The nature of noise

What is noise? This is one of those questions like ‘What is time?’, to which one might answer: ‘I know exactly what it is until someone asks me’. But it’s worse than that, for ‘noise’ means two quite difierent things. For a scientist, noise is extraneous acoustic or electromagnetic energy. When used for communication, noise is whatever is not the signal, in other words does not carry information (whether the information is a voice down a wire, or a structural element in an ultrasonically scanned foetus). The larger the amount of noise, the more diffcult it becomes to unweave the signal from it—hence the concept of signal to noise ratio.

Once such noise has been identifled, it can be fairly easy to remove. Noise in a circuit that occurs at particular frequencies (a 50 Hz hum from a mains electricity supply, for example) can often be dealt with by fllters, so that only frequencies of interest remain. Unfortunately, in electronics, a very common form of noise is white noise, a signal that fllls all frequencies of interest and sounds like a hiss when heard through a loudspeaker. This cannot be flltered out so easily, but one may instead fllter all but the frequency bands that carry the signal; there will still be white noise in those bands, but the overall signal to noise ratio will improve.

For most of us most of the time though, ‘noise’ is any sound which is unwanted by the person who is exposed to it. A trumpeter may be very pleased with his output and certainly not think of it as noise, but a neighbour might well disagree, however technically excellent it may be. And both are right—one person’s music is another’s noise. This disagreement goes to the very heart of why noise is so diffcult to quash. But fortunately, while almost every sound has at least a few people who love it, and there are those to whom almost anything but silence is anathema, it is also true that sounds which many of us think of as noise do tend to share three characteristics: suddenness, loudness, and tunelessness. To flnd out why this is so, we need to consider why we can hear anything at all.

Ancestral legacies

We hear because our ancestors did, and they heard because their deaf relatives died. Among the many sounds with which they would have been surrounded, a few were matters of life and death. A sudden roar, the twang of a bow string, the snap of twigs underfoot, the crack of thunder overhead were all danger signs, and what they have in common is suddenness—hence today our instant reactions to sudden noise. Sudden silence, like the hush of bird song when a predator is sighted, meant as much as sudden sound and hence causes the familiar deafening silence when a clock stops ticking, an air conditioner switches ofi, or persistent rain suddenly ceases. In fact, any unexpected change to the sounds around us can be an annoying noise: every driver knows the irritation and concern caused by a new and unidentiflable engine noise, which an unsympathetic passenger may not be able to pick out at all, however well the driver imitates it.

But why are we annoyed by those sounds which seem unrelated to any ancestral sound that might signal danger? People talking in the street outside at midnight? MP3 players you can just hear across a train carriage?

The answer lies in a key social function of sound: to demonstrate power, in particular by claiming ownership of space. We are used to the idea that each of us has a personal bubble of space that we carry round, to which the uninvited are unwelcome. Someone who invades that space with their sounds can be as annoying as someone intruding into it physically, and the main point in both examples is intent. How much more annoying is the person sitting next to you if they phone someone to talk nonsense than if they receive a call of real concern?

Whether another person is regarded as noisy or not depends too on the relationship between them and the listener. Historians Shane and Graham White, for instance, in their studies of slavery in the US, found that the sermons of African American evangelist preachers, while appreciated by congregations, were referred to simply as noisy by white American Christians who were concerned because the sounds, and perhaps their messages too, spread far beyond the conflnes of their venues. Cultural norms are important too: acceptable sound levels in libraries, during mealtimes, and at funerals vary greatly in difierent countries.

The sea of other people’s noise in which one is often submerged on public transport, in public spaces, or open-plan offces can be a disturbing environment. The individual sounds can be neither identifled nor located, but the hearing system constantly tries to do both, making them as hard to ignore as to escape.Even when one does manage to ignore such sounds on a conscious level, the mental processing continues, as does the stress that sometimes results—and the high blood pressure that this stress can cause.

Often, the only way to reclaim one’s private sound bubble is to flll it with a sound fleld of one’s own making, by use of an MP3 player or smartphone. According to Michael Bull, Professor of Media and Film at the University of Sussex (UK): ‘iPod use can usefully be interpreted as a form of pleasurable toxicity within which the “total mediated” world of users lies a dream of unmediated experience—of direct access to the world and one’s emotions’.

There are some sounds which cannot be ignored, even though they do not really exist. These so-called earworms are usually brief snatches of music, often banal ones, which insist on repeating themselves in the mind. Advertising jingles are especially good at worming their way in. Though so widespread and seemingly simple, earworms are actually very hard to explain. Neurologist Oliver Sacks suggests that the unusual way music is remembered may be signiflcant. Recalling a scene or event involves reconstructing it, which in turn means that the result may be very difierent in difierent ‘replays’. In recalling a piece of music, however, something much closer to a direct copy of the original is preserved, an ‘almost defenceless engraving of music on the brain’,

as Sacks puts it, over which we have relatively little control. Perhaps the fact that tunes, and hence this ‘engraving’ process, arrived so late in the evolution of the mind is a reason why such memories are hard to control.

Music can have even more profound efiects on the mind—in some people it can cause epileptic seizures, while in others it can have a signiflcantly calming efiect, and can even alleviate pain and high blood pressure. It can be especially efiective in the treatment of some mental problems, and has been used in this way since World War II. Music has also been shown to be of great relief to some who sufier from Parkinson’s disease.

Deafening

The problem—indeed the tragedy—of noise-induced hearing loss is that it may not happen until years after the damage has been done. Also, while we react immediately to dangerously high temperatures or intense light, our defensive reactions to noise are far poorer.

Our eyes are equipped with a range of protective adaptations, most obviously eye lids and the contracting pupil (or expanding iris). Why does the ear have neither? The reason for the lack of ear lids is that while deafness may shorten an animal’s life, it won’t shorten it anything like as much as being successfully crept up on by a hungry tiger. And we do have an equivalent to the pupil, though it’s not particularly efiective: the acoustic refflex is provided by two muscles in each ear called the stapedius and the tensor tympani. When a loud sound arrives, the stapedius pulls the ossicle called the stapes (‘stirrup’, named from its shape) away from the oval window, while the tensor tympani pulls on the malleus (hammer) ossicle, which is attached to the eardrum and hence stifiens the latter. As a result, sounds are mufied. The stapedius usually tenses when we speak, to stop us annoying ourselves with our own voices; the tensor tympanum does the same job when we eat, to suppress our own chewing noises.

The acoustic refflex is one cause of temporary threshold shift (TTS), in which sounds which are usually quiet become inaudible. Unfortunately, the time the refflex takes to work (called its latency) is usually around 45 milliseconds, which is far longer than it takes an impulse sound, like a gunshot or explosion, to do considerable damage. The refflex is only one contributor to TTS; the other mechanisms are still unknown.

Where the overburdened ear difiers from other abused measuring instruments (biological and technological) is that it is not only the SPL of noise that matters: energy counts too. A noise at a level which would cause no more than irritation if listened to for a second can lead to signiflcant hearing loss if it continues for an hour. The amount of TTS is proportional to the logarithm of the time for which the noise has been present—that is, doubling the exposure time more than doubles the amount.

However, very loud impulse sounds, like gunshots, will also cause instant damage, even if the total energy is low. (Reputedly, a trained audiologist can identify the permanent damage caused by a single shot flred by a patient with no ear protection.) Recovery times increase with the amount of shift; a TTS of 40 dB can take weeks to recover from. The amount of TTS reduces considerably if there is a pause in the noise, so if exposure to noise for long periods is unavoidable (at a football match, say), there is very signiflcant beneflt in removing oneself from the noisy area, if only for flfteen minutes.

It is on the evidence of many careful experiments on the efiects of noise on hearing that noise regulations, now statutory in workplaces and elsewhere in many countries, have been deflned. They place limits both on peak noise levels and durations of exposure to specifled lower levels.

One problem with identifying how many people have noiseinduced hearing loss is that everyone’s ability to hear highfrequency sounds declines with age: newborn, we can hear up to 20 kHz, by the age of about forty this has fallen to around 16 kHz,

and to 10 kHz by age sixty. Aged eighty, most of us are deaf to sounds above 8 kHz. The efiect is called presbyacusis (literally ‘old man hearing’). Since noise-induced hearing loss usually also impacts higher frequencies and, being cumulative, becomes more common with age, we don’t really know how much loss is avoidable.

If there has been a downside to the often laudable progress made in tackling dangerously high levels of noise, it has been the relative neglect of noise which is ‘only’ loud enough to annoy people, despite their vast numbers: the 2008 Ipsos MORI (UK) National Noise Survey found that 26 per cent of respondents were annoyed by neighbour noise, for example. Those afiected lose sleep, they lose concentration, they lose patience—they lose the joy from their lives (39 per cent said their quality of life was adversely afiected). Such noise also exacerbates mental health problems, community tensions, and social isolation.

There are few who would deny that the best approach would be to remove noise from the environment in the flrst place. How this can be done, and how efiectively, depends crucially on the source. To most inhabitants of the developed world and many elsewhere, the following sources are likely to be the main ones:

·Air traffc

·Industrial

·Neighbour

·Neighbourhood

·Rail

·Road traffc

In addition, shipping noise and wind-farm noise afiect some areas signiflcantly, while most areas are blighted by construction noise from time to time.

Of these groups, neighbour and neighbourhood noise are particularly diffcult to deal with, partly because the maker does not class them as noise but the listener does. This diffculty has led sound historian Karin Bijsterveld to conclude that noise abatement is in the grip of a ‘paradox of control’, in which the promises of experts and politicians to control noise actually amount to passing the buck to neighbours to do so, or developing such complex formulae for limiting other forms of noise, such as that from aircraft, that few can understand, engage with, or critique them.

Fighting back

With the exception of some neighbour noise and some warning signals, no one would be likely to complain if noise were to be eliminated. Can this be done?

In those (unfortunately rare) cases in which noise is highly regular, consists of one or few frequencies, and is bothersome only in a small and well-deflned location, active noise cancellation (ANC) can be highly efiective. ANC relies on the fact that sound waves consist of compressions followed by rarefactions. On a pressure plot such as Figure 1, this is clear from the fact that the line goes flrst above and then below the midpoint. If a second sound wave were superimposed on the flrst such that at every point at which there is a compression in the original wave there was an equally intense rarefaction in the new one, perfect silence would result, the acoustic energy being converted entirely to heat. However, in practice this can only be achieved over very limited areas, such as plane cockpits, the driver’s head area of some cars and—perhaps most usefully—the space inside ear defenders (Figure 23).

Many highly efiective technological solutions to noise have been developed. The problem is that all of them have been applied already. Today’s vehicles are extremely quiet compared to their mechanical powers, far quieter than their much feebler ancestors ever were, but this fact is overwhelmed by the vastly greater number of them in use today. But it is still worth looking at how today’s machines are kept so quiet, since the same principles will almost certainly be applicable in future, especially as new materials become available.

23. Active noise cancellation.

The flrst principle of noise control is to identify the source and remove it. For perhaps the single most important noise source today, the internal combustion engine (ICE), this is impossible: the noise is made by the explosions within, which are intrinsically noisy. However, as explained in Chapter 3, our hearing systems do not measure the total energy of a sound—they are tuned to respond most strongly to frequencies around 4 kHz, and it is usually possible to shift the output frequency of ICEs to below this. The main means are the adjustment of the timing and the rate of fuel injection, to modify the combustion of the fuel so that such high-frequency components are reduced.

Having dealt as far as possible with the noise source, the next step is to contain it. Again, ICEs are challenging since they have to be connected to the air around them, to draw in oxygen and expel exhaust gases, and air paths are noise routes. To reduce noise emissions, mufiers are the answer. (A silencer is really the same thing, but is more often so-called when used to quieten flrearms.)

There are two basic mufier principles, and most ICEs use both. Absorptive silencing is achieved simply by lining the pipe with absorber so that any waves that do not travel straight down are converted to heat. Those which do travel straight down are also reduced, since their edges are retarded by the duct lining. But this effect is only slight at low frequencies, which are better dealt with by reactive mufiers. These contain Helmholtz resonators tuned to the most troublesome frequencies. Sounds at these frequencies are thus enhanced inside the mufier, allowing their energies to be absorbed there. A few mufiers also use ANC.

Over the next few decades, it seems likely that ICEs will largely give way to electric versions. There is no technical reason why these should not be almost silent, but a vehicle which approaches silently could be deadly, so all will be provided with specially generated running and warning sounds (which might also be of value for bicycles—especially those with no bells). There is no reason for these to be unpleasant, however—even an alarm need not be alarming. To attract attention, ‘shushing’ sounds can do the same job with less annoyance.

Jet engines, on the other hand, are probably here to stay for the foreseeable future. In their case, the source of the noise is not the burning gases themselves, but what happens when they emerge from the engine at high speed and mix with the relatively slow-moving air outside the plane. The turbulence this causes is the source of almost all the noise. Enormous efiorts have been made to smooth the mixing, mainly by increasing the circumference of the jet by corrugating it and by cocooning the exhaust jet in air of intermediate speed. No more can be done in these directions without an excessive loss of thrust, but other approaches have been considered. One idea that has met with modest success is to mount engines above the wings. Another is a strategic change, such as the use of airships for the transport of goods where transit times are not an issue.

All but the slowest or most antiquated moving vehicles are now streamlined, which avoids the turbulences that would produce noise as well as increasing effciency. Streamlining is achieved partly by designing the vehicle with a suitable shape so that the airstream slips past it without the sudden change of direction or speed that would result in turbulence, and partly by smooth surfacing. Nonetheless, as far as car drivers are concerned, the aerodynamic noise which streamlining is intended to deal with is still the most signiflcant at cruising and higher speeds. The principal sources are usually the wing mirrors and the A pillars (the ones which hold the sides of the windscreen in place).

For those on the pavement, the loudest noise is frequently made by the tyres, especially on concrete roads. Low-noise surfaces can do a great deal to reduce this. The best available are layers of asphalt with many air-fllled pores (about 25 per cent of volume), but such surfaces are costly and not very robust. A less efiective, but more durable and cheaper alternative, is the use of thin (~2 cm) asphalt layers.

Time was when the design of machines did not consider quietness at all, and noise management was applied only as a sticking plaster afterwards or as unreliable rules of thumb (like stretching a few strings under concert hall ffloors or furnishing theatre stages with vases). Now, machines are usually well designed and carefully engineered to be at least fairly quiet—though regular maintenance of industrial machinery and careful installation of domestic appliances are essential.

When noise can be neither avoided not contained, the next step is to keep its sources well separated from potential sufierers. One approach, used for thousands of years, is zoning: legislating for the restriction of noisy activities to particular areas, such as industrial zones, which are distant from residential districts. The flrst recorded example dates from around 700 bce, when tinsmiths, blacksmiths, carpenters, potters, and even roosters were banned from the city centre of Sybaris, a Greek colony on the Aegean coast.

Where zone separation by distance is impracticable due to historical or geographical considerations, sound barriers are the main solution: a barrier that just cuts ofi the sight of a noise source will reduce the noise level by about 5 dB, and each additional metre will provide about an extra 1.5 dB reduction. Also, the barrier must be massive enough to prevent much sound passing straight through it; about 10 kg under each square metre is usually suffcient. Since barriers largely refflect rather than absorb, refflected sounds need consideration, but otherwise design and construction are simple, results are predictable, and costs are relatively low.

Temporary fflexible versions of noise barriers, known as acoustic fencing, can be placed around construction sites and, if carefully positioned and arranged to avoid leaving gaps, can be very efiective, reducing noise levels by up to 30 dB in some cases. Unlike permanent noise barriers, acoustic fencing achieves noise reduction primarily by absorption.

The big problem with noise barriers is their unattractiveness to the eye. This can be ofiset by introducing a screen of vegetation between barrier and observer. If space permits at least 10 metres of vegetation, the plants will help reduce the noise somewhat too. At low frequencies (around 25 Hz), this is mostly due to the extra ground absorption caused by fallen leaves, and at high frequencies (above ~1 kHz) it is caused by foliage. It is best if the leaves are half as long as the sound waves, so, for both these reasons, some deciduous content is essential, (a mix of trees and bushes) reaching as high as possible (up to about 10 metres, after which refflection from large branches reduces the efiect). On windy days, the sound of such vegetation is another bonus, as a distraction from the noise.

The addition of natural sound sources (soundscaping) is especially efiective in urban parks. In such spaces the commonest way to improve a soundscape is by adding fountains, but birdsong (real or recorded) can work well too.

Quietening homes

Of all spaces, the most precious is home, because we relax and sleep there, and because it is an extension of our personal space. Keeping noise out is, therefore, a high priority: the best way to do this is through good acoustic design, including selection of the right materials, and, crucially, their proper installation and maintenance. This is much cheaper, more aesthetically pleasing, and more successful that retrofltting noise solutions into a pre-existing dwelling.

Whether retrofltted or not, the basic approaches to home sound reduction are simple: stop noise entering, destroy what does get in, and don’t add more to it yourself. There are three ways for sound to enter: via openings; by structure-borne vibration; and through walls, windows, doors, ceilings, and ffloors acting as diaphragms. In all three cases, the main point to bear in mind is that an acoustic shell is only as good as its weakest part: just as even a small hole in an otherwise watertight ship’s hull renders the rest useless, so does a single open window in a double-glazed house. In fact, the situation with noise is much worse than with water due to the logarithmic response of our ears: if we seal one of two identical holes in a boat we will halve the infflow. If we close one of two identical windows into a house, the infflow will again half—but that 50 per cent reduction in acoustic intensity is only about a 2 per cent reduction in loudness.

The second way to keep noise out is double glazing, since single-glazed windows make excellent diaphragms. Structure-borne sound is a much greater challenge, though if the source is within the dwelling, the room which it (or (s)he) occupies, can be treated. But this will not be cheap and will usually require treatment of the ceiling as well as all walls, the ffloor, and the door.

One inexpensive, adaptable, and efiective solution—though maybe not the most attractive—is the hanging of heavy velour drapes, with as many folds as possible. If something more drastic is required, it is vital to involve an expert: while an obvious solution is to thicken walls, it’s important to bear in mind that doubling thickness reduces transmission loss by only 6 dB (a sound power reduction of about three-quarters, but a loudness reduction of only about 40 per cent). This means that solid walls need to be very thick to work well.

A far better approach is the use of porous absorbers and of multi-layer constructions. In a porous absorber like glass flbre, higher-frequency sound waves are lost through multiple refflections from the many internal surfaces. Using stud-mounted panels with air gaps between them and the wall will deal with lower frequencies, through refflection of lower frequency waves at the air/solid interface (yet another example of an impedance mismatch). A well-fltted acoustically insulated door is also vital.

The fioor should not be neglected: even if there are no rooms beneath, hard ffloors are excellent both at generating noise when walked on and in transmitting that noise throughout the building. Carpet and underlay are highly efiective at high frequencies but are almost useless at lower ones, so if you have ffloor-mounted loudspeakers or an extrovert baritone in the home, something special will be required, and again there is no real alternative to bringing in an expert.

Though rarely a problem in the home, the most signiflcant issue in the design of public spaces is reverberation. Surprisingly, though it has been a bugbear for architects, performers, and audiences alike for millennia, no way to quantify it was developed until 1898, when Wallace Clement Sabine deflned the reverberation time of a room as the period necessary for the intensity of the sound to decline to one millionth (fi60 dB) of its initial level (Box 13). Even more usefully, Sabine derived an empirically based equation that allowed this time to be calculated from the size and shape of a room—so architects could predict it at last.

Box 13

Sabine’s equation: T=0.161(V/A), reverberation time in seconds T, room volume V, total absorption area A. This equation is still used today except for highly absorbent rooms.

However much care, money, and expertise is spent on an individual noise problem, the results will be inefiective unless both the root cause and the context of that problem are investigated flrst: it may be pointless to replace standard doors with acoustic ones if the actual problem is that they are routinely left open, and there is little point in reducing the speed limit on a trunk road if the efiect is to send traffc down residential streets. Such holistic approaches to noise reduction extend beyond acoustic treatments: the beneflts of fltting sound absorbing panels must be balanced against their visual impact, the introduction of artiflcial sounds to distract park users from traffc noise may annoy more people than it pleases, and moving people from open-plan offces to quieter cellular ones may reduce team working. Similarly, while well-built houses can keep noise out, this necessarily involves keeping the inhabitants in, cutting ofi the sounds of nature too, and leading to alienation. As David Hendy puts it: ‘whenever we withdraw into separate soundscapes . . . we make strangers of each other’.

The wider view

Despite the great importance to us of our hearing systems, their performance is not usually a matter of life and death. This is not the case for some underwater species. The prime importance of sound to many marine animals, combined with the effciency with which sounds travel underwater, means that some marine creatures are devastated by the efiects of noise, in particular whales and dolphins. The scale of the efiects is unclear but some are subtle: since whales communicate with their calves by sound signals, the mother/child link can easily be broken by extraneous sounds and the calf separated. In other cases, the startle efiect of unfamiliar loud sounds on whales and dolphins can be so extreme that rapid surfacing occurs, leading to death from the formation of nitrogen gas bubbles in the liver and other organs (the condition divers describe as ‘the bends’).

There are many contributors to underwater noise pollution. As well as the sounds of shipping, sonar systems, blasting, and signalling all add their load. Fortunately, there is now widespread recognition of the efiects on animals; more in fact than on divers, who routinely risk severe hearing damage (although, as the efiects are often the same as those of pressure changes in the ear, the scale of the problem is hard to establish).

The interdisciplinary and public nature of attempts to battle noise, both on land and underwater, make collaboration at all levels vital, and cannot succeed without the active support of international organizations such as the World Health Organization, national governments, local authorities, specialist organizations like the UK Noise Abatement Society, and the public. One activity which should bring all these players together is noise action and awareness weeks.

As technologies have evolved, cities have grown, and vehicular travel has increased, noise sources have proliferated and noise is now a fact of life for many, whether made deliberately and selflshly via loudspeakers or unavoidably and carelessly through engines and other hardware. Although there is much that science can do to help, it must be accompanied by both education and legislation which ensures that noise is universally regarded as the pollutant that it is.

Ultimately, noise-making must become unacceptable, rather than just annoying. While this may seem impractical, exactly such a change has taken place in the UK regarding smoking—and this took only a few years. The change came about through a combination of new laws implemented through both national and local government, and a loud and clear message that smoking causes serious illness. Given the same approach to noise, just as radical a change might be accomplished.

Of all disciplines, sound has by far the widest variety of interested parties. Understanding it enhances the work of actors, advertisers, aerospace engineers, automotive engineers, anthropologists, architects, artists, broadcasters, builders, communication engineers, composers, designers, ecologists, educators, electronic engineers, environmental health offcers, fllm-makers, historians, marine biologists, musicians, physicians, physicists, politicians, prospectors, psychologists, seismologists, sociologists, town planners, zoologists, and many others.

Hitherto, many of these domains have been separate, but over the past few decades approaches like those of soundscaping, sound studies, and the design of ever more intimate and personalized media systems have started to bring together expertise from across these flelds. Such collaborations are often hampered by the narrow training ofiered to most practitioners, but meanwhile they are making people more aware of the vast range and power of the subject—of the multitude of practical and emotional beneflts a knowledge of sound can lead to, and of the many new applications and insights the bringing together of its many branches may yield. Such developments are to the beneflt of everyone. For all of us, the future sounds good.