The Acoustics of The Violin
The Acoustics of The Violin
The Acoustics of The Violin
BY
ERIC JOHNSON
This Thesis is presented in part fulfilment of the degree of Doctor of Philosophy at the University of Salford.
Acoustics
TABLE OF CONTENTS
Chapter
1:
The Violin.
page 1 1 2 5 11
15
Introduction A Brief History of the Violin Building the Violin The Best Violins and What Makes Them Different
References
Chapter
2:
Experimental the
Methods.
16 16
Measuring
Frequency
Holography
The Green's References Chapter, 3: Function Technique
24
34 40 41 41 53
Dynamics
of
the
Bowed String.
References
Chapter 4: An Overview of Violin Design.
58
59 59 69 73 76 79 84 85 85 93 99
106
The Violin's. Design The Bridge as a Transmission Element The Function of the Soundpost and Bassbar Modelling the Helmholtz and Front Plate Modes Other Air Modes in, the Violin Cavity References Chapter 5: Modelling the Response of the Violin.
the
Model Violin
Design
Chapter
6:
Mass-
Production Techniques
Applications
Manufacturing References
I
Y., a
ii
ACKNOWLEDGEMENTS
as its
this
it
is
diff
icult' Ideas
to and
all
those so
cdnspicuous list work by is by by the the of R. time Ford times violins and have the been
the
are itself
but
means
grant
was
O'Connor, considerable suggestions. thanked ideas. for The Prof. with Jackson
give
valuable to be and
many of
always,
of-interest 'Vernon,
Valiance,
" D.
discussions A.
Finally, advised
uncounted concerning
me
on
question
photography.
To these thanks*
To those away place, from the
'and the
many others
who helped
me in
this
work
I give
my
me in
a more
personal
way
during and
four
years
States, for
a friendly kindness.
hospitable
there
no way
iii
ABSTRACT
is
highly science,
which,
quite -
aid. of
sophistication. requires
considerably to
another, of
be fashioned that
mass-produced of these
instruments instruments
improvement
features
of
the is
response
upon
which
the
developed
questions the
As it"is is the
extremely that
change
using
such
low A.
response
production by
violin,
micro-processor
machinery
concludes
work..
CHAPTER 1
THE VIOLIN
PAGE
Introduction.
Scientific uncommon today. been looked details climbing inevitably In this of at, its a
of
the
violin every
family aspect it
are of this is
by
no subject
means has
nearly advance
the
are
the
over
ridges,
problem
brings
to view it
another, is not
which
are
govern bowing
consequences which
determine to to for
problems goal is
to find those
improve students,
majority
violins, add to
goal
mass-produced price.
characteristics in the
a way which
step are
does not
this
their
most
market
obvious
component
makes
response Finally,
assembly, possibility
while of
applying
prediction
technique
a mass-poduction
situation
is
briefly
explored.
F this perhaps is approach stressed also very the too importance much. No yet of the steady-state the'transient of most
doubt the
of a violin
important,
quality
CHAPTER 1
THE VIOLIN
PAGE
is
so
low the
that, situation
even}
ignoring
large
be improved. importance of
analysis at this
be carried be addressed.
time
the
most general
should
A' Brief
History
of
the
Violin.
The origins among scholars. rebec, violins of string While rarely seen lute, viol,
of It
the
violin
have
long
been
the
subject
of
assumed to have evolved seems most likely from these that earlier
viols
developed
families
'examples
everyone It
familiar a flat
with back,
the five, of
viol
is
arched
belly
a violin. to
which-the to lead
proved of all
but
a recent
revival
medieval
to
lute of
but
applied the
a named
makers
stringed violin.
instruments)
Duiffopruggar Frenchman, is probably to
developing
became of the a
originator Vuillaume
due
violins
made
which They
were were
many
be copies
instruments
by Duiffopruggar.
modelled
on
his
viols, Acnati
exist
today
[2]. of
Gaspar Brescia
da Salo have
(1540-1609),
Andrea
and Maggini
Ch AFTER
THE
V1ULIN
MAGI:
all
been still
hivewi exist,
credit, it cau
violins to however, of
to
the
first
two
named
question created Guarnaeris, in both Cremona, valued and Perhaps or to many copy
What rapidly,
the
names school
the of
Amatis,
standing by these
who with
craftsmanship today.
musicianship their
equaled fur us
dominance
uutortuiiate
lutIhicrs
+gl,
these
Ey d ., i.. _ 'h'71
master originality
?b
violins and
r {. *_
Kh J"' yjz
experimentation,
had so the quickly violin,
which developed
u>v
".v
9"
1
IpS..
,4
were were
discarded. of the
of
exceptionsC(k f: ze School
r; 4r, . t. '
led
by
Jacob
(1621-1683), instruments beauty and and violins universally for man y years. tonethe of in but old were
L'} Y
both these
form
Figure
1.1:
Modern (right)
copies and
of
violins (left).
by
Stainer t-holes.
differently
shaped
CHAPTER
THE VIOLIN
PAGE
Before
investigation
begining
of of one
the
the the should
` '`r
J
sGw
acoustics violin
Al to
spire
look
a few
at
moments
t 114"
two
r
:+14
! `.
ifs
=
t3
rte,
,' ' 1
violins
in f] Gures
which
1.1
appear
and
1.2,
which
few
to
moments
apprec
in
Late ly
the made
beauty violin.
of It
a we li is to , f <t5"i,
`"t
"'t
VL
depth the
careful or the
Figure (left)
1.2: is
The much
arching more
of
the than
Stainer that of
copy the
abrupt
Stradavarius
copy
(right).
the
the of is
acoustic old
properties, Cremonese
genuine their
violins
demand.
CHAPTER 1
THE VIOLIN
PAGE
Building
the
Violin.
1.3 Most
illustrates of the to
which blocks, to
are
used
present
structural
strength
FRONT PLATE
F-HOLE% BASS-BAR
F-HOLE
... fl,.
rTf. n.
BACK PLATE
FIGURE 1.3a" Exploded view of a violin.
BRIDGE
BASS-BAR SOUND-POST
view of a violin.
CHAPTER
THE
VIOLIN
PAGE
sound-box. fase 1.nating The sycamore easily of such it the from the as
method
of
constrict
ion
of
the
box
is
itself
begins
of
ribs
frum that to an
which are
I mm. so of a mould
the glued.
inside
blocks form to
to of
and
pieces
remain
firmly
glued
the
removed
str1p5 they to
ions, of the
have are
form known
been cut
the as
into
six bouts,
pieces
sect
tf
4'
a rt
with { C-shaped lutlifer s of then it wood very Alq" over rounded until the snugly Pure are luthier slab. 1.4_ held on traces A nearly an their inside complete mould outline onto set of ribs the a T "r
fits the
while
mould. peices
in the coat I- lie luthier tiro
a maple
remaining
then way. not bent Next to
rubs blocks, so
the that
mould the
wi th glue will.
soap, not
being to
very it.
stick
glue to
is the
used glue
to
join
the
bouts
the the
blocks pieces
joints ten
glue
hardened, '
changes
The
strength
ribs the spruce, of of This are surface
may occur
constructed of the
a much longer
that This and however, block, is copy 7 the is
period.
latter project so' that mayare cut to the be so 30
above of
to from the
the
inside
32 mm. at
taper in
great
violins perform
luthiers, task
instruments
this
the
lining for
strips, the
which plates,
are have
to
provide into
wider place,
surface reduced
and 'glued
section.
wood is secure
impossible linings
the of
notches
in
and corners
at
two
step
in on the
Much
the this
carved
the
appearance projection
small
the back of
tongue, and
sycamore
the
spruce or then
plate
piece for
wood,
pieces.
a plate edges
adjacent, then
slabs In this
planed appears
are
way
the
wood
nearly
symmetric
the
ciLA.P1'I: I( I
THE VIOLIN
PAGE
plate also
is
This the of
center. of the
ensures
iY
narrowest grain A
t
the
is
`.
14
^l 711
dusircd
next
plate
made on
outline
the slat),
is
Y"tsb. vftr"' + W1
which saw,
is knife,
then
by
r
'^t' I I,
t'lil.'i1
l ruilcI s
wi 01
tu
the
tlhe
slab
working
t skill
curved is
r.,
needed
YR
here
it
for
single
carrying
large the the has cut the to the These TI [e been and widest remove edges two wide
away F.i ; ure luthier 1.5: to Template8 check the are arching used of the by the splinter work.
spoil
plates.
lengthwise cut again portion the in gouge which to match matched of the marks the bring the arching of the three violin's a template, to lower a the template. and to l i ng upper a arch between After bouts, a the doing small
and prrrf.
sink will
shallow eventual
groove ly be
cut
set.
important luthier
and thick,
edge
without
flawing
of strips
plate. is
consisting emphasizes
wood, More
shape.
importantly
CHAPTER 1
THE VIOLIN
PAGE
from
propogting
from
the are
plate's used
'edge to complete
into the
the
vibrating
region. surface of
and glass-paper
outside
to
check
the
is
next
which
centre
thickness,
4mm. to must
finally
the
cut the so
bass to must
bar cut be
knife be place
must in
slightly done, of
'plate.
This
belly holes
placed
unlined into
later stuck.
small
the
drilled
through
groove
the
These'
and are serve
located
locating from the
where points
purfling which
into
small
pegs
The
ribs
removed
mould,
glued this violin is an of
the
on with the
other
the is
set
pegs attached,
of
in
rib
place the
liners
to
glued
keep the cut,
in
place,
from edges
and then
warping. rounded, of
the
belly
After
back is
and the
body little
fashioning the
it
provides in carving
craftsman the
opportunity it is
demonstrate no consequence
beautiful '
scroll,
of
violin's
acoustics.
provides a well
the
luthier
with oil-varnish a
another
chance
to
show
for
applied it
use.
enhances
besides
providing
of they
with
Spirit
protective
varnishes but the are these claims
coating
also lack of used the
qualities
varnishes.
countless
people
the
varnish
used
by
the
Italian
masters
was not
the
secret
of
their
CHAPTER
Tlll:
VLOLIN
PAGE
10
success, its
for
the
varnish are of
does minor
the
violin, when
damping,
frequency
response
the
plates
many
violintap-tones
the best [ruin When
..
r 4 rar ., Ir
employ
obtain sound instruments.
to
help
fl
'r
making k'.
-A f
final plate a
they loosely
{Late
thumb and
between
about
finger,
C 1?
\V iP
4/5 plate
of
the length,
the it
sharply
as:
I
wi th
demonstrated
knuckle
in
*, eFigure to 1.6: Tap for by tuning pitch. tones th& make it I possible of roson-
This of
by carving U. 1 mm. of few Of may until the square the be the front
compensate plates to
wooden ance
a specific
Many of good is
This use
tones
perhaps n thud
important applied. It
aspect is
of by no
making
which
scientific
has
CHAPTER 1
THE VIOLIN
PAGE
11
only
one.
Most chapters,
of
the but
important before it is
work
in
this to
field build
is on to
possible it is
mass-produced that
instruments
necessary
determine
the
universally
admired.
The Best
Violins
One of
the old Italian
the
first
violins
attempts
was of, over arbitrarily the relative
to
scientifically
by F. Saunders instruments, response as of was in
explain
[5].
the
esteem
frequency
of
made many
-fine the
in
frequency ranges
which
a the
assisted
instruments
blind
conclusions prime
importance - an
power, of
the-lowest he makes
of
Another
interest:
"Many violins, including two three the seven of or _, Strads for which we have curves... show a weakness in the range 1300 to 1800 or 2000 cps., to a drop of 4 to amounting 8 dB. This to have no important appears... effect on the reputation of the violins concerned. "
Subsequent Other frequency who used research at have bands, spectral has revealed the notably [7], These the irony of by this a statement. small and the shared number Meyer of [6], frequency the same
violin
response
experiments
CHAPTER 1
THE VIOLIN
PAGE
12
as did features
tests:
the
too in
and This
which to
many quality is of
shown the
feature of
the,
determining
researchers postulated is the, the there Recent fast, bowed
a violin!
question
early Meinel
response an
have
output
for
sound of the
concentrated
importance
response. used tried use to third to describe [6]. were All and were only used octave correlate their In order bands some instruments to to make describe to of record the with these the response of terms many which
response as nine in a
required of the
those in
subjective of
every-judge which
the
and
subjective
appears
Sc most important from conclusions about violin and used frequency Jansson to rate in the response which qualities are "Long of
arising averaged
a study
spectra"
CHAPTER 1
THE VIOLIN
PAGE
13
Experimenter Saunders
Bands 5
Frequency 196-
range Arbitrarily
Remarks selected.
Lottermoser
et
al
Chosen to correspond to vowel formants of U, 0, A, E, I, the S, and the second nasibilant sal formant. 194-12K 1809K Spectral 1/3 fifths. bands. frequency researchers scale. to
Octave
[9]. this violin bands process as from whole loudly about the tone scales were played over three octaves in of
By in
comparing a
LTAS of
twenty-two ,
which
of to
the be
frequency found
bands by their of
which Hz., a of low rapid the
to
those
predecessors,
accuracy
most high
a definate
characteristics up and the of of to
order
importance.
appeared response drop in their 1.3 4kHz. by only tests kHz., The two but
500
around above
then
judging
violins group
was
done of
have are
been
violins,
findings
still
are
and the
the bridge
work is
may
progress.
It
will for
demonstrated
primarally
responsible
CHAPTER 1
THE VIOLIN
PAGE
14
Subjective
term
Characteristic
Bands at 250,500,800,
frequency
response
Soprano
and 1250 show strong the 1250 Hz. band dominant. with fall in this category. violins
High
output
between in
Body resonance
Frequencies ritating tone from
1.2
to
2KRz.
cause
this
ir-
quality.
Tight,
thin
may cause
below
terms these
and the
the the
production lowest of
results. come to
of
sound
at
frequencies ranges
information
so
it
is the
the
frequency
As more
yield
its nor
about will
light,
understanding this
dull
mystery,
which
envelope
amazing
man's
genius
400 years
ago.
CHAPTER 1
THE VIOLIN
PAGE
15
[1] (1972).
Sheila
Nelson,
The
Violin
and Viola,
Ernest
Bean Ltd.,
London,
the
Violin,
reprinted
in
[3] "The John Schelleng, acoustical JASA, vol. 44, pp. - 1175- 1183, (1968). [4] Woodhouse, "The M. McIntyre and'J. instruments", Interrdisciplinary Science (1978). [5] violin "The Saunders, F. family", JASA, vol. mechanical 17, p. 169,
effects
of
violin
varnish",
acoustics Reviews,
of stringed 3, vol.
musical 2, no.
action (1946).
of
instruments
of
the
[6] "Resonanzen W. Lottermoser E. Meyer, and von Gergendecker 13, no. 7, pp. 185-189, (1959). Boden", Z. Instrumentebau, vol. [7] reported [8] Soviet [9], rated (1979). [10] scientific (1957). H. Meinel, in Soviet "Scientific Physics: principles Acoustics, vol. of violin construction", (1960). 6, pp. 149-161, quality",
und
H.
"Regarding Meinel, quality the, sound ofI -violins basis for violin JASA, vol. 29, construction", pp.
and a 817-822,
CHAPTER 2
EXPERIMENTAL
AND THEORETICAL
METHODS
PAGE
16
Measuring
the
Frequency
Response.
the
frequency the
response
of
can
be
Ideally the
way in
which
held. violin
experimenters
or wheel, is hand It limit by hand. is difficult. and is in possible the string a the bowing
some form
bowing,
bowed excitation,
a swinging the pendulum
either
or of a
construction requires
such a
only
first
force
'A raucous, avoids. content the Saunders, the violin's' results curves", " may as and bowingwho
produced
the
although on all
of the of
response, be
to
such
reproduced
response used in
made
known,
have I
been
pitch, form of
than
the
frequency is used
is the for
essentially
one
a violin
CHAPTER 2
PAGE
17
each the
pitch loudness, in
is
of a
which
contributes,
to are
useful
many
much, the
frequency of
information indiviual
be This plotted spacing filter
making modes.
at
resonant loudness
frequency curve
"equal
usually even
" must
apart. of. a such as
discreet is to wide
with
frequency in
details resonance
occurs are they action. of the have way, testing. frequency heavily The also but
maybe:
details violin
probably may
of
to
insights
violin's
bridge in
problems. the
measuring the
response removed when the response for when order body driving
throughout This is
strings
stringstill
strings curve.
are
damped of
presence obscures
strings a large
and portion
bridge, of
them,
making to
holograms. the violin's response, transducer, with had a high a very magnetic minor. The effect, beset mounting by on force was applied MM0002, to the
measure
by an
B&K type
small only
resonance had
transducer the.
the
itself, workers to
curve. prompted
have standardize
similar techniques
which Figure
an effort
CHAPTER 2
PAGE
18
.,
degree
to which may be
(a)
affected. these 10 dB
In
response
was
mounted different
on
foam
supported
transducer
With the
-mounting.
first two
on threads at the corners, (b)when resting on a foam block, and (c) when lightly clamped.
methods was
the
fixed
adjustable The-measurements'-took change only in the either form the of place temperature, mounting the curves one after pressure, which display.
curves is the second by the resonance interaction
the
other
with
or humidity, responsible
it
is large
is
discrepencies
In form a two of
which
these
appears of
double and to to be
peak the
support impossible
was
clamped, frequency
influencing
violin's
response,
figures
demonstrate.
CHAPTER 2
EXPERIMENTAI,
AND THEORETICAL
METHODS
PAGE
19
1.37
467
497
FREQUENCY in Hz.
I IGURE . the frunt pinto res, unnn, o and that a poorly dethe effect mount ran r igrxxd trunsducor have on 'he frequency rcw, ponso kurve b) 2.3 TIw req un nrar
Figure mounted
2.2:
The
vi
H1in for
front driving;
plate it.
and
the
transducer
The its of as
clamping
of
the
within corners
the
could
also
affect
drastically Light
altered clamping
shown
rnounting as it
the proved
violin to the
advantage to and this
and be
used of the
icon bttwoen
addi easily tiona1
transducer
in the that
assembly laboratory
transferred violin
holography to the
both
the
transducer system
anti-vibration its
frcciIity.
overall
demonstrated
CHAPTER 2
PAGE
20
over methods
the of
testing. The remainder equipment recording response in figure the curves 2.5. . used of the in
All
electronic
260 437 497 Frequency in the and lower 600 Hz. anechoi shielded situated
equipment
outside chamber cables
through noise range floor of was interest Pa. -aregenerator throughout. typically when used-5 measured throughout to drive to -10 on
the
walls. the
frequency could
frequency-
sinusoid
high-sensitivity the
microphone,
I %Y, type
4165,
was
used,,
to
output. was
in a B&K 2610
V---VV
B&K mm0002 Q^
o00
&K
Sine 1
Magnetic
Transducer. & Kze, Microphone.
Random Generator.
B&K 2131 Narrow
Band Filter.
amp-
Figure
2.5:
The
equipment
used
for
making
lifier
or
a B&K
response
measurements.
CHAPTER 2
PAGE
21
2131 narrow-band A white response displaying to 600 Hz. and the the
-" record for required 1 kHz. a.. frequency averaging from and 190
noise. the
signal. analyzer
mode.
frequency Hz.
band-width. the 0
when
measured, is noise
selected.
instrument
sensitive often
virtually
. response store it in
the
noise
subsequent
mesrement
within
machine. of the violin was often measured sources set long of at of less noise doors than could which time over was
certain the
anechoic
averaging
therefore five
used
2048
samples
about
minutes
being
during digital
as linear
it
small
recorded whole
Fourier
any frequency in
window begining
component
signal can
have a
the
number
transform. the
wavelengths
A Nanning of the
the
uses and
to thereby curve
recorded
possible
to
output-
it
to a level
recorder
or an x-y
plotter.
CHAPTER 2
PAGE
22
useful
feature of
of
the
a number interuption.
samples, This
easier
measurements. in the it experimental should be work made or a the in method of measuring or the
whether
be, used, of of
reverberent Applied
Although octave
resonances in the
they
do.. not
frequency meaningful
by a violin
impossible
to obtain
90
co ti c80
J
a N
c
1 1'
200
300
400
500
600 in
700 Hz.
800
900
Frequency 2. b: The response of room to a white reveberant 2.5 Hz. bands. noise
of
tests at were
were
made
to
see as distance
if low
they. as-' of
microphone
one
from
the
center
of
the
belly,
in
the
hemisphere
into
which
it
would
CHAPTER 2
PAGE
23
radiate
if
A frequency position, of
Was axis
then
normal 2.7
center" that, at
Comparison
figure
shows taken
although high
directionality in the
greatly range
frequencies,
shape of
sound
does
not-greatly of this
experimental frequency
confirmation response
measurements
40 A It It
1
30
1 d1
1 1lit
1tq V1!
t 1
AI
a`, 0
20 1%V 1
c 10
1230 Frequency in kHz. Figure 2.7: Sound- power (--) and ( position a single at response
Once a frequency frequencies output of and the response curve had been were
frequency ).
recorded, using the the resonant sinusoidal counter, which Noise was 10 dB
half-power
points
determined amplifier,
generator
and the
measuring
frequenc}4 filter
meter. typically
much grtater
instruments
the
CHAPTER 2
PAGE
24
as this
equipment
was
only
used
to
measure
resonances
it
no difficulties. of the error resonance involved has a slope to obtain generator 3-dB-down for from frequencies mostly were due accurate, to the to within in peak. for dB the while of
difficulty at the
zero
much, of to the
accurately as one
sine-random the
varied points.
locate
a large
number
produced Results
figures obtained
Q accurate the
to about
techniques with no
were
although
violin the
made frequency
to
and rapid it to
modes
measurement
holography vibration.
move
individual
the
of
laboratory"
behavior
Holography
has stringed
proven
to
for
modelling shapes be
mode could
although vibration
was
modes
Jansson [5].
established It is
this described
acoustician
below.
CHAPTER 2
PAGE
25
Although as 1947, it
holographic not It
process
as laser
early had a
been
developed single
special of
of a laser proved
beam,
coherent, breakthrough
There
frequency
beam
to be the
that
are-many
made holography
types of laser
possible.
in use, but the most common is the
gas 633
type, nm.
which An of
produces
a by
electrons fall
to
normal
release neon.
characteristic range.
633 neon
excited
these
atom, to
-a
sustain
to maintain
a large
proportion
excited
discharge
tube so nm. must will in
tube.
which this light occurs travelling wave. uniform to has mirrors along A phase escape at the each axis portion non-diverging a partially with end a of
that form
a standing be of is allowed
necessarily wave,
standing
silvered
allowed
mirror.
to decline
If,
however,
below' a
the
number
of
photons
the
in
the
cannot
tube
sustain
is
-certain
level
process
lasing
action
This of of
limits that in
a photon
a high-level nameLight
by another by Stimulated
Amplification
Emission
Radiation.
CHAPTER 2
PAGE
26
is
a type
of
photographic of light
in'
both
the
phase
information with
light
information
about
recorded
on a photographic
Object
once more into reconstruct will and image the appear retains may be from
Object Beam
-'-
the
original
Reference Beam Prism emulsion
in
three
dimensions
angles.
Real Image
The theory
the formation may
which describes
and reconstruction be to simply, figures
both
of
holograms with
Figure 2.9: The reconstruction process.
reference
2.9.
The light distribution be referred upon the
photographic object
emulsion
from
the as
object
(this
will
to as the
(2.1)
S(r)= S, (?)exp[jwt+
j(f))
where plate,
phase
r is
the
position S0(i)
surface
vector
xi
+ yj
and- lies
in
the the
plane'
of
the and
and where
along the
and exp[jwt
of an
+ j$(=)J
represent
thin
amplitude
infinitesimally
emulsion.
The reference
beam, which
is,
simply
diverged
and
collimated
CHAPTER 2
EXPERIMENTAL
AND THEORETICAL
METHODS
PAGE
27
laser
beam
=which
retains
its
coherence,
may also
be represented
in
the
general
form
of
equation
(2.1).
This
gives
(2.2)
where
ej
accounts
the
path,
produced
The photographic which over infinity, irradiates a large this it, number gives
so the
[S(r)+R(r)]' number
be integrated approaches
periods.
(2.3)
where
a coinplex
conjugate. fringes of will light, A appear, will. properly proportional of the two depend whose upon
the'emulsion
intensity
exposed. is directly
developed to the
interference
beams
(2.4)
with into
when
plate
is
object
information, image
nor
R*S(r),
occur term
three-dimensional
formed.
CHAPTER 2
PAGE
28
is but
the with
original
object
beam multiplied -$ (r), image point. which and The object light is with and
screen includes
focal
(r) ] + 4D the
be
viewed
may be focussed
a camera. which of are produced onthe platein the are ve ryclose of If, mean, the object A its a
and bright
a line of
greatest
amount as
time 2.10,
two extremities density contain object, extreme greater consequently function information but
figure
the The
as the
than on two
at
form
fringes
the
predominant
YAt -A
on reconstruction. For images, the and moment only their will be images object) these interaction considered. at a have point two main upon If
Figure t
+ y3 on a planer
CHAPTER
EXF'ERLMENTAL
AND
THEORETICAL
METHODS
PAGE
29
different -A and +A
displacements n which destrucwi 1 spot. the a li Ilk Table difference 1: Roots between of J0(k) subsequent and roots. the are 1 2.4048 5.5201 8.6537 11.79I5 14.9309 18.0711 21.2116 2 3 4 5 6 7 2.4048 3.1153 3.1336 3.1 378 3.1394 3.1402 3.1405 k,, roots of . lo (k) k,, - kr, _,
Frence dark on
on
reconwith displacea
/2cos
or In frin-
so-on. dark
areas
where
interferThese
t t' f`t
fringes with
points peak U,
displaceetc.
F w
Xcosm,
This case in
is
exactly
the
a
F
t F
distinct formed to
2.1 I:
plate'
A
s
vi hr; it il)n
Lowest no(Ie.
ho Iog, ram
of
are
produce
CHAPTER 2
PAGE
30
hologram a change
The hologram the
with in
fringe
of
equal .
of
contrast
and representing
X/2cos$ of
series way in not
images the
in
the appear,
which be
fringes
analysis to
included the
suffice peak
displacements Bessel
k, X/ncosO
, with J0(k),
successive appear
zero-order
function,
is the
very
nearly
the of
same as first
of table
double clearly
exception with
the
changes
increasing as
and 2.11.
figure
contrast to remove in
by employing
a stroboscopic
used
research. While vibration it is a deterministic of necessary may be any to element isolate vibration in the by the may produce equipment will holograms, degrade it. from which it - a random
which setting
up the the
laboratory step
used provide
throughout such
important
vibration
inner table,
tubes
were
inflated
and a
slate Its
bed, mass of
improved were
made 25U
when placed
an
gave
was
isolation,
placed
upon these.
was of further kg.,
excellent
vibration with
which miss
optical
benches,
a combined
13U
upon
the
table
to
provide
a firm
mounting
system
for
the
equipment. analyzer
A were
B&K type
8306 accelerometer
and the
2131 narrow
band
CHAPTER 2
PAGE
31
"to
measure
40 30 Figure 1: The vibration isolation provided by the anti-vibration table in the holography laboratory. ------------------
on 20
10
laboratory, on the
suflow enerfloor,
Frequency
in Hz.
be caused about
by lorries on piles of
passing carpet
membersisolation
leapt
was measured
appears with
in
figure
A Scientifica 10 mW. holograms was used. 'z laser The its parts to and a
laser, nm.
a continuous
which
although
mW. model
power of
supply table.
which,
due
to
vibrations
from using
the
A shutter
camera,
mounted
The beam was'split which moving strengths surface, was it specially across of -should the the the
This
two
beams of
which, between
have
a ratio
10: 1, 1: 1 ratio
stronger. emulsion as it
expect
as the levels
respond levels,
logarithmically a constant
does at'higher
intensity'
CHAPTER 2
PAGE
32
to
be
-present a ratio
to of
prevent about
complete ideal
destructive [7].
interference.
In
practice
4: 1 is
Two
Egure 2.13: The holographic apparatus. Abbreviations: laser M. mirrors (M), beam splitter (B). violin (V), diverging lens(D). holographic plate (H), and shutter (S).
mi r-
rors, (25x
microscope jectives),
simple holder
oband a
plate
completed initial
Oscillator
Power supply
the equipment, which is shown in figure 2.13. both it and spatial a lens noise
to mount
Anotherthe
plate
holder
easilyAnother of-
Previously addition
trivial is
in the
laser
to keep of
two
attached a piece
held room this to and
violin
to
compare
already
stated,
the to
to
adjust the
at
conditions.
Oscillator ready
Huirhead
resonance,
Any stray
light
which
could
reach
the
plate,
would
of,
course
CHAPTER
EXPERIMENTAL
AND
T11EORKT1. CAL
METHODS
Y,A(; f;
33
it
and the
degrade from
the stray
were a and
on
plate both
to
form
a canopy
apparatus
windows.
and and, electronic after 10E7'5 image in a plates, with ,ah: ith a of five stop-watch series ten to minute Kudrit of fit were trial teen deve iy used to control the with was time ion And In found Kodak final,
produce doveloppr.
good Fixing
Kodak
r, "tut
Figure
L. 14_
The
I() 1 graph
is
apparatus.
rinse the
in
water
with
a When
few
drops
of the to
wetting and
agent Kodak At
using; f ive tu
one plates
point but
was image
bleach
was
urnot
icahle.
of object
the
image beam
was with
usual the
ly
clone
by holding curtain
the obscuring
plate the
blackout
CHAPTER 2
PAGE
34
any
hologram
of to
interest produce
required, which
a may
throughout
The
Green's
Fuction
Technique.
Function
Technique; which
in
which
a system
is
treated is
as very
functions the
together of
vibrations to a, point
a violin. it
determining to find
a the This
system's from
response
input
input
simply
exactly
''the
Fourier in be
impulse there
In the will
response is
system cannot
domain,
treated!
but
no reason
domain
possibilies,
similarly this
remarks be applied
which'follow, torvibrating
which later
technique
present
systems
and
used
to model
the'violin. of
+oo
The' convolution
(2.5)
g(l)
f f('C)h(t-T)dT'= -470
output in the time
f(t) 9 h(t) ,
domain The the of a linear filter response impedance. theorem is in terms be in
which of the
the
impulse
impulse system
may Then, e
convolution
used. - This
(2.6) '
G(w) = F(w)H(w)
r
where
H(w) =g:[h(t)},
etc.
CHAPTER 2
PAGE
35
is
reduced is easily
use of any
the
transform An
output this
input degree
example which
of
single
system
appears
is
2.15.
f,
The equation
with the
of motion
functional
mg + cg
dropped. fa 6(t)
To
find and
the gs h(t)
(2.5).
by exp(-jwt) t
H(w)-
1/(k In
the
impulse
response. like this one, only the time system of both of "r . and
an example
lumped is
and frequency the time the t, is problem and input g(r the t). , is
a continuous
position, point
t),
system. system
To find
the
to a point the is
input
, which
Green's
function, of
requires a continuous
virtually system
as before. form
The equation
of motion
(2.7)
g(r, t) +A
function here
ZO-
t) t) Bf(ro, g(r, =
f("r0 t), , t) which the Green's with function the is
a delta
for
by g("ro
conforms
symbols
system separable in
one into
already that
equation then
ordinary
differentials.
For
a simple
would
to
relate
"ro,
r,
to,
and t,
by using
the
Fourier
transform
in
both
CHAPTER 2
PAGE
36
time
and
space.
The
violin to complex
is find
system,
which
with
systems
is
much easier
calculate to determine
response. but in
course,
are of
instance
no
concern.
of
is, is
for
most
written as
expanded
a This
together g(rt)=exp(jwt)
Fourier
terms
a1 complex.
(2.7)
may then
(z. s
(jw)ma
+A (r)
terms
r" () la 4 a1 = Bf(r)5(r-
o)
j
and
with
the
time terms
dependant in equation
eliminated. to yield
Next
let
,;
combine
(2.8)
(2.9)
Each side
properties gives
dv=0 if
This
CHAPTER 2
PAGE
37
(2.11)
g(o,tl=
A-1a fv $Z(r)dV.
aiiejwl =B
where
only
to
impose to
the the be
to
response
system it
will
any a may
frequency
wave with
complex number
and phase. in
be an infinite"
of waves
travelling
one direction:
(2.12)
Sf (x, t)=Y(acos(wt+kx+4)+bsin(wt+kx
+ -(WI
Rewriting
this
using leaves
Euler's
equation
and
factoring
out
the
time
dependenttterm
(2.13)
27
n
JWI1
jikx"
e)
bnej(kx+ehlJ+ejwtfln +
j(kx+e,
)-
b, 6j(kx+$^
our
the this
exp(jkx) becomes
and exp(-jkx)
terms,
and redefining
as
(2.14)
Qel(wt"kx)
bej(wt"kx) +
or
acos(wt+kx+4)
with
the
proper
choice in
of
phase 2.16,
$. driven of its
f(t)
figure
a point
one third
Figure 2.16: The amplitude of vibration of a stringy driven at L/3 with w.: 1.5irc/L
equation
of
motion
for
string
r,
CHAPTER 2
EXPERIMENTAL
AND THEORETICAL
METHODS
PAGE
38
(2.15)
at
with f unction
the
mass
per
unit
length then
and
T the
tension.
equationof
motion-may
be written
directly
(2.12)
as
(2,16)
g(x, t)(xl
n; EC 6A
orthogonal
c2=T/ e.
setare
of
which
fit
the
boundary when
f=(ni/L) into
series
produced that x-L 4L/3. but have the held there the
ends of Zj
the
by
and ZZ,
then
boundary With
produced of
center
may be written
(2.17a)
and (2.17b) --
F(-L/2) Z1
T_ (-L/2)
=jwO-L/2)
jwO(L/2) 2 &ven
Now let
4; =sin
where this
the into
in/2 the
term
allows
both
and
odd symmetry,
boundary
conditions
to give
(2.18a)
gcos(-qL/2+h1/2)=LT
sin(-qL/2
+ri/2)
CHAPTER 2
EXPERIMENTAL
AND THEORETICAL
METHODS
PAGE
39
and
(2.18b)
(2.19)
which,
(2.20)
rL lW
1 -1T JT
z, +z2) Z1 Z2
1 J
of ,a compliant A spring-like
termination
for
the at
string
can
easily
be
makes are on q.
denominator
equation easily
More complicated
impedances
this of the
type violin is
of
of in
analysis chapter to 5.
it
possible forces
of is
to that
improve are
a function
system to
= it
form mode use in
possible
periodic of
predict
a continuous impossible to
excitation. an object of
it be
calculate
shapes the
determined
experimentally
equations
CHAPTER 2
EXPERIMENTAL
AND THEORETICAL
METHODS
PAGE
40
[1] strings",
C.
"On the Raman, mechanical Bull., Assoc. Cult. Indian Saunders, in the R. study Watson, and violins", of
theory of vibrations of 15, pp. 1- (1918). vol. W. Cunningham, 12, pp. JASA, vol.
bowed
Hutchins, the
and L. [4] W. Reinicke Cremer, interferometry to the vibrations of JASA, vol. 48, pp. 988- 992, (1970). [5]
"Application bodies of
of string
holographic instruments",
"Resonances E. H. Sundin, Jansson, Malin, N. -E. and of a body studied interferometry by hologram violin and acoustical methods", Physica 243- 256, (1970)., Scripta, 2,. pp. vol. [6] of [7] D. Gabor, "Microscopy the Royal Society, series A. by reconstructed wave-fronts", 197, pp. 454- 487, A, vol. Prodeedings (1949). to 954-
"An Porter S. George, and holography", American Journallof practical 959, (1975).
3
.,
My thanks
to Roger
Darlington Theoretical
this
McGraw-Hill
CHAPTER 3
PAGE
41
would At
the it
motion appears of
of
the merely
bowed string to be a in
first
string,
sort
problem that
encountered the
equations. static it is
would is
appear
overcome,
leaving bow.
string what
by the
bow pressure normal increase overcome dependent truel produce and has One force in
string
players is a the
larger period
-would
a conclusion
watch
on the
bow to
speed,
some dynamic
been ignored.
to
was the
first
to
shed light
on this of
an optical string
to study
the motion
experiments
C4
concluded
of
any point
time
os
portions
the end, string wave by two
related
t t
length. whole
represented
very
3.1: Helmholtz motion of a bowed striAg, ob served at a point xe. The string flies back during the interval if and is captured by the bow over tc.
CHAPTER 3
PAGE
42
lines which
with travels
of
slope on a
envelope is bowed
when the
string
The slope velocity Eigute 3.2; The slope dispictured above continuity describes a parabolic curve as It travels around the bowed string. discontinuities the string.
from
the
bow
travels
phase, It then termination when then it
the
acceleration
towards the
passes along the end the of the by are, the and
discontinuity
bridge, the string, the bow. the of reflects point, from and or for from it initiates the nut with nearly sticking. (the string's opposite
and
as
it
bowing
continues at again
slipping the to
bow is
travel model. in
string. effects
course,
then
so much
depends
on subtleties
be ignored. is the each easy to see how a musician model is in of string can control The amplitude. the amplitude between are not, the force dynamic at the level bridge in
motion. to its
speed distance
interval motion
and so
release. easily
Some described
factors the
Helmholtz C. V.
fifty
later, He used
a detailed describe
of
velocity
CHAPTER 3
PAGE
43
into
far two
greater state
his
predecessor. of sliding
By
(a coefficient velocity
acting a
when there of
between are
two not
coefficient equal
static
and assuming
firmly
representation
of Helmholtz
motion.
(3.1)
,.
and Ttension. even those with
modes,
acting
string (3.1)
equation with
taken,
amplitudes of forces
spectrum position..
violin
body is
this
as
requires
the many . at to
be "richer",
would Clearly of
produced something
the
stroke.
be accounted
Helmholtz
the
bowed string. Arthur non-linear Benade and this suggests lead that seems a the bowing one process [3j. He is
highly that
good
states
CHAPTER 3
PACE
44
the the
existance bowing
of point Losses
in
the
string not
motion have
at been
should
when bowing
at
a node.
and in
-8-3. W1ULll t. C .. VL
tl. 6S4c e
x,l tt
the
help the
to
excite
zjtt)
modes, the
non-linearity is for in
bowing
process responsible
which
the as
studied the
principles in with force like thoseor in very figure small, 3.4. then this
constant NM(v),
speed whose
positive ,U1-0.5
[exp(-v/10) friction as
- exp(-v/100)] [4].
The equation
system
(3.2)
where series the 'Sform.
NB
for expansions
expanded to
into change
possible of equation
right-hand
side
CHAPTER 3
PAGE
45
(3.3)
1P6-'v CV,
O`yl1
NB +
then These k(t)
all k
is
of
as being right
a hand into
frequency of equation
term, (3.3).
powers powers
equal the
on the
frequencies
solution.
equation
(t) must arise.
so that
include With
must
of (t)
Obviously
frequency
combinations
represented
as. a Fourier
Series,
+00
(3.4)
X (t)
QrelWrt 7=-0o
so that equation
frequencies
it
on any periodic be
in
form although
ways.
at it
all,
the
power
series
in New
solved,
different
becomes very
When P--2,
lengthy.
appear,
many
(3.5)
(t)
_Z7am0n2lw(m*n)t m 1
and
when
P= 3,
(3.6) m 71 p
and so on. some resonant at is Physically frequency, of this
a aflap eJw(m++p)t
the
system
vibrates of that
not
only
at or it the
any multiple
frequency (3.3)
two or more frequencies. the motion tedious. the of the simple 3.1
Using
equation
out
Table
two normal in
parameters which
given Helmholtz
The origin
frequency
components
observed
a string
bowed at
CHAPTER 3
PAGE
46
now due to
be
obvious:
Order of Coefficient 1
tow it obtained
is
combination
1
By including only
in of has reduced, is if, lengthy. rather but the
460
290 340 630 920 120 510 '750 460-170 2*170 460+170 2*460 290-170
terms
table
up the
to
third number
order
2 2 2 2 3 3
possible been,
460-34U
3 340+170 3*170 920-170 290+460
even' What
'
3 3
would two than system were system dealt has related a a
800 1090
1380
3
degree-of-freedom one When of the a
3.1:
which
harmonically
parameter
problem.
as a string is possible
very
nearly, same analysis is frought may has as that with be applied difficulties. more effectively but the It to the By
to use the
system, in as the
but time
problem field
recent
shown
[5,61,
approach to
some valuable
demonstrate of point
affect motion
quality
bowing
may be written
(3.7)
jaeyWt_ i! i NA(')
Z71 the point the r1impedance Green's
z NAf 2
of function the
Q? eiw(mrn)t ICI? l
nth string and If mode, where the which the frequency may be operator terms
with
using as
approach,
[f'exp(-Y1)-oe'exp(-c()].
CHAPTER 3
PAGE
47
are
equated,
then
the
normal
modes
are
defined
by
the
series
of
equations
(3.8)
for
all
a., are
equation this
(3.8),
scaling
to match
velocity, change
the
when the
or velocity
immediately
apparent for a
that string
the
coefficients is dependant V
a., are on It is
Z, which speed
through of
and N.
combination is il
as these be
string
N or decreasing
that
combination so
important, hear a
more high
30
components, already
phenomenon
From
25; JL Relative value of Nf 1l1
expect low
high'bow combination
Figure bow
speed alter the amount of second harmonic present in the steady-state Nr=N(Ue'''vibration. oce^'1 with '=bow velocity, N=pressure.
exert
vibrating Helmholtz
exhibit Nl is
CHAPTER 3
PAGE
48
increased
40
string as where
motion''
will
C6
20 J
N ti,
change in figure
content
10 13
two notes
bowed on a
01 5 Harmonic Figure 3.6: Harmonic content of 'a note bowed near the bridge Q, and near the fingerboard
are
of to
which
equations
(3.8a)
Z-, - NA["
NAr its term "negative ' is the
only
one which it
supplies have a
energy greater
to
the value
resistance", energy
must
Otherwise
input resistance
cannot
balance
the
losses. how
total up.
negative
which
controls
builds is also
to NAr
above
which
the
solutions cannot
w in
equation
become purely
imaginary
and oscillations
of is
these
phenomena
occur
in
the
bowed
although "minimum J.
the bow
somewhat minimum
value
Schelleng
work
on this [7].
excellent
Bowed String
Player"
maximum longer a rasping
T
bow force dislodge growl is the ensues. reached string If the when frone flyback the the Helmholtz bow's portion discontinuity grasp. of the When this Helmholtz
occurs
CHAPTER 3
PAGE
49
of
the
cycle must
with
xo the
the
bow The
velocity
be V L/x, is then
discontinuity of
with
Es
ce
characteristic is
impedance then
a string.
The
condition
maximum
bow pressure
(3.9)
Nmax = ZZcV L1
(/tds
). ud -,
when
the
just
overcome.
forces bow
released sent
and another
wave is on the
out.
As the the
multiple Helmholtz
Schelleng on the
dependent
(3.10)
Z_ 1Lz12r(, Lts-P-d) x Nm = ,
as the loss term for the of first string mode. This
with
r defined of N,;
estimate
frequency
with
when these
been exceeded,
The placement limits so
(3.7)
differ pressures
(3.8).
their. dependance which on bow
and
speeds
may be used
increases illustrated
as the in
bow is
the
bridge. of
These
are
figure
(3.8)
range
pressures,
CHAPTER 3
PAGE
50
80 (a) 60
40 -
20 i
(b) 40-
. _j a:
20-
(c) 40
20
Frequency
In
Hz.
Figure 3.7: Multiple slip motion of a bowed string. demonstrated by these A response curves. 'cello bowed faster to progressively was pro(a) odd harmonics duce oscillations with attenuated, (b) every third four harmonics harmonic attenuated prevailing, . and (c) the first
are at
against very
it
may meet. a
such a high
necessary
this
1.000
u_ "100
_
I;
3.8 :
bow
l ice
a
15 cc
Urnn' b
Muttipi Multipi slip
010
"001 01
wf w fO 0
" nO , O-
q/
pressure at a constant
speed as a function of 1.6 position 171. .
0.2 Relative
CHAPTER 3
PACE
51
due to best
the string
large
amount
of, combination
produced. narrow
very
bowing musicians
process play
rich
Lesser range of
expression, bowing
parameters shows, and be that the if
consequences
exceeding
The
the
limits.
may affect the in with or real the part sound of in some is will heavily other significant be ways too. then shifted. and slowly that of sharp of is high due the
bowing (3.8)
NAt'
peak
reponse by bowing
observed of a violin
the
lowest effect
observes
harmonics to string
pitch. interesting
discussion [5,61.
domain
extremely
There bowed string. causes series, decreases. which With the the
are
which
the
of for
the it
be the to
most
partials effect It is
of being the
string
deviate
the as
harmonic so to
makes it an
possible
combination the is
inharmonic
series
reinforce
partials
will is
amount this
transmitted between
To maximize must
transfer
impedance This
match
requires
the
mass of
consequently
to be as large
as possible.
Material Silver Brass German Silver Gut Steel Aluminium Table 3.2:
Density, gm/cm3 1U. 5 8.6 8.4 1.37 7.8 -2.7 various material'
Inharmonicity E/Pz U* DO 1.24 1.53 2.1 3.1 9.6 used for strings
/1010
Inharmonicity
[7]. "
String
type
frequency perturbation Bx103 /n20.02 0.25 1.3 0.12 0.08 0.05 0.13 0.23 0.06 of using
stiffness -, n,,,, parameters 28 13 7 '16 19 22 16 28 21 considered evaluating strings. materials for table their strings 3.2
are to
then
the be when
E, steel Violin Violin D. gut G, gut Violin Violin D, wound gut Violin G, wound gut gut G, wound 'Cello Cello C, wound gut Cello C, ` on steel Acceptable value Table 3.3: Inharmonicity strings
materials
for Several
used in with
along
density upon
and Young's which the Young degree solid best solid and so
inharmonicity
Modulus, stiffness
depends. the
showed of
that
it
is
the
ratio
of ratio
E/p? '
which the
in
part basis
in
table
has a very
tensile
a dense
wire
the 3.3
mass compares
increases various
stiffness.
Table for
using
Young's
expression
st ring
inharmonicity
CHAPTER 3
PAGE
53
with
f77 the
frequency string
of
the the
nth
and
for is of
a the the
as in
table. falls
n m;n in
the
partial
which
about tones
between
G-string,
response
rapidly
vibrations can
'too.
The
by Helmholtz
motion,
contact
between
modes the
much of
bow hairs
that treated
of
torsional in the
will-be is most
least closely
important
waveform
associated
Helmholtz
motion.
The wolf-note.
of
the which,
string at
makes times, in
it are
it,
to from part
are for
forms motion
still.
such although
remain
as these, there
contact
unpleasant are many
the
the
cycle.
known
as wolf-notes,
possible
CHAPTER 3
PAGE
54
them to
take
these
string
all
owe their
mode and lumped quite two and its
origin
coupling
to
the
to
same phenomenon.
the bridge was and used plate earlier. between is
A single analogous If-the the both these course are two the to-the
parameter similar
system a doublet
be formed
frequencies other
simultaneously are by
impedance bowed
frequencies unaffected
generally
coupling
between
in
3rd
2 94 96 104 106 194 196 204 206 294
(3.7), excited
is
a bowed series of
If two of
forms the
a harmonic non-linearities
bowing
process.
where would as is of
the
by enough occur action to for them, the to be excited to related tones which take affected the tones combine new such with a an
Beats heterodyne
inharmonically
roughness three
harmonics,
starting is
impedance. 1 of second
column the
produces
new frequencies
CHAPTER 3
PAGE
55
the
original It is
list
to
produce
additional
ones, these
to observe of an
beats
the
frequency
recorded a
80
60 ah
v
instrument Figure additional wolf is It under will String in is just appear 3.9
c 40
a.
20
to conditions,
the occur
bowed
resonances
whenever
(3.12)
) jT/w7 w= nii(L-
Z,
the and
of
the to to
plate same
at
the
tuned
frequency
distinct, differently
roots there
With distinct
(3.12). defines
occurs
which
frequencies
(3.13)
w3 - wlntrc -w L
P= -nnc/L, PI) and B= L/27 (3.12) Q= -(Lk (2P3are
kL+ T+ mL
+ t)/Lm,
cnirk mL
and
Then, AL/3
defining (3Q to
and of
with the
9PQ + 2711), by
solutions
equation
given
root
if roots
B /4 if
+A
/27
unequal
B /4 +A
If the three
the
string are
are
tuned
to exactly
w,
then
invariably
roots
(3.14)
a=Tl Lm
as
move more closely several all of which will reduce the action The remedies are p in
suggests-
avoiding
wolf-notes
considerably. at be all a
of so
solution. while
same
could
be said such to
string reduce
pc is
again
which
a vibrating
string
go
to
extreme
lengths
to
ensure
the so
between is
be excited. the
sucessful
'cello,
which
is
problem) or
properly,
changes
resonance solution to
enough must
wolf-note
another
the
wolf.
playing
A(440), only
violins,
CHAPTER 3
PAGE
57
while
on the
a of
mere
14
cm. is
of
string the
vibrates most
at
this to on a
by far
suggests.
Playing but
higher it is
to disappear essential
unfortunately upper
musically
'technically
to use these
positions
on the
G and D-strings. happen if a tuned done with with vibration 'cellos? absorber were attached impedance much larger frequencies. the resistive resonance to then than If
bridge
violin has
alone
at
these
vibration
absorber the
between
absorber's
to eliminate to one of
short
the
and the
tailpiece, It
such
a vibration performer
absorber. without
may be response
applied' at other
by the
sacrificing
instrument's
frequencies. may occur plate Cy,., to avoid be done on has, mode, this the at other this frequencies is the most been besides troublesome. determined before for [9]. described angle [ there seems to
of Those
that
of
the
but
problem
has not of
perception published of
however, is the
predicting
which of
considerable have
Although remains
have
causes learn
wolf-note this
much
about
presence, a string
phenomenon.
due interaction motion to
Bowing
the with see a
an effect
on its on at
transverse waves
waves
polarized
900
to
near-infinite
CHAPTER 3
PAGE
58
at set
the
bridge
frequency
they
vibrate
at
ulay
up a stable
polarized a
motion interaction
the
torsional types
modes at of wolf-notes of
or be
time
'domain
modelling the
the
mechanics tones.
complex" it first
appeared
to be!
H. Helmholtz, On'the Sensations Dover Books, reprint edition, C. Raman, On the is referred of print
A. Benade, New Yord,
by
Alex.
book,
long
Fundamentals (1976).
Acoustics,
Oxford
University
Lazarus, Thesis, in Cremer, L. reported Schwingungen selbsterregten 30, pp. 119- 136 (1974). , [S] string [6] string", [7) pp.
[$] VOL.
[4]
H.
M. McIntyre dynamics", R.
and J. Acustica,
"On pp.
the 91-
of
bowed
"Self43, pp.
sustained 109-120,
the
bowed
"The
bowed string
and the
JASA, 'vol.
F 53,
of
plain
wire
piano
strings",
JASA,
[91 326-
J. 338,
Schelleng, (1963).
"The
violin
as
a circuit",
JASA, vol.
35, F
pp.
CHAPTER 4
AN OVERVIEW OF VIOLIN
DESIGN.
PAGE
59
Design.
the
previous the is
chapter force
that, on
except the
for bridge
the
special the
due to
transverse It is
string
bowing
to deal to
remainder how
converts chapter of
violin
the
action
several bridge, in
the in
air-cavity, All of
the the
plates, to are
detail.
f igure
the keep
reasons in mind
which the
the
one in be
to
must
chapter large
-2 Kitz.
1 [1]. as
Briefly, possible
again
response from
fall 1.2 for and the
as
about at
low
and
this
of f 2
, once
41Qiz.
above second
between responsible
coincides
in
in, by the
n,
and
ng,
and
its
presence to not a of
in
the
musician where
as contributing efficiency is
nasal prime
and uniform should energy. behavior. pure tones solution notes, but
obtained as is
convert This is
accomplished at
by designing this
Except were
used as the
unsatisfactory different
output of
wildly the
the
properties
bowed string
CHAPTER 4
AN OVERVIEW OF VIOLIN
DESIGN.
PAGE
60
ear
response. of a harmonic frequencies if is the series very near itself bowed are to
missing that is of
assigns
present even is if
string lies
will
pitch
little a note
frequency. summation of
each harmonic's
so that
though is
unimportant, or
resonances a loud
both
response loudness plotted. lowest frequency sound although is the region loud is when
curves Below
are the
60
resonance
c 50
J CL N
very
little
40
radiated 250
Ejgure
1K notes
and 1KHz.
in
this
4.2: The frequency response below maximum- loudness curve --of a poor quality violin.
sound the
quite
violin
bowed.
CHAPTER 4
AN OVERVIEW OF VIOLIN
DESIGN.
PAGE
'61
Curves
of
this limit
type
are
at give
the
a, definite
which of
example for
not
does not
same degree
previous critical,
example. as is not a
resonances will
indeed This
prove. in and
conclusion 1
unanticipated
described response
Chapter violin
showed quality
a correlation [1].
between
low frequency
sound Violin
also to
is exploit
well
response. besides to portion string's ensures resonance included oscillating often tone large would the
Vibrato, contributing transient of cycle, that peaks in regime. enough. fall in the also more are the The frequency to excite variation a particular resonances. of high string modes is
r to QO N It
plays an important role in violin technique. The bands show how the frequency variations of a note with vibrato allow the first and to a much higher level. sixth modes to be excited
figure
4.3: Vibrato
violin
adjacent
a violin in an
regard which
for
its collapsed a .
The design
always
remain
compromise
between
CHAPTER 4
AN OVERVIEW OF VIOLIN
DESIGN.
PAGE
62
static seeking It
and the is
requirements.
The
master
craftsman
is
always
obvious on a
that
the
violin
be
sounding-board construction on-a enough Another bridge. the force: have dire bridge modern'
mounted which
This about
Newtons is light
instrument, into of is
retain by is
surface of
motion
strings. of the
down-bearing of the
to are or
the
angle
and
reduce
reduce
string
bridge,
consequences. which to its a vibrating tension plate string may exert on the the bridge driving is and
The force proportional static forces Lowering violin enjoys. KHz. a rigid point high only the due to
between
on the the
to which
drive the
the bridge 3
the 4.4
mechanical
Figure bridge
Below
about
which the
(1)
CLO
the
figure, of
(2)
point,
vibrating height is
larger pivot.
component
directed between
mechanical advantage as this vector representation shows. F is 32% greater than F. in this example.
A compromise
CHAPTER 4
AN OVERVIEW OF VIOLIN
DESIGN.
PAGE
63
the
bridge's
mechanical
advantage
and the
down-bearing
force
of
the
strings It great
must be sought. is the arching which allows Viols, the with violin their to flat lacked viol withstand plates, the family such a
down-bearing string
by
the
violins benefits of
arching
of
by luthiers results
comparable
beautifully
the
Italian
Usable
materials
E 10 *10 11.0
E *10 . 33 44 . 1.5
gm/cm -. 13
10.3
10.3
1.4
5.6
20 .
sandwich
maple sycamore poplar white pine Aluminum magnesium fiberglass
40.31
>. 4
56 . 56 . 55 . 50 . 2.7 1.7 1.9
composite
on a core Table 4.1:
1.4
721
12 .
which violin
Materials for
and Dy stiffnesses.
are
the
in a
which suitable
the
need for
is
the main in
a low of
density
extremely Add to
are
materials. along
this
having most
different
velocities
perpendicular
axes
the
suitable several
natural types
material of wood,
must
be spruce.
Mechanical
for 'have
in vibrate is time. is
making with a
than
down-bearing
a long
allows
to use a larger
The third
point
more obscure. Any bowed instrument reached bridge so that waist for its of one at a time. This highest violin. length. radiating violin. of seperate A noticable the sudden the narrow-waisted vibrating increase change in in areas, the , plates. 'is its their own is and also must It be designed is for, this the so that reason the that strings the top can -, be of the
body must
reasons which
sudden
surface
emitted
feeble -a
holograms
orientation that
But what
becomes clear
that rather in
resonance surprising
explained grain.
difference
wood across
the
51,
F .;
A'
..:. N
5UU Hr.
[2
,11
at
635
liz.
[2,21
at
800
liz.
and ,
13,1 1 at
9L0
liz.
CHAPTER 4
is
medium,
the
across direction
the
grain so
in that
resonance from
much closer is
to
the to
[1,1] the
Holograms
mode. in
4.5
show that
would mean
much closer
be about of a the violin. modal uniform two
These
normally
a uniform the
dimensions to increase
be used range,
in
the in
weak,
frequency
giving
a more
assembled
instrument.
Table velocities
4.1 in
lists the
some
structures
which
have
wave prove to a
than
used
in
advance. in
fibers the is
modulus of Glass
properties of this.
each.
Fiber-glass tensile
has a high is
course
brittle. in a matrix
extremely
a light, Boron
used,
somewhat
offset the
unidirectionaly as form it of
a material alone
matrix been
the for
"grain". golf-clubs,
used
frames,
to of
blades,
ratio are core is used is
even violins
important, to form
in a
composite this
a sandwich to a
On bending
subject
considerable
j.
CHAPTER 4
AN OVERVIEW OF VIOLIN
DESIGN.
PAGE
67
compresive able
force
normal
to
its
surface Cardboard
yet
many light
materials
are used
has been
successfully
as a core In over
have greatly
advantage piece as
spruce:
wood.
potential and it
the
damping that an
selection
substitute of
available
entirely
surprising qualities,
valued
aesthetic its
finish,
as for
attributes. for size energy the a great offset the and and plate plate shape yet size has been selected, which still will resist lower it needs the force (1,11 radiating to be
greatest of the
may well
resonant area
advantage,
moment. resonant
thickness
has
much nearer
sounds
a violin
color size. of
lacks
[1,11 the
bouts, the
plate's the
edges by the
displaced plate,
increased.
A rectangular
clamped
four
CHAPTER 4
AN OVERVIEW OF VIOLIN
DESIGN.
PAGE
68
sides, are
almost
twice
when two of
these
boundaries
benefits by using
foot the
bass-bar
beneath
plate the
so that it would
lowest
bottom
violin's
be handicapped reduced.
output located
thereby
somewhere of
second large
harmonic gap in
excite filled
between
220 and'440
Hz.
is
manner. is made into or air a cavity, lute, such as the f-holes is the of formed plates. a
a Helmholtz by the of
resonator motion of
enclosed analogous
driven
excited,
one degree
freedom
frequency hole
by changing violin it is
area.
been shown to
produce
it
that
the
design
of
a violin
is
several sumarized
These
and with
a characteristic possibilities of of
a
infinitely which
been
instruments
values have
best
well
combination
over two and
aesthetic It is
copied to their
centuries.
a fitting
tribute
work.
CHAPTER 4
AN OVERVIEW OF VIOLIN
DESIGN.
PACE
69
Desireable
Consequences resist
Indesireable
Consequences
mass vibrating -Increases freqs. resonance -Raises -Increases -Increases -Increases force of -Less force violin vibrating bending mass moment
freqs. resonance -Lower area radiating -Increase mech. advantage -Better -More bow clearance -Decreases force of down-bearing the strings
able to resist -Better force the bridge resonance -Helttnholtz f req s. resonance -Lower radiating"surface -Greater A summary of of the the violin, contradictory especially
damping freqs.
Table mine
4.2: the
which plate.
deter-
design
The Bridge
as a Transmission
Element.
the 1
spacing KHz.
of the
resonances modal
is
critical is the
at high
low enough of
density in
be interested
trend
the
frequency
particulars. of excitation increases smaller say, into is energy the plate must area. one the divide This would lower
regions, figure to
be divided in air of
the
wavelength
much
remains
nbar
another. into
efficiency regions.
quickly
divides
CHAPTER 4
AN OVERVIEW OF VIOLIN
DESIGN.
PACE
70
trend
towards
acoustic
output
with
frequency response
should curve of
by the or
strings frequency
in
figure by
4.6 spruce
below. also
exhibited
this
output irate tating sound occurs acoustic around Hz., to but 4 KHz. It how then which is and its freedom a is it possible to obtain showed were of the so peak in response in
reduced
to olimthe irrinasal which with output 1200 from a 2
the the
important
good
design
action
as a
0.0 8
ca
0.04
0 o.: -C cd 0.08 .d cn 0 0 0.12
gives this
range.
needs its
made bridge
reach
at
4.7
shows essentially
bridge base
de-
a mass
connected
to
it
by a narrow
section
which
acts
CHAPTER 4
AN OVERVIEW OF VIOLIN
DESIGN.
PAGE
71
as-'a
stiffness
-element. motion
This
simple three to
is of in the
translational
the
because frequency
continuous these
than
modes. characteristics element figure properly peak is display 2 4.8. so at about are With of such a, and bridge the KHz. the
.f Q
should
important
4.10.
A is
F=figure4,8: The three lowPst bridge modes, which may be represented by a single degree-of-freedom lumped parameter system.
peak in is
transmission to
output region. of
be enhanced
The response a -dramatic quality. kHz., lighter this but" effect Take, with for
the on a
bridge violin's
can
example,
two bridges
their
resonances
at
3 The
greatly would a
different the
bridge would
produce violin is
response nasality
while be well
make
sound favored
increased. against an
by soloists but is
as it
bridge, around of
decrease
3 kHz.,
would
such a case.
The number
possibilities
which
the the
great is
variety found
been bridge it order was and Musical an
of [8].
shapes
in
bridge
far of it
has the
that in that
motion second
"indirect force
passages a string
lowest
differential
}"
(4.1)
AS=
Ay-
x2+AyL
1+ftIy/Ex)2 X =A
the binomial
As, "Ax,
0 this is
becomes a differential
which,
using
expansion,
(4.2)
dS= 1+ (dyldx)2/2
On integration
this
yields
(4.3)
L+ Azf
sin(wt)
4L
The change
and
kW
in
length
of
is
related
force
to
the
tension
downwards
by Young's
is
Modulus
the
component
dynamic
directed
(4.4) -
CHAPTER 4
AN OVERVIEW OF VIOLIN
DESIGN.
PAGE
73
typical
violin
E-string
L-32.5
cm., It is
cm.,
.. 16 to note
a , frequency
(4.5)
f(t)=TW 2
indirect excitation If the with the all a force "may applied.
E (Atr/2L)2'sin(e)cos
(2wt)
This direct
be by
ttolnlioltz tension
then 7.3x10s
the A
second ..
For
be
(the this
amplitudes larger
indirect musical
there value
phenomenon.
would
be of
this musical
indirect consequences to
small,
have already
been shown
importance.
The Function
tr
of
the
Sound-post
and the
Back Plate.
It help
that the
the
sound-post
serves force,
not; but_also
only acts
to as
support for
belly
bridge.
(See figure
is
properties It is
by a jam fit
the
r, ''
CHAPTER 4
AN OVERVIEW OF VIOLIN
DESIGN.
PAGE
74
quite
secure may
(unless be but
the ignored
strings for
loosened). lowest
Its
importance
as
a at
as a direct
between post
Replacing
or moving
may
character to itwill in
Such is soul of of
importance the belly violin. and so What, response? of the plate about moving mode be
the
frequencies. violin's
sound-post would
be similar
bridge
excitation but
second
would all.
be very
efficient
mode would
hardly
With
the
effectively. is
bridge strings
for
frequency t" one foot impedance frequency, frequency, simple but front it
however, very
post the
the
which plate's
was
matched
impedance
the
by the far it
bridge. from
of
course which
that the
impedance
presents
performance. the post has expect arises that if ran the back its is
characteristics plate.
with not
as an impedance which is
device
The fact
maple,
much more
CHAPTER 4
AN OVERVIEW OF VIOLIN
DESIGN.
PAGE
75
dense
than
spruce, the
lends
support of
to
this radiated
idea. from
It
would the is
to, measure some' this tested In ribs, such of work. with this
sound but
back
the
such
a project violin
beyond of the
Instead,
a good
1920's
plate fitted
and
the
with
mould, Caulking
constructing
violin,
sound-post
place. sults
30 cr m a 20 it t it vvr
figure suggest
It v
I V'1 1
10 v
dB that
1230 Frequency In kHz. Egure 4.10 : The sound-power on both sides of a violin mounted in a large baffle. resonance. requirements could the that benefit test its violin main its Clearly of the a balance back at must be struck
study. of great
back
as an impedance contribution
device,
instruments
may be a noticable
I
CHAPTER 4 AN OVERVIEW OF VIOLIN DESIGN. PAGE 76
ti
of
the It
other is
the its
bass-bar, main
is
less role
dynamic
and
lower
to in
end of applied.
f-holes?
to which
S, "4
Modelling
the
Helmholtz
and Front
Plate
Modes.
of
the
sound-post, the
and the
in to
the it
coupling
models.
violin's modes as
action
by
plate
Electric thesis
analogues
used and
in
his
work,
a mixture
of mechanical
acoustic
Schelleng
Hz. to so that
limited
of all
the
'the
frequency
belly's
range
resonances
in
his
only as
model
the a mass, plate and the
to
below
one
600
lowest lumped
needed
This
appropriate To determine
to
the plate's
region
another on it
a smaller frequency.
smaller
effect
resonant
and
the
half-powerpoints
is
possible
to
determine data
all
of
the from
equivalent an assembled
for
the
plate.
As this
was obtained
effects
CHAPTER 4
AN OVERVIEW OF VIOLIN
DESIGN.
PACE
77
of
the
sound-post, air
Helmholtz of
back
plate,
radiation intrinsic
may be are
impedance, properties.
to
the'cavity's
The the in
modes were
resonator the the holes
in than
the the
usual
way: of shape
much range
wavelength their
frequency
piston
impedance
accounts
usually The
f-hole.
lumped It vibrates
r f
into is
of cavity air
the
that
mode.
This
analogous 4.11.
to a base excited
mechanical
system,
such as that
shown in
figure
Schelleng's
Cf Mf Kf CL
electrical
ana-
logue
frequency
poduced
rescurves to line figure the in for velit is first between is not, be the the much dot
Kh
f(t) Mh
Ch
Frequency
41 " A simple model of the violin b elow Subscripts h and f refer to the f-holes and Figure
respectively. each mode's Using to the mechanical his sum of analogue results. these levels. before and a digital acoustic which relationship computer output must
duplicate the
simply to
pressure
The phase it is
interact. of sound
Since in air
shapes
If the
it
was possible of
to air its
find
area the
such velocity
that of
it the to
when it driving
the SPL.
would
be simple at a
estimate center.
point
them and
product
volume front
product
the
and the
plate
information vibration
shape
readily
available
integration
cm? - . on axis
one meter
well be in.
below
the
the
air
to compress
cavity,
these energy
near
shuttle radiation
back
and
taking are
also so that
velocity
quadrature
above of
the
air
phase
so too and
phase "h"
output
approximately
sum of
curves
and "p"
CHAPTER 4
AN OVERVIEW OF VIOLIN
DESIGN.
PAGE
79
figure
4.11.
The phase difference Helmholtz on each between resonance the two resonators through at frequencies two have
close little
to
the
passes
90 and the
effect There
other. that this as form in of coupling figure, in it is of benefit. the Below impossible case. will two the to
Output
is.
the
between
the
range
Strong produce
the
output
below
Schelleng's so much 'about simplifying The mutual as are " the the-effect plate are it
subject action
a degree of
surfaces of
plate
resonances
coupling To be must
major of the
importance. violin, a
use for at
model
incorporate
of
these
features.
Other
Air
Modes in
the
Violin
Cavity.
of
the
belly that
for in
many
more
than
those that
the
has
to
imagine first,
what third,
these fifth,
lowest
few
modes modes
turns
the
and sixth
CHAPTER 4
AN OVERVIEW OF VIOLIN
DESIGN.
PAGE
80
are
similar
to
the
and are
[2,1,0] the
subsections
seven
modes appear
4.12
p..
23467 1090
1190
1290
1610
1800
1910
the
third the
and second to
a node at
little that
the `standing
mostof
these
air This
radiate
to
the
outside
environment
(131.
these of
range will
couples
very-strongly
when using
approach
a pressure all,
mode at the
plate
If the then
would
the plate the the air
drive
mode
the cavity
pressure.
be driven
for
contribution
17.
the
upper
half this
plate the
would case:
to
that of
of the
the plate
lower drives
half. the It is
half at the
figure
reveals.
area on the
with
volume
velocity the
which
drives of
integrating
product at
magnitude
a point,
graphical front
back. of
plates
to use both the air latter, the interaction mode. with air
the with
effective the
Helmholtz
the
former the
indicating interacts.
plate
second
with
the cavity's
the way in
air" affects
mode the
Rayliegh
estimate
resonant
dimension. measurement
resonant equation
frequency is
The pertubation
(4.8)
AM=
AS cos(21Tx/L) _1 L0 So
cross-sectional constant gives
dx
where
So
is for
the tubes
mean
Rayliegh but
developed Jansson's to
equation
has, demonstrated
method
results
when applied
CHAPTER 4
AN OVERVIEW OF VIOLIN
DESIGN.
PAGE
82
-2.5sin(x-7.7)17/13.5 -sex
`2.3sin(x-21.2)tt/9
"
Inside
length
32.5 cm.
used to calculate frequency. the
The perturbations are predicted. directly by the shown, in using with compliance of
in
finding
the
effective 4.3.
of
the
4.14
and table
model,
as will
apparent
evaluated.
Perturbation
.. 1 f'
limits 0 to 5
Af/f "2.9
in %
2 3 4
-0.721Exp(x-30)
-11
change -1.9% Table 4,3: The change , In the 11,0,01 resonance frequency due to the cavity shape.
: x.
It
will
also was
be useful no problem
to
be able with
the
losses
of
this
mode
the
largest
but
here
losses,
absorption
CHAPTER 4
AN OVERVIEW OF VIOLIN
DESIGN.
PACE
83
treated
as a rectangular
room,
this
absorption
should
be given
by
(4.9)
'
where. Pr,, cc 4 is '--'The samples Values 'losses calculated defined by the is
the
power
lost
to
the
walls,
P2'
is
the
pressure,
and
acoustic
absorption coefficient
coefficient. is difficult tube from ignored Q-21TfW/P, test the to measure to with small
An were
proved
through in
f-holes' of
a Q-value,
(4.'10)
V .W=11
determined data 73., for in a violin
2
this
which
Considering agreement
between value
of wood the Q of
between
experimental
54 is
satisfactory.
Most of is now at
the
needed John
to model
the will
violin
below to
600
Hz.
hand.
be extended
include plate.
the'coupling
between will
belly,
back is
be the
way in
which
treated, of the
possible before it
to accurately is assembled.
predict
response
completed
r
...
CHAPTER 4
AN OVERVIEW OF VIOLIN
DESIGN.
PAGE
84
and 55,
"Objective Yankovskii, B. " Physics, Acoustics, vol. F. 81Saunders, 98, (1937). "The mechanical
tone of violin appraisal 11, pp. 231- 245, (1966). action of violins",
quality",
JASA, vol.
9,
19,
de pp.
D.; -Haines and N. Chang, "Making better Engineering, Mechanical graphite composites", 25-27, "(1976). [61 i"C. 138, pp. "The Hutchins, 73- 86, (1962). physics of violins",
musical vol.
from 3, pp.
Scientific
American,
vol.
of
the
violin
bridge",
For a further bridge dynamics discussion the -theviolin of of. "C , "Violin frequency to M. Hacklinger, timbre and bridge refer response", (1978). Acustica, 39, pp. 323--330, vol. [9] Fundamentals : A. ; Benade, , (1976). New York, Press, [10] E. Johnson, Applied-Acoustics, [11] ,'J. 326- 338, [12] vol. E. ,, 37, Schelleng, (1963). of Musical Acoustics, Oxford University
"Decoupling the back and front plates of a (1981). 14, 157161,. pp. "-. vol. "The violin as a circuit", JASA, vol.
violin",
35,
pp. -: -;
properties
of
cavities",
Acustica,
[131 :. F. Saunders, 498, (1953). [141 Lord Rayliegh, York, ' (1945). [15] ; =. U. JASA, vol.
' t. r.
on violins",,,
JASA, vol.
25, Dover --
pp.
491-
Theory
of
Sound,
second
Books,
New
"On the theory Ingard, 25, pp. 1037- 1061, Beranek, Acoustics,
resonators",
rJrr
[16]
L.
McGraw-Hill,
New York,
(1954).
....
The Model.,
model
for. violin
it
first
step
vibrations
elucidated f-holes
Helmholtz volume of
formed violin
the,
into
further
insights model
examined.
time the
more closely
approximates ,
of.
chapter which
is occur
the in
the
violin.
these some
the
front
the
model
[1,0]
always of -:,.the
the
length.
to a lesser It
4) dictates in
the-resonant a violin-shaped
a short will
probe
into
and, an experiment in This 5.1 resonance may where " be the Just
nature
acoustic in
output the to
remains curve
observed ,.
response
contribution.,
by each mode
indicated.
CHAPTER 5
PAGE
86
like
the doublet
air to
mode draws be
energy
from The
the
plate's
formed. in plate, it
sometimes range
resonanceperhaps not
this
although the
between direct
through have to
the
sound-post
be included
"which 'the
imaginary
materials
plates that
insure violin'which
ao .i""
different.
a DacK
is
figure
possess
ptaLe
resonance
did
not
degree in
of the no
Frequency
coupling.
that 'either
sound-post frombe
this
'that
Figure 5.1: Resonance which commonly appear in the vlolln's tow frequency range. peaks
coupling and
provided is
on the
and
-There beginning resonance boundary volume of they known. the for ribs
which
should
the
frequency of
the the-air
conditions the
plate
stiffness
resonator.
'-provided and
change This
resonance jTheir
frequencies -effect on
obstacle.
resonance
can be calculated
by treating
them as a spring-like
begins idealized
all with
of
thel
a right-regular surfaces.
as models 5.2.
vibrating mixture
strange
of mechanical approach to
and acoustic the problem. equation describe the the the wave easily
results
components
used as it
green's may be
function, used to
kf mf
cf mh
Ch
(2.16), pressure
throughout is possible
the This
standing is most
be
in
developed.
one - dimension
pars-
then
extended Newton's
to law, as
particle
of
air,
may bewritten
(5.1) ..
with particle as dP/dx the
dP dxS=pSdxd2 i dt dx
pressure gradient, From this S the
F ,t
cross sectional of motion area, and e the may be written
displacement.
the'equation:
()d s: 2
Ze d2e cz _ d tz dx2
between
.,.
and the-relationships
displacement,
velocity:
and pressure
as
(5.3)
u= jwe
,P
jfcu
and ,
dX
jwpu. =,
CHAPTER 5
PACE
88
into
three
with a point
the
vector at
notation r. is
xt+
y3, +
equation-of
source
(s. 4)
P2r) OP(Fr) +d dr
jwps(-ra)ulr)dV c
velocity, same form function as and pressure equation in (2.8)
A--c2
..
Theas, - .
green's
written
from
(5.5)
g(r, rw) - -J CS
violin there drive, are, are however, driven.
-K-)
no internal by, It , the is air therefore sources; modes . only and the -these to the plates are
In-: the
-. or
the- cavity
perform volume
integration in the
previous-equation.,
;".:..;
(5.6)
P(r, row)
10 -'wCS
is
piston this
which
introduced
related
integral, point
a reference
surface. on both
point air
depend will
4 mode
and the
be necessary area of
to adopt the
to describe f, b, or h,
equivalent
mth surface,
CHAPTER 5
PAGE
89
d rives; f?
or 1u(?
is
driven
by the this
nth
air
mode,
either
0 (5.6)
or
1,
eg.:
S;
0 as
c)IuFdS.
Using
notation
equation
may be expressed
-cS
(kZ Kz1
the
In boundary of the
order
to determine mist
shape
the
on the
problem. (see
x'direction or
for
z 'directions, be
pressure
which the Zo in zx
end,
impedances and
particle
use of
the-relationships
(5.3),
then
dP(0)- . wp P(0) So dx Zo
where: So- and SL are change An If sign of the areas
(5.8b)
dP(L) dx
x-axis
jwPP(L)SL ZL
at x-O and x'L. considerations. in much chapter the 2, ' The
to
the
the
pressure
string"of;
equations
(5.9a).
and
Ksin(4)=
cos(4)So
(5.9b)
-Ksin(o)=JWPcosfKL+flSL
ZL
CHAPTER 5
PAGE
90
Rearranging
terms
gives
the
equations
(5: 1o)
and
(5.10b)
jwpS, /Zv
and the
nature as
function
to express
(5.10a) +
$kjwpS0/ZOK.
(s. la)
and (5: 11b
jwp so + SL L Zo ZL
[Zo So
LL +
Kn = (ntt/L)Z-
L
dimensions
ZL
Extending
these
results
to
three
yields
(5.12a)
K=
52 Izs wvp 1 V
Sb ++
Sho
Zc
zbbh
and
(5.12b)
K = MIL
where this
S is area
the of
total the
area
of
the
cavity of the
average due
over to
impedance of the'walls.
mostly
absorption
and the
equations
(5.12)
are as
so =1 in
that
the and
with
length bracketed
chapter K1
possible
'ignore
when K or
CHAPTER 5
PAGE
91
appear this
-in
the
of loss
equation terms
(5.7) in Z.
imaginary
part
of is
includes to
mostly'due" which
the
-term
way in
the
air on
modes the
is
now and
known, f-holes,
the still
forces, to the
plates in the
be found. pressure
approach.
,.,,,,
which
Once again,
fora plate
the first
is
step is to write
the
equation
of
nation,
5*`13)ph
z+
4rP()(r-),
at
where acting E-on Young's the plate the the cavity
120-v
Modulus, at plate
1
Ratio, r If'
+ _E_5(_'0)
ds
l) - Poisson's a point
velocity using
pressure
function:
(5.14)
u(r) =L Ph
u(r) must'be
(r ( AA(`oWZG
obtained of surface and the P, h for
where
m could the
on this may. It be is
pressure $ (rm),
surface
n indicating'theair'mode. area as
pssible
rewrite
the
equivalent
(5.15)
Smt1
n)VRm)dS
due of as
to
the
nature This
of
orthogonal makes it
functions, possible to
demonstrates write
the
reciprocity.
equation
(5.16)
with .
only
In
shown for
was
each surface
directly do this design be the of
in.
from as the the the shape
deriving
equation
obtained
shape
plates. as of
mass, the
equations equations
the it is
are,
some
violin's
force,
zero
center
the"f-holes. '
With
these
simplifications
velocities'are
[2]:
u= bFfftS, . OSb
+SrISb
lj'prf
(ShoSfISaa1) J{/A ,
CHAPTER 5
PAGE
93
and where
,,
--JW(keq-W2meq
jWE)/A, 4
oon=-jw2p/ICSA
(k2
K', )) ;
and
"A.,
=1/2 when
(5.18)
Evaluating
the
model
utilizing' a violin it is
the
nadel
it
by,
predicting
by.. comparing to
were
made in the
However, equations
may be
appear
must
violin. measure, the parameters response for. the for., with would coupling the have violin fully
between
largely
reassembling are,
however,
different so that
the of
ribs the
similar sound-post
the of the
to minimize at a time
the
plates
were joint
tested
while the
The glue
between
CHAPTER 5
MODELLING
PAGE
94
and and of so
the the
ribs
was
not
broken response
between of is the
measuring assembled as
plate (The in )
the that
back
minimal
had back
plate
frame, plate.
to
modes 5.4.
be seen
Flure
5.4:
Setting
up a test
in
the
anechoic
chamber.
Once from plate glued the had in The same plate a small as was
the frame
been front
complete to measure in
tested
frequency Once
occured
mass
to
the
resonant
frequency
CHAPTER 5
PACE
95
and
were resonance
and
the
loss
expression integrating
m, 1Wm
with
the
pressure
and with
displacement vibration graphical computer'with although still digitizer fringe holograms. integration a great
performed in
'by -time,
' information
manually' location
' A"computer
plate
along' "
Parameter 5o Sn1 , m
,-
k
f Table at low 5.1:
gm/s gm/s
Hz The parameters with
2.2x10 * 2130
415
4.5x10 * 11700
515,
-----200*
" _ -----the violin A
1 ,
which its characterize
sound-post
removed.
denotes
value.
which
describe from
the
f-holes
were 'all'
affected of
accounted phase
mass loading.
CHAPTER 5
PACE
96
the
f-hole
velocity incompressible
so that flow
and of to the
a be
mass-loading is applied + 16
this
near-field
mrod' aS[1
term
in in
in
surface the
the
surface. -the,,
of
drive
already on
included only
Therefore
losses
one side
of each
to be considered. measured areas in directly were from 1650 cm! the test
surface
length,
chapter of
computer to the
at of
features the to
Helmholtz
plate
mode, air in
and the
the is
[1,0,0] plotted. is
violin the
plotted
areas three
difference
importantly, Hz.
-resonant - experiments
frequencies of t is
a -few the
similar
results:
model
a good. description`of
is
a good point
at: which
to once again
consider
the
effect
of
CHAPTER 5
PAGE
97
40
m30
V1\/'\ j%_
;f\
//1\
0.20 10
\I
V
200
Ej
300
500
---a sound-
the
ribs
on the enter
resuiLS
violin's the
J_ lA
response. model
_--1-. VSr1VUb _
It a
earlier although
i1 tinny
that they
wh4i-h
they would
that/
need not
aiieCL Lne
as
seperate
TL... `11%'
..... W 7.
impose measure
on the its
front
plate
are
implicitly
the
method
to a
impedance. were
important
way in
affect
as a means of as the
transferring of the
energy model
between is already
accuracy in to
changes to be
the are
occur
A half-inch cavity, at a
a short
probe
inserted point
[1,0,0] in figure
minimum of the
appears
predicts figure,
levels
a seemingly
CHAPTER 5
PAGE
98
experimental
conf irn this.
results
_ 120 cm V 110
J/
^A ` .0/ 1
These
exper-
-which without in
were a posi-
100
demonstrate
90
200
300 Frequency
400 in dB.
500
600,
the for
model
is prethe
It only is some of
is if
equally
valid
for
no motion
occurs often
make
observations be measured
the the
violin, complete
parameters
area
for
each values
plate will
in
the
their
different prevents
the the
modes the
resonant the
each plate
affected
modes
when
stiffness it
and error. is possible whole to violin's and the individual vary any one in actual is of no of
known it
effect by
on the
response
a
would
impossible this
testing%n response
Of course, luthier
method
who wants
to know how
plates
will
CHAPTER 5
PAGE
99
behave left
when the
he next
finally chapter.
glues
them together,
but
this
problem
will
be
for
Investigating
Violin
Design.
of
violin
were any
discussed change. It
without is now
results the
possible
do exactly the
using
requirement but
ones,
degree (5.8)
In figure plate
response constant
resonance
40
m30
n. 20
10
300 ' Frequency 400 in Hz. 500
figure 5.5: The test violin (---) und the Increased output brought about by a 200/o decrease in equivalent mass and stiffness(-). while the effective of mass and stiffness 2 to 3 dB occurs per is cent have been decreased between of the the main roughly 20%. -and like plate be the
two resonances mass may sen of the the plate use total
Twenty this to
only
a small more in
remove
wood
must if
dB increase
output
little
CHAPTER 5
PAGE
100
--I
violin
_1I
coilapsesa
.1
1_I
Another
me-
thod
creasing put
for
the
involves
radiation.
also raises
fr
It
the
aac
volume a
must
to 40%
counter produces
Figure a2 dB dB be an
change
region
between
Helmholtz
peak.
seem to
iy^
two figand
one Figure violins 510: An with
200
400 in Hz.
500
The resonances. that the solid Helmholtz and plate resonances should occur at 290 435 Hz. to the and provide most uniform response
CHAPTER 5
PAGE
101
other
copy for
of
These G-string,
large
German violins which figure should It is (5.9). be put is the curves the for not
have a at all
having
considering basic
feature of
to
the
spacing are of
primary (5.10) to
produced loudness,
calculated
Helmholtz
to
change
spacing best
between
many combinations 290 and 440 Hz. resonance spacing prove to peaks could
response in
as well different,
lowest figures
be a good
the
[1,0,0]
mode in response,
the
model
brings (5.11)
about
a" remarkable A
frequency
as figure
demonstrates.
40 A / 30 v c 20
10
250
LL 30 0 Frequency
400 in Hz.
500 brought .
600 about
T '
M
doublet minimum
is"formed at-the
between [1,0,0]
the
air
plate, function
with
a pronounced as a vibration Sfj cmL at is shows both too an 470 the area
resonant
absorber-can large.
double quite
response
which
sound-post-brings,, from-54, to
-about , is
responsible violin,
reducing latter
equivalent curve
19 cut`- -4n
the-test
and this
and even with an output doublet level well the which case, enough
such similar
a value. to the
between bf -a
frequency held
response -. often
although to
none of a
been discussed
produce
.not
only
the
area
Sri
which
affects in the
the sound-box
doublet, are
for equally as
length
and the
damping
present
(5.13).
in of
the these be a
would by
----
cx=0.34.
varnished. though violin much ceptable vagaries weather, certainly poor the it such . would less to of
Ejgure 5.13: Changing the cavity length by the response of the s 1Q/o greatly alters test violin. (---) "10 /o , (-) -10 b . 40
i I
r i 1
r II
11 11 V
CL V) 20
response region of
NO
resonIncreasing
400 Frequency in Hz
500 .
600
damping reduces of
Helmholtz In addition
between
CHAPTER 5
PAGE
104
is
then
necessary-but
too
much,
as
-the.
figure
shows,
is
'counter
productive. The cavity, consequently doublet, this violin curves explain instrument While be said particular. like to size case. with in the length controls the [1,0,01 of the is resonant two peaks of equal frequency that form and the in test the to one
although
were
produced
length are of
(5.13) length
small,
perhaps violin
why
a full-size
to another. the plate other Ideally violin, (5.14) bass if sizes of violins of the are very similar, family, the the same cannot viola exactly difficult the in
about
be constructed
be nearly
differently are
violin
it
compromises luthier,
reached
design tried
modern
Carleen
situation
more accurately acceptance weaknesses colors. It-has cavity possibility appear that modes is by is of
Whether
composers and
violas
exploited
unique
that
direct the
radiation of in a
of
sound
response
explored opening
left"or'lower
; right
corners
would
CHAPTER 5
PAGE 105
the
volume would a
to
compensate the
for
the is great
shift not
in
the
produce,
response be of
would at
interest
higher the
over extra
addition
radiator
40-
m 30
'.
: r :. ,
j
lF d ' J ! Xt
5_20
10
250 300 Frequency 400 in Hz. 500 600
Figure 5.15: Adding a hole to radiate the 11.0.01 air mode changes the ordinary violin---.
In
this
the
violin
at
low about
ensuing
equations of the
design.
The function it
sound-post for
in
the
this
response
from this.
point
impedance which
function the
approach. time
computer
CHAPTER 5
PAGE
106
prohibitive third,
for
those
who could
benefit
from is
such
an in
approach. chapter 6.
The
and most
intriguing,
possibility
explored
J.. 338,
Schelleng,,,. (1963).
s"The
violin
as
a circuit",
JASA, vol.
35,
pp.
"Air E. Johnson, between coupling (1979). Institute of Acoustics, . C. pp. Hutchins, -"The physics 73- 86, (1962). of
violin
plates",
Proceedings
of
violins",
Scientific
American,
vol.
of
the
violin
a family
Today,
vol.
[6] "Resonances body studied by A. Gabrielsson al., of a violin et. hologram interferometry Physics Scripta,, and acoustical methods", vol. 2, pp. 243- 256, (1970).
.,_ .
CHAPTER 6
MASS-PRODUCTION APPLICATIONS.
PAGE
107
Manufacturing
Techniques.
For is so
the
vast
majority related
the
sound
of
remotely
by their instrument.
different
of
expected,
many give explorations way, the Most attention selection produced specified and may in detail, It deal adjust would of
of musical
enjoyment
because When
unpleasant
viewed
It
is
the
luthier which
to his
as well
as his
wood,
makes these
or by hand, to be well-made,
be copies has
perfect
in
every
no
attention
their to be
dynamic otherwise
required
to employ
tap-tones
accordingly. some mass-produced have the correct are violins spacing poor can be very of goodby But, chance on the of in needed
made instruments
as the
expense
an-automated before
predicting together
violin quickly
put
any adjustments
and cheaply. The model which was developed in chapter 5 was meant to meet this
CHAPTER 6
MASS-PRODUCTION APPLICATIONS.
PAGE
108
a the
obstacle of the to
remains direct
to coupling the
be
to
necessary violin
measure
parameters sound-post It is of
assembled
and the
position. to predict
no use in shapes
applications. frequency
response
as a point approach,
impedance but
function
a further
itself
advantages. 4 the little to back plate radiation remove element in was revealed when compared the to direct as to an impedance It device should by
the'front. of the
sound-post an is
an internal remain
perform while
element and
position
tested serves
adjusted.
A cross-piece,
such as illustrated
6.1,
admirably easily
in
this
for
it
is
easily
and
may
itself tics. be
be
best
as it both
made of
from
material only
modulus the,
elasticity.
section
need be used
obtain
dynamic
CHAPTER 6
MASS-PRODUCTION APPLICATIONS.
PAGE
109
which could
are
This
isimportant of the
reduce
mode. similar
Additionally, to that
material plates in
stresses
cause
cracks
extremes
a uniform spruce
to be used
easy
to
a good choice about 700 Hz. as of front a layer was the plate of
material. its
The bar
where
impedance in A
should 4.
plate,
chapter
dimensions of the
match
between
bar material
impedances to a
may be surface
obtained latter.
by adding
to match
the
front
plate
at at
its
second
resonance it
as ideally two
Some
compromise
between many
different is beyond is
chosen. of such
demonstrated
clearly. A reasonably cross-piece spruce violin plate, the the and good factory this prototype of made violin is 0.3 was fitted in with just 6.2. such A curved to the front of a
figure was
dimensions shortened
attached it
impedance
front
-plate
parameters response,
on and
frequency This
which
6.3,
frequency of the
response resonance
certainly low
encouraging. is not
Although ideal,
the
spacing
frequencies
CHAPTER 6
MASS-PRODUCTION
APPLICATIONS.
PAGE
110
Figure
6.2:
The
prototype
violin
with
its
cross-bar
in
position.
these
could
easily above
be 1 KHz. curve
modal
density
is
quite
high
at the be
frequencies sort of
a properly so highly
designed in chapter
bridge, 1 could
response
which
11
_'
40
possible
c
curately
30
estithe freq-
mate
20
uency of the
400 in Hz.
5 50
with
response to support
CHAPTER 6
MASS-PRODUCTION APPLICATIONS.
PAGE
111
by using possible components. With presents rigid between levels. frequency difference effect
the
crossbar the
to
replace of
the the
traditional completed
it
is its
to predict
response
the itself.
removal The
of
the
sound-post plate
an
intriguing
possibility and
back of
be made plates
coupling
would
possibility
responses is may be it
compared
Unfortunately curves,
a small the
noticed
between
and although 40
Figure 30 64:
probably
1\ \
:: L
The theoretical frequency response of violins with maple--spruce backs, the latter without a sound-post. and
'importo 'the, to
change. 220 300 Frequency 400 in Hz. 500 0 600 /\I V dB 0
enough
overcome resistance
such
aa.
a
VSGO
Ana1-h-f-4p llG
a..
. rn
important finished
to maple
the back
whether
novice
or virtuoso,
and a well
to behold! with traditional small,, are a cross-bar way is might could doubtful. be There compete with
Whether violins of
a violin
carefully
noticable
compared. for
as the the-
plate
from the
frequency
response
CHAPTER 6
MASS-PRODUCTION APPLICATIONS.
PAGE
112
wet
Raw materials
sand bouts
Store
Raw materials
Store
Assemble front plate and ribs Determine eq. area by raster scan Adjust
Measure impedance
Final
Determine
Evaluate
poor involved
corrective
action in manufacturing
front
plate a violin.
is the It be controlled. computer The entire process may frequency asseman of the response to predict ability
bled over violin which gives such a system an advantage forms of violin manufacture. traditional
use of advanced
this
great
advantage techniques
at
a low cost
it
is
necessary
to use some
production labour to
and equipment. versatility of production a single, same milling make groove cutting micro-processors centers unchanging machine a have reality. as violin plate by a
costs make
automated be assigned
size, and
thickness, microprocessor.
directed these
operations
to build but it
a violin
are
should
be emphasized center is
pieces
through of
a production
fixed.
on the
} necessary as
instructions efficient
therefore production
programmer.
CHAPTER 6
MASS-PRODUCTION APPLICATIONS.
PAGE
113
is
described portion,
very with
briefly which
with
the
details
of
all
but
involved,
the
production
engineers. spruce and maple for the plates and for the crossis
cut
computer. too,.,: or cut finished testing. must required blocks, are if initially
models
properly areas
components assembly-or
storage
When be formed
these into
cross-bar.
Vacuum heat
wood is to
to be easily to the
means
be to the
return
moisture
content 'a
processor
holding
proper sound-
moisture post
plate, of
next the
finally the
the
major
assembly
of
process.
the process and the the then modelling to evaluate is the used violin. the a device over of this to scan" to measure holographic the data done the to
portion frequency to
response measure
is
necessary The
which
plate. as that
impedance by
measured Computer
interpretation for
analog
and
CHAPTER 6
MASS-PRODUCTION APPLICATIONS.
PAGE
114
technique
over
plate
when it part
is of
driven the
at
its
resonant work
equivalent
must
be
reconstructed, by hand or
integrated must
either
which
photograph
possible. of
Rather
intensity
interference
done in to the
holography, can be
responds measure
amplitude. a
solid-state
combined be possible
interfaced
impedance at
vibration large
any point
information the
available violin in
to
the
computer
the
response
perhaps the
are each 'plate to
were used
curve the calculate response. of any
to extend
changes for the
modes
providing
alterations
chosen.
CHAPTER 6
MASS-PRODUCTION
APPLICATIONS.
PAGE
115
the
predicted assembly
of back
the is
violin glued
is on,
acceptable it is
it
goes on and
cut
the on
assembly a few
begining
be
applied of
in violins.
industries
revolutionize
production
no doubt might
scoff
at the
many of intention
the
ideas to
applaud
factory-made modelling in
instruments, by computer,
the testing,
analysis and
construction, of argue of
finally.
suggestion one
change with
design
would years
be dismissed of
Stradivarius, there
the of
And yet
body spirit,
beauties likely
Perhaps is the to
another's
work violin
mass-produced today.
with
a richness
fear
for
his sound,
craft, it is
for the
a who
music, of
individuality,
creates i to'enjoy of a
beauty.
machine to
of music,
people
man's
CHAPTER 6
MASS-PRODUCTION APPLICATIONS.
PAGE
116
"Small mechanical [1] impedance head for use with I. Firth, (1976). 348349, 35, Acustica, instruments", pp. vol.
musical