ÐÓ°ÉÔ­´´

Noisy neighbours are a familiar irritation; the din of heavy machinery
is a more hazardous one. Engineers have adapted sound to play many beneficial
roles, from surveying the seabed to monitoring air pollution, but they have
been much less successful in controlling its harmful effects. One of the
main reasons is the difficulty of pinpointing sources of noise and assessing
the strength of the sound produced. This is changing, however.

In January 1991, the International Standards Organisation (ISO) introduces
the first guidelines for measuring sound based on a technique that directly
evaluates the energy of a noise source. At the moment, many engineers try
to quantify acoustic output in terms of sound pressure – the same effect
that enables us to hear. This is like trying to rate the heat output of
an electric fire in terms of the temperature produced at some point around
it. To go with the ISO standard, the International Electrotechnical Commission,
which sets standards for electrical equipment, will announce the requirements
at the end of next year for instruments that measure sound energy. The fact
that the two standards will appear in the wrong sequence – a standard for
the measuring instrument should logically precede one for the technique
that uses it – reflects the urgency with which acoustical engineers view
the need to bring more order to their field.

Unwanted sound, or noise, damages the psychological and physical health
of many people, particularly in industrialised countries. But reducing noise
levels is not straightforward. Insultating the walls of your home is often
not the best way of muffling a neighbour’s music or DIY activities; the
noise may not be coming into your living room through a party wall, but
by a more circuitous route. In a busy factory or on a large industrial complex,
where hundreds of machines may be creating a reverberant ‘bath’ of sound,
pinpointing the main sources of noise is extremely difficult.

Even when engineers can identify the source of noise, making it noticeably
quieter requires a disproportionate amount of effort because the human response
to sound is more logarithmic than linear: to halve the loudness of the sound
a person hears, the energy of the generated noise must be reduced by a factor
of 10. This is made doubly difficult because only a tiny fraction of the
operational power of a noise source is converted into sound – and reducing
something that is already small is a problem. The fraction varies from about
0.1 per cent for pneumatic drills to about 0.001 per cent for turbojet engines.
Even hi-fi loudspeakers, specially designed to generate sounds, convert
only 1 or 2 per cent of the electrical power supplied into noise. Football
supporters may seem a noisy bunch but during a cup final at Wembley stadium
in London they generate barely enough sound energy – less than 5000 joules
– to fry an egg.

ÐÓ°ÉÔ­´´s first devised – and patented – a theoretical technique for
measuring sound energy precisely in 1932. But it was not until the 1980s,
following developments in solid state electronics and the introduction of
tougher legislation to control noise in the 1970s, that technology caught
up with the science of acoustics. Until then, little had been done to implement
research into the practical problems of measuring sound so that a noise
source could be pinpointed and modified.

The choices that have been used up to now are crude, or expensive and
time-consuming. The owners of noisy machinery simply insulate their equipment
or provide workers with ear muffs. In some cases, they send the machinery
to acoustics laboratories, or they build their own test rigs to discover
how they should quieten its operation. This is particularly difficult in
the case of old or heavy components of production lines.

Now engineers are perfecting a technique, known as ‘sound intensity
measurement’, that pinpoints major sources of noise without the need for
special acoustics facilities. They have developed a sound intensity meter
for recording the direction of flow and the strength of sound energy at
different points in a noisy environment. This information yields a detailed
‘acoustics map’ that reveals where noise is coming from. Though designed
to help factory owners to suppress noisy machinery effectively, the technique
has a wide range of uses. For instance, it should enable car manufacturers
to design quieter vehicles and it should help neighbours to get on better
with one another.

Until the development of sound intensity measurement, acoustical engineers
made do with conventional microphones that record the very small fluctuations
in air pressure that we recognise as sound. These microphones measure the
magnitude of the pressure and the frequency of the fluctuations. Pressure
fluctuations are also responsible for the irreversible damage that can be
done to our hearing. Changes in pressure of just 1 newton per square metre,
equivalent to only 10 millionths of atmospheric pressure, constitute a serious
hazard to people continuously exposed to them during a working day. The
hazard is greatest at frequencies between 1 and 4 kilohertz. Like our hearing,
which is most acute between frequencies of 1 and 2 kilohertz, the sensitivity
drops off at lower and higher frequencies.

Fluctuations in air pressure became the traditional basis for evaluating
noise levels. The most widely used unit of measurement is the A-weighted
decibel, or dB(A). This takes into account that the human response to sound
is more logarithmic than linear and that the ear is not equally sensitive
to all frequencies in the audible range. For instance, EC regulations introduced
this year require employers to assess noise levels where workers are likely
to be exposed to 8.5 dB(A) or more. Machinery suppliers are required to
provide noise ratings for equipment that is likely to generate more than
85 dB(A). But this approach is not entirely satisfactory.

The level of sound pressure produced by a noise source varies with the
position around the source, with the distance from the source and with the
acoustic characteristics of the surroundings. In the open air, sound pressure
is reduced by about 6 dB(A) when the distance from a source of noise, such
as a loudspeaker or a lawn mower, is doubled. For extended sources, such
as a pipeline or a stream of traffic, the pressure falls by between 4 and
5 dB(A) when the distance from them is doubled. In enclosed spaces, such
as factories, the prssure level also depends on the ability of the surroundings
to absorb sound; the reflective effect of walls at home can be quite marked,
as every bathroom Pavarotti or Te Kanawa knows. The presence of other objects
besides the source of noise also influences the level and distribution of
sound pressure.

While sound pressure may be the easiest quantity to measure, it is not
very useful for diagnosing the precise cause of the noise, or for locating
the major regions of noise radiation in complex sources, such as industrial
machinery. This is because pressure, like temperature, is a scalar quantity;
it has magnitude but no direction. A source of noise affects a region, or
field, around it, in the same way as a magnet or an electric current does.
At any point in space around noisy equipment, the sound pressure field contains
contributions from many different sources and its net result depends on
the relative strengths of each one. A distant, strongly radiating component
can produce as much sound as a nearby, weakly radiating one. The presence
of sound-reflecting surfaces, such as floors and walls, increases the confusion;
while multiple reflections make it imposible to identify individual sources
of noise, as a visit to an indusrial production line demonstrates.

Though industry still uses sound pressure levels, measured in decibels,
to evaluate the noisiness of its equipment, a better measure, and one that
is less sensitive to the acoustic environment, is the rate of generation
of sound energy, or the ‘sound power’ of a source. Like all mechanical powers,
this is measured in watts; for example, domestic vacuum cleaners generate
between 100 and 1,000 microwatts of noise, while male conversation produces
about 50 microwatts.

Power of sound intensity

Similarly, the rate and direction of flow of sound energy is a more
revealing and discriminatory measure of the distribution and strengths of
sound sources. This quanity, known as the sound intensity, is measured in
watts per square metre; for example, the intensity of male conversation,
1 metre in front of the speaker’s mouth, is about 4 microwatts per square
metre, while that of an airliner, taking off about 100 metres away, is 1
watt per square metre.

Sound intensity is a vector quantity; it possesses both magnitude and
direction. Like meteorologists representing wind on a weather map with arrows
of varying lengths, engineers can plot the distribution of sound intensity
vectors in a sound field to spot the major sources of noise. Once they have
located the source, they can determine its sound power and evaluate how
effectively design modifications will cut down its noise.

Calculating the sound power of a source involves measuring the sound
intensity at a large number of points around it. Multiplying the intensities
by the areas over which they act and adding together the products gives
the result. Only the sound power of the source within the region measured
contributes to this sum; the intensity meter records sound energies radiated
by other sources but, by the time the meter has moved all the way round
the source being investigated, the accumulative effect of these other energies
is zero.

In the past, such sound measurements could be performed only in an acoustics
laboratory, isolated from other noise sources. Apart from the expense, many
industrial sources could not be tested in this way because they were too
large, too dangerous or too heavy to be moved or fitted in a laboratory;
often they could not be separated from ancillary equipment essential to
their operation. A major advantage of the new measurement technique is that
noise sources may be evaluated in their normal operating environments, so
there is no need for special test rooms.

But what sort of device do engineers need to measure sound intensity?
This is the key quantity in determining the location of a source of noise
and its sound energy. The sound intensity at a point in space is the product
of the sound pressure and the speed of the oscillating air particles at
that point, which is typically of the order of fractions of a millimetre
per second.

Efforts to develop a reliable and practical sound intensity meter began
in the US in the early 1930s when Harry Olson, an acoustical engineer working
for the Radio Corporation of American, patented the first one. Olson’s device
incorporated two microphones: one was a pressure microphone in which a piezoelectric
crystal converted sound pressure into fluctuating voltages that could be
measured; the other was a ribbon, or ‘sound particle velocity’, microphone
in which sound pressure caused a thin metal strip to vibrate in a magnetic
field. This generated a fluctuating voltage that was proportional to the
differences in pressure, or pressure gradient, and to the velocity of the
particles.

From pressure and particle velocity measurements, Olson hoped to derive
the sound intensity field. His device, however, was not reliable over a
wide range of frequency, atmospheric temperature and humidity, and it was
not developed commercially. Another problem was that the phase responses
of the two microphones did not match precisely. This artificial phase difference
threatened to swamp any real phase difference between his two recordings.
Such a difference would occur, for instance, if the maximum pressure did
not coincide with the maximum particle velocity. When the artificial phase
difference is insignificant, the sound intensity is simply the product of
the pressure and the particle velocity.

It was not until 1975 that a practical instrument was first produced
by B G van Zyl, a young graduate at the National Physical Research Laboratory
of the Council for Scientific and Industrial Research in Pretoria, South
Africa. Helped by developments in solid state electronics and in manufacturing
technology, van Zyl’s device worked over a broad range of frequencies. It
had a wide dynamic range, so it could measure quite and loud sounds; and
it was stable in different atmospheric conditions, so it could be used outside
the laboratory. The device incorporated two conventional pressure microphones
in which sound waves vibrate a metal diaphragm, a few micrometres thick.
As the diaphragm moves, it changes the electrical capacitance between the
diaphragm and a stationary metal plate, which generates a fluctuating voltage.
From two readings of sound pressure, and knowing the distance and phase
difference between them, van Zyl could calculate sound intensity.

Between 1975 and 1980, more compact and sensitive instruments were developed
in various countries, including Switzerland, France, the US and Britain.
For instance, researchers at the Unversity of Southampton produced a portable
instrument based on the then newly available switched capacitor, or charge
transfer, device. An oscillator and microchip in the device replaced arrays
of capacitors and resistors in the electrical filter circuit, which selects
the operating frequency range. The development promised a sound intensity
meter that would be cheap, accurate and stable. At the time, however, British
industry did not appreciate the scientific and commercial potential of sound
intensity meters, and the instrument never went into production.

The most commercially successful device was developed in the early 1980s
by Bruel & Kjaer, a Danish manufacturer. Though the device was unwiedly,
it used digital circuitry based on microchips instead of the analogue circuitry
that had been used until then. It also allowed engineers to measure sound
at frequencies up to 8 kilohertz – the previous limit was 5 kilohertz. Bruel
& Kjaer’s device incorporated two condenser microphones, facing each
other 12 millimetres apart. The microphones are described as nominally identical
because the artificial phase difference is insignificant. As with van Zyl’s
device, engineers derive sound intensity from two readings of sound pressure.

In 1982, Norwegian Electronics introduced an intensity probe that works
on a completely different principle; the device measures sound particle
velocity by recording the disturbances that the noise causes to an ultrasonic
beam it generates. The probe contains two piezoelectric cystals. One crystal
fires a high frequency beam of sound at the other, but the passage of the
beam is affected by the fluctuating air movement in between them. This yelds
a signal that is proportional to the sound particle velocity; engineers
use a conventional pressure microphone to measure the sound pressure.

Further improvements in manufacturing technology and materials science,
plus developments in electronic circuitry, have intensified the demand for
international standards for instrument performance and measurement procedures.
Portable devices, capable of picking out and measuirng the sound intensity
of a single source of noise in the din of an industrial complex, are becoming
commonplace.

Measuring up to noise

Intensity meters have revolutionised the measurement of sound. The sound
power outputs of sources of all kinds and sites can now be evaluated in
situ, outdoors or indoors. Expensive acoustic test rooms are no longer needed.
Sources of noise in vehicles may be placed in rank order of importance.
Direct measurements can be made of the sound insulation performance of partitions
such as walls, doors and industrial noise enclosures, and weaknesses in
seals and closures may be rapidly detected. It is possible to determine
whether the principal path of transmission of noise between adjacent dwellings
is via the party wall, up through the floorboards or down and in at the
windows. This reduces greatly the possibility of unnecessary financial outlay
on party wall treatment where it would be ineffective.

By measuring the rate of flow of sound energy into objects, instead
of from sources, intensity probes can help to determine the effectiveness
of sound absorbers in recording or broadcasting studios, and the influence
of seating, for instance, in concert halls. Taking measurements around a
chair bathed in sound and recording the sound intensity close to the chair’s
surface enables engineers to determine how much noise is being relfected
or absorbed, as heat, by the chair.

Using sound intensity measurements to detect sources of noise is not
always successful. This is because sound energy in reverberant enclosures
flows along tortuous paths. When sound energy is very close to vibrating
solid surfaces, it can leave one region and re-enter the structure by another,
which is known as the nearfield region. In the hands of an expert, however,
a sound intensity meter can yield valuable results. Because sound fields
are so complicated, especially where they interact with solid objects to
create diffraction effects, the ability to map intensity vector fields,
and the paths of energy flow, allows teachers and students of acoustics
to appreciate the behaviour of sound better. As a result of the technical,
economic and educational benefits that it affords, sound intensity measurement
promises to make a significant impact in the battle to create a quieter
world.

* * *

Sound units for reflecting what we hear

Sound is the result of fluctuating movements of air particles, which
generate very small fluctuations of air pressure about the steady atmospheric
pressure. These fluctuations can make our eardrums vibrate, stimulating
nerve impulses to the brain so that we hear sound. They can also cause a
sensitive steel diaphragm in a conventional microphone to vibrate, converting
the pressure waves into electrical signals; the signals can be recorded
or, after amplification, sent to a loudspeaker to make another diaphragm
vibrate.

To quanfity the ‘strength’ of a continuous sound, engineers must measure
the average magnitude of these pressure changes with time. Because the mean
value about atmospheric pressure is zero, engineers square the signal from
a microphone to make it positive at all times and calculate the mean value
of the squared signal. This they term the ‘mean square value’.

Another problem is that the ear is not a linear device: at any one
frequency, increasing or decreasing the air pressure by a particular amount
does not affect our perception of the sound by the same ratio. For instance,
a threefold reduction in sound pressure, or a tenfold reduction in sound
energy, only halves the loudness of the noise we hear. To compensate for
the fact that the human response to sound is more logarithmic than linear,
and to make sound pressure readings reflect more closely what we hear, engineers
use a logarithmic scale to measure the ‘sound pressure level’ in units known
as decibels (dB). As a result, a noise of 60 dB is twice as loud as one
of 50 dB.

A further refinement takes account of the fact that the ear is not equally
sensitive to all frequencies in the audible range (from 20 to 20 000 hertz).
A sound pressure at a frequency in the middle of the audible range is given
a higher value than the same pressure at either end of the range. The unit
of measurement is the A-weighted decibel, or dB(A).

Typical values of sound pressure levels range from between 100 and 110
dB(A) for jets flying overhead at 300 metres to between 40 and 50 dB(A)
in residential districts during the daytime. Lawn mowers and food blenders
produce between 80 and 90 dB(A), while dish washers and vacuum cleaners
generate between 60 and 75 dB(A).

Noise is becoming an increasingly important criterion in the selection
of industrial equipment and plant. Legislation relating to noisy equipment
and vehicles has significant commercial and economic implications for their
manufacurers and operators. Last year, member countries of the European
Community agreed on a new directive to control noise at work. The European
Council of Ministers will review the directive in January 1994.

Frank Fahy is professor of engineering acoustics at the Institute of
Sound and Vibration Research of the University of Southampton.

Further reading Sound Intensity, by F J Fahy. Published in 1989 by Elsevier
Applied Science, London & New York. Price 45 Pounds.