You are on page 1of 37

Distance Learning Program for

Professional Education in Acoustics

Module 1
General Principles of
Acoustics

Acoustics and Vibration Unit


UNSW, Canberra ACT 2600
Australia

Issue date January 2016


General Principles of Acoustics Section 1

The program is based on a similar program that has been offered via Universities
and the UK Institute of Acoustics (IOA). This first module, General Principles of
Acoustics, maintains the structure of the IOA module but has been revised and
updated as well as including Australian references as appropriate. This material has
been produced under the guidance of an Advisory Committee comprising members
of Australian Acoustical Society and Association of Acoustical Consultants.

The module comprises 9 Sections of Notes plus Tutorials and Experiments that are
provided as separate pdf files

All correspondence on this module to be forwarded to


Acoustics and Vibration Unit
UNSW, Canberra
Canberra ACT 2600
Australia

Tel 02 6268 8241


Mob 0402 240009
Email avunit@adfa.edu.au

© Copyright UNSW
This material has been copied by or on behalf of the University of
New South Wales pursuant to Part VB of the Copyright Act 1968 (the
Act). The material in this communication may be subject to copyright
under the Act. Any further copying or communication of this material
by you may be the subject of copyright protection under the Act.

GPA Jan 2016


General Principles of Acoustics Section 1

1 Basic Concepts in Acoustics


Section 1 introduces the basics concepts and principles of acoustics. It is important
to have an understanding of these basic principles. The notes attempt to minimise
the equations focussing on the important ones. If you do find the mathematics a little
daunting remember that the aim is to gain an overview so try to focus on what the
equation shows about relationships between the variables.

1.1 Wave motion - a qualitative view


Acoustics is the science of sound, and sound is a wave motion. In this section we
shall try to understand what happens in a very simple wave motion. It will be an
exercise that requires you to use your imagination.

The type of wave we are going to describe are waves on the surface of water, simply
because such waves are familiar and can be seen. Imagine a stone being dropped
into the middle of a very large pond. We see ripples moving outwards over the
surface of the water. The stone has disturbed the water and the ripples are the
visible signs of the disturbance travelling outwards from the source of disturbance
over the surface of the water. The waves travelling outwards away from the source
are called progressive waves. If the pond is very large we do not need to worry
about waves being reflected at the edge of the pond.

Now imagine a cork on the surface a little way from where the stone is to be dropped.
When the stone is dropped an initial ripple starts its outward journey and after a little
while (dependent on how fast the ripples travel and the distance of the stone from the
cork) it reaches the cork. The cork now starts to bob up and down and continues to
do so after the initial ripple has moved on and successive ripples take its place. The
up and down motion of the cork is an example of a type of motion called vibration.
The cork was still (i.e. undisturbed) but now it is bobbing up and down. This requires
energy. The cork has gained energy from the wave. This is part of the energy
transferred into wave motion by the stone hitting the surface of the water. If we
spread lots and lots of corks all over the surface of the pond at different distances
from the stone and in different directions, we would see lots and lots of corks bobbing
up and down. The collective motions (vibrations) of all these corks, with each related
but slightly different, is a representation of the wave generated by the falling stone.

Returning to our single cork, eventually the bobbing up and down motion dies away,
just as the ripples also die away. This is because there is always some sort of
frictional mechanism involved in the movement and, furthermore, the disturbance
caused by the stone is a 'one off’. The energy given to water by the stone is
gradually used up in overcoming these frictional forces. If there was no friction then
we would have perpetual motion, but there is always some frictional mechanism
involved in any type of motion, and so perpetual motion is impossible.

The frictional process involved in vibration motions is called damping. Because of


its 'one off and dying away nature' the motion of the cork is called a damped transient
vibration, and the cork represents a transient wave disturbance.

Notes based on the General Principles of Acoustics Module of the Institute of Acoustics, UK
General Principles of Acoustics Section 1

Now imagine a second cork further away from the stone than the first. It will also
start bobbing up and down when the initial ripple reaches it, but not as vigorously
(i.e., with lower amplitude). There are two reasons for this. Firstly because, as the
wave (i.e., the ripple on the surface) travels further from its source, the energy in the
wave spreads out more, or disperses. Therefore the cork that is further away
receives a lesser disturbance, and hence less energy, than the cork which is closer.
Secondly, as the wave travels through the water it is subjected to the same sort of
frictional mechanisms as affected the motion of the individual cork. In this case
however, we refer to the frictional process as absorption and we say that some of the
energy in the wave is absorbed. In both damping of the vibration or absorption of the
wave motion, the energy used in overcoming the frictional mechanism is turned into
heat, in the same way that friction between our hands when rubbed together also
generates heat.

1.2 Wavefronts, wave velocity and rays


The outermost ripple on the pond is called the wavefront - it tells how far the wave,
and the disturbance it creates, has reached. In front of the wavefront the water is, as
yet, undisturbed. In this example the wavefronts are circular and the waves are
called two dimensional, because they only exist on the two dimensional surface of
the pond. Later you will be asked to imagine their three dimensional equivalent, the
invisible spherical ripples or wave-fronts which travel through the air from a point
source of sound like an expanding balloon.

The velocity of the wave is the speed at which the wavefront moves forward,
measured in metres per second, written as either m/s or ms-1. The energy in the
wave is travelling in a direction which is at right angles to the wavefront, along an
imaginary ray. In this case the rays representing the direction of wave travel are the
radii to the circular wavefronts (with centres at the point at which the stone was
dropped), showing that the wave is spreading out equally in all directions.

Rays

Wavefronts

Figure 1.1 Sketch illustrating rays and wavefronts for circular and spherical waves.

GPA Jan 2016 Section 1 page 1.2


General Principles of Acoustics Section 1

1.3 Understanding and describing wave motion


Can we understand how the motions of the various corks are related to each other?
If we know exactly what is happening to one cork at a particular moment, can we
easily work out what is happening in the case of a nearby cork at the same moment,
or a little bit later?

Before we attempt to visualise what is happening we need to make a couple of


simplifications, and more imagination is needed. To get rid of the complication
caused by the wave spreading out, let us imagine the stone is being dropped into the
water of a long straight canal, and let us further assume that there is no friction either
in the water or at the sides of the canal. If we can forget about the width of the canal
(which is very short compared to its length) then we can imagine that the ripples from
the stone will form straight lines moving away from the stone along the length of the
canal. We now have a model of a plane wave - in other words, a one dimensional
wave with plane wavefronts, which is travelling in one direction only - down the canal.
The one dimensional plane wave is the simplest of all waves to describe, whether
qualitatively (as we shall now proceed to do), or mathematically, which we shall
attempt later.

Ray

Wavefronts
Figure 1.2 Sketch illustrating rays and wavefronts for plane waves.

In order to help us describe the wave, imagine a line of corks spread at equal
distances apart down the centre line of the canal. If we focus attention on any one
cork, it performs a vertical 'bobbing up and down' motion, which is cyclical, i.e. after a
while the motion repeats itself. Because there is no damping (as there is no friction)
this motion never dies away and repeats indefinitely. And, because there is no
spreading out of the wave energy and no absorption (as there is no friction), all of the
corks will go through exactly the same cycle of motion with the same amplitude of
motion.

However, because the wavefront which set the corks in motion takes some time to
travel from one cork to the next, the motion of adjacent corks will be a little out of step
(i.e. out of phase) with one another.

1.3.1 Variation of cork displacement in time and in space


We can visualise what is happening along the line of corks at any moment. It is
helpful to try to separate out the time and space variations, and to think of a picture in
time and a picture in space. In doing so we shall learn about two very important
properties of the wave: its wavelength and its frequency. First of all, let us think of

GPA Jan 2016 Section 1 page 1.3


General Principles of Acoustics Section 1

how the position of any one cork varies with time. The following graph shows how
the vertical displacement of the cork (the 'bobbing up and down'), y, varies with time.

y Amplitude

Time

Period, T

Figure 1.3 Variation of displacement at a fixed location with time for a wave.

After a time T, called the period of the motion, the cycle repeats. The frequency, f,
of the vibration or wave is the number of cycles of motion occurring in a second:
f=1/T
Frequency is measured in cycles per second or Hertz (abbreviation Hz.).

The amplitude of the vibration is the maximum value of y, i.e. the maximum
movement of the cork from its original 'at rest’ position, and is also known as the
'peak' value of the vibration. The second way of describing what is happening is to
think about the position of every cork in the line at one particular moment.

y Amplitude

Distance

Wavelength
λ
Figure 1.4 Variation of displacement at a fixed time with distance for a wave.

The above graph shows the variation of displacement, y, with distance along the
canal for the corks at a fixed time. It shows that, at any particular time, the different
corks are at different points in their cycle of vibration, with the difference depending
on how far apart they are. However two corks which are a certain distance apart are
at the same points in their cycle, i.e. they are in phase. This distance is called the
wavelength of the wave, usually represented by the Greek letter λ.

GPA Jan 2016 Section 1 page 1.4


General Principles of Acoustics Section 1

1.3.2 The transition from ripples on ponds to sound waves in air


Consider the movement of the air particles in a perfectly smooth-walled pipe, tube or
duct (the equivalent of our one dimensional canal for the water waves). There is a
sound source at one end of the tube which sends sound waves down the tube - in
the sketch below the source is shown as a vibrating piston. This source causes the
air particles next to it to vibrate to and fro and this disturbance (the to and fro motion)
passes down the tube. As a result of the to and fro motion sometimes the air
particles are bunched together, causing a very slight increase in pressure in the tube
(termed a compression) and sometimes causing them to be spaced further apart,
causing a very slight reduction in pressure (termed a rarefaction). These very small
fluctuations in pressure constitute variations in sound pressure caused by the
passage of a sound wave. Hence the motion of the piston source causes a series of
compressions and rarefactions to propagate as a waveform along the tube, as
indicated in the sketch below.

Extracted from IOA Diploma Notes


Figure 1.5 Representation of a longitudinal wave in a tube with a piston at one end.

But there is an important difference between the motion of the water particles on the
surface of the canal and that of the air particles in the tube, to do with the direction of
particle motion relative to the direction in which the wave is travelling. In the case of
the waves on the water surface, the water particles move to and fro in a vertical
direction whilst the wave travels in a perpendicular direction, i.e. horizontally along
the canal. Such waves are called transverse waves. In the case of the air particles
in the tube, the to and fro motion of the particles is in the same direction as that in
which the wave is travelling, i.e. along the tube. Such waves are called longitudinal
waves, and the difference between these two types of wave is shown schematically
in the next figure.

GPA Jan 2016 Section 1 page 1.5


General Principles of Acoustics Section 1

Direction of
Direction of
particle motion
particle motion

Transverse wave Longitudinal wave

Figure 1.6 Direction of particle movement for transverse and longitudinal wave.

The following graphs show respectively how sound pressure at one position in the
tube varies with time, and how it varies along the tube at any fixed time, for the
simplest case when the motion of the piston produces a pure tone (i.e. a single
frequency sound). As with the similar graphs used earlier to describe the motion of
corks in the water, they are used below to explain the frequency and wavelength of
the sound.

Amplitude or Peak Value


Sound
Press.
p

Time, t

Peak to
Peak value Period T

Figure 1.7 Variation in pressure with time at one position.

The graph of sound pressure, p, versus time, t, at one position in space may be
represented by the equation:
p = Psin(ωt)
where ω = 2πf = 2π / T is the angular frequency, and P is sound pressure amplitude.

GPA Jan 2016 Section 1 page 1.6


General Principles of Acoustics Section 1

Sound Amplitude or Peak Value


Press.
p

Distance, x

Wavelength
λ
Figure 1.8 Variation in pressure with distance at any one time.

The graph of sound pressure, p, versus distance from the source, x, at one moment
in time may be represented by the equation:
p = Psin(kx)
where k = 2π / λ is termed the wave number, and P is the sound pressure amplitude.

Note that, in general, the sound pressure at any position, x, and time, t, is given by
the formula p = Psin(ωt – kx).

1.3.3 The magnitude of sound pressures - RMS and peak values


If a pure tone is being played over the radio and the volume is turned up, the
amplitude of the sound pressure is be increased - the sound becomes louder. The
amplitude is thus a convenient measure of the magnitude of the sound and can be
related to its intensity and loudness, terms which will be discussed later.

With a more complicated waveform, however, it is not so easy. One might think that
the magnitude of the peak pressure of the waveform would be the value which would
be most useful. However, the sound pressure might be near to the peak value for
only a small fraction of the duration of the sound, and might not be very closely
related to the subjective impression of the sound. Perhaps an "average" sound
pressure would be a better measure of the "size" of a sound? However if we look at
the sinusoidal waveform of a pure tone, we see that, taken over a complete cycle, the
average sound pressure, including rarefactions and compressions, is zero. This is
true for all waveforms, not just a pure tone. We need an "average" which takes into
account the magnitude of the sound pressure fluctuations but not their sign (positive
and negative) so that the compressions and rarefactions do not average out. There
are various possible ways of obtaining a "non-zero" average sound pressure, but the
one most commonly used is the root-mean square (abbreviation RMS) sound
pressure. This can best be described by looking at the waveform shown in the
diagram below.

GPA Jan 2016 Section 1 page 1.7


General Principles of Acoustics Section 1

sound pressure squared

Sound
Press. Peak Sound
p Pressure

Mean p squared

RMS Sound Pressure

Time

Figure 1.9 Comparison of the pressure and the pressure squared signal with time.

In effect the sound level meter first "squares" the signal, that is multiplies it by itself.
This has the effect of producing a pressure squared waveform, which is always
positive - remember that in algebra minus one multiplied by minus one gives plus
one. The next stage is to take the average (or mean value) of this pressure squared
waveform - called the "mean pressure squared". Finally, by taking the square root of
this value, we get back to a pressure - the root mean square or RMS pressure
(strictly the square root of the mean pressure squared) The process is illustrated in
the diagram above.

Most sound level meters have electronic circuits which convert the microphone signal
into an RMS value corresponding to the RMS sound pressure. The RMS pressure is
used because it can be related to both the average intensity and the loudness of the
sound. For a pure tone it can be shown that the peak pressure and the RMS
pressure are simply related by:
prms = ppeak / √2 = 0.707 ppeak.
For more complex signals, there is no such simple relationship between the two.

Despite what has been said above, there are occasions when it is important to
measure the peak value of a complex sound waveform, or the peak to peak value.
This may be the case in particular for loud impulsive noise, such as gunfire,
explosions or punch presses, and most sound level meters allow for the
measurement of peak as well as RMS sound pressures. The crest factor is the ratio
of the peak amplitude of a waveform to the RMS value. It is a measure of the
sharpness of the peak, so that short intense impulses will have high values of crest
factor.

GPA Jan 2016 Section 1 page 1.8


General Principles of Acoustics Section 1

1.4 The physics of sound waves


So far we have described what happens in a wave. In order to understand the “why”
of wave motion we have to look at the mechanics of the situation. Sound is a
particular form of wave motion, and the important characteristics of waves in general
are that they:
(1) involve the transmission of a small change; or disturbance, in some physical
property, and
(2) also involve the transmission of energy although, in the case of sound waves,
the amounts of energy involved are very small. This is because the forces and
pressures involved are so minute, requiring an instrument of great sensitivity,
like the human ear, to detect them.

Sound waves in air belong to a family of waves known as elastic waves which,
unlike electromagnetic waves for example, require a physical medium, i.e. a solid,
liquid or gas, in which to exist and propagate. In other words, sound cannot travel
through a vacuum. To support the propagation of elastic waves the medium needs
three essential features - it has to have mass (i.e. inertia), elasticity and some form of
damping.

The physical nature of one dimensional sound waves can be modelled in a simple
and idealised way by considering the medium to consist of a series of interconnected
masses and springs, rather like a long line of railway trucks with buffers in between
them. Putting aside the property of damping for the moment, the wave motion may
be understood in terms of the interplay between the inertia of the masses, i.e. their
resistance to change of motion, and the forces exerted by the springs (aptly called
restoring forces in this case) which always try to restore the masses to their original
position if they are disturbed by some external event. If one of the masses is
disturbed, i.e. given a push or a jolt, it will start to move and this movement is
transmitted, via the adjacent springs, to the neighbouring masses, which in turn pass
the disturbance on to their neighbours, and so on.

This is the wave motion, and it is characterised by the following important features:
• Each mass moves to and fro, i.e. vibrates, about its original rest position.
• The disturbance passes along the line with a characteristic speed or velocity.
This wave velocity, c, depends on the size of the masses and the stiffness of
the springs or, more generally, on the density of the medium and its modulus
of elasticity.
• Assuming that there is no damping, each of the vibrating masses vibrates in
exactly the same way except that, because the wave velocity is finite, there is
a small time lag, or delay, between each vibrating motion. Thus the form, or
shape, of the wave is preserved as it propagates and this exactly what is
seen, for example, with waves on the surface of water, as well as waves
travelling along a rope or long coiled spring.
• If, instead of a single transient impulse, the disturbance is a sustained
vibration then each mass vibrates with the same amplitude and frequency but,
because of the time delays, the cyclic motion of each mass is slightly out of
step, i.e. out of phase, with its neighbour.
• This phase difference increases with increasing separation between masses
until the cumulative effect, for two masses a certain distance apart, is a phase
difference corresponding to one complete cycle, i.e. 360 degrees – in which

GPA Jan 2016 Section 1 page 1.9


General Principles of Acoustics Section 1

case these two masses are vibrating in step, or in phase (i.e. with a phase
difference of zero degrees). The distance separating these two masses is
known as one wavelength.
• In reality there is always some friction or damping in the process, otherwise
there would be perpetual motion. Consequently there is some attenuation or
loss of amplitude as the wave propagates through the medium.
• It is important to distinguish between the motion of the disturbance travelling
through the medium, which is characterised by the wave velocity, and the
individual to-and-fro motions of the masses, which is characterised by their
instantaneous displacement and velocity at any point in their cycle.

1.4.1 The relationship between sound speed, frequency and wavelength


In order for masses which are one wavelength apart to be in phase, the wave must
travel one wavelength in the time that it takes for any one of the masses to complete
one cycle of motion. Since the number of such cycles completed in one second is
the frequency of the wave, and wave velocity is the distance travelled by the wave in
one second, it follows that frequency, f, wavelength, λ, and wave velocity, c, are
related by the well-known equation:
c=fλ
For sound waves in air, the speed of sound ranges between 330 and 340 metres per
second, depending upon air temperature. Thus for a frequency of 100 Hz, at the
lower end of the audio range, the wavelength will be about 3.3 metres (about the size
of a small car) whereas at the much higher frequency of 1000 Hz it is about 0.33
metres – so the lower the frequency the greater the wavelength, and vice versa.

Example:
A plane sound wave in air has a frequency of 660 Hz. Taking the velocity of sound in
air as 330 m/s, what is the phase difference:
(a) between two points in the wave separated by a distance of 0.125 metres,
at the same moment?
(b) at the same position, but separated by 0.0001 seconds?

Firstly calculate the wavelength = sound velocity/frequency = 330/660 = 0.5 metres.

(a) In terms of phase, a distance of one wavelength, 0.5 metres, corresponds to a


phase difference of 360 degrees. Therefore a distance of 0.125 metres corresponds
to a phase difference of 360 x (0.125 / 0.5) = 90 degrees.

(b) At a frequency of 660 Hz, there are 660 cycles per second so the time period for
one cycle is 1/660 = 0.00152 seconds. In terms of phase, a time period of one cycle
of the vibration, i.e. 0.00152 seconds, corresponds to a phase difference of 360
degrees. Hence a time separation of 0.0001 seconds corresponds to a phase
difference of 360 x (0.0001 / 0.00152) = 23.7 degrees.

GPA Jan 2016 Section 1 page 1.10


General Principles of Acoustics Section 1

Exercise:
a) What is the frequency of the sound if the phase difference between two points that
are 0.3 metres apart is 180 degrees?
Answer 550Hz.
b) For a sound wave in air with a frequency of 1,000 Hz, what is the phase difference
at the same position, but separated in time by 0.0003 seconds?
Answer 108 degrees.

Note that the frequency of a sound wave is determined only by the source of the
sound, but the sound velocity depends on the medium through which the wave is
travelling. These two factors then determine the wavelength in the medium,
according to the equation c = f λ. If the sound moves from one medium to another,
e.g. from air into water, the frequency will remain the same in both media but,
because of the difference in sound velocities, the wavelengths in the two media will
be different.

1.4.2 The motion of particles in a wave, displacement velocity and acceleration


Each individual particle involved in sound wave propagation through the medium
performs a to-and-fro oscillation, in other words it vibrates. If the sound is a pure
tone, i.e. a single frequency sound, then the vibration is the simplest possible and is
called simple harmonic motion. A harmonic motion is one that is repetitive, and the
simplest type of repetitive motion is one in which the graph showing how the
instantaneous particle displacement, x, varies with time, t, is sinusoidal, that is it
follows the shape of a sine wave. We demonstrated earlier, but it may also be
proved by a more mathematical treatment of the subject, that this graph may be
represented by the equation:
x = Xsin(2πft)
where f is the frequency of the vibration and X is the displacement amplitude.

Note that the use of the term 'displacement' to represent the movement of the
vibrating particle (rather than distance) implies a displacement from a fixed rest
position (where x = 0) and that that a direction is also involved because the
displacement may be either positive or negative.

In order to find out how fast the particle is moving, i.e. its velocity, we need to find out
the rate at which its displacement is changing with time. The use of the term velocity,
rather than more simply speed, indicates that the direction of movement is included
because velocity is speed in a given direction.

In order to find the velocity averaged over a certain time interval, or over a certain
distance, simple mathematics will suffice: i.e. velocity is change in displacement
divided by the time interval. However, in order to find the instantaneous velocity at
any one moment it is necessary to calculate the average over an infinitesimally small
time interval. This process of finding instantaneous rates of change by considering
changes over infinitesimally small intervals is fundamental to the branch of
mathematics called calculus, where it is termed differentiation and represented by
differential coefficients such as dx/dt.

GPA Jan 2016 Section 1 page 1.11


General Principles of Acoustics Section 1

Indeed, the instantaneous particle velocity, v, at time t is the infinitesimal rate of


change with respect to time of displacement, x, at time t, and is calculated by
differentiating the expression x = Xsin(2πft) as follows:
v = dx/dt = 2πf Xcos(2πft) = Vcos(2πft)
where the velocity amplitude, V = 2πfX. The appearance of the cosine instead of the
sine function indicates the quarter cycle (90 degree) change of phase between the
displacement and velocity waveforms.

Acceleration is the rate at which velocity changes with time, and provides yet another
way of describing the motion of the vibrating particle. The instantaneous particle
acceleration, a, is the instantaneous rate of change of velocity, v, with respect to
time, and is calculated by differentiating the expression v = Vcos(2πft) as follows:
a = dv/dt = – 2πf Vsin(2πft) = – Asin(2πft)
where the acceleration amplitude, A = 2πfV. The sine function proceeded by the
minus (–) sign in this equation indicates that there is a half cycle (i.e.180 degree)
phase change between the acceleration and displacement waveforms.

It is possible to combine the relationships between X and V, and between V and A, to


obtain a third relationship between A and X, namely:
A = 4π2f2X

Example:
a) Calculate the velocity and acceleration amplitudes of a particle which is vibrating
with displacement amplitude of one micron (10-6 m) at a frequency of 100 Hz.

Velocity amplitude, V = 2πfX = 2πx100x10-6 = 6.28x10-4 m/s or 0.63 mm/s.


Acceleration amplitude, A = 2πfV = 2πx100x6.28x10-4 = 0.40 m/s2.

b) Calculate the velocity and displacement amplitudes of a particle which is vibrating


with acceleration amplitude of 10 m/s2 at a frequency of 1000 Hz.

Velocity amplitude, V = A / (2πf) = 10 / (2πx1000) = 0.00159 m/s or 1.59 mm/s.


Displacement amplitude, X = V / (2πf) = 0.00159/(2πx1000) = 0.25x10-6 m or 0.25
microns.

Exercise:
a) Calculate the velocity and acceleration amplitudes of a particle which is vibrating
with displacement amplitude of one mm (10-3 m) at a frequency of 125 Hz.
Answer: V = 0.78 m/s A = 616 m/s2.

b) For a particle with a velocity amplitude 1 m/s and displacement amplitude of 1 mm


determine the frequency of vibration.
Answer: f = 160 Hz

GPA Jan 2016 Section 1 page 1.12


General Principles of Acoustics Section 1

1.4.3 Summary.
It is worth gathering these three important formulae together:
Displacement, x = Xsin(2πf t) = Xsin(ωt)
Velocity, v = Vcos(2πf t) = Vcos(ωt)
Acceleration, a = – Asin(2πf t) = – Asin(ωt)
where ω = 2πf = angular frequency, in radians per second

Note that the velocity waveform is 1 / 4 cycle (90 degrees) out of phase with the
displacement waveform. The acceleration waveform is 1 / 4 cycle (90 degrees) out
of phase with that of the velocity, thus 1 / 2 cycle (180 degrees) out of phase with the
displacement waveform. These relationships are shown schematically in Figure 1.10
where, for Displacement amplitude X:
Velocity amplitude V = 2πfX
Acceleration amplitude A = 2πfV = 4π2f2X

A
X Acceleration
v, Displacement Amplitude
V
x, Amplitude Velocity
a Amplitude

Time, t

Velocity
v Displacement
Acceleration x
a

Figure 1.10 Variation in displacement, acceleration and velocity with time.

GPA Jan 2016 Section 1 page 1.13


General Principles of Acoustics Section 1

1.5 Acoustic pressure, intensity and impedance


In our simple mass-spring model of a medium propagating a one-dimensional elastic
wave, the particles (i.e. masses) in their undisturbed state are a constant distance
apart. A disturbance temporarily either crowds the particles more closely together, or
moves them further apart. This results in small changes of pressure in the medium,
and the wave may be described in terms of a series of compressions (i.e. increases
in pressure) or rarefactions (i.e. reductions in pressure) passing through the medium.
The small change in pressure caused by the passage of the sound wave is called the
acoustic pressure, or the sound pressure. A sound wave in air therefore consists of
a series of small fluctuations in the steady atmospheric pressure which would
otherwise prevail in the absence of the sound wave.

The sound pressure is obviously related to the motion of the particles in the medium
which cause it, and is most easily related to the particle velocity. These two
quantities are related by the specific acoustic impedance of the wave, z, which is the
ratio of the acoustic pressure, p (measured in Pascals, abbreviation Pa), to the
acoustic particle velocity, v (measured in m/s): hence z = p / v.

The response of the human ear to sound is discussed in Section 3, but it is


interesting at this point to consider the ear’s range of sensitivity. The healthy young
ear can just detect a sound pressure of 20 micropascals (abbreviation μPa) and
withstand a pressure of 200 Pa; so its sensitivity range is a ratio of 107:1. In terms of
frequency range, the healthy young ear can hear frequencies as low as 20 Hz up to
as high as 20 kHz, but with varying sensitivity across this range.

For a wave travelling in one direction, i.e. a plane wave, the specific acoustic
impedance depends only upon the nature of the medium, and may be shown to be
given by z = ρc, where ρ is the density of the medium in kg/m3 and c is the velocity of
sound in the medium in m/s. For air the value of ρc is 415 Nsm-3 at 20 degrees
Celsius, although a value of 420 Nsm-3 is commonly used and, by way of
comparison, the acoustic impedance of water is a much larger 1.5x106Nsm-3.
Combining the original definition z = p / v with the relationship z = ρc for plane waves,
we find that:
p = ρc v = z v

Vibrating particles possess mechanical (i.e. potential and kinetic) energy, and the
transmission of a disturbance through a medium involves the flow of energy. The
rate of flow of energy, in other words the energy transmitted per unit time, is
measured in Watts (abbreviation W). At any point in the medium, the sound
intensity, I, in any given direction is defined as the rate of flow of energy per unit area
in that direction and is measured in W/m2. For a plane wave, and at a point where
the acoustic pressure is p and acoustic particle velocity v, the sound intensity is given
by:
I=pv
Using the previous specific acoustic impedance relationship between p and v, this
becomes:
I = p2 / ρc = ρc v2

GPA Jan 2016 Section 1 page 1.14


General Principles of Acoustics Section 1

Example:
1) Calculate the acoustic particle velocity and acoustic intensity at a point in a plane
wave where the sound pressure is 0.001 Pa (use a value of 420 Nsm-3 for the
specific acoustic impedance of air).

Particle velocity, v = p / ρc = 0.001 / 420 = 2.38x10-6 m/s.


Acoustic intensity, I = p2 / ρc = (0.001)2 / 420 = 2.38x10-9 W/m2.

2) Calculate the sound pressure and the acoustic particle velocity at a point in a
plane wave where the sound intensity 1x10-6 W/m2.
p2 = Iρc = (1xl0-6) x420 = 4.2x10-4 Pa2
so therefore p = √(4.2x 10-4) = 0.02 Pa, and
V = I / p = 1x10-6/0.020 = 5.0x10-5 m/s.

Exercise:
a) Calculate the acoustic intensity at a point in a plane wave where the sound
pressure is 20 μPa (note this corresponds with the pressure that can just be detected
by the human ear).
Answer: I = 0.9x10-12 W/m2.

b) For a plane wave in air, what is the sound pressure when the sound intensity is
3x10-3 W/m2?
Answer: p = 1.1 Pa.

1.6 The relationships between sound power, intensity and pressure


for a simple point source under free-field conditions
Reminder: Sound intensity is sound power per unit area. Therefore if sound power
W passes through an area S then the sound intensity, I, is given by I = W / S.

If we assume an idealised model of a sound source, i.e. a simple point omni-


directional source which radiates sound energy equally in all directions, and we also
assume free field conditions, i.e. with no reflecting or absorbing surfaces nearby, the
wave-fronts radiated by such a source will be spherical. The surface area of a
sphere of radius r is 4πr2. Therefore, substituting this for S in the above equation, we
find the sound intensity at a distance r from a point source of power W is I =W / 4πr2.

Note that when there are reflecting/absorbing surfaces present the sound field is
more complex and the relationships in this section do not apply. These more
complex reverberant and diffuse fields will be discussed in a later Section of the
Module.

GPA Jan 2016 Section 1 page 1.15


General Principles of Acoustics Section 1

Example:
What is the acoustic intensity, and the corresponding sound pressure, at a distance
of 10 metres from an idealised point source of sound power of 1 Watt under free field
conditions?
I = W / 4πr2 = 1.0 / (4πx102) = 7.96x10-4 W/m2

In order to obtain the sound pressure we assume that, at a distance of 10m from the
source, the spherical waves will have flattened out sufficiently to be considered as
plane waves, in which case I = p2 / ρc
p2 = Iρc = 7.96x10-4x420 = 0.33 Pa2
so p = √(0.33) = 0.58 Pa

Exercise:
What is the sound power for an idealised point source that, under free field
conditions, at a distance of 10 metres has acoustic intensity of 10-12 W/m2?
Answer: W = 1.26x10-9 W

What is the acoustic intensity of this sound at 20 m and compare this with the
acoustic intensity at 10 m from the source?
Answer: I = 2.5x10-13 W/m2 which is a quarter the sound intensity at 10 m.

1.6.1 The inverse square law


The equation I = W / 4πr2 tells us that the acoustic intensity I is inversely proportional
to the distance r from the source. Therefore if the distance from the source doubles,
the sound intensity reduces to one quarter of its value, because the sound power has
been spread over a sphere of four times the area. This is known as the inverse
square law relating sound intensity and distance from a spherically-radiating source.
Provided that the intensity I1 at one particular distance r1 is known, the relationship
can be used to find the intensity I2 at any distance r2 via the equation I2 / I1 = (r1 / r2)2.

Example:
a) If the sound intensity at a distance of 10m from the source is 7.95x10-4 W/m2, what
is the sound intensity at a distance of 25 m?
I1 = 7.95x10-4 W/m2 r1 = 10 m r2 = 25 m
-4 2
I2 / (7.95 x10 ) = (10/25)
from which I2 = 1.27 XI0-4 W/m2.

b) How far will it be before the intensity drops to 5x10-5 W/m2?


I1 = 7.95x10-4 W/m2 r1 = 10 m I2 = 5x10-5 W/m2
-5 -4 2
(5.0 x10 ) / (7.95 x10 ) = (10 / r2)
0.0629 = 100 / (r2)2
(r2)2= 100 / 0.0629
from which r2 = 39.9 m from the source.

GPA Jan 2016 Section 1 page 1.16


General Principles of Acoustics Section 1

Exercise:
If the sound intensity at a distance of 10 m from the source is 10-12 W/m2, what is the
sound intensity at a distance of 7 m from the source?
I1 = 10-12 W/m2 r1 = 10 m r2 = 7 m
Answer: I2 = 2.04x10-12 W/ m2.

Since intensity is proportional to pressure squared, and inversely proportional to


distance squared, it follows that pressure squared is inversely proportional to
distance squared and therefore sound pressure is inversely proportional to distance.

This can be used to find the sound pressure p2 at any distance r2 provided that the
sound pressure p1 at one particular distance r1 is known, i.e. p2 / p1 = (r1 / r2).

Exercise:
If the sound pressure at a distance of 10m from the source is 20x10-6 Pa, what is the
sound pressure at a distance of 20 m from the source?
p1 = 20x10-6 Pa r1 = 10 m r2 = 20 m
-
Answer: p2 = 10x10 6 Pa.

1.7 The decibel scale


The decibel scale is used for comparing and measuring powers (electrical as well as
acoustic) and related quantities such as sound intensity and pressure, so it is
necessary to explain the relationships between these quantities.

Sound sources have a tremendous range of sound powers, varying from about 10-9
W for the human voice when whispering, to millions of Watts radiated by a space
rocket during launch. The human voice radiates about 20x10-6 W during
conversation, increasing to about l0-3 W when shouting. A pneumatic drill used for
road breaking radiates about 1 W, whilst a typical figure for the noise output of a jet
airliner is about 5 x104 W.

Audible sounds can have a very wide range of intensities, from 10-12 W/m2 (the
threshold of hearing for an average person) to more than 100 W/m2 (approaching the
threshold of pain) - a ratio of more than a million million to one.

Sound intensity is a useful quantity because it can be related to the sound power of
the noise source, and is one of the important factors in subjectively assessing the
'loudness' of sound. However, although sound intensity is important as the basis of
many prediction calculations, sound pressure is a more useful quantity in practical
terms and is the quantity measured when using the microphone of a sound level
meter.

The decibel scale is a logarithmic (abbreviation log) scale for measuring or


comparing energies, powers or related quantities such as sound intensity, and is
used extensively in acoustics. There are a number of options for logarithmic scales
and the one used for sound is the logarithm to base 10, denoted by the symbol lg, in
which case lg10 = lg101 = 1 The use of base 10 means that each time a value is
multiplied by 10 its logarithm increases by 1, so that:

GPA Jan 2016 Section 1 page 1.17


General Principles of Acoustics Section 1

lg1 = lg100.= 0 lg10.= 1 lg 100 = lg102 = 2 lg 1000 = lg 103 = 3 etc.

If you need revision on logarithmic scales – what they mean and how to use them,
then refer to the Annex to this chapter before proceeding

One reason using a logarithmic unit for sound measurement is that of convenience.
For example, by using a logarithmic scale, the very large range of audible sound
pressures corresponding to an even larger range of sound intensities is compressed
into a much more manageable range of about 120 dB.

Another reason is that the human response to sound is also logarithmic, with each
tenfold increase (i.e. 10 dB) in sound intensity being judged, on average, to double
the loudness of the sound. Hence a hundred-fold (102) increase, i.e. 20 dB, would
produce a four-fold increase in loudness and a thousand-fold (103) increase, i.e. 30
dB, would increase the loudness by a factor of 8. A 1 dB increase is only just
noticeable under the most favourable listening conditions but a 3dB increase,
corresponding to a doubling of sound intensity, produces a small but noticeable
increase of loudness in most situations.

Measurements on the decibel scale are calculated via the equation:


N = 10 lg(I2 / I1)
which means that N dB separates two sounds of intensities I1 and I2. Noting that
sound power, W, is directly proportional to sound intensity, l, and using relationships
to sound pressure, p, this equation can be expressed as:
N = 10 lg(W2 / W1)
N = 10 lg(p22 / p12) = 10 lg(p2 / p1)2 = 20 lg(p2 / p1)
where W1 and W2 are the sound powers, and p1 and p2 the sound pressures, for the
two sounds.

1.7.1 Reference values for decibel scales


The statement that a sound of intensity I2 produces a noise level which is 10 dB
higher than a sound of intensity I1 uses the decibel scale in a relative way, without
assigning an absolute value to either of the two levels. Whilst this may be useful, it is
preferable to have an absolute value to describe the sound. The use of
internationally agreed reference values gives an absolute value to quantities
measured on a decibel scale. Reference values are used in place of I1, W1, and p1 in
the above equations to ensure that, providing there are no measurement
inaccuracies, a measurement anywhere in the world of the same sound will have the
same dB value. Reference values are given the subscript ‘0’ or ‘ref’, so that they
may be expressed as I0, W0, and p0 or, on other occasions, as Iref, Wref or pref.

The reference value for airborne sound pressure, p0, is 20x10-6 Pa, i.e. 20 μPa. This
again relates the decibel scale to the subjective reaction to sound, as it represents
the threshold of hearing at a frequency of 1 KHz for the average young person with
normal hearing. Using this value as the reference means that p2 = p0 for any sound
at the threshold of hearing which, as lg1 = 0, corresponds to 0 dB.

Values of sound pressure measured on a decibel scale relative to this reference


value are termed sound pressure levels, and are denoted by SPL or Lp. Similarly

GPA Jan 2016 Section 1 page 1.18


General Principles of Acoustics Section 1

the reference value, W0, for sound power levels (denoted by Lw) is 10-12 W, and the
reference value, I0, for sound intensity levels (denoted by LI) is I0-12 W/m2.
Therefore for any sound, the appropriate absolute decibel level scales are:
sound pressure level Lp=20 lg(p / p0) = 20 lg(p / (20x10-6)) dB re 20x10-6 Pa.
sound power level LW=10 lg(W / W0) = 10 lg(W / (1x10-12)) dB re 10-12 W.
sound intensity level LI=10 lg(I / I0) = 10 lg(I / (1x10-12)) dB re 10-12 W/m2.
These three scales represent three different physical quantities, i.e., sound pressure,
power and intensity, although in each case the measurement is in dB. In order to be
precise, the reference value should be quoted as shown above to make clear what
scale is being used. In practice, it is usually assumed that the scale being discussed
is clear from the context, so the reference value is often not included. For this reason
it is important to avoid the use of “sound level”, as this is ambiguous unless there is a
statement at the beginning of a document noting something like:
…..in this report the use of the term “sound level” refers to sound pressure
level re 20 x10-6 Pa.

It is also important to be aware that logarithmic scales are also used in electronics
and for sounds underwater. The reference values for the decibel scale in these fields
are not the same as described above for airborne sound. In particular it should be
noted that the reference for sound pressure levels underwater is 1x10-6 Pa.

It is also important to realise that sound power level, sound pressure level and sound
intensity level for any sound source will not necessarily have the same numerical
value in all situations.
• The sound power level relates to the sound energy inherent in the source
under particular operating conditions and remains essentially independent of
location. For this reason the sound power level is quoted in specifications for
items.
• The sound pressure level for the same source varies depending on the
distance from the source and the acoustic characteristics of the space into
which the sound is propagating. All sound level meters are essentially sound
pressure level meters as the single microphone responds to the sound
pressure. Sound pressure level is most commonly used and is the basis for
criteria such as for occupational and environmental noise.
• Similarly, the sound intensity level for the same source varies with distance
and the acoustic characteristics of the space but also includes an specification
of direction. There are sound intensity meters with special microphone
probes, usually comprising a pair of microphones. These are particularly
useful for identification of specific source locations within large machines, or
when trying to locate pathways for sound through partitions.

GPA Jan 2016 Section 1 page 1.19


General Principles of Acoustics Section 1

Example:
1) Determine the sound pressure level for a point with sound pressure of 5x10-3 Pa.
Lp = 20 lg(p / p0) = 20 lg(5x10-3 / 20x10-6) = 20 lg(2.5x102)
= 20 (0.398+2) = 47.96 dB re 20x10-6 Pa.

2) Calculate the sound intensity level where the sound intensity is 8.5 x 10-7 W/m2.
LI = 10 lg(I / I0) = 10 lg(8.5x10-7 / 1x10-12) = 10 lg(8.5x105)
= 10 (0.929+5) = 59.29 dB re 1x10-12 W/m2.

3) Calculate the sound pressure and sound intensity at a point in a plane wave at
which the sound pressure level is 75 dB. Take the specific acoustic impedance of air
as 420 N/m2.
Lp=20 lg(p / p0) from which p / p0 = 10(Lp / 20)
p = p0 10(Lp / 20) = 20x10-6x10(75 / 20) = 20x10-6x103.75 = 20x10-2.25
= 20x 0.00562 = 0.112 Pa
and
I = p2 / ρc = (0.112)2 / 420 = 0.0126 / 420 = 3x10-5 W/m2.

4) Calculate the sound pressure and sound intensity at a point in a plane wave at
which the sound intensity level is 75 dB.
LI=10 lg(I / I0) from which I / I0 = 10(LI / 10)
(LI / 10)
I = I0 10 = 1x10 x10(7 5 / 10)
-12

= 1x10-12x107.5 = 1x10-4.5 = 3.16x10-5 W/m2


and
I = p2 / ρc from which p2 = I ρc
p = √(3.16x10 x420) = 0.115 Pa.
-5

Exercise:
Calculate the sound pressure and sound intensity at a point in a plane wave at which
the sound pressure level is 65 dB.
Answer p = 0.0356 Pa I = 3.02 x10-6 W/m2.

1.7.2 Combining sound pressure levels


When a number of noise sources are operating simultaneously it becomes necessary
to consider how individual sound pressure levels are combined. Since decibel values
are based on logarithms, we cannot simply add the levels. Therefore, if the sound
pressure level at a point from one source is 80 dB, then the sound pressure level at
the same point from two co-located identical sources will not be 160 dB! It is the
sound pressure that doubles, not the sound pressure level. So to find the overall
sound intensity or sound pressure level at a point from a number of individual sound
levels it is necessary to calculate the sound intensity or sound pressure squared
value for each source, add these up and then determine the overall sound level.

As shown in the examples above the sound pressure for any one sound pressure
level can be found from the following:

Lp1 = 20 lg(p1 / p0) = 10 lg(p1 / 20x10-6)2 p12 = (20x10-6)2x10(Lp1/ 10)


in which case the combined or total sound pressure squared, pT2, from n sources is
given by:

GPA Jan 2016 Section 1 page 1.20


General Principles of Acoustics Section 1

pT2 = (20x10-6)2x(10(Lp1 / 10)+10(Lp2 / 10)+10(Lp3 / 10)+…+10(Lpn / 10))


and the total sound pressure level, LpT, for the n sources is given by:
LpT = 10 lg[(20x10-6)2(10(Lp1 / 10)+ 10(Lp2 / 10)+ 10(Lp3 / 10)+…+10(Lpn / 10)) / (20x10-6)2]
which becomes
LpT = 10 lg(10(Lp1 / 10)+ 10(Lp2 / 10)+ 10(Lp3 / 10)+…+10(Lpn / 10))

While this may seem tedious there are some “rules” which can be derived to make
the task easier.

Consider the case when the only two sound pressure levels are equal, i.e. the
difference in sound pressure levels is 0 dB so that Lp1 = Lp2, then:
LpT = 10 lg(10(Lp1 / 10)+10(Lp1 / 10)) = 10 lg(2x10(Lp1 / 10))
= 10 lg(2)+10 lg(10(Lp1 / 10))
= 3+Lp1.
So irrespective of the actual values of the sound pressure levels at a particular point,
the combination of two identical sound pressure levels is 3 dB greater than one
sound pressure level. Just as 50 dB plus 50 dB leads to a total of 53 dB, similarly 80
dB plus 80 dB gives 83 dB. Working through the mathematics in a similar manner
you can find what the total sound pressure level will be when the dB difference in
sound levels is 1, 2,…, etc.

This leads to a simple way of determining the overall sound level at a point for pairs
of sounds by using the following Table. It gives the amount that must be added to
the highest level based upon the difference between the two levels. Although this is
only an approximate method, it gives results that are accurate to the nearest dB,
which is satisfactory for most purposes. It is important to remember that you always
add the value from the Table to the highest of the pair of sound levels.

Table for combining decibels.


Difference in sound levels, dB to be added to the
dB highest sound level
0, 1 3
2, 3 2
4, 5, 6, 7, 8, 9 1
10 and greater 0

GPA Jan 2016 Section 1 page 1.21


General Principles of Acoustics Section 1

Example:
a) What is the total sound level at a point when the individual sound levels at that
point are 60 and 64 dB?
For 60 and 64 the difference is 4 so from the table you add 1 to the higher level
giving a total of 64+1=65 dB.

b) If the source that was 64 dB has an enclosure over it, so that its level is reduced
by 8 dB to 56 dB, what is the overall sound level from the two sources?
For 60 and 56 the difference is 4 so from the table you add 1 to the higher level
giving a total of 60+1 = 61 dB. So in this case the overall sound level has been
reduced by 4 dB as the remaining source of 60 dB has now become the dominant
noise.

c) Now return to the original two sources, namely 60 and 64 dB. If the source that
was 60 dB has an enclosure over it, so that its level is reduced by 8 dB to 52 dB,
what is the overall sound level from the two sources?
For 64 and 52 the difference is 12 and from the table you add 0 to the higher level
giving a total of 64 dB. So in this case the overall sound level has been reduced by
less than 1 dB.

This highlights the importance of first reducing the noise from the louder sources as
little overall benefit is gained by reduction of the quieter ones.

Such a procedure may be used for the addition of more than just two sound levels at
a point as long as you start with a pair of levels to determine the sub-total, and then
add the next sound level to that sub-total until eventually you have the total sound
level of all the sounds. This is more accurate if you first arrange the sound levels in
ascending numerical order and start combining from the bottom of the sequence.

Example:
What is the overall sound level at a point if four sources are operating together: when
individually they contribute sound levels of 82, 88, 76 and 84 dB at that point?

First arrange the sound levels in ascending numerical order: 76, 82, 84 and 88 dB.
Taking the first pair, 76 and 82 the difference is 6, so from the table the amount to be
added is 1 giving a total of 82+1=83 dB.
This 83dB is then combined with the next level in the sequence, 84.
For 83 and 84, the difference is 1, so from the table the amount to be added is 3
giving a total of 84+3=87 dB.
This 87dB represents the combination of 76, 82 and 84 dB and has to be combined
with the remaining 88dB to complete the calculation.
For 87 and 88, the difference is 1, so from the table the amount to be added is 3
giving a total of 88+3=91 dB.
Thus the combination of 76, 82, 84 and 88 dB is 91 dB.

GPA Jan 2016 Section 1 page 1.22


General Principles of Acoustics Section 1

Exercise:
What is the overall sound level at a point if four sources are operating together: when
individually they contribute sound levels of 72, 78, 66 and 74 dB at that point?
Answer 81 dB.

1.7.3 Averaging sound pressure levels


Sometimes it is necessary to find the average value of a number of sound pressure
level measurements. A good example would be in building acoustics where, in order
to find a representative value of the sound pressure level in a room, measurements
are taken at a number of different positions within the room, and an average value is
calculated.

Again, the logarithmic basis of the dB must be considered and it is not accurate to
just average the numerical value of the dB levels. It is necessary to determine the
average sound intensity or the average value of p2.
For sound level Lp1 =10 lg(p1 / 20 x10-6)2 p12 = (20x10-6)2x10(Lp1 / 10).
So the average sound pressure, pAv2, for n measurements is given by:
pAv2 = (20x10-6)2x(10(Lp1 / 10)+10(Lp2 / 10)+10(Lp3 / 10)+ … +10(Lpn / 10)) / n
and hence the total sound pressure level, LpAv for the n measurements is given by:
LpAv = 10 lg[(20x10-6)2x(10(Lp1 / 10)+ … +10(Lpn / 10)) / n) / (20x10-6)2]
which means:
LpAv = 10 lg[(10(Lp1/ 10)+10(Lp2 / 10)+ … +10(Lpn / 10)) / n)].

Example:
Calculate the average of the following four decibel levels: 82 dB, 84 dB, 86 dB, and
88 dB.
LpAv = 10 lg[(10(82 / 10)+10(84 / 10)+10(86 / 10)+10(88 / 10)) / 4)]
= 85.6 dB.
Note that arithmetic average of these levels is 85 dB.

Exercise:
Calculate the average of four decibel levels: 82 dB, 84 dB, 86 dB, and 92 dB.
Answer LpAv = 87.8 dB.
Note that arithmetic average of these levels is only 86 dB.

GPA Jan 2016 Section 1 page 1.23


General Principles of Acoustics Section 1

1.7.4 Inverse square law and decibels


The inverse square law has already been discussed in relation to the sound
distribution from a point source, so that sound intensity I at distance r from a source
of sound power W is given by:
I = W / 4πr2
This can be used to find the intensity I2 at any distance r2, provided that the intensity
I1 at one particular distance r1 is known, via the formula I2 / I1 = (r1 / r2)2. As intensity
is proportional to pressure squared, this can be expressed as:
(p2 / p1)2 = (r1 / r2)2

Converting from pressure to pressure levels and then rearranging, this becomes
(20x10-6)2x10(Lp2 / 10) / (20x10-6)2x10(Lp1 / 10) = (r1 / r2)2
in which case
Lp2 – Lp1 = 20lg(r1 / r2).

A useful relationship for reduction in sound level with a doubling of distance, i.e.
when r2=2r1, can now be determined as:
Lp2 – Lp1 = 20lg(r1 / 2r1) = 20lg(2-1) = – 6 dB
which is commonly known as the “6dB per distance doubled” rule.

Example:
If the sound pressure level for a point source is 89 dB at 10m what will be the sound
pressure level at a distance of 160 m?
Lp2 = Lp1 +20lg(r1 / r2) = Lp1 – 20lg(r2 / r1)
In this case: Lp1 = 89 dB, r1 = 10m and r2 = 160m so that:
Lp2 = 89 – 20lg(160/10) = 89 – 24 = 65 dB.

Checking against the 6 dB per doubling of distance rule:


10 to 20m would be 89 – 6 = 83dB, 20 to 40m would be 83 – 6 = 77dB.
40 to 80m would be 77 – 6 = 71dB, 80 to 160m would be 71 – 6 = 65dB.

Exercise:
If the sound pressure level for a point source is 78 dB at 120 m, what will be the
sound pressure level at a distance of 20 m?
Answer Lp2 = 93.6dB.

GPA Jan 2016 Section 1 page 1.24


General Principles of Acoustics Section 1

1.7.5 Relationship between sound pressure level and sound power level
For a source of known sound power level, an estimate of the sound pressure level at
distance r in a free field can be obtained from the two relationships I=W / 4πr2 and
(assuming the spherical wavefront is sufficiently flattened) I=p2 / ρc as follows:
Lp =10lg(p12 / p02) = 10lg(p12 / (20x10-6)2) = 10lg(p12 / 400x10-12)
Noting that p12 = Iρc = W1x420 / 4πr2 this becomes
Lp =10lg(W1 / 10-12)+10lg(420 / (400x4πr2)
As the reference for sound power level is 10-12 W the expression becomes
Lp = LW – 20lg r – 11

Example:
Calculate the sound pressure level generated at a distance of 10 metres from a
sound source of sound power level of 120 dB under free-field conditions.
Lp = LW – 20lg r – 11
In this case LW = 120 dB and r = 10 so that:
Lp = 120 – 20lg(10) – 11 = 89 dB.

Exercise:
Calculate the sound pressure level generated at a distance of 40 metres from a
sound source of sound power level of 120 dB, under free-field conditions.
Answer Lp= 77 dB.

1.8 Complex waveforms and frequency analysis


The pure tone is a single frequency sound, and is the simplest sort of sound. It is
produced, for example, by a tuning fork or by a loudspeaker fed with a sinusoidal
voltage signal. Most sounds have a waveform which is more complicated than the
simple sine wave, and they can be considered to contain more than one frequency.

The next simplest type of sound is that produced by a musical instrument, playing
one single note. The waveform is harmonic, that is it repeats itself, but in a more
complicated way than the sine wave of the pure tone waveform. The French
mathematician, Fourier, showed that such a waveform could be "built-up" or
"synthesised" by combining together a number of simple sinusoidal waveforms. The
frequencies of these components are the fundamental frequency (the repeating
frequency of the complex waveform) and multiples of it, called harmonics. The
diagram below shows a harmonic waveform (a “square wave”) and illustrates how it
may be produced from a fundamental and two harmonics.

GPA Jan 2016 Section 1 page 1.25


General Principles of Acoustics Section 1

Result of addition of 3 sine waves

Waveform to be
synthesised.

Figure 1.11 Synthesis of a harmonic wave form, showing how a square


waveform can be built up from a series of sine waves. With the
fundamental plus the first two harmonics the resultant shape is
approaching the shape of the square wave. The addition of more
harmonics would improve the approximation to the square wave
shape.

The reverse process to Fourier synthesis is Fourier analysis, in other words the
"breaking down" or analysing of complex waveforms into their component
frequencies. Complicated waveforms can be analysed numerically using computers,
or the analysis can also be performed using frequency analysis instrumentation
attached to sound measuring equipment. These frequency analysers consist of a
series of electronic filters which allow sounds within particular frequency to be
measured.

1.9 The frequency spectrum of sounds


A frequency analysis may be displayed on a graph called the frequency spectrum
showing the amplitude of the different frequencies or frequency bands. The
frequency spectrum and the waveform can be considered as two alternative ways of
describing a sound.

Just as the waveforms of sounds vary from single pure tones to more complex
noises, so do frequency spectra vary, although they can be broadly categorised into
the following four types.
● Pure-tone waveforms are the simplest and their frequency spectra consist of
one single line indicating the amplitude of the single frequency.

● Harmonic waveforms also have a relatively simple line frequency spectrum,


showing the fundamental frequency and the harmonics. The spectrum also
indicates the relative amplitudes of the fundamental and harmonics, and thus
the spectra of the same note played on two different musical instruments may
be different.

GPA Jan 2016 Section 1 page 1.26


General Principles of Acoustics Section 1

Time Waveforms Frequency Spectra


Pure Tone
Sound
Press. Sound
Press.

Time

f Frequency

Harmonic Wave Sound


Sound
Press. Press.

Time

f 2f 3f Frequency

Sound Sound
Press. Random Noise Press.

Time

Frequency

Sound
Sound Press.
Press.

Impulse
Impulse

Time

Frequency

Figure 1.12 Time curves versus frequency spectra for different waveforms.

GPA Jan 2016 Section 1 page 1.27


General Principles of Acoustics Section 1

● Random waveforms are generated by processes which are random, i.e. never
exactly repeat themselves, and hence the waveforms are not repetitive.
Examples are noise produced by traffic, by wind and by many sorts of machine.
Even when the noise is produced by some repetitive event such as the rotation
of an engine or of gear teeth, the noise often has a random component because
the machine process never repeats exactly from cycle to cycle due to slight
changes in speed, load and other conditions. Random noise contains a little of
all frequencies and so its frequency spectrum is a continuous curve, called a
broad-band spectrum.

● Transient waveforms die away quickly to zero over time because their sound
energy is dissipated by frictional processes. They arise from sources which
provide only short bursts of acoustic energy, such as a plucked violin string, or
the impact of a hammer blow onto a rivet. The simplest transient waveform
resembles a sine wave but with an amplitude which decays over time.
Examples are the noise produced by impulses from punches, presses,
hammering and mechanical handling of goods. The frequency spectra of
transient sounds are complicated. For a plucked violin string the spectrum will
obviously contain the fundamental frequency and its harmonics whilst for a
repeated transient, such as the impacts between teeth in a gear mechanism,
the repetition rate of the impacts and its harmonics will also be important.

1.9.1 Frequency analysis and frequency spectra


In most cases noise has a broad-band spectrum, i.e. it contains a mixture of all
frequencies, but with some frequencies more than others. To investigate its
frequency content in more detail, we split the frequency range of the noise into bands
using frequency filters (an ideal filter is shown below) and measure sound pressure
level in each band. This is called frequency analysis, and the graph of sound
pressure level versus frequency across the bands is called the frequency spectrum of
the noise. There are two ways of splitting the frequency range into bands; either on a
constant bandwidth basis, or on a constant percentage bandwidth.

Pass Band – zero


attenuation of signal

Zero transmission Zero transmission


of signal of signal

fl fu
Frequency

Figure 1.13 An ideal filter.

GPA Jan 2016 Section 1 page 1.28


General Principles of Acoustics Section 1

1.9.2 Constant bandwidth filters


In the constant bandwidth method, each frequency band has the same width, so that
if for example the bandwidth was 100 Hz, then one of the bands might be from 100
Hz to 200 Hz, and others might be from 200 to 300 Hz or from 1100 to 1200 Hz or
from 10000 to 10100 Hz. The centre frequency of each band is half way between its
upper and lower cut off points, so that for the 100 to 200 Hz band the centre
frequency is 150 Hz.

This constant bandwidth approach is usually only used for “narrow-band frequency
analysis”, in which the main purpose is to identify with precision a particular
frequency component in the spectrum which might be causing a disturbance, such as
an annoying hum, whine or a vibration. Knowing the specific frequency can aid the
diagnosis of the noise source, for example by linking it to a particular fan, motor or
turbine via the rotational speed or number of rotors.

Typical constant bandwidths might be 300 Hz, I00 Hz, 10 Hz, 3 Hz, or 1 Hz. Some
filter banks might be arranged in contiguous frequency bands, e.g. 100 Hz to 200 Hz,
200 Hz to 300 Hz, 300 Hz to 400 Hz, etc. For others the centre frequency may be
continuously variable, allowing a “frequency sweep” in which the bandwidth remains
the same but the centre frequency changes.

1.9.3 Constant percentage bandwidth filters


Covering the entire audio range (20 to 20,000 Hz) using constant bandwidth filter
banks would involve either a small number of bands with very wide bandwidths, or a
great number of bands of narrow bandwidth. Constant percentage bandwidth
measurement is often a more convenient way to perform frequency analysis, and is
similar to the way the ear responds to broad-band sound. The commonly used
constant percentage bandwidths are based on octave and third octave frequency
bands, although one sixth, one twelfth and one twenty-fourth octaves are also
available.

Octave frequency bands are constant percentage bandwidths covering all


frequencies between a lower frequency fl and an upper frequency fu related by:
fu = 2 fl
The internationally-defined system of octave bands for sound measurement
comprises a series of contiguous bands with centre frequencies fc, each one octave
apart. Hence octave bands commonly used in noise measurements and
assessments are centred on frequencies of 63 Hz, 125 Hz, 250 Hz, 500 Hz, 1 kHz
(i.e. 1000 Hz), 2 kHz and 4 kHz.

When we plot an octave band spectrum, with sound pressure level in dB on the
vertical scale and octave band centre frequency on the horizontal scale, we assign
equal intervals of space to each octave along the horizontal axis. In doing this we
are, in effect, plotting frequency on a logarithmic (abbreviation log) frequency scale.

Unlike constant bandwidth filters, constant percentage bandwidth filters get wider (in
terms of the frequency range within each band) as frequency increases. For
example, the octave band with 1 kHz centre frequency covers a range of frequencies
ten times that of the octave band with 100 Hz centre frequency.

GPA Jan 2016 Section 1 page 1.29


General Principles of Acoustics Section 1

An important property of a constant percentage bandwidth filter is that its centre


frequency fc is not half way between the upper and lower cut-off frequencies. In fact
the relationship between fc, fl and fu is a geometric average rather than an arithmetic
average, so that:
fc 2 = fl fu
If the constant percentage bandwidth filter is also an octave band filter, then we know
that fu = 2fl
Hence fc2 = fl 2fl = 2fl2 and fc2 = (fu / 2) fu
in which case fu = fc√2 and fl = fc / √2.

Thus for the octave bands with centre frequency:


1000 Hz fl = 707.1 Hz and fu = 1414.2 Hz.
2000 Hz: fl = 1414.2 Hz and fu = 2828.4 Hz.

Exercise:
Calculate the upper and lower frequencies for the octave bands with centre
frequencies 100 Hz and 10 KHz.
Answer: fc= 100 Hz fl = 70.7 Hz and fu = 141.4 Hz.
fc= 10KHz fl = 7.1KHz and fu = 14.1 KHz.

When more precise information is required, one third octave analysis is undertaken.
As the name implies each octave is simply split into three constituent bands. The
one third octave bands are also described by their centre frequency so that, for
example, the one third octave bands of 100, 125 and 160 Hz cover the same
frequencies as the octave band with centre frequency 125 Hz. This is shown below
for a part of the octave band sequence:

Centre Frequency, Hz
Octave Band 125 250 500 1K
1/3 Octave Band 100 125 160 200 250 315 400 500 630 800 1K 1.25K

The geometric average relationship fc2 = fl fu remains true for third octave bands, but
the relationship between fl and fu is a little more complicated, because in this case fu
= 21 / 3 fl. From these two relationships it can be shown that:
fu = 21 / 6 fc and fl = fc / 21 / 6.

1.9.4 Frequency weightings


Sometimes the additional detail of a frequency spectrum is not needed, but it is
required to describe the noise by a single number which in some way takes into
account the broad spectrum characteristic of the noise, e.g. that it is predominantly
high or low frequency in character. Several single figure frequency weightings have
been devised for this purpose, with the best-known being the A and C frequency
weightings. These are described in more detail in later sections.

GPA Jan 2016 Section 1 page 1.30


General Principles of Acoustics Section 1

1.9.5 White and pink noise


These are two special types of broadband noise that both have flat frequency spectra
and they are both used in building and electro-acoustic applications. White noise is
defined as having equal energy per constant bandwidth, or equal energy per Hz,
across the whole frequency range. Thus when its spectrum plotted on a linear
frequency scale it is flat, i.e. a horizontal line graph. If white noise is plotted on an
octave (log) frequency scale, i.e. measured in octaves or third octaves, then the
spectrum slopes upwards at 3 dB increase per octave because, as noted earlier,
octave bandwidths increase with frequency.

Pink noise has equal energy per percentage bandwidth, in other words equal energy
per octave or one third octave band, across the whole frequency range. Therefore
the spectrum appears as a straight line when plotted on an octave frequency scale
and with a slope downwards at 3 dB per octave when plotted on a linear frequency
scale.

Sound Sound
Pressure Pressure
Level Level 3 dB per
octave

Frequency, White Noise. Frequency,


linear scale octave scale

Sound Sound
Pressure Pressure
Level 3 dB per Level
octave slope

Frequency, Pink Noise. Frequency,


linear scale octave scale

Figure 1.14 White noise and pink noise characteristics.

1.10 Types of elastic waves in solids and fluids


There are a wide variety of elastic waves which can be transmitted through matter.
In an unbounded or infinite expanse of solid, two types of waves are possible,
compressive waves (called P waves) and shear waves (called S waves). They are
important in transmitting vibrations through the ground such as from trains, road
traffic, construction and demolition activities, and are also responsible for transmitting

GPA Jan 2016 Section 1 page 1.31


General Principles of Acoustics Section 1

shocks from earthquakes. In a compressive wave, the to and fro motion of the
particles of the medium are in the same direction as the direction of travel of the
wave itself - i.e. they are longitudinal waves. In the case of shear waves the particle
vibration is perpendicular to the direction of wave travel - i.e. they are transverse
waves.

If the solid is bounded or finite, as in the case for a beam, rod or plate, several other
types of waves are possible such as bending (or flexural) waves, or torsional
waves. These waves may be a combination of both compressive and shear waves.
Yet other elastic waveforms, surface waves, propagate only on or near to solid or
liquid surfaces, and ripples on the surface of a pond or waves on the ocean are
obvious examples. One type of surface wave – the Rayleigh wave - is important in
transmitting vibration through the ground over relatively short distances, whilst very
high frequency ultrasonic surface waves in solids are important in electronic and
telecommunication engineering applications. Bending waves in plates are examples
of transverse waves, as are elastic waves on a wire or string and all types of surface
waves.

Shear forces applied to a fluid, i.e. a liquid or a gas, cause fluid flow to occur. Thus
liquids and gases cannot propagate shear waves, and the only types of elastic waves
that can travel through the bulk of a fluid are compressive ones. That explains why
sound waves in air are compressive, and hence longitudinal, waves.

1.11 The speed of elastic waves in fluids and solids


Wave propagation is described mathematically by a second-order partial differential
equation called the wave equation - the solution of which, for waves in an elastic
medium, shows that wave speed through the medium is:
c = √(K / ρ)
where the medium’s density is ρ kg.m-3 and its elastic modulus K is given by:
K = (stress ∕ strain)
= (force per unit area) / (fractional change in deformation) N.m-2.
Note that strain - being a ratio of lengths, areas or volumes - is dimensionless, so the
dimensions of elastic modulus are the same as those of stress or pressure, i.e. N.m-2.

For compression waves in a fluid, the elastic modulus is termed bulk modulus, and
the definition of strain is the fractional change in volume of a fluid element subjected
to a uniform fluid pressure acting equally in all directions. In a gas such as air there
are, in principle, two possible versions of bulk modulus, namely the isothermal and
adiabatic. If heat flows within the gas are rapid enough to keep pace with the
pressure fluctuations caused by the sound wave, then gas temperature remains
constant and the isothermal bulk modulus is applicable. Alternatively, if heat flows
cannot keep pace with pressure fluctuations, the compressions and rarefactions take
place under adiabatic (i.e., non-constant temperature) conditions and the adiabatic
bulk modulus is applicable. In normal circumstances, adiabatic conditions hold for
sound propagation through gases, and the adiabatic bulk modulus is γP where γ (the
Greek letter gamma) is a constant for a particular gas (1.4 for air) and P is the gas
pressure (atmospheric pressure for air). So the speed of sound in air is c = √(γP / ρ).

GPA Jan 2016 Section 1 page 1.32


General Principles of Acoustics Section 1

Example:
Calculate the speed of sound in air at sea level.
Typical values at sea level for the density of air ρ and atmospheric pressure P are
respectively 1.2 kg m-3 and 101x103 Pa
c = √(1.4 x 101000/1.2) = 343.3 m.s-1.

Exercise:
Go to http://www.aeromech.usyd.edu.au/aero/info/atmtab.txt and use the standard
atmosphere calculator accessible from this page to determine the speed of sound at
the 600 m elevation of Canberra and at the 2034 m top of the chair lift at Perisher.

From the science of thermodynamics it can be shown that, under normal


circumstances, any gas at temperature T, pressure P and density ρ obeys a law
which specifies that the quantity P / (ρT) depends only on the molecular make-up of
the gas, and is otherwise a constant. However, in applying this law, T must be
expressed on the absolute temperature scale and so measured in degrees Kelvin,
which are related to degrees Celsius by the formula:
degrees Kelvin = degrees Celsius+273.

It follows from this gas law that, for any gas of a specified molecular make-up, P/ρ is
directly proportional to absolute temperature T, and so the speed of sound in the gas
is directly proportional to the square root of its absolute temperature. Therefore the
following relationship can be used to determine the effect of absolute gas
temperature on the speed of sound:
c2 / c1 = √(T2 / T1)
where c1 and c2 are respective sound speeds at temperatures T1 and T2 degrees
Kelvin.

Example:
If the speed of sound in air is 330 m s-1 when the air temperature is 0 degrees
Celsius, what will it be at an air temperature of 20 degrees Celsius?
c1 = 330m s-1 T1= 0+273 degrees Kelvin, T2= 20+273 degrees Kelvin
c2/c1 = √(T2 / T1)
c2 = 330√(293/273) = 342 m s-1.

Exercise:
If the speed of sound in air is 330 m s-1 when the air temperature is 0 degrees
Celsius, what will it be at – 20 degrees Celsius?
Answer 318 m s-1.

1.12 Dispersive and non-dispersive waves


The velocity of sound in air does not vary with the frequency of the sound, a fact of
crucial importance to theatre and concert goers since it means that audience in both
front and rear seats will receive the same mixture of different frequencies generated
by the performers on stage. In this respect, air is a non-dispersive medium for
sound waves. However, some types of elastic waves are dispersive, most notably
bending waves in solid plates and beams, for which the wave speed increases with
frequency. These types of dispersive waves may be important for the transmission
of sound in buildings and other structures. ###############

GPA Jan 2016 Section 1 page 1.33


General Principles of Acoustics Section 1

ANNEX A
Review of Logarithms
There are numerous examples in acoustics where a physical parameter has an
extremely wide range of values. In such circumstances it follows, when presenting
data for graphical inspection that give an equal step along the axis of graph for
values of the variable from 1 to 10 as for values from10 to 100, or from 100 to 1000
etc, will give proportionately more space to the lower values.

A form of scaling that involves equal movements along the axis for equal
multiplicative (instead of additive) changes of the basic variable is said to be a
logarithmic one and the most common is to the base 10, i.e. related to multiples of
10. Thus a change from 24 to 96 would involve the same step along the axis as a
change from 240 to 960. Figures 1 and 2 show a set of six ordered data points:
1, 10, 100, 1,000, 10,000, 100,000
plotted by magnitude, on linear and logarithmic scales respectively, along the vertical
axes, and by position number in the ordered set along the horizontal axes Note that
on the linear scale most of the data is crammed very close to the horizontal axis,
whilst on the logarithmic scale individual data points tend to be more distinctly
separate.

120000 1000000

100000 100000

80000 10000

60000 1000

40000 100

20000 10

0 1
0 2 4 6 8 0 2 4 6 8

Figure 1 Linear scales. Figure 2 Logarithmic scales.

By way of a definition, the logarithm of a number to the base 10 is the power to which
10 must be raised to give that number. Thus 1,000 is 103 and so the logarithm of
1,000 to the base 10 is 3. Similarly, 100,000 is 105 and so the logarithm of 100,000
to the base 10 is 5.

The base 10 has been, and continues to be, the most commonly used base for
logarithms. The previously-used symbol for logarithms to base 10, log10, has now
been internationally replaced with the symbol lg, which implies to the base 10. Thus
the aforementioned definition of logarithms to base 10 may be expressed
symbolically as:
if N = lg(a) then a = 10N

Another logarithmic base used extensively, especially in calculus but not generally in
acoustics, is e = 2.718…. Logarithms to base e, previously written as loge, are now

GPA Jan 2016 Section 1 page 1.34


General Principles of Acoustics Section 1

are written ln. The n in this symbol is a reference to the Scottish mathematician John
Napier (1550-1617), who is often credited with the discovery of logarithms.

The logarithm for integer powers of 10 is straightforward so, for example:


lg(100) = lg(102) = 2
lg(1) = lg(100) = 0.
Otherwise more complex calculations are necessary, although it is important to note
that the logarithm of any non-positive number (i.e. including 0) does not exist, at least
outside of rather abstruse mathematical contexts. But with the aid of a calculator, it
is relatively easily shown that any positive number can be expressed as a power of
10. Hence:
2 = 100.301 and so lg(2) = lg(100.301) = 0.301
8 = 100.903 and so lg(8) = lg(100.903) = 0.903
whilst for larger numbers:
200 = 102.301 and so lg(200) = lg(102.301) = 2.301.
This latter logarithm can also be calculated in another way as:
200 = 2x102 = 102.301
giving the same 2.301 value for the logarithm.

The previous example clearly demonstrates the general rule:


lg(a x 10b) = b+lg(a)
which, in a generalised form, brings us to the first rule of logarithms:
lg(a x b) = lg(a)+lg(b)
from which the second rule of logarithms follows:
lg(ab) = lg(a x a x a ……) = lg(a)+lg(a)+lg(a)+….= b x lg(a).
Finally, noting that a/b=ab-1 and applying both these first and second rules, we obtain
the third rule of logarithms:
lg(a/b) = lg(a x b-1) = lg(a)+lg(b-1) = lg(a) – lg(b).

Antilogarithms.
If we know that lg(a) = N, then, from the definition of a logarithm, we can determine
the parameter a by expressing it as a power of the base 10. In other words, restating
the definition:
if lg(a) = N then a = 10N.
This mathematical operation of inverting a logarithm, commonly known taking the
antilogarithm, may be expressed as:
if lg(a) = N then antilg(N) = 10N = a
antilg(lg(a)) = a.
The latter equation clearly shows why the expression taking the antilogarithm is
appropriate!

GPA Jan 2016 Section 1 page 1.35

You might also like