <= 2, 0 <= r,s,t,u <= 1
is required for pfa to yield meaningful results. this
restriction implies that n is restricted to the range
1 <= n <= 720720 (= 5*7*9*11*13*16)
pfacr:
Because pfacr uses pfacc to do most of the work, n must be even
692
pfarc:
Because pfarc uses pfacc to do most of the work, n must be even
pfamcc:
To perform a two-dimensional transform of an n1 by n2 complex array
and n2 slow:
pfamcc(isign,n1,n2,1,n1,z); (to transform 1st dimension)
pfamcc(isign,n2,n1,n1,1,z); (to transform 2nd dimension)
pfa2cc:
Only one (either the 1st or 2nd) dimension of the 2-D array is transformed.
If idim equals 1, then n2 transforms of n1 complex elements are performed;

else, if idim equals 2, then n1 transforms of n2 complex elements are
performed.
Although z appears in the argument list as a one-dimensional array,

z may be viewed as an n1 by n2 two-dimensional array: z[n2][n1].
Valid n is computed via the "np" subroutines.
To perform a two-dimensional transform of an n1 by n2 complex array

and n2 slow: pfa2cc(isign,1,n1,n2,z); pfa2cc(isign,2,n1,n2,z);
pfa2cr:
If idim equals 1, then n2 transforms of n1/2+1 complex elements to n1 real
elements are performed; else, if idim equals 2, then n1 transforms of n2/2+1
complex elements to n2 real elements are performed.

n2 or an n1 by n2/2+1 two-dimensional array of complex elements.

693
pfa2rc:
If idim equals 1, then n2 transforms of n1 real elements to n1/2+1 complex
elements are performed; else, if idim equals 2, then n1 transforms of n2
real elements to n2/2+1 complex elements are performed.

n2 or an n1 by n2/2+1 two-dimensional array of complex elements.

References:
Temperton, C., 1985, Implementation of a self-sorting
in-place prime factor fft algorithm: Journal of
Computational Physics, v. 58, p. 283-299.
Temperton, C., 1988, A new set of minimum-add rotated

rotated dft modules: Journal of Computational Physics,
v. 75, p. 190-198.
Differ significantly from Press et al, 1988, Numerical Recipes in C, p. 417.
694
POLAR - Functions to map data in rectangular coordinates to polar and vise versa
recttopolar convert a function p(x,y) to a function q(a,r)

polartorect convert a function q(a,r) to a function p(x,y)
void recttopolar ( int nx, float dx, float fx, int ny, float dy, float fy,
float **p, int na, float da, float fa, int nr,
float dr, float fr, float **q);
void polartorect ( int na, float da, float fa, int nr, float dr, float fr,
float **q, int nx, float dx, float fx, int ny,
float dy, float fy, float **p)
recttopolar:
Input:
nx number of x samples
dx x sampling interval
fx first x sample
ny number of y samples
dy y sampling interval
fy first y sample
p array[ny][nx] containing samples of p(x,y)
na number of a samples
da a sampling interval
fa first a sample
nr number of r samples
dr r sampling interval
fr first r sample
Output:
q array[nr][na] containing samples of q(a,r)
polartorect:
Input:
na number of a samples
da a sampling interval
fa first a sample
nr number of r samples
dr r sampling interval
fr first r sample
nx number of x samples
fx first x sample
695
ny number of y samples
dy y sampling interval
fy first y sample
q array[nr][na] containing samples of q(a,r)
Output:
p array[ny][nx] containing samples of p(x,y)
Notes:
The polar angle a is measured in radians,
x = r*cos(a) and y = r*sin(a).
recttopolar:
Linear extrapolation is used to determine the value of p(x,y) for
x and y coordinates not in the range corresponding to nx, dx, ....
polartorect:
Linear extrapolation is used to determine the value of q(a,r) for
a and r coordinates not in the range corresponding to na, da, ....
696
PRINTERPLOT - Functions to make a printer plot of a 1-dimensional array
pp1d printer plot of 1d array

pplot1 printer plot of 1d array
void pp1d (FILE *fp, char *title, int lx, int ifx, float x[]);
void pplot1 (FILE *fp, char *title, int nx, float ax[]);
pp1d:
Input:
fp file pointer for output (e.g., stdout, stderr, etc.)
title title of plot
lx length of x
ifx index of first x
x array[lx] to be plotted
pplot1:
Input:
fp file pointer for printed output (e.g., stdout, stderr, etc.)
title title of plot
nx number of x values to be plotted
ax array[nx] of x values
Notes:
Are two subroutines to do this really necessary?
697
QUEST - Functions to ESTimate Quantiles:
quest returns estimate - use when entire array is in memory

questalloc returns quantile estimator - use before questupdate
questupdate updates and returns current estimate - use for large
numbers of floats, too big to fit in memory at one time
questfree frees quantile estimator
quest:
Input:
p quantile to be estimated (0.0<=p<=1.0 is required)
n number of samples in array x (n>=5 is required)
x array[n] of floats
Returned: the estimate of a specified quantile
questalloc:
Input:
p quantile to be estimated (0.0<=p<=1.0 is required)
n number of samples in array x (n>=5 is required)
Returned: pointer to a quantile estimator
questupdate:
Input:
q pointer to quantile estimator (as returned by questalloc)
n number of samples in array x
Returned: quantile estimate
questfree:
q pointer to quantile estimator (as returned by questalloc)
Notes:
quest:
The estimate should approach the sample quantile in the limit of large n.
The estimate is most accurate for cumulative distribution functions

that are smooth in the neighborhood of the quantile specified by p.
This function is an implementation of the algorithm published by
698
Jain and Chlamtac (1985).
questalloc:
This function must be called before calling function questupdate.
See also notes in questupdate.
questupdate:
The estimate should approach the sample quantile in the limit of large n.
The estimate is most accurate for cumulative distribution functions

that are smooth in the neighborhood of the quantile specified by p.
This function is an implementation of the algorithm published by

Jain, R. and Chlamtac, I., 1985, The PP algorithm for dynamic
calculation of quantiles and histograms without storing observations:
Comm. ACM, v. 28, n. 10.
Reference:
Jain, R. and Chlamtac, I., 1985, The PP algorithm for dynamic
calculation of quantiles and histograms without storing observations:
Comm. ACM, v. 28, n. 10.
699
RESSINC8 - Functions to resample uniformly-sampled data via 8-coefficient sinc
approximations:
ress8c resample a uniformly-sampled complex function via 8-coeff. sinc approx.

ress8r resample a uniformly-sampled real function via 8-coeff. sinc approx.
void ress8r (int nxin, float dxin, float fxin, float yin[],
float yinl, float yinr,
int nxout, float dxout, float fxout, float yout[]);
void ress8c (int nxin, float dxin, float fxin, complex yin[],
complex yinl, complex yinr,
int nxout, float dxout, float fxout, complex yout[]);
Input:
nxout number of x values at which y(x) is output
dxout x sampling interval for output y(x)
fxout x value of first sample output
Output:
yout array[nxout] of output y(x) values: yout[0] = y(fxout), etc.
Notes:
left and right values yinl and yinr, the output x values defined
by nxout, dxout, and fxout are not restricted to lie within the range
of input x values defined by nxin, dxin, and fxin.
The maximum error for frequiencies less than 0.6 nyquist is

less than one percent.
700
RFWTVA - Rasterize a Float array as Wiggle-Trace-Variable-Area.
rfwtva rasterize a float array as wiggle-trace-variable-area.
Function Prototype:
void rfwtva (int n, float z[], float zmin, float zmax, float zbase,
int yzmin, int yzmax, int xfirst, int xlast,
int wiggle, int nbpr, unsigned char *bits, int endian);
Input:
n number of samples in array to rasterize
z array[n] to rasterize
zbase z values between zbase and zmax will be filled (see notes)
yzmin horizontal raster coordinate corresponding to zmin
yzmax horizontal raster coordinate corresponding to zmax
xfirst vertical raster coordinate of z[0] (see notes)
xlast vertical raster coordinate of z[n-1] (see notes)
wiggle =0 for no wiggle (VA only); =1 for wiggle (with VA)
wiggle 2<=wiggle<=5 for solid/grey coloring of VA option
shade of grey: wiggle=2 light grey, wiggle=5 black
nbpr number of bytes per row of bits
bits pointer to first (top,left) byte in image
endian byte order =1 big endian =0 little endian
Output:
Notes:
The raster coordinate of the (top,left) bit in the image is (0,0).
In other words, x increases downward and y increases to the right.
Raster scan lines run from left to right, and from top to bottom.
Therefore, xfirst, xlast, yzmin, and yzmax should not be less than 0.
Likewise, yzmin and yzmax should not be greater than nbpr*8-1, and
care should be taken to ensure that xfirst and xlast do not cause bits
to be set outside (off the bottom) of the image.
Variable area fill is performed on the right-hand (increasing y) side

of the wiggle. If yzmin is greater than yzmax, then z values between
zmin will be plotted to the right of zmax, and z values between zbase
and zmin are filled. Swapping yzmin and yzmax is an easy way to
reverse the polarity of a wiggle.
701
The variable "endian" must have a value of 1 or 0. If this is
not a case an error is returned.

Added solid/grey shading scheme, wiggle>=2 option for peaks/troughs
702
RFWTVAINT - Rasterize a Float array as Wiggle-Trace-Variable-Area, with
8 point sinc INTerpolation.
rfwtvaint rasterize a float array as wiggle-trace-variable-area, and

apply sinc interploation for display purposes.
Function Prototype:
void rfwtvaint (int n, float z[], float zmin, float zmax, float zbase,
int yzmin, int yzmax, int xfirst, int xlast,
int wiggle, int nbpr, unsigned char *bits, int endian);
Input:
n number of samples in array to rasterize
z array[n] to rasterize
zbase z values between zbase and zmax will be filled (see notes)
yzmin horizontal raster coordinate corresponding to zmin
yzmax horizontal raster coordinate corresponding to zmax
xfirst vertical raster coordinate of z[0] (see notes)
xlast vertical raster coordinate of z[n-1] (see notes)
wiggle =0 for no wiggle (VA only); =1 for wiggle (with VA)
wiggle 2<=wiggle<=5 for solid/grey coloring of VA option
shade of grey: wiggle=2 light grey, wiggle=5 black
nbpr number of bytes per row of bits
endian byte order =1 big endian =0 little endian
Output:
Notes:
The raster coordinate of the (top,left) bit in the image is (0,0).
In other words, x increases downward and y increases to the right.
Raster scan lines run from left to right, and from top to bottom.
Therefore, xfirst, xlast, yzmin, and yzmax should not be less than 0.
Likewise, yzmin and yzmax should not be greater than nbpr*8-1, and
care should be taken to ensure that xfirst and xlast do not cause bits
to be set outside (off the bottom) of the image.
Variable area fill is performed on the right-hand (increasing y) side

of the wiggle. If yzmin is greater than yzmax, then z values between
zmin will be plotted to the right of zmax, and z values between zbase
703
and zmin are filled. Swapping yzmin and yzmax is an easy way to
reverse the polarity of a wiggle.
The variable "endian" must have a value of 1 or 0. If this is

not a case an error is returned.
The interpolation is by the 8 point sinc interpolation routine s8r.

Memorial University of Newfoundland: Tony Kocurko, Sept 1995.
Added sinc interpolation.
added solid/grey color scheme for peaks/troughs wiggle=2 option
704
SBLAS - Single precision Basic Linear Algebra Subroutines
(adapted from LINPACK FORTRAN):
isamax return index of element with maximum absolute value

sasum return sum of absolute values
saxpy compute y[i] = a*x[i]+y[i]
scopy copy x[i] to y[i] (i.e., set y[i] = x[i])
sdot return sum of x[i]*y[i] (i.e., return the dot product of x and y)
snrm2 return square root of sum of squares of x[i]
sscal compute x[i] = a*x[i]
sswap swap x[i] and y[i]
int isamax (int n, float *sx, int incx);
float sasum (int n, float *sx, int incx);
void saxpy (int n, float sa, float *sx, int incx, float *sy, int incy);
void scopy (int n, float *sx, int incx, float *sy, int incy);
float sdot (int n, float *sx, int incx, float *sy, int incy);
float snrm2 (int n, float *sx, int incx);
void sscal (int n, float sa, float *sx, int incx);
void sswap (int n, float *sx, int incx, float *sy, int incy);
isamax:
Input:
Returned: index of element with maximum absolute value
sasum:
Input:
Returned: sum of absolute values
saxpy:
Input:
sx array[n] of elements to be scaled and added
705
sy array[n] of elements to be added
Output:
sy array[n] of accumulated elements
scopy:
Input:
sx array[n] of elements to be copied
Output:
sy array[n] of copied elements
sdot:
Input:
Returned: dot product of the two arrays
snrm2
Input:
Returned: square root of sum of squares of x[i]
sscal:
Input:
Output:
706
sx array[n] of scaled elements
sswap:
Input:
Output:
sx array[n] of swapped elements
sy array[n] of swapped elements
Notes:
Adapted from Linpack Fortran.
snrm2:
This simple version may cause overflow or underflow!
707
SCAXIS - compute a readable scale for use in plotting axes
scaxis compute a readable scale for use in plotting axes
Function Prototype:
void scaxis (float x1, float x2, int *nxnum, float *dxnum, float *fxnum);
Input:
x1 first x value
x2 second x value
nxnum desired number of numbered values
Output:
nxnum number of numbered values
dxnum increment between numbered values (dxnum>0.0)
fxnum first numbered value
Notes:
scaxis attempts to honor the user-specified nxnum. However, nxnum
will be modified if necessary for readability. Also, fxnum and nxnum
will be adjusted to compensate for roundoff error; in particular,
fxnum will not be less than xmin-eps, and fxnum+(nxnum-1)*dxnum
will not be greater than xmax+eps, where eps = 0.0001*(xmax-xmin).
xmin is the minimum of x1 and x2. xmax is the maximum of x1 and x2.
708
SGA - Single precision general matrix functions adapted from LINPACK FORTRAN:
sgefa Gaussian elimination to obtain the LU factorization of a matrix.

sgeco Gaussian elimination to obtain the LU factorization and
condition number of a matrix.
sgesl Solve linear system Ax = b or A’x = b after LU factorization.
void sgefa (float **a, int n, int *ipvt, int *info);
void sgeco (float **a, int n, int *ipvt, float *rcond, float *z);
void sgesl (float **a, int n, int *ipvt, float *b, int job);
sgefa:
Input:
n dimension of a
Output:
info index of last zero pivot (or -1 if no zero pivots)
sgeco:
Input:
n dimension of a
Output:
rcond reciprocal condition number (see notes below)
Workspace:
z array[n]
sgesl
Input:
a matrix[n][n] that has been LU factored (see notes below)
n dimension of a
b right-hand-side vector[n]
job =0 to solve Ax = b
=1 to solve A’x = b
709
Output:
b solution vector[n]
Notes:
These functions were adapted from LINPACK FORTRAN. Because two-dimensional
Elements of a are stored as follows:

a[0][0] a[1][0] a[2][0] ... a[n-1][0]
a[0][1] a[1][1] a[2][1] ... a[n-1][1]
a[0][2] a[1][2] a[2][2] ... a[n-1][2]
. .
. . .
. . .
. .
a[0][n-1] a[1][n-1] a[2][n-1] ... a[n-1][n-1]
Both the factored matrix a and the pivot indices ipvt are required
to solve linear systems of equations via sgesl.
sgeco:
Given the reciprocal of the condition number, rcond, and the float
epsilon, FLT_EPSILON, the number of significant decimal digits, nsdd,
in the solution of a linear system of equations may be estimated by:
nsdd = (int)log10(rcond/FLT_EPSILON)
This function was adapted from LINPACK FORTRAN. Because two-dimensional

710
SHFS8R - Shift a uniformly-sampled real-valued function y(x) via
a table of 8-coefficient sinc approximations.
shfs8r shift a uniformly-sampled real function via a table of 8-coeff. sinc

approximations.
void shfs8r (float dx, int nxin, float fxin, float yin[],
float yinl, float yinr, int nxout, float fxout, float yout[]);
Input:
dx x sampling interval for both input and output y(x)
fxout x value of first sample output
Output:
yout array[nxout] of output y(x) values: yout[0] = y(fxout), etc.
Notes:
left and right values yinl and yinr, the output samples defined by
dx, nxout, and fxout are not restricted to lie within the range of
input sample locations defined by dx, nxin, and fxin.
The maximum error for frequencies less than 0.6*nyquist is less

than one percent.
711
SINC - Return SINC(x) for as floats or as doubles
fsinc return float value of sinc(x) for x input as a float

dsinc return double precision sinc(x) for double precision x
Function Prototype:
float fsinc (float x);
double dsinc (double x);
Input:
x value at which to evaluate sinc(x)
Returned: sinc(x)
Notes:
sinc(x) = sin(PI*x)/(PI*x)
712
SORT - Functions to sort arrays of data or arrays of indices
hpsort sort an array of floats by the heap sort method

qkisort sort an array of indices i[] so that
a[i[0]] <= a[i[1]] <= ... <= a[i[n-1]]
uses the "quick sort" method
qkifind partially sort an array of indices i[] so that the
index i[m] has the value it would have if the entire
array of indices were sorted such that
a[i[0]] <= a[i[1]] <= ... <= a[i[n-1]]
qksort sort an array of floats such that a[0] <= a[1] <= ... <= a[n-1]
qkfind partially sort an array of floats so that the element a[m] has
the value it would have if the entire array were sorted
such that a[0] <= a[1] <= ... <= a[n-1]
void hpsort (int n, float a[]);
void qkisort (int n, float a[], int i[]);
void qkifind (int m, int n, float a[], int i[]);
void qksort (int n, float a[]);
void qkfind (int m, int n, float a[]);
hpsort:
Input:
n number of elements in a
a array[n] to be sorted
Output:
a array[n] sorted
qkisort:
Input:
n number of elements in array a
a array[n] elements
i array[n] indices to be sorted
Output:
i array[n] indices sorted
qkifind:
713
Input:
m index of element to be found
a array[n] elements
i array[n] indices to be partially sorted
Output:
i array[n] indices partially sorted sorted
qksort:
Input:
a array[n] containing elements to be sorted
Output:
a array[n] containing sorted elements
qkfind:
Input:
m index of element to be found
a array[n] to be partially sorted
Output:
a array[n] partially sorted
Notes:
hpsort:
The heap sort algorithm is, at worst, N log_2 N, and in most cases
is 20% faster. Adapted from Standish.
qkisort, qkifind, qksort, qkfind:

n must be less than 2^NSTACK, where NSTACK is defined above.
qkisort:
This function is adapted from procedure quicksort by
Hoare, C.A.R., 1961, Communications of the ACM, v. 4, p. 321;
the main difference is that recursion is accomplished
explicitly via a stack array for efficiency; also, a simple
insertion sort is used to sort subarrays too small to be
partitioned efficiently.
714
qkifind:
This function is adapted from procedure find by
Hoare, C.A.R., 1961, Communications of the ACM, v. 4, p. 321.
qksort:
This function is adapted from procedure quicksort by
Hoare, C.A.R., 1961, Communications of the ACM, v. 4, p. 321;
the main difference is that recursion is accomplished
explicitly via a stack array for efficiency; also, a simple
insertion sort is used to sort subarrays too small to be
partitioned efficiently.
qkfind:
This function is adapted from procedure find by Hoare 1961.
Reference:
hpsort:
Standish, T. A., Data Structure Techniques, p. 91.
See also Press, W. A., et al., Numerical Recipes in C.
quick sort:
Hoare, C.A.R., 1961, Communications of the ACM, v. 4, p. 321.
715
SQR - Single precision QR decomposition functions adapted from LINPACK FORTRAN:
sqrdc Compute QR decomposition of a matrix.

sqrsl Use QR decomposition to solve for coordinate transformations,
projections, and least squares solutions.
sqrst Solve under-determined or over-determined least squares problems,
with a user-specified tolerance.
void sqrdc (float **x, int n, int p, float *qraux, int *jpvt,
float *work, int job);
void sqrsl (float **x, int n, int k, float *qraux,
float *y, float *qy, float *qty,
float *b, float *rsd, float *xb, int job, int *info);
void sqrst (float **x, int n, int p, float *y, float tol,
float *b, float *rsd, int *k,
int *jpvt, float *qraux, float *work);
sqrdc:
Input:
x matrix[p][n] to decompose (see notes below)
n number of rows in the matrix x
p number of columns in the matrix x
jpvt array[p] controlling the pivot columns (see notes below)
job =0 for no pivoting;
=1 for pivoting
Output:
x matrix[p][n] decomposed (see notes below)
qraux array[p] containing information required to recover the
orthogonal part of the decomposition
jpvt array[p] with jpvt[k] containing the index of the original
matrix that has been interchanged into the k-th column, if
pivoting is requested.
Workspace:
work array[p] of workspace
sqrsl:
Input:
x matrix[p][n] containing output of sqrdc.
n number of rows in the matrix xk; same as in sqrdc.
k number of columns in the matrix xk; k must not be greater
716
than MIN(n,p), where p is the same as in sqrdc.
qraux array[p] containing auxiliary output from sqrdc.
y array[n] to be manipulated by sqrsl.
job specifies what is to be computed. job has the decimal
expansion ABCDE, with the following meaning:
if A != 0, compute qy.
if B, C, D, or E != 0, compute qty.
if C != 0, compute b.
if D != 0, compute rsd.
if E != 0, compute xb.
Note that a request to compute b, rsd, or xb automatically
triggers the computation of qty, for which an array must
be provided.
Output:
qy array[n] containing Qy, if its computation has been
requested.
qty array[n] containing Q’y, if its computation has
been requested. Here Q’ denotes the transpose of Q.
b array[k] containing solution of the least squares problem:
minimize norm2(y - xk*b),
if its computation has been requested. (Note that if
pivoting was requested in sqrdc, the j-th component of
b will be associated with column jpvt[j] of the original
matrix x that was input into sqrdc.)
rsd array[n] containing the least squares residual y - xk*b,
if its computation has been requested. rsd is also the
orthogonal projection of y onto the orthogonal complement
of the column space of xk.
xb array[n] containing the least squares approximation xk*b,
if its computation has been requested. xb is also the
orthogonal projection of y onto the column space of x.
info =0 unless the computation of b has been requested and R
is exactly singular. In this case, info is the index of
the first zero diagonal element of R and b is left
unaltered.
sqrst:
Input:
x matrix[p][n] of coefficients (x is destroyed by sqrst.)
n number of rows in the matrix x (number of observations)
p number of columns in the matrix x (number of parameters)
y array[n] containing right-hand-side (observations)
717
tol relative tolerance used to determine the subset of
columns of x included in the solution. If tol is zero,
a full complement of the MIN(n,p) columns is used.
If tol is non-zero, the problem should be scaled so that
all the elements of X have roughly the same absolute
accuracy eps. Then a reasonable value for tol is roughly
eps divided by the magnitude of the largest element.
Output:
x matrix[p][n] containing output of sqrdc
b array[p] containing the solution (parameters); components
corresponding to columns not used are set to zero.
rsd array[n] of residuals y - Xb
k number of columns of x used in the solution
jpvt array[p] containing pivot information from sqrdc.
qraux array[p] containing auxiliary information from sqrdc.
Workspace:
work array[p] of workspace.
Notes:
!!! WARNING !!!
These functions have many options, not all of which have been tested!
(Dave Hale, 12/31/89)
This function was adapted from LINPACK FORTRAN. Because two-dimensional

arrays cannot be declared with variable dimensions in C, the matrix x
Elements of x are stored as follows:

x[0][0] x[1][0] x[2][0] ... x[p-1][0]
x[0][1] x[1][1] x[2][1] ... x[p-1][1]
x[0][2] x[1][2] x[2][2] ... x[p-1][2]
. .
. . .
. . .
. .
x[0][n-1] x[1][n-1] x[2][n-1] ... x[p-1][n-1]
sqrdc:
Uses Householder transformations to compute the QR decomposition of an n by p
718
matrix x. Column pivoting based on the 2-norms of the reduced columns may be
performed at the user’s option.
After decomposition, x contains in its upper triangular matrix R of the QR

decomposition. Below its diagonal x contains information from which the
orthogonal part of the decomposition can be recovered. Note that if
pivoting has been requested, the decomposition is not that of the original
matrix x but that of x with its columns permuted as described by jpvt.
The selection of pivot columns is controlled by jpvt as follows.

The k-th column x[k] of x is placed in one of three classes according to
the value of jpvt[k].
if jpvt[k] > 0, then x[k] is an initial column.
if jpvt[k] == 0, then x[k] is a free column.
if jpvt[k] < 0, then x[k] is a final column.
Before the decomposition is computed, initial columns are moved to the
beginning of the array x and final columns to the end. Both initial and
final columns are frozen in place during the computation and only free
columns are moved. At the k-th stage of the reduction, if x[k] is occupied
by a free column it is interchanged with the free column of largest reduced
norm. jpvt is not referenced if job == 0.
sqrsl:
Uses the output of sqrdc to compute coordinate transformations, projections,
and least squares solutions. For k <= MIN(n,p), let xk be the matrix
xk = (x[jpvt[0]], x[jpvt[1]], ..., x[jpvt[k-1]])
formed from columns jpvt[0], jpvt[1], ..., jpvt[k-1] of the original
n by p matrix x that was input to sqrdc. (If no pivoting was done, xk
consists of the first k columns of x in their original order.) sqrdc
produces a factored orthogonal matrix Q and an upper triangular matrix R
such that
xk = Q * (R)
(0)
This information is contained in coded form in the arrays x and qraux.
The parameters qy, qty, b, rsd, and xb are not referenced if their
computation is not requested and in this case can be replaced by NULL
pointers in the calling program. To save storage, the user may in
some cases use the same array for different parameters in the calling
sequence. A frequently occuring example is when one wishes to compute
any of b, rsd, or xb and does not need y or qty. In this case one may
equivalence y, qty, and one of b, rsd, or xb, while providing separate
arrays for anything else that is to be computed. Thus the calling
719
sequence
sqrsl(x,n,k,qraux,y,NULL,y,b,y,NULL,110,&info)
will result in the computation of b and rsd, with rsd overwriting y.
More generally, each item in the following list contains groups of
permissible equivalences for a single calling sequence.
1. (y,qty,b) (rsd) (xb) (qy)
2. (y,qty,rsd) (b) (xb) (qy)
3. (y,qty,xb) (b) (rsd) (qy)
4. (y,qy) (qty,b) (rsd) (xb)
5. (y,qy) (qty,rsd) (b) (xb)
6. (y,qy) (qty,xb) (b) (rsd)
In any group the value returned in the array allocated to the group
corresponds to the last member of the group.
sqrst:
Computes least squares solutions to the system
Xb = y
which may be either under-determined or over-determined. The user may
supply a tolerance to limit the columns of X used in computing the solution.
In effect, a set of columns with a condition number approximately bounded
by 1/tol is used, the other components of b being set to zero.
On return, the arrays x, jpvt, and qraux contain the usual output from
sqrdc, so that the QR decomposition of x with pivoting is fully available
to the user. In particular, columns jpvt[0], jpvt[1], ..., jpvt[k-1]
were used in the solution, and the condition number associated with
those columns is estimated by ABS(x[0][0]/x[k-1][k-1]).
720
STOEP - Functions to solve a symmetric Toeplitz linear system of equations
Rf=g for f
stoepd solve a symmetric Toeplitz system - doubles

stoepf solve a symmetric Toeplitz system - floats
void stoepd (int n, double r[], double g[], double f[], double a[]);
void stoepf (int n, float r[], float g[], float f[], float a[]);
Input:
n dimension of system
r array[n] of top row of Toeplitz matrix
g array[n] of right-hand-side column vector
Output:
f array[n] of solution (left-hand-side) column vector
a array[n] of solution to Ra=v (Claerbout, FGDP, p. 57)
Notes:
These routines do NOT solve the case when the main diagonal is zero, it
just silently returns.
The left column of the Toeplitz matrix is assumed to be equal to the top
row (as specified in r); i.e., the Toeplitz matrix is assumed symmetric.
721
STRSTUFF -- STRing manuplation subs
cwp_strdup - duplicate a string

strchop - chop off the tail end of a string "s" after a "," returning
the front part of "s" as "t".
Input:
char *str input string
Output:
none
Returns:
char * duplicated string
Notes:
This local definition of strdup is necessary because some systems
do not have it.
Author: John Stockwell, Spring 2000.
722
SWAPBYTE - Functions to SWAP the BYTE order of binary data
swap_short_2 swap a short integer

swap_u_short_2 swap an unsigned short integer
swap_int_4 swap a 4 byte integer
swap_u_int_4 swap an unsigned integer
swap_long_4 swap a long integer
swap_u_long_4 swap an unsigned long integer
swap_float_4 swap a float
swap_double_8 swap a double
void swap_short_2(short *tni2);
void swap_u_short_2(unsigned short *tni2);
void swap_int_4(int *tni4);
void swap_u_int_4(unsigned int *tni4);
void swap_long_4(long *tni4);
void swap_u_long_4(unsigned long *tni4);
void swap_float_4(float *tnf4);
void swap_double_8(double *tndd8);
Notes:
These routines are necessary for reversing the byte order of binary data
for transportation between big-endian and little-endian machines. Examples
of big-endian machines are IBM RS6000, SUN, NeXT. Examples of little
endian machines are PC’s and DEC.
These routines have been tested with PC data and run on PC’s running
several PC versions of UNIX, but have not been tested on DEC.
Also, the number appended to the name of the routine refers to the
number of bytes that the item is assumed to be.
Authors: Jens Hartmann, Institut fur Geophysik, Hamburg, Jun 1993

John Stockwell, CWP, Colorado School of Mines, Jan 1994
723
SYMMEIGEN - Functions solving the eigenvalue problem for symmetric matrices
eig_jacobi - find eigenvalues and corresponding eigenvectors via
the jacobi algorithm for symmetric matrices
sort_eigenvalues - sort eigenvalues and corresponding eigenvectors
in descending order
void eig_jacobi(float **a, float d[], float **v, int n);
void sort_eigenvalues(float d[], float **v, int n);
(inspired by Press et. al., 1996)
Macro used internally

define ROTATE(a,i,j,k,l) g=a[i][j];h=a[k][l];a[i][j]=g-s*(h+g*tau);\
a[k][l]=h+s*(g-h*tau);
void eig_jacobi(float **a, float d[], float **v, int n)

eig_jacobi - find eigenvalues and corresponding eigenvectors via
the jacobi algorithm for symmetric matrices
Function Prototype:
void eig_jacobi(float **a, float d[], float **v, int n);
{
int j,iq,ip,i;
float tresh,theta,tau,t,sm,s,h,g,c,*b,*z;
/* allocate space temporarily

b=alloc1float(n); b-=1;
z=alloc1float(n); z-=1;
/* initialize v to the identity matrix

for (ip=1;ip<=n;ip++) {
for (iq=1;iq<=n;iq++) v[ip][iq]=0.0;
v[ip][ip]=1.0;
}
/* initialilize to the diagonal on matrix a
b[ip]=d[ip]=a[ip][ip];
z[ip]=0.0;
}
/* main iteration loop

for (i=1;i<=50;i++) {
724
sm=0.0;
for (ip=1;ip<=n-1;ip++) {
for (iq=ip+1;iq<=n;iq++)
sm += fabs(a[ip][iq]);
}
/* normal return
if (sm == 0.0) {
z+=1; free1float(z);
b+=1; free1float(b);
return;
}
/* tresh values for first 3 sweeps and therafter
if (i < 4)
tresh=0.2*sm/(n*n);
else
tresh=0.0;
for (ip=1;ip<=n-1;ip++) {
for (iq=ip+1;iq<=n;iq++) {
g=100.0*fabs(a[ip][iq]);
if (i > 4 && (float)(fabs(d[ip])+g) == (float)fabs(d[ip])
&& (float)(fabs(d[iq])+g) == (float)fabs(d[iq]))
a[ip][iq]=0.0;
else if (fabs(a[ip][iq]) > tresh) {
h=d[iq]-d[ip];
if ((float)(fabs(h)+g) == (float)fabs(h))
t=(a[ip][iq])/h;
else {
theta=0.5*h/(a[ip][iq]);
t=1.0/(fabs(theta)+sqrt(1.0+theta*theta));
if (theta < 0.0) t = -t;
}
c=1.0/sqrt(1+t*t);
s=t*c;
tau=s/(1.0+c);
h=t*a[ip][iq];
z[ip] -= h;
z[iq] += h;
d[ip] -= h;
d[iq] += h;
a[ip][iq]=0.0;
/* Jacobi rotations
725
for (j=1;j<=ip-1;j++) {
ROTATE(a,j,ip,j,iq)
}
for (j=ip+1;j<=iq-1;j++) {
ROTATE(a,ip,j,j,iq)
}
for (j=iq+1;j<=n;j++) {
ROTATE(a,ip,j,iq,j)
}
for (j=1;j<=n;j++) {
ROTATE(v,j,ip,j,iq)
}
}
}
}
b[ip] += z[ip];
d[ip]=b[ip];
z[ip]=0.0;
}
}
/* this will not happen, hopefully
fprintf(stderr,"jacobi iteration does not converge\n");
}
void sort_eigenvalues(float d[], float **v, int n)

sort_eigenvalues - sort eigenvalues and corresponding eigenvectors
in descending order
void sort_eigenvalues(float d[], float **v, int n);
{
int k,j,i;
float p;
for (i=1;i<n;i++) {
p=d[k=i];
for (j=i+1;j<=n;j++)
if (d[j] >= p) p=d[k=j];
if (k != i) {
d[k]=d[i];
726
d[i]=p;
for (j=1;j<=n;j++) {
p=v[j][i];
v[j][i]=v[j][k];
v[j][k]=p;
}
}
}
}
727
TEMPORARY_FILENAME - Creates a file name in a user-specified directory.
Function prototype:
char *temporary_filename(char *tempfile);
Input:
tempfile pointer to directory prefix string (eg. /usr/tmp/)
Output:
tempfile pointer to filename string (eg. /usr/tmp/1206aaa)
Notes:
temporary_filename creates a file name by appending a sequence of
numbers and letters (which is created by tmpnam) to the prefix string
passed as its argument. On return the input argument points to the
(now augmented) prefix string.
It is duty of the calling program to provide room for the augmented

string. The resulting string is typically used as a name for a
temporary file; in this case it is the calling program’s job to make
sure that the supplied prefix ends with a slash.
This routine was written to supplement the ANSI C function tmpnam

which also creates a temporary filename, but within a fixed directory,
usually the /tmp directory. Unfortunately, some /tmp directories are
too small to hold typical seismic data sets, so this routine allows
the user to specify a directory with sufficient capacity. Also note
that on many systems, the tmpfile() call avoids this problem by
simulating a temporary file with a memory buffer. However, this is
not a panacea as the file size might exceed available memory and on
some systems this call does actually create a file (again, usually in
tmp).
Author: Jack K. Cohen, Colorado School of Mines, 12/12/95
728
TRIDIAGONAL - Functions to solve tridiagonal systems of equations Tu=r for u.
tridif Solve a tridiagonal system of equations (float version)

tridid Solve a tridiagonal system of equations (double version)
tripd Solve a positive definite, symmetric tridiagonal system
tripp Solve an unsymmetric tridiagonal system that uses
Gaussian elimination with partial pivoting
void tridif (int n, float a[], float b[], float c[], float r[], float u[]);
void tridid (int n, double a[], double b[], double c[], double r[], double u[]);
void tripd (float *d, float *e, float *b, int n);
void tripp(int n, float *d, float *e, float *c, float *b);
tridif, tridid:
Input:
a array[n] of lower sub-diagonal of T (a[0] ignored)
b array[n] of diagonal of T
c array[n] of upper super-diagonal of T (c[n-1] ignored)
r array[n] of right-hand-side column vector
Output:
u array[n] of solution (left-hand-side) column vector
tripd:
Input:
d array[n], the diagonal of A
e array[n], the superdiagonal of A
b array[n], the rhs column vector of Ax=b
Output:
b b is overwritten with the solution to Ax=b
tripp:
Input:
d diagonal vector of matrix
e upper-diagonal vector of matrix
c lower-diagonal vector of matrix
b right-hand vector
n dimension of matrix
Output:
b solution vector
729
Notes:
For example, a tridiagonal system of dimension 4 is specified as:
|b[0] c[0] 0 0 | |u[0]| |r[0]|

|a[1] b[1] c[1] 0 | |u[1]| = |r[1]|
| 0 a[2] b[2] c[2]| |u[2]| |r[2]|
| 0 0 a[3] b[3]| |u[3]| |r[3]|
The tridiagonal matrix is assumed to be non-singular.
tripd:
Given an n-by-n symmetric, tridiagonal, positive definite matrix A and
n-vector b, following algorithm overwrites b with the solution to Ax = b
Authors: tridif, tridid: Dave Hale, Colorado School of Mines, 10/03/89

tripd, tripp: Zhenyue Liu, Colorado School of Mines, 1992-1993
730
VANDERMONDE - Functions to solve Vandermonde system of equations Vx=b
vanded solve Vandermonde system of doubles

vandef solve Vandermonde system of floats
void vanded (int n, double v[], double b[], double x[]);
void vandef (int n, float v[], float b[], float x[]);
Input:
v array[n] of 2nd row of Vandermonde matrix (1st row is all ones)
b array[n] of right-hand-side column vector
Output:
x array[n] of solution column vector
Notes:
The arrays b and x may be equivalenced.
Reference:
Adapted from Algorithm 5.6-2 in Golub, G. H., and Van Loan, C. F., 1983,
Matrix Computations, John-Hopkins University Press.
731
WAVEFORMS Subroutines to define some wavelets for modeling of seismic
data
ricker1_wavelet Compute the time response of a source function as

a Ricker wavelet with peak frequency "fpeak" Hz.
ricker2_wavelet Compute a Ricke wavelet with a given period, amplitude

and distorsion factor
akb_wavelet Compute the time response of a source function as

a wavelet based on a wavelet used by Alford, Kelly,
and Boore.
spike_wavelet Compute the time response of a source function as

a spike.
unit_wavelet Compute the time response of a source function as

a constant unit shift.
zero_wavelet Compute the time response of a source function as

zero everywhere.
berlage_wavelet Compute the time response of a source function as a

Berlage wavelet with peak frequency "fpeak" Hz,
exponential decay factor "decay", time exponent
"tn", and initial phase angle "ipa".
gaussian_wavelet Compute the time response of a source function as a

Gaussian wavelet with peak frequency "fpeak" in Hz.
gaussderiv_wavelet Compute the time response of a source function as a

Gaussian first derivative wavelet with peak frequency "fpeak"
in Hz.
void ricker1_wavelet (int nt, float dt, float fpeak, float *wavelet);
void ricker2_wavelet (int hlw, float dt, float period, float ampl,
float distort, float *wavelet);
void akb_wavelet (int nt, float dt, float fpeak, float *wavelet);
void spike_wavelet (int nt, int tindex, float *wavelet);
void unit_wavelet (int nt, float *wavelet);
void zero_wavelet (int nt, float *wavelet);
void berlage_wavelet (int nt, float dt, float fpeak, float ampl, float tn,
732
float decay, float ipa, float *wavelet);
void gaussian_wavelet (int nt, float dt, float fpeak, float *wavelet);
void gaussderiv_wavelet (int nt, float dt, float fpeak, float *wavelet);
Authors: Tong Fei, Ken Larner

Author: Nils Maercklin, February 2007
733
XCOR - Compute z = x cross-correlated with y
xcor compute z= x cross-correlated with y
Function Prototype:
void xcor (int lx, int ifx, float *x, int ly, int ify, float *y ,
int lz, int ifz, float *z);
Input:
lx length of x array
ifx sample index of first x
x array[lx] to be cross-correlated with y
ly length of y array
ify sample index of first y
y array[ly] with which x is to be cross-correlated
lz length of z array
ifz sample index of first z
Output:
z array[lz] containing x cross-correlated with y
Notes:
See notes for convolution function conv().
The operation "x cross correlated with y" is defined to be:
ifx+lx-1
z[i] = sum x[j]*y[i+j] ; i = ifz,...,ifz+lz-1
j=ifx
This function performs cross-correlation by

(1) reversing the samples in the x array while copying
them to a temporary array, and
(2) calling function conv() with ifx set to 1-ifx-lx.
Assuming that the overhead of reversing the samples in x is negligible,
this method enables cross-correlation to be performed as efficiently as
convolution, while reducing the amount of code that must be optimized
and maintained.
734
XINDEX - determine index of x with respect to an array of x values
xindex determine index of x with respect to an array of x values
Input:
nx number of x values in array ax
ax array[nx] of monotonically increasing or decreasing x values
x the value for which index is to be determined
index index determined previously (used to begin search)
Output:
index for monotonically increasing ax values, the largest index
for which ax[index]<=x, except index=0 if ax[0]>x;
for monotonically decreasing ax values, the largest index
for which ax[index]>=x, except index=0 if ax[0]<x
Notes:
This function is designed to be particularly efficient when called
repeatedly for slightly changing x values; in such cases, the index
returned from one call should be used in the next.
735
YCLIP - Clip a function y(x) defined by linear interpolation of the
uniformly sampled values: y(fx), y(fx+dx), ..., y(fx+(nx-1)*dx).
Returns the number of samples in the clipped function.
yclip clip a function y(x) defined by linear interplolation of

uniformly sampled values
Function Prototype:
int yclip (int nx, float dx, float fx, float y[], float ymin, float ymax,
float xc[], float yc[]);
Input:
nx number of x (and y) values
fx first x
y array[nx] of uniformly sampled y(x) values
ymin minimum y value; must not be greater than ymax
ymax maximum y value; must not be less than ymin
Output:
xc array[?] of x values for clipped y(x)
yc array[?] of y values for clipped y(x)
Returned: number of samples in output arrays xc and yc
Notes:
The output arrays xc and yc should contain space 2*nx values, which
is the maximum possible number (nc) of xc and yc returned.
736
YXTOXY - Compute a regularly-sampled, monotonically increasing function x(y)
from a regularly-sampled, monotonically increasing function y(x) by
inverse linear interpolation.
yxtoxy compute a regularly sampled function x(y) from a regularly

sampled, monotonically increasing function y(x)
Function Prototype:
void yxtoxy (int nx, float dx, float fx, float y[],
int ny, float dy, float fy, float xylo, float xyhi, float x[]);
Input:
nx number of samples of y(x)
dx x sampling interval; dx>0.0 is required
fx first x
y array[nx] of y(x) values; y[0] < y[1] < ... < y[nx-1] required
ny number of samples of x(y)
dy y sampling interval; dy>0.0 is required
fy first y
xylo x value assigned to x(y) when y is less than smallest y(x)
xyhi x value assigned to x(y) when y is greater than largest y(x)
Output:
x array[ny] of x(y) values
Notes:
User must ensure that:
(1) dx>0.0 && dy>0.0
(2) y[0] < y[1] < ... < y[nx-1]
737
ZASC - routine to translate ncharacters from ebcdic to ascii
zasc - convert n characters from ebcdic to ascii format
Input:
nchar number of characters to be translated
ainput pointer to input characters
Output:
aoutput pointer to output characters
Function Prototype:
int zasc(char *ainput, char *aoutput, integer nchar);
Notes:
translated by f2c. Horribly inefficient, but little used
Author: Stew Levin of Mobil, 1997
738
ZEBC - routine to translate ncharacters from ascii to ebcdic
zebc - convert n characters from ascii to ebcdic format
Input:
nchar number of characters to be translated
ainput pointer to input characters
Output:
aoutput pointer to output characters
Function Prototype:
int zebc(char *ainput, char *aoutput, integer nchar);
Notes:
translated by f2c. Horribly inefficient, but little used
Author: Stew Levin of Mobil, 1997
739
CHECK - CHECK triangulated models
badModel bad Model flag

checkVertexUse check Vertex Use
checkEdgeUse check Edge Use
checkFace check Face
checkModel check Model
void badModel(void);
void checkVertexUse (VertexUse *vu);
void checkEdgeUse (EdgeUse *eu);
void checkFace (Face *f);
void checkModel (Model *m);
checkVertexUse:
Input:
vu Pointer to VertexUse
checkEdgeUse:
Input:
eu pointer to EdgeUse
checkFace:
Input:
f pointer to Face
checkModel:
Input:
m pointer to Model
Notes: Routines for checking triangulated models.
740
CIRCUM - define CIRCUMcircles for Delaunay triangulation
circum - compute center and radius-squared of circumcircle of 3 (x,y)

locations
circumTri - compute center and radius-squared of circumcircle of
triangular face
void circum (float x1, float y1, float x2, float y2, float x3, float y3,
float *xc, float *yc, float *rs);
void circumTri (Tri *t);
circum:
Input:
x1 x-coordinate of first point
y1 y-coordinate of first point
x2 x-coordinate of second point
y2 y-coordinate of second point
x3 x-coordinate of third point
y3 y-coordinate of third point
Output:
xc pointer to x-coordinate of center of circumcircle
yc pointer to y-coordinate of center of circumcircle
rs pointer radius^2 of circumcircle
circumTri:
Input:
t Pointer to Tri
Returns:
xc x-coordinate of circumcircle
yc y-coordinate of circumcircle
rs radius^2 of circumcircle
Author: Dave Hale, Colorado School of Mines, Fall 1990.
741
COLINEAR - determine if edges or vertecies are COLINEAR in triangulated
model
edgesColinear see whether or not two edges are colinear
vertexBetweenVertices determine whether or not a vertex is on a line
between two other vertices
int edgesColinear (Edge *e1, Edge *e2);
int vertexBetweenVertices (Vertex *v, Vertex *v1, Vertex *v2);
edgesColinear:
Input:
e1 pointer to first Edge
e2 pointer to second Edge
Returns: (int)
1 if colinear
vertexBetweenVertices:
Input:
v pointer to first Vertex in question
v1 pointer to first reference Vertex
v2 pointer to second reference Vertex
Returns: integer
1 if v=v1 or v=v2 or if v is between v1 and v2
0 otherwise
742
CREATE - create model, boundary edge triangles, edge face, edge vertex, add
a vertex
makeModel Make and return a pointer to a new model

makeBoundaryEdgeTri Create a boundary edge and triangle
makeEdgeFace Create an edge by connecting two vertices
makeEdgeVertex Create an edge connecting an existing vertex (v1) to a
new vertex
addVertexToModel Add a vertex to model, and return pointer to new vertex
insideTriInModel return pointer to triangle in model containing
specified (x,y) coordinates
Model *makeModel (float xmin, float ymin, float xmax, float ymax);
void makeBoundaryEdgeTri (Vertex *v, Edge **enew, Tri **tnew);
void makeEdgeFace (Vertex *v1, Vertex *v2, Edge **enew, Face **fnew);
Vertex* addVertexToModel (Model *m, float x, float y);
Tri* insideTriInModel (Model *m, Tri *start, float x, float y);
makeModel:
Input:
xmin minimum x-coordinate
ymin minimum y-coordinate
xmax maximum x-coordinate
ymax maximum y-coordinate
Returns: pointer to a new Model
makeBoundaryEdgeTri:
Input:
v specified boundary Vertex
Output:
enew new boundary Edge
tnew new boundary triangle
Notes:
The specified vertex and the adjacent vertices on the boundary
are assumed to be colinear. Therefore, the resulting
boundary triangle has zero area, and is intended to enable
deletion of the specified vertex from the boundary.
makeEdgeFace:
743
Input:
v1 First Vertex
v2 second Vertex
Output:
enew new Edge
fnew new Face
Notes:
The vertices must be adjacent to a single common face.
This face is closed off by the new edge, and a new edge and
a new face are made and returned.
addVertexToModel:
Input:
m model
x x-coordinate of new vertex
y y-coordinate of new vertex
Notes:
If the new vertex is close to an existing vertex, this function returns NULL.
insideTriInModel:
Input:
m Model
start triangle to look at first (NULL to begin looking anywhere)
x x-coordinate
y y-coordinate
Notes:
Points on an edge of a triangle are assumed to be inside that triangle.
An edge may be used by two triangles, so two triangles may "contain"
a point that lies on an edge. The first triangle found to contain
the specified point is returned.
744
DELETE - DELETE vertex, model, edge, or boundary edge from triangulated model
deleteVertexFromModel Delete a vertex from model

killModel Delete a model along with everything in it
killEdge Delete an edge
killBoundaryEdge Kill a boundary edge
void deleteVertexFromModel (Model *m, Vertex *v);
void killModel (Model *m);
void killEdge (Edge *e, Face **fs);
void killBoundaryEdge (Edge *e);
deleteVertexFromModel:
Input:
m pointer to Model
v pointer to Vertex to be deleted
killModel:
Input:
m pointer to Model
killEdge:
Input:
e Edge to delete
Output:
fs surviving Face
killBoundaryEdge:
Input:
e boundary Edge
Notes:
Killing a boundary edge is typically done after a new boundary vertex
is inserted on an existing boundary edge.
745
DTE - Distance to Edge
distanceToEdge - return distance to edge from specified (x,y) coordinates
Function Prototype:
float distanceToEdge (Edge *e, float x, float y);
distanceToEdge:
Input:
e edge to which distance is to be computed
x x-coordinate
y y-coordinate
746
FIXEDGES - FIX or unFIX EDGES between verticies
fixEdgeBetweenVertices Fix edge(s) between vertices, creating new

colinear edges as necessary.
unfixEdge unfix edge
fixEdgesBetweenVertices:
Input:
v1 pointer to first Vertex
v2 pointer to second Vertex
Returns:
0 if unable to fix edges
1 otherwise
unfixEdge:
Input:
e edge to be unfixed
Returns:
0 if unable to unfix edge
1 otherwise
747
INSIDE - Is a vertex or point inside a circum circle, etc. of a triangulated
model
inCircum determine whether or not a vertex is inside a circum circle

inCircumTri determine whether or not a vertex is inside a circum circle of
a triangle
in3Vertices determine whether or not a vertex is inside triangle (v1,v2,v3)
inTri determine whether or not a vertex is inside a triangle
int inCircum (float x, float y, float xc, float yc, float rs);
int inCircumTri (float x, float y, Tri *t);
int in3Vertices (float x, float y, Vertex *v1, Vertex *v2, Vertex *v3);
int inTri (float x, float y, Tri *t);
inCircum:
Input:
x x-coordinate of vertex
y y-coordinate of vertex
xc x-coordinate of center of circumcircle
yc y-coordinate of center of circumcircle
rs radius^2 of circumcircle
Returns:
1 if x,y inside of circumcircle
0 otherwise
Notes:
A vertex exactly on the edge of a circumcircle is taken as being outside
inCircumTri:
Input:
t pointer to Tri
Returns:
1 if x,y inside of circumcircle of a triangle
0 otherwise
Notes:
A vertex exactly on the edge of a circumcircle is taken as being outside
in3Vertices:
748
Input:
v1 pointer to first Vertex
v2 pointer to second Vertex
v3 pointer to third Vertex
Returns:
1 if x,y inside of v1,v2,v3
0 otherwise
Notes:
A vertex exactly on an edge of the triangle is taken as being inside
inTri:
Input:
t pointer to Tri
Returns:
1 if x,y inside a triangle
0 otherwise
Notes:
A vertex exactly on the edge of a triangle is inside
749
NEAREST - NEAREST edge or vertex in triangulated model
nearestEdgeInModel Return pointer to edge in model nearest to

specified (x,y) coordinates
nearestVertexInModel Return pointer to vertex in model nearest

to specified (x,y) coordinates
Vertex* nearestVertexInModel (Model *m, Vertex *start, float x, float y);
Edge* nearestEdgeInModel (Model *m, Edge *start, float x, float y);
nearestEdgeInModel:
Input:
m model
start edge to look at first (NULL to begin looking anywhere)
x x-coordinate
y y-coordinate
Returns: pointer to nearest Edge
nearestVertexInModel:
Input:
m model
start vertex to look at first (NULL to begin looking anywhere)
x x-coordinate
y y-coordinate
Returns: pointer to nearest Vertex
Author: Dave Hale, Colorado School of Mines, Fall 1990
750
PROJECT - project to edge in triangulated model
projectToEdge - Project point with specified (x,y) coordinates to specified

edge
Function Prototype:
void projectToEdge (Edge *e, float *x, float *y)
Input:
e edge to which point is to be projected
x x-coordinate before projection
y y-coordinate before projection
Output:
x x-coordinate after projection
y y-coordinate after projection
751
READWRITE - READ or WRITE a triangulated model
readModel Read a model in the form produced by writeModel

writeModel Write a model to a file
Model *readModel (FILE *fp);
void writeModel (Model *m, FILE *fp);
readModel:
Input:
fp file pointer to file containing model
writeModel:
Input:
m pointer to Model
fp file pointer
Author: Jack K. Cohen, Center for Wave Phenomena, 09/21/90

Modified: Dave Hale, Center for Wave Phenomena, 11/30/90
Converted representation of model from ascii to binary for speed.
Added code to read attributes.
Modified (writeModel): Craig Artley, Center for Wave Phenomena, 04/08/94
Corrected bug; previously the edgeuses and vertextuses of the
exterior face were not written.
752

Selfdocs 600dpi A4

Uploaded by

Copyright:

Available Formats

Selfdocs 600dpi A4

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Selfdocs 600dpi A4

Uploaded by

Copyright:

Available Formats

Complete Listing of CWP Free

April 10, 2007

Center for Wave Phenomena

MAKEVEL - MAKE a VELocity function v(x,y,z)

To search on a program name fragment, type:

For more information type: program_name <CR>

Items labeled with an asterisk (*) are C programs that may

ctrlstrip <dirtyfile >cleanfile

Usage: downfort < infile.f > outfile.f

Usage: fcat file1 file2 ... > file3

This program belongs to the Center for Wave Phenomena

Usage: isatty filedes

See: man isatty for further information

This program belongs to the Center for Wave Phenomena

Note: These results will be in limits.h on most systems

This program belongs to the Center for Wave Phenomena

Usage: pause [optional arguments]

Usage: upfort < infile.f > outfile.f

Reverse of: downfort

a2b <stdin >stdout outpar=/dev/null

outpar=/dev/null output parameter file, contains the

b2a <stdin >stdout

outpar=/dev/tty output parameter file, contains the

cshotplot <cshot1plot [optional parameter file]

dzdv <infile afile=afile dfile=dfile>outfile [parameters]

Author: CWP: Zhenyue Liu, 1995

farith <infile >outfile [optional parameters]

op=noop noop for out = in

The option "cartprod" (cartesian product) requires also that the

For file operations on SU format files, please use: suop, suop2

AUTHOR: Dave Hale, Colorado School of Mines, 07/07/89

ftnstrip <ftn_data >c_data

Caveat: this code assumes the conventional Fortran format of header

ftnunstrip <stdin >stdout

outpar=/dev/tty output parameter file, contains the

Notes: This program assumes that the record length is constant

grm <stdin >stdout [parameters]

If the XY calculation is bypassed and XY specified, the values

D. Palmer, "The Generalized Reciprocal Method of Seismic

h2b <stdin >stdout outpar=/dev/tty

Note: that the binary file may need to be transposed using

Note: output will be floats with the values 0-255

hti2stiff [optional parameter] (output is to outpar)

Credits: Andreas Rueger, CWP Aug 01, 1996

Reference: Andreas Rueger, P-wave reflection coefficients for

Required paramters: <none>

vp=4.5 p-wave velocity uncracked solid

AUTHOR:: Andreas Rueger, Colorado School of Mines, 10/10/96

kaperture [optional parameters] >stdout

Default case: both=0 nfreq=2

kaperture nshot=NSHOT ngeo=NGEO nstep=NSTEP |

both=1 nfreq=NFREQ > 2

When in doubt to the size of NPAIRS, redirect output of kaperture to

linrort [optional parameters]

hspace1=ISO medium type of the incidence halfspace:

medium parameters of the 2nd halfspace follow the same convention

vp2=2.5 vs2=1.2 rho2=3.0

(note you do not need "gamma2" parameter for evaluation

in the case of a_file=-1: