Lecture 1

Acoustics of Speech and Hearing

6.551/HST 714J

LECTURE 1: The Quantification of Sound and the Wave Equation

Required Reading: Denes & Pinson Chapters 1-3;
Supplemental reading: Kinsler et al (KFCS) pg 99-111
1. The Basic Physical Attributes of Sound: What is Sound? How is it Produced? How does it
Propagate?
Sound is a propagating mechanical disturbance in a
medium. Propagation of sound does not occur via net
translocation of matter, it is the mechanical disturbance
that propagates. Most of the course will be spent
discussing sound in fluid media such as air and water. The
picture on the right shows a cam driven piston at the end of
a rigid tube. When the piston is set into oscillation,
successive layers of the air medium within the tube are also
set into motion causing local increases and decreases in the
density of air (coded in shades of gray). Associated with
the net back and forth oscillations of the air particles are
increases (condensations) and decreases (rarefactions) in
local pressure. The velocity with which the particle
oscillates is the particle velocity, while the velocity with
which the disturbance in pressure and velocity moves down
the tube is the propagation velocity of sound. Note that the
variations in density have been greatly amplified for
viewing purposes (after HF Olsen, Music Physics and
Engineering Dover Press 1967).

Image removed due to copyright

considerations.
Source: Olsen, H. F. Music Physics
and Engineering. Dover Press, 1967.

A. Sound Pressure, p(t), is the variation about the baseline pressure that results from the alternating
condensations and rarefactions of media that describe the propagating sound wave. The units of
sound pressure are pascals, where 1 Pa = 1 newton/m2. A sound pressure of 1 Pa at 1000 Hz is of
uncomfortable but not painful loudness. This loud pressure is equivalent to 1/100,000 of an
atmosphere and 50,000 times the lowest sound pressures that are audible. Sound Pressure is a scalar
quantity.
B. Particle Velocity, v (t)is a vector quantity that describes the alternating average velocity of
motion of a particle of medium. The units of particle velocity are m/s.
An acoustic particle is a volume element large enough to contain millions of molecules so that the
fluid may be thought of as a continuous medium, yet small enough that all acoustic variables may be
considered nearly constant throughout... (KFC&S, page 99).
Particle size depends on the medium and the frequency. A medium excited by large wave length
sounds can be broken into larger (more voluminous) particles than a medium excited by sounds with

9-Sept-2004

Lecture 1

Acoustics of Speech and Hearing

6.551/HST 714J

smaller wave lengths. Whats important is that within the particle the sound pressure and the average
motion of the particles is constant. A particle of free air exposed to a sinusoidal sound pressure of 1
Pa, moves back and forth with a velocity amplitude of about 2 mm/s. How does that compare with
the propagation velocity (speed of sound) in air?
The definition of an acoustic particle is relevant to an important issue in acoustics, i.e. whether we
can consider a system to be made up of a collection of lumped elements which behave as particles
or distributed systems which behave like continuous media and support wave motion. We will
revisit this distinction regularly throughout the coming term.
C. Scalars and Vectors
Scalars describe nondirectional physical processes like pressure. Vectors describe the
magnitude and direction of directional physical processes like force and velocity. A vector can be
broken into its three-dimensional components, e.g.
(1.0)
v = i x v x + i y v y + iz v z
Where: i x , i y and iz are unit vectors in the x, y and z directions, and
v x ,v y and v z are scalars that define the magnitudes of the x, y and z component vectors.

D. Sound Frequency, f, describes the temporal variation of a pure tone, e.g.

p(t) = Acos(2f t + ), where 2f =, the radian frequency

(1.1)

E. The Density of the Sound Conducting Medium,, is the mass per unit volume of the medium with
SI units of kg/m3. The density of air at Standard Temperature & Pressure (20C and 1 atm) 0 is
about 1.21 kg/m3. The density of gases goes up as pressure increases and goes down as temperature
increases.
F. Linear Acoustics
For most, if not all, of this course we will deal with sound flow through fluid media (mostly
air) in a regimen know as linear acoustics. As sound travels through a medium there are temporal
and spatial variations in the pressure, density, particle velocity and temperature associated with the
sound.
sound pressure:
particle velocity:

sound density:
sound temperature:

p(x, y,z, t ) = pTotal (x, y,z, t ) P0 ,

v (x, y, z,t) = v x (x, y, z, t)ix + v y (...)iy + vz (...)iz ,
( x, y, z, t) = Total (x, y,z,t) 0 and
T(x, y,z,t) = TTotal (x, y,z,t) T0

In linear acoustics the sound-induced variations in pressure p, density and temperature T are small
compared to the baseline value of these quantities P0, 0 and T0.
G. The Bulk modulus, B of a material is the pressure difference associated with a fractional change in
the volume of the material. If we consider a collection of air particles of volume V1 at pressure P1,
9-Sept-2004

Lecture 1

Acoustics of Speech and Hearing

6.551/HST 714J

and we then change the pressure acting on the volume to P2 with a resultant new volume of V2, the
bulk modulus is:
P2 P1
B=
with units of pressure.
(1.2)
(V1 V2 ) V2
For ideal gases at 1 atmosphere with isothermal compression Boyles law tells us B=1 atm =105 Pa.
In cases of adiabatic compression of an ideal gas at P0 = 1 atm: BA=P0 = 105 Pa. For diatomic
gases like N2 and O2 , the ratio of specific heats, = 1.4 . For monatomic gases like He = 1.67 .
Air, which is mostly diatomic gases has a specific heat of = 1.41 such that at 1 atm and 20 C,
BA= 1.41 105 Pa .
H. The Propagation Velocity of Sound, c, depends on the stiffness and density of the soundconducting medium, i.e.
c = B A 0 with units of m/s.

(1.3)

Why do we use the adiabatic compressibility?

At standard temperature and pressure (20C and 1 atm) c 340 m/s.
I. The wavelength of sound , depends on sound frequency and the propagation velocity,
(1.4)
= c f with units of meters per cycle
Since the human ear is sensitive to sound pressures of frequencies varying from 20 Hz to
20,000 Hz (aka 20 kHz), the wavelengths of the sounds to which we are sensitive vary from over 10
meters to less than 2 cm. This large variation from wave lengths that are much bigger than the
objects around us to wave lengths that are smaller than the objects around us is one of the challenges
of acoustics.

Frequency (Hz)

Wave length in
air

Comparable
Structure

Wave length in
water

Comparable
Structure

31.5

10.8 m

class room

47m

Olympic pool

100
315
1,000

3.4
1.08
0.34

bed room
torso & head
head

15
4.7
1.5

small yacht
small boat
small human

3,150
10,000
31,500

0.108
0.034
0.0108

vocal tract
ear canal
human TM

0.47
0.15
0.047

tuna
mackerel
anchovy

100,000

0.0034

ossicle/gnat

0.015

chum

J. The Characteristic Impedance, z0, is another property of the sound conducting medium that
depends on the stiffness and density of the medium, i.e.
z 0 = B A 0 = 0c .

(1.5)

The characteristic impedance of a media relates the sound induced variations in pressure and particle
velocity. For the special case of a plane wave propagating in the x direction in free open space
9-Sept-2004

Lecture 1

Acoustics of Speech and Hearing

6.551/HST 714J

p(t) = z 0v x (t) .

(1.6)

The unit of characteristic impedance is the rayl, named after Lord Rayleigh, where 1 rayl = 1Pa-s/m,
i.e. the ratio of a pressure and a velocity.
How does the characteristic impedance compare with electrical impedance and mechanical
impedance?
All of the formulae we have introduced thus far contain only real numbers. This simplification
is consistent with the plane-wave open-space constraint on Eqn 1.6, where that equation specifies
that the patterns of temporal variations in particle velocity and pressure are identical. As we will
soon see, there can be significant differences in the temporal patterns of velocity and pressure in
sound resulting in descriptions of impedance that depend on complex numbers.

100,002

100,000

A
B
S
O
L
U
T
E
P
R
E
S
S
U
R
E
(P
a
)

2. Simple Sounds: the Pure Tone.

Figure 1.2
The total pressure as a
function of time is
pT(t)=P0+p(t),
at a point x in space in the
presence of a propagating
plane wave produced by a
sound source playing a
continuous tone. The sound
pressure is zero at time t =... 0,
0.5, 1 ms ...

99,998

0.5
TIME (ms)

What is the frequency of the tone in Fig 1.2?

What is the peak amplitude of the sound pressure in Fig 1.2?
What is the root-mean-square or rms amplitude of the sound pressure in Fig. 1.2?
To compute rms amplitude: square the waveform, average it over some characteristic time,
and then take the square root of the average.
Describe the wave in Figure 1.2 in terms of a cosine function of time.
What is the amplitude/magnitude of the cosine function?
What is the frequency?
What is the phase angle?
A. Magnitude and Phase: Complex notation
The fact that sinusoids can vary in magnitude and phase can be coded in terms of a complex
number, e.g. a + jb, where a is the real part of the number, b is the imaginary part of the number and
9-Sept-2004
4

Lecture 1

Acoustics of Speech and Hearing

6.551/HST 714J

j = the imaginary number 1 . I use italic script to represent variables where lower-case variables
generally refer to real constants or functions of time and upper-case variables generally refer to
quantities that vary with frequency. (An exception to this rule is BA.) I denote complex variables
with an underbar, e.g. B = a + jb.
We can convert between the rectangular notion for complex numbers B = a + jb, and a
polar notation B = B e jB where the magnitude of the complex number B is

a2 + b2 ,

(1.7)

B = atan(b/a).

(1.8)

|B| =
and the angle of the complex number B is

We can convert a complex magnitude and angle back into the rectangular real and imaginary
components using Eulers Equations, such that for B = a + jb:
Real{B} = a = B cos(B), and
.
(1.)
Imaginary{B} = b = B sin(B)
B. Eulers Equations: the relationship between complex exponentials and sinusoids.
i. Eulers equations:
where j = 1

e j = cos + j sin
e j = cos j sin

ii. A specific example:

j t + )
Ae (
= A(cos(t + ) + j sin(t + ))

iii. Separation of (ii) into two complex

amplitudes, one that is constant and one
dependent on time

Ae j e jt = A(cos(t + ) + j sin(t + ))

iv. A complex amplitude B with magnitude |B|

and phase angle B

Be jt = B (cos(t + B) + j sin(t + B))

v. Description of a simple cosine function by a

Real Be jt = B cos(t + B)
complex exponential
Row (i) notes Eulers basic equations where a complex exponential amplitude of magnitude 1 and
angle noted as ej has a real part of cos and an imaginary part of sin.
Row (ii) gives another example of a complex exponential of magnitude = A and angle = (t+).
Note that the angle argument has a component that varies with time t and radian frequency
=2f, and a component that is constant .
In Row (iii) the complex exponential from the previous row is split into two components by simple
j t + )
algebra, i.e. Ae (
= Ae j Ae jt .
In Row (iv) we see the combination of two complex numbers B = B e jB and e jt yielding an
argument to the cosine and sine terms made up of the time varying t and the constant B.
Row (v) describes a cosine function of magnitude |B| and angle B as the real part of the product of
two complex exponentials, one constant and the other time dependent.
C. The Specific Acoustic Impedance ZS
9-Sept-2004

Lecture 1

Acoustics of Speech and Hearing

6.551/HST 714J

With the introduction of complex amplitudes we can now give a more general description of
acoustic impedance. Suppose a sinusoidal source of frequency f produces a sound wave that passes
through a point Q while propagating in direction x. We can define the temporal variations in the
sound pressure and the x-component of the velocity at point Q in terms of complex exponential
amplitudes, where =2f, and
p(t) = P cos(t + P ) = Real Pe jt , and

v x (t) = V cos(t + V ) = Real Ve jt .

(1.10&b)

The specific acoustic impedance relating sound pressure and particle velocity at point Q is
defined by the ratio of the complex amplitudes P and V, i.e.
(1.11)
ZS = PV .
Note that ZS is:
-complex (it has a magnitude and an angle),
-independent of time,
-like the characteristic impedance of the medium has units of rayls, and
-unlike the characteristic impedance that describes the propagation of sound in an infinite
expanse of a medium, the specific acoustic impedance can describe relationships between
p(t) and v(t) where the two time functions are out of phase. This is important when
describing sound near objects, i.e. not in the free-field.
D. Sound Intensity and Power.
A more general metric of the amplitude of a sound wave is the Intensity or energy per unit
time per unit area (joule/s/m2=watt/m2). This quantity is sometimes called power density. Why?
The average intensity of a sound wave is a real quantity related to the product of the sound
pressure and the particle velocity. We can define the instantaneous intensity in the x direction:
ix(t)=p(t) vx(t).
(1.12)
In the sinusoidal steady state:
i x (t ) = P cos(t + P ) V cos(t + V ) .
(1.13)
Using the identity
cos(a + b) + cos(a b)
, leads to
cos acosb =
2
PV
PV
i x (t) =
cos(2t + P + V ) +
cos(P V ) .
(1.14)
2
2
Since we are really interested in the intensity averaged over some time, we only need
consider part of Eqn 1.14. In particular note that the first term on the right-side of (1.14) is a cosine
function with a frequency that is twice the frequency of the pressure and velocity variation. The
temporal average of such a sinusoidal function is zero for each cycle, and that term does not
contribute to the average intensity which is a constant completely defined by the second term on the
right,

PV
cos(P V ) .
(1.15a)
2
Note that the average intensity depends greatly on the phase relationship between P and V . If
P=V then I =|PV|/2. If P andV differ by /2 then I =0.

I=

(1.15a) can also be written in terms of complex exponentials as

9-Sept-2004

Lecture 1

Acoustics of Speech and Hearing

{ }

1
I = Real PV * ,
2
*
where V is the complex conjugate of V , i.e. V * =|V | e jV .

6.551/HST 714J

(1.15b)

We can also use the Specific Acoustic Impedance relating P and V to define sound intensity,
where: Z S = P V . Substituting this relationship into 1.15b and realizing that V V*=|V|2 yields:
1
1
1 2
1 2
I = Real PV * = V Real Z S = P Real S .
2
2
2
Z

{ }

{ }

(1.16)

In the case of a plane wave Z S = z0 = 0c and is real.

The average sound power is I (with units of power/area) times the area of the power
collecting surface that is orthogonal to the direction of propagation of the wave.

E. The decibel, dB, as Measure of Pressure and Intensity.

The decibel is a logarithmic description of the ratio of two energy levels. The use of
logarithmic scales to describe sound pressure and intensity stems from two facts:
(1) The human ear responds to a wide range of sound intensities (about 6 orders of magnitude) and
humans generally like to describe quantities in as few significant units as possible. (2) Our ability to
distinguish differences in the intensity of two sound signals, can be roughly described in terms of a
threshold fractional difference in the intensity between the two, i.e. sounds of intensity that are
different by more than 25% are generally distinguishable while sounds of closer relative intensity are
not. This sensitivity to fractional changes is just what logarithmic scaling is all about, and is a
common approximation for human sensation. (Youll hear more about Webers and Fechners laws
of psychophysical detection later in this course.)
The bel is a unit named after Bostons own Alexander Graham Bell, where 1 bel describes an
order of magnitude change in energy. The decibel or dB breaks the bel into ten pieces where in the
case of sound intensity (which is proportional to energy):
I
The dB value of Intensity = 10log10 ,
(1.17)
I0
where I0 is some arbitrary reference intensity level.
A common intensity reference level in Speech and Hearing Science is the Sound Pressure Level
(SPL) reference of 10-12watts/m2.
Equation 1.17 can be used to compute the dB level re some reference for any quantity that is
proportional to energy, e.g. power or sound pressure. However intensity and power are proportional
to the square of sound pressure (see 1.16), and therefore to convert pressure measurements to some
dB term:
P 2
P
= 20log10
.
(1.18)
The dB value of Intensity = 10log10
PRef
PRef
The sound pressure reference level for Sound Pressure Level is 2x10-5 Pa. You should use equation
1.16 to convince yourself that the intensity and pressure references for SPL are consistent with each
other.
9-Sept-2004

Lecture 1

Acoustics of Speech and Hearing

6.551/HST 714J

3. Periodic Complex Signals

Two Cycles of a Periodic Wave Form

The signal has an rms amplitude of 0.27 Pa.

What is its peak amplitude?

0.8
0.6
SOUND PRESSURE (Pa)

Figure 1.3 Shows two periods of a periodic

complex acoustic signal that can be
characterized as an exponentially decaying
tone. The period of each repetition is 10 ms, the
period of the tonal component of the signal is 1
ms.

0.4
0.2
0
-0.2
-0.4
-0.6
-0.8

0.005

0.01
TIME (SECONDS)

0.015

0.02

What are the frequencies within the complex signal of Figure 1.3?
What is the repetition frequency?
What is the frequency of the tonal component?
Are there other frequencies?
A. Fouriers Theorem applied to periodic signals
Any periodic signal of period T can be reconstructed from the sum of a static component and
a series of sinusoidal components that are harmonics of the repetition frequency. In general:
For p(t) with period T,

(1.19)
1
p(t) = P0 + P n cos n2 + P n

T
n=1
Fourier analysis of the time wave form in Figure 1.3 yields the magnitude and angle spectra below,
where the lines on the left show the magnitude of the complex Fourier components Pn and the pluses
on the left show the angles of the same components.

Figure 1.4 Bode plots of the Magnitude and Angle of the Fourier Components
that describe p(t) in Fig. 1.3
ANGLE SPECTRUM OF 2.2

MAGNITUDE SPECTRUM OF 2.2

-1

ANGLE (radians)

PRESSURE MAGNITUDE (Pa)

2
10

10 -2

-1

-2

-3

10
10 2

9-Sept-2004

FREQUENCY (Hz)

FREQUENCY (Hz)

10

10 3

Lecture 1

Acoustics of Speech and Hearing

6.551/HST 714J

B. Summing sound pressures of different frequencies

The rms value of p(t) in Figure 1.3 can also be computed from the magnitude spectrum by
summing the intensities of the spectral components of Figure 1.4 components and converting back
to pressure.
1
2
2
Prms
0c = P n / 0c

2
1
2

Prms =
P n = 0.27 Pa

(1.20)

As a simpler example, suppose we have two sound sources one of 1000 Hz and another of 1500 Hz
that each produce tonal amplitudes of 2 Pa at some point Q, such that at Q:
p1 (t ) = 2cos(21000 t ), p2 (t ) = 2cos(21500 t )
The rms sound pressure when either source is active by itself is 1.41 Pa.
What is the rms sound pressure when both sources are turned on?
Since it is the intensity that adds: the total intensity is:
ITotal =
Or simply

1
2

P12 1
+
z0 2

P22
rms
; PTotal = 2 z0 ITotal ; PTotal
= z0 ITotal ;
z0

(1.21)

rms
2
2
= P1,rms
+ P2,rms
= 2.0Pa .
PTotal = P12 + P22 , PTotal

4. The Propagation of Sound in Time and Space: The Wave Equation.

A wave equation describes the variation of an acoustic variable in time and space. The wave
equations we will be using use the basic assumptions of linear acoustics (i.e. the sound-induced
variations in p,T and are small compared to P0,T0 and 0) and also assumes that air (and water)
are inviscid (i.e. that sound propagation is primarily determined by the density and compressibility
of the fluid and is not dependent on fluid viscosity).
The one dimensional wave equation for sound pressure:
2 p(x,t) 0 2 p(x,t)
1

(1.22)
=
, where 0 = 2 .
2
2
B A t
BA c
x

The derivation of the one dimensional wave equation for sound in a fluid includes:
(i) An acoustic version of Newtons second law
p( x,t)
v ( x,t)
= 0 x
x
t
(ii) an acoustic version of the conservation of mass:
(x,t)
v (x,t)
=
0 x
t
x

(iii) a relationship between compressibility and sound-induced changes in density and pressure:

(x,t) = p(x,t) 0 .
BA

Those of you interested in the full derivation can look at the following pages.
9-Sept-2004

Lecture 1

Acoustics of Speech and Hearing

6.551/HST 714J

Topic 2: The Equations of Linear Acoustics; Derivation of the wave equation

Advanced Reading: Beranek Chapter 2; Kinsler et al. Chapter 2
Another view: Fletcher, Chapter 6

A. Define acoustic variables

p(x, y,z,t) = pTotal (x, y,z,t) P0 ,

v (x, y, z,t) = v x (x, y, z, t)ix + v y (...)iy + vz (...)iz ,
( x, y, z, t) = Total (x, y,z,t) 0 and
T(x, y,z,t) = TTotal (x, y,z,t) T0

sound pressure:
particle velocity:

sound density:
sound temperature:

B. The 'linear' assumption

The primary assumption of linear acoustics is that all variations in acoustic scalar quantities
are small relative to the static equilibrium quantities, i.e.
p(x, y, z, t) << P0 ; (x, y, z, t) << 0 ; T(x, y, z, t) << T0
C. The 'inviscid' assumption
Another common simplifying assumption is that the viscosity of air is so small that we can
consider the medium inviscid. This assumption is generally valid except for tubes of very small
cross-sectional area, when determining the 'Q' of a resonant acoustic system (which depends strongly
on damping) or considering the propagation of sound over some large distances.
D. Derivation of the one-dimensional wave equation
Assume a sound wave propagating in the x direction down a long rectangular duct of crosssectional area S (Figure 1.2.1), with a 'wave-front' orthogonal to the long axis. Under these
conditions, the sound pressure and particle velocity within any slice where x is constant are
invariant and we can describe the system in terms of variations in a single dimension:

p(x, y, z, t) p(x, t) , and

v (x, y, z, t) v x ( x, t) i x + 0 i y + 0 iz v x (x, t).

(1.2.1a)
(1.2.1b)

S=z y
y

( 0 ,0 ,0 )

x+x

Figure 1.2.1: A long duct of height y, width z and undetermined length. Our derivation of the wave
equation is based on a section of duct described by the interval x to x+x.
In order to characterize the two unknowns p(x, t) and v x (x,t) , we need two constraining
equations. We use Newton's second law as one constraint and the conservation of mass together
with the elastic properties of the medium for the second constraint.
9-Sept-2004
10

Lecture 1

Acoustics of Speech and Hearing

6.551/HST 714J

1. Constraint 1: Newton's second law:

Consider a small section of the duct of length x.. Newton's second law reads:

Force = m

dv x (t)
dt .

The net force acting on the air in the small section of the duct is:
Force = S (Difference in pressure at x and x )
= S ( p(x,t) p(x + x,t))

(1.2.2)

(1.2.3)

The right side of equation (1.2.2) can be described in terms of the volume of the section of duct
(Sx) and the total density of air, leading to:
dv (x,t)
.
(1.2.4a)
S ( p(x,t) p(x + x,t ) = Sx Total (x,t) x
dt
Substituting Total(x,t)=0+(x,t) into (1.2.4a) yields:

dv x (x,t )
(1.2.4b)
dt
Dividing each side of (1.2.4b) by the volume (Sx), assuming (x,t) << 0 and taking the limit as x
goes to zero, yields:
S(p(x,t ) p(x + x,t )) = Sx ( 0 + (x,t ))

p(x, t)
dv (x, t)
= 0 x
.
x
dt

(1.2.5)

The derivative on the right of (1.2.5) can be rewritten in terms of partial derivatives as

dv x (x, t)
v (x,t) vx (x, t)
= v x (x, t) x
+
x
dt
t , and
since we assume vx(x,t) is small:

dv x (x, t) v x (x, t)

.
dt
t

(1.2.6)

The final result relates pressure and velocity via two first order differential equations, one in space
and the other in time:

p(x, t)
v (x,t)
= 0 x
.
x
t

(1.2.7)

The acoustic version of F=ma!!

9-Sept-2004

11

Lecture 1

Acoustics of Speech and Hearing

6.551/HST 714J

2. Constraint Two:
a. The conservation of mass.
In a sound field we can also define a flux J (the mass flowing per unit area per unit time) as
the product of the total density and the particle velocity

J = Total v .

(1.2.8)

(The units of flux are kg/m2/s).

In the one dimensional system of Figure 1.2.1, we need only worry about the x component of the
vectors J and v ; and we can rewrite (1.2.8) as:
(1.2.9)
J x ( x,t) = Total ( x,t) v x ( x,t) .

According to the conservation of mass, the net flux into a volume (Sx) during time t must equal
the change in mass defined by the product of the volume and the change in total density during time
t:
(J x (x, t) J x (x + x, t))t S = (T (x, t + t) T (x, t))x S . (1.2.10)
Dividing both sides of (1.2.10) by x t S, and taking the limit as both x and t go to zero yields;
J x ( x,t)
( x,t)

.
(1.2.11)
= Total
x
t
Substituting (1.2.9) into (1.2.11) and noting that

Total
= ;gives us our final description of the
t
t

conservation of mass or continuity equation for a one-dimensional linear acoustic system:

(x,t)
v x ( x,t)
.
(1.2.12)
=
t
x
Given a box of medium, the change in density as a function of time is proportional to the net
velocity of particles entering and leaving the box.
0

9-Sept-2004

12

Lecture 1

Acoustics of Speech and Hearing

6.551/HST 714J

b. The elastic properties of the medium

In general the density of a gas (like air) depends on pressure. For example, the Ideal Gas Law
relates the total pressure pT and volume Vol of the gas to some physical constants, n the number of
moles of the gas, R the universal gas constant and TTotal the absolute temperature:
pTotal(x,t) Vol=nRTTotal(x,t) .
(1.2.13)
The gas law can be rewritten in terms of density :
RTTotal
pTotal = Total
,
(1.2.14)

where M is the molecular weight of the gas.

If TTotal were to remain constant (isothermal conditions) then the relationship between pTotal and
Total is especially simple, such that:

pTotal Total
=
.
p0
0

(1.2.15)

However, the rapid changes of pressure associated with sound at most audible frequencies does not
permit heat exchange either within the sound field or between the sound field and the environment.
Therefore the relation between pTotal and Total must be defined for circumstances of no heat flow
(adiabatic conditions). For an ideal gas under adiabatic conditions:

pTotal Total
=
,
p0
0

(1.2.16)

where is the ratio of specific heats and equals 1.41 for an ideal diatomic gas. Equation 1.2.16 can
be rewritten to include acoustic variables:

0 +
p0 + p 0 +
=
or p0 + p = p0

p0
0
0
A Taylor series expansion of the right side of (1.2.17) yields:
2 2 pTotal
p
p0 + p = p0 + Total +
+ ...
Total 2 2Total

(1.2.17)

(1.2.18)

Since is small the higher terms can be ignored, leaving

p
( x,t)

or (x,t) = p(x,t) 0
(1.2.19)
p(x,t) = (x,t) Total = B A
Total
0
BA
p
where B A = 0 Total . For an ideal gas, B A = p0 . For a diatomic gas at 1 atmosphere of pressure
Total
BA=1.4x105 Pa.
Directly from the gas law: P = n/V RT, or n/V = P/(RT).

9-Sept-2004

13

Lecture 1

Acoustics of Speech and Hearing

3. Synthesis:
Newtons 2nd Law:

6.551/HST 714J

p(x,t)
v (x,t)
.
= 0 x
x
t
(x,t)
v ( x ,t)
.
=
0 x
t
x

Conservation of Mass:

(x,t) = p(x,t)

Elasticity Relationship:
Substituting 19 into 12 yields:

0
.

(1.2.7)

(1.2.12)
(1.2.19)

v x (x,t)
1 p(x,t)
.
(1.2.20)
=
x
B A t
Equations (1.2.7) and (1.2.20) describe the partial derivatives of pressure and velocity with respect
to x in terms of two equations, two properties of the media (0 and ) and two unknowns (the partial
derivatives with respect to time). Taking the partial of both sides of (1.2.7) with respect to x and
substituting (1.2.20) into the result yields a one-dimensional "wave equation" in x and t,

2 p(x,t) 0 2 p(x,t)
1

=
, where 0 = 2 .
2
2
B A t
BA c
x

(1.2.21)

One possible solution for (1.2.21) is in terms of two wave functions traveling in opposite directions:

p(x,t) = f + (t x /c ) + f (t + x /c ) ,
v x (x,t) =

and

Things to note about 1.2.22

1). z0 = B A 0 = 0c ,
2).
3.)
4.)
5.)
6.)

9-Sept-2004

1 +
f (t x /c ) f (t + x /c ) ,
z0

(1.2.22a)
(1.2.22b)

The wave functions f+ & f- are functions in time and space but the units to function
argument is seconds.
In the forward traveling wave f+, for any t the argument (t-x/c) is smaller for larger xs.
In the backward traveling wave f- for any t the argument (t+x/c) is larger for larger xs.
While the scalar pressures produced by the two oppositely traveling waves add, the
directional velocities are of opposite sign and subtract.
While the pressure and velocity components of each traveling wave are related by z0,
the sum terms p(t,x)/vx(t,x) may not be.

14