3  Basic acoustics

Part 1

It would not be far from the truth to say that the secret of a good recording lies in having satisfactory acoustics to start with. But we should make two points: first, it is possible to achieve acceptable recordings in less than ideal acoustic conditions, and perhaps this book will help to show how that can be done; second, what may be good acoustics for one purpose may not be good for another. Again this will, we hope, be made clear. Nevertheless ‘good acoustics’ are a great help and the aim of this chapter is to explain what is meant by the term and how it may be possible to achieve if not good, then at least adequate, acoustics for particular purposes.

Briefly, we can say that good acoustics enhance, or at least do not impair, recorded sound quality; bad acoustics can reduce the intelligibility of speech, allow distracting background sounds to be present, adversely affect the quality of music, and so on.

We shall look at two aspects of acoustics. These are:

1.   Sound isolation – the prevention or reduction of external noise in the recording venue.

2.   Internal acoustics – the results of the behaviour of sound waves within this venue.

Sound isolation

It is unrealistic on financial grounds to try to soundproof totally even a professional studio. In practice, broadcasting and recording companies generally adopt criteria which establish how much background noise is acceptable for any particular type of programme material. For example, to aim for an absolutely silent background in a studio to be used for rock music is pointless. On the other hand, it is essential for radio drama studios to have no intrusive noises as these can destroy the illusion: a distant audible police siren cannot help a play set in the eighteenth century! But the siren would probably be utterly inaudible in the rock music studio. There are two categories of background sound which we can consider here:

1. Airborne sound

This means sound which has travelled through the air for the great majority of its journey. The most important methods of preventing its ingress are:

1.   Heavy (i.e. massive) walls. More details are given in Part 2 of this chapter, but they can be summarized by saying that a wall made of a single thickness of brick will reduce external sounds by some 45 dB(A) on average. This may sound impressive until it is realized that a moderately busy road could produce a sound level of 80 dB(A) at the outside of the walls. This, reduced by 45 dB, leaves 35 dB(A), which is almost certainly going to be picked up by microphones. In particular, any fluctuations are going to be more distracting than a steady noise. Doubling the mass of the wall has a very modest effect – usually an increase in sound reduction of around 5 dB – so a double thickness brick wall would provide only about 50 dB of sound reduction. Cavities in walls add to the effect, of course.

2.   Double or triple glazing of windows. Up to a point, domestic double glazing is not all that effective in acoustic terms as the air gap, which is typically a matter of millimetres thick, does not possess sufficient compliance (i.e. ‘springiness’). Vibrations in one sheet of glass are easily transmitted to the other. For really good sound insulation the sheets of glass need to be some 200 mm apart, although spacings of 60–80 mm can be fairly useful. For this reason ‘secondary double glazing’ may be better, provided all gaps are sealed (see below). Notwithstanding what we have just said, domestic double glazing, with its relatively few millimetres of gap, is better than single glazing!

3.   Sealing of all air gaps. This is vitally important, as sound waves can pass through extremely small apertures. The fact that domestic double glazing gives an improvement in sound insulation is probably more due to the careful sealing than to the fact that there are two sheets of glass. In professional studios all doors have, or should have, ‘magnetic seals’, in principle not unlike the seals around refrigerator doors. Double doors, providing a kind of ‘sound lock’, are common. Further, care has to be taken to seal all places where services such as water, gas and electrical trunking enter the studio.

Reduction of airborne sound in the non-professional environment

Away from a proper studio, what can be done? The short answer is usually not very much. If a room is to be used as a temporary studio the following suggestions may be useful:

1.   The obvious, but possibly easy to overlook, first step is to use a room on the quietest side of the building, and if traffic noise varies with the time of day, then quiet times should be chosen if possible.

2.   It may sometimes be very worthwhile putting strips of adhesive foam round the edges of all opening windows, unless they are very well-fitting ones. It can happen that an almost dramatic improvement in sound isolation may be achieved for a very modest cost.

3.   Noise from adjacent rooms can be a problem. If reducing the noise at source (always the first course of action) cannot be done then temporary sealing round the doors can be tried. Tightly folded sheets of newspaper pushed into the spaces round the door may help – until someone needs to go through the door!

4.   Other gaps should be sealed with whatever materials come to hand. Bitumastic sealants can sometimes be very effective for permanent scaling of gaps.

The Haas Effect, mentioned in Chapter 2, must be borne in mind here, as sound may be leaking through a gap but because the first-arrived sound is through, say, a wall, it may not be apparent that the wall is not the major source of trouble.

What are known as flanking paths may be mentioned here. These are routes round, typically, a wall. The wall may be thick, but a possible sound path could be over the top of the wall and through relatively thin ceiling materials.

2. Structure-borne sound

Here the sound vibration travels through the actual fabric of the building, and it is well known that bricks and concrete are good conductors of vibration. (An electric drill with a masonry bit at work on a wall can be heard throughout a house.) Not only that, but water pipes or other rigid material extending through the building can be serious offenders. This kind of noise transmission is the most difficult to try to cure. In fact, cure is often virtually impossible – the problem has to be anticipated and dealt with at the design stage. In the professional world, the following steps are taken:

1.   Studios are sited as far away as possible from likely sources of vibration.

2.   They are isolated from the rest of the building by making the studios independent ‘boxes’ within the building. They stand upon resilient bases of steel springs or some form of synthetic rubber. (Some modern concert halls, such as the Birmingham Symphony Hall, are acoustically isolated in this way.)

3.   Care is taken that the studio walls which may be part of a cavity wall system are connected to the other brickwork by flexible tie rods.

4.   Flexible seals are used at junctions between studio floors and where they meet, for example, a corridor.

5.   Water pipes, etc. have flexible sections.

All this is very expensive, and for obvious reasons can rarely, if ever, be applied as a remedial treatment to an existing building. It is interesting to note that many recording studios built in the last decade or two have been conversions of older buildings, but in rural areas where vibration through the ground, often caused by heavy traffic, is unlikely to be a problem.

As a further precaution against trouble from structure-borne sound, any machinery that has to be near a studio, for example fan motors for ventilation and conditioning, are fitted with anti-vibration mounts (AVMs) so that the motor is, as far as possible, isolated from the floor on which it stands.

Reduction of structure-borne sound in the non-professional environment?

It is clear that very little can be done about structure-borne sound. Noise from machinery can of course be stopped by switching off the machine – but this may not always be practicable. Microphones can often pick up severe rumble by sound transmission through the floor and via a table to the microphone stand. It can happen that this is very disturbing at the playback of a recording when it may not have been obvious at the time of making the recording. Suitable resilient supports for the microphone help, but putting the first rubber mat that comes to hand under the microphone stand may not be the answer. The best kind of sound-insulating support is one where the microphone and its stand ‘float’ on the mat, so that the stand sinks slightly into it but does not compress it fully. A little experimenting with different materials is often worthwhile.

Figure 3.1 Structure-borne sound

Internal acoustics

This is basically a matter of sound reflections from walls, floor, ceiling and other surfaces, together with, as we shall see, absorption of sound at such surfaces. We shall consider two of the aspects which, in the context of this book, are the most important. These are:

1. Standing waves

When there are two parallel facing surfaces, such as opposite walls, it is possible for there to be a number of modes of vibration, called standing waves, as shown in Figure 3.2. If the wavelength of a sound created between these surfaces is such that multiples of the half-wavelength correspond exactly to the distance between these surfaces, then it is possible for an acoustic resonance to develop.

There are many instances of these resonances. Organ pipes are good examples. Also, the note which is produced when one blows across the top of an empty bottle is the result of a resonance, although the mechanism by which it is produced is very different from that which can occur between two walls.

Figure 3.2 Standing waves

The term standing waves for the vibration patterns shown in Figure 3.2 is somewhat inaccurate, because the only things that are stationary are the patterns of minima and maxima. These are called nodes and antinodes and are marked N and A in the diagram. Figure 3.2 shows the pressure variations (there are other ways of representing standing waves) so that an antinode is a point where there is the greatest pressure variation and a node is where there is zero (or at least minimum) pressure variation. (There are other ways besides the sound pressure that can be used to represent standing waves. For instance, the air particle displacement can sometimes be useful.)

It is possible to detect standing waves in most rooms if a steady pure tone (i.e. one with no harmonics) is played. Walking about in the room, especially if one ear is covered tightly with the hand, shows variations in the loudness as one moves through the nodes and antinodes. This effect is best noticed when the wavelengths are in the region from about 0.1 to 1 m – a frequency range from around 3 kHz down to roughly 300 Hz.

Standing waves can also be observed in water when the conditions are right. Circular wave patterns in a vibrating cup of liquid are good examples of a rather special set of standing waves.

In acoustic terms, standing waves are a bad thing as there will generally be different nodes and antinodes for different frequencies. Thus, a microphone placed between two parallel reflecting surfaces is likely to give a distorted version of what it should be picking up because it will be at nodes (minimum loudness) for some frequencies and antinodes (maximum loudness) for other frequencies.

Standing waves are reduced by having non-parallel surfaces and/or irregularities in walls, ceilings, etc. Also, the presence of sound absorptive materials (see later) has a beneficial effect, up to a point. Irregularities, to be effective, need to be large – a figure of one seventh of the longest wavelength is sometimes quoted for this – so taking the lowest speech frequencies as being about 300 Hz, with a wavelength of roughly 1 m, the surface irregularities for speech should be about 15 cm. Ingenuity can help in the non-professional environment. Cupboard doors ajar and the tops of trestle tables set at an angle against a wall are possibilities.

Figure 3.3 Reverberation

2. Reverberation time

This is the second important aspect of internal acoustics that we shall deal with. In any room, sound energy dies away rather in the manner shown in Figure 3.3. The curve is somewhat idealized as the decay is always much more irregular than this. This is a result of multiple reflections from all the surfaces, the sound losing energy at each reflection.

Reverberation time, which we shall denote by RT, is defined as the time it takes for the sound to decay through 60 dB. This may be made a little clearer by a look at Figure 3.4. Here the decay of a sound is shown with a decibel scale for the vertical axis – in Figure 3.3 the vertical axis was pressure. The most obvious change is that the sound decay now appears as a straight line (in practice never quite as straight as this!).

Figure 3.4 Reverberation time

More details and methods of calculation are given in Part 2 of this chapter. At the moment, though, it is enough to say that 60 dB represents roughly the difference between a moderately loud sound and virtual inaudibility.

Two important factors which affect the reverberation time are:

1.   The volume of the room or studio.

2.   The amount of sound-absorbing material present.

Taking these in turn, it is not difficult to see that the bigger the room the longer it takes sound to reach the various surfaces at which energy will be absorbed. If, for instance, 2 dB of energy is lost at each reflection, then a loss of 60 dB will take place after 30 reflections. The time for 30 reflections to occur will be longer in a large room than a small one.

Sound-absorbing materials present in a room will clearly help the reverberant sound to die away more quickly. We might say briefly at this stage that all materials absorb sound to some extent, although materials like glass and hard woods have only a very small absorption. Porous and fibrous materials are likely to be good absorbers. Hence carpets and curtains and soft furnishings are helpful, although not likely in themselves to provide all the absorption which might be needed. More details are given in Part 2 of the chapter.

Reverberation time is very important in any sound recording work. For a recording to sound convincing, the RT should be appropriate to the kind of material being recorded. Examples are:

1.   Speech. Too long a value of RT gives the speech a distant, ‘echoey’, effect. Too short and the speech quality is apt to sound dry and lifeless, besides which too short an RT is unpleasant for the speaker. Generally, an RT of around 0.3–0.4 of a second is considered satisfactory.

2.   Orchestral and choral music. Around 2 seconds is preferred, although there can be some latitude in this. Most traditional concert halls have RTs of this order. Too short an RT is apt to result in an inadequate blend of individual notes. However, much chamber music and baroque music written for small ensembles may often benefit from a shorter time. It is worth noting that, with very large halls (the Royal Albert Hall in London is a good example), speech on the platform can seem to be in a very ‘dead’ (i.e. short RT) acoustic because the low sound levels are not able to reach the distant reflecting surfaces. With an orchestra, though, the Royal Albert Hall's full RT of over 2 seconds is apparent!

3.   Pop and rock music. A short RT is needed here, at least in a studio, as it is normal practice to multi-mic the band – that is, each instrument has a microphone. A long RT implies that there is considerable reflection of sound in the studio and as a consequence each microphone will be likely to pick up quite a lot of the sound from other instruments, which is undesirable. Studios for this sort of music have RTs of, typically, around 0.5 second.

4.   Drama (sound only). This presents a problem as outdoor conditions may need to be simulated and this requires a short RT. At other times it may be necessary to represent something like a church interior, and this calls for a long RT. It might seem that in the latter case the addition of artificial reverberation could solve the problem, but this may not be satisfactory as an actor often needs a reasonably good simulation of the acoustic conditions in order to help with the delivery of his/her lines. Some typical RTs are given in Part 2 of the chapter.

An interesting, and sometimes very important, point about realism in reverberation is in the Initial Time Delay (ITD). In any place where there is reverberation (and this means pretty well everywhere!) a listener will hear first the direct sound from the source and then there will be a small time interval before the reverberation arrives. This interval is the ITD. The ear and brain take note of an ITD – even if it is too short to be consciously perceived (less than about 30 ms) – and use it to help assess the size of the environment.

In a small room the ITD will be short, perhaps only a very few milliseconds; in a large hall it will be perhaps a few tens of milliseconds.

This means that, in sound, to simulate accurately a particular environment not only does the RT have to be correct, but the ITD must also be appropriate. Artificial reverberation devices usually have a time delay control which in effect varies the ITD.

The next question is, what can be done to give a conventional room or hall the right sort of RT for a particular recording? Let us say straight away that the achievement of the ideal acoustic conditions is a very expensive task (some might say an almost impossible one, even with plenty of money available!) and in what might be termed ad hoc circumstances the range of possibilities is limited. To a large extent good microphone technique can be a great help, and this will be covered in later chapters. The following suggestions may be helpful:

1.   Speech. The usual fault with simple equipment is allowing too much reverberation to be picked up. This may be a matter of microphone positioning as much as anything, but that will be gone into later. The non-professional recordist often has little choice of ‘studios’ and an office or a sitting room may be all that is available. The average sitting room has an RT of around 0.5 second – rather on the long side for speech recording. Offices may be very reverberant if they are large and have little in the way of soft furnishings; they can be quite dead, acoustically, if they are of the plush, executive variety.

Any or all of the following can help to increase the sound absorption, if that is what is needed: curtains drawn across windows, rugs, blankets, etc. laid on uncarpeted floors, plenty of coats on all hooks, tables covered with thick cloths, rugs or blankets.

2.   Music. Choral or orchestral recordings are probably going to be made in sizeable rooms – halls, churches and so on. The RTs may be theoretically too long, especially in the absence of an audience, but it is generally accepted that, for music (classical music anyway), too long an RT is preferable to one that is too short. The opposite tends to be true for speech.

It is worth noting that it is possible to have a small room with many hard surfaces, yet which has a longer RT than a much larger room with many absorbent surfaces. However, the pattern of reflections in the small room will give it a ‘small-room sound’.

A final point here – and this should be seen as being applicable to all recording work – it is very desirable indeed that test recordings are made and listened to critically before proceeding with the proper recording. The ability of the ears to discriminate against unwanted sounds has already been referred to. A microphone doesn't have this ability!