CHAPTER 13

Control Rooms

Although the function of almost all sound studio control rooms is to produce recordings (or live broadcasts) which will ultimately be heard in domestic rooms, there are other demands made of the monitoring systems and acoustics that lead to some very different solutions than the use of hi-fi loudspeakers in ‘standard’ listening rooms. Control room monitoring usually needs to be capable of a degree of resolution that will show up any errors in the proceedings to such an extent that the recording personnel will always be one step ahead of the consumers. Control rooms must also be practical working environments. Several people may need to be able to work in different places in the rooms whilst listening to substantially the same musical and tonal balances. This implies that position-dependent differences within the working area should be minimised.

Music recording also frequently tends to take place at higher levels than will be used for most domestic reproduction or standard listening tests, which can further alter the balance of compromises. The reasons for this are numerous, and many are discussed in Chapter 15. However, in general, it can be said that most recording studio control rooms should be more acoustically dead than most domestic rooms. As can be seen from Figure 4.17, the reverberant/reflected energy in domestic rooms can substantially enhance the subjective loudness of a loudspeaker. An increase of 5 or 6 dB is shown, which is not untypical, so loudspeakers for use in rather dead control rooms need to be able to supply the extra power.

In fact, this 5 or 6 dB, plus the few dB s extra that are needed for other reasons, such as greater listening distances and the need to inspire musicians, explains why loudspeakers for recording use may need to be able to supply up to 100 times the output power of domestic loudspeakers. However, one hundred times the power is only four times as loud, and much of this can be soaked up in highly controlled rooms. In engineering terms, the 100 times (20 dB) power increase poses many more electro-mechanical design problems than ‘only’ four times the loudness would suggest. (An increase of 10 dB [at least at mid frequencies] is generally considered twice as loud as the original.) Increased power usually means increased size, both to move more air and to lose the additional waste heat. Many loudspeakers in professional use are typically only 0.5 to 5% efficient, therefore 500W into such a loudspeaker would produce between 475 and 497.5W of heat in the voice coil. This must be dissipated in the metal chassis and by means of ventilation. Such problems can also lead to the use of different construction materials, and this can also move studio loudspeakers one step further from their domestic counterparts.

13.1 The Advent of Specialised Control Rooms

As mentioned in Chapter 12, back in the days when most recording was done in mono, judicious movement of the loudspeakers and/or listening positions could often achieve a more or less desirable sound, both in terms of critical distance considerations and the avoidance of troublesome nodes or anti-nodes in a room. When stereo arrived, a new set of restrictions arrived with it. The listening position became a function of the loudspeaker distances and the subtended angles between them. Thus, it was no longer a relatively simple matter of adjusting the positions of a single loudspeaker and/or listening position, but the moving of a whole triangle, formed by the two loudspeakers and the listener. In other words, all three items may need to be moved not only in relation to their individual responses in the room, but whilst maintaining relative angles and distances between them. They also introduced the necessity to maintain their stereo imaging, which can be very fragile. Moving loudspeakers away from troublesome modes may disrupt the stereo balance or drive them into impractical locations. Improving the siting of one loudspeaker may, through the need to maintain the triangulation, move the other loudspeaker into an acoustically worse position, or into an inconvenient place, where it may need to be moved every time the door needs to be opened. This would clearly not be practical in professional use. These requirements suggested a need for the suppression, to a very high degree, of the temporally and spacially dependent characteristics of the room, but the better solutions took a long time to develop.

13.1.1 Geometrically Controlled Rooms

Commercial recording studios put up with a lot of bad rooms until the early 1970s, when serious efforts were made, on an international scale, to try to find control room designs which could be relied upon to produce recordings which generally travelled well; both to the outside world and between studios. This was an era when work really began to travel from studio-to-studio, and even country-to-country, during its production. One of the first big commercial efforts to produce acoustically standardised ‘interchangeable’ rooms, was by Tom Hidley at Westlake Audio in California, USA. Soon, rooms of this type were in use worldwide, and were designed to have reverberation times of less than about 0.3 s. They incorporated large volumes of ‘bass traps’ in an effort to bring the low frequency reverberation time into relative uniformity with the mid-band times, and also to try to avoid the build up of low frequency standing waves, or resonant modes.

An attempt to reduce further the resonant mode problem was effected by the use of entirely non-parallel construction, to deter the formation of the more energetic axial modes. Monitor equalisation was de rigueur as the rooms were ‘sold’ on their flat frequency response at the listening position, and third-octave equalisation was employed to achieve that goal. The rooms were generally quite well received at the time, and were a significant improvement on much of what was then current. Nevertheless, their responses were by no means as subjectively similar as their pink noise, real-time, third-octave spectrum analysis led so many people to expect. It soon became widely apparent that the control of reverberation time and a time average of the combined direct/reflected (loudspeaker/room) response was not sufficient to describe the sonic character of a room. Some people already knew this, but they were a small minority and few of them held any significant sway in the recording world. A typical design of an early Westlake-style room is shown in Figure 13.1.

Figure 13.1:

Typical room design of Westlake/Hidley style of the late 1970s.

13.1.2 Directional Dual Acoustics

Also during the mid-1970s, Wolfgang W. Jensen, in Europe, was producing rooms as depicted in Figure 13.2. These rooms used ‘sawtooth’ absorbers, which tended to absorb much of the incident wave from the loudspeakers. They had reflective surfaces at angles where they could reflect back sounds created within the room, such as by the speech and actions of the personnel, but they would not cause reflexions directly from the monitors. Total absorption of the incident wave was not intended, because the Jensen rooms still sought to maintain a room decay time on the low side of the ‘standard’ domestic range, in accordance with well-established recommendations. Reflective rear walls were quite common in these rooms, though most examples had a rather absorbent rear wall.

Figure 13.2:

(a) Typical room in the style of Jensen. The reflective side panels are relatively lightweight, and act as low frequency absorbers. Their angling also prevents ‘chatter’ between the hard-surfaced side walls. (b) View towards the rear of the Jensen designed control room at Sintonia, Madrid, Spain, showing the exposed entrances to the sawtooth absorbers on the side wall. (c) View towards the front of the Sintonia control room, showing only the reflective surfaces on the side wall.

The rooms of Jensen were interesting because they made a clear distinction between the acoustic conditions required for the monitoring (relatively dead) and the perceived acoustic needs for a sensation of comfort within the room (relatively live). The absorber openings were facing the loudspeakers only, although the wooden panelling, used to provide reflective energy to speech within the room, would also act as a general low frequency absorber. The rooms were clearly bi-directional in the same way as the recording room shown in Figure 7.3, and the much later BBC rooms shown in Figure 12.5. The Jensen rooms, like the Westlake rooms, usually featured flush mounted (built-in) monitors as standard, which were normally of their own designs and custom tailored to the rooms. This was an important advance.

13.1.3 The LEDE

In the late 1970s, Don and Carolyn Davis were keenly investigating many acoustic and psychoacoustic phenomena with the then newly developed Time Energy Frequency/Time Delay Spectrometry (TEF/TDS) measurement systems. TDS measurements made at Wally Heider Studios, Los Angeles, USA, and at RCA and Capitol Records, in Hollywood, had given them a lot to think about, leading them to the concept of the ‘Reflection Free Zone’ and the ‘Live-End, Dead-End’ (LEDE) control rooms.

Concurrently, Carolyn (Puddie) Rodgers was presenting new ideas about how certain room reflexion characteristics could confuse the ear by giving rise to response filtering which closely mimicked the pinnae (outer ear) transformations used by the brain to facilitate spacial localisation.¹ This work gave some very explicit explanations of the psychoacoustic relevance of the Energy/Time Curve (ETC) responses in the above-mentioned room measurements, and reinforced the concepts of the Davis’ LEDE principles. Don Davis and Chips Davis (no relation) then wrote their seminal paper on the ‘LEDE’ concept of control room design,² and these rooms came into very widespread use in the subsequent years. The concepts were further developed by Jack Wrightson, Russell Berger and other notable designers. Sadly, however, Puddie Rodgers died of cancer before ever seeing the fruits of her labours fully mature.

The LEDE rooms rely on some psychoacoustic criteria such as the Haas effect and the directional aspects of human hearing. The Haas effect,³ otherwise known as the precedence effect, manifests itself when two short sounds are heard in rapid succession. The human auditory system appears to suppress the separate identity of the second sound and give precedence to the first. The pair of sounds is perceived as a single sound coming from the direction of the first sound, but with greater loudness than the first sound alone. For this effect to be operative there must be a minimum separation of about 1ms between the two sounds, or the ear will tend to confuse the source position if the two are not co-located. Beyond about 30 or 40 ms, the two sounds are heard as separate events. The effect can be overridden if the second sound is more than 10 or 15 dB higher in level than the first. Haas stated that within the 1–40 ms, or thereabouts, time zone, where the effect is in operation, the second sound to arrive would need to be at least 10 dB higher in level than the first sound if it were to be heard to be separate and equal in level. The ‘just detectable’ threshold for the second sound to arrive is about 4 to 6 dB above the level of the first sound, depending upon conditions.

Figure 13.3 shows the LEDE concept, where the front half of the room is largely absorbent. Its geometry is designed to produce a zone free of early reflexions around the principal listening position. The idea is to allow a clean first pass of the sound, directly from the loudspeakers, and then to allow a suitable time interval before the first room reflexions return to the listeners’ ears. The rear half of the rooms are made diffusively reverberant, allowing the perception of a room ‘life’, which should then not unduly colour the perception of the directly propagated information. The rooms require a diffuse reverberation/decay characteristic, and proprietary diffusers such as those developed by Dr Peter D’Antonio, and marketed by his company RPG, are widely used in such rooms.⁴ Strong, discrete, specular reflexions are to be avoided, as they produce position dependent colouration and general response flatness irregularities.

Figure 13.3:

Concept of a Live-End, Dead-End room. This figure shows the concept of the Live-End, Dead-End stereo control room, in which a reflexion-free zone is created around the listeners. The rear half of the room is made to be highly acoustically diffusive, usually without specular reflexions, to simulate a sense of ambience within the room that does not obscure the general clarity of the monitoring. The front half of the room is acoustically dead, to prevent any room effect from returning from the frontal direction which could spacially superimpose itself on the direct monitor sound. It should be noted, however, that some of the early proponents of the concept stated that the room is defined more by the characteristics of its ETC response rather than by the physical distribution of hard and soft surfaces. These, strictly speaking, need not go hand-in-hand, although very often they do.

There was a period of time when specularly reflective panels were used at the rear/side corners of many LEDE control rooms. These were known as ‘Haas kickers’, and were intended to prolong the Haas effect, but their subjective effects failed to live up to expectations and they were generally soon abandoned. Despite the Haas effect, a reflexion coming from the opposite rear corner to the source did not help the stereo imaging. LEDE rooms will be discussed in Chapter 17 at much greater length.

13.1.4 The Non-Environment

After growing criticism of his 1970s designs, Tom Hidley took a break between 1980 and 1983. During this lay-off, he thought over the old problems and came up with a new concept that he called ‘the Non-Environment’ room principle. In these rooms, he made the front wall maximally reflective, and, other than for a hard floor, made all the other room surfaces as absorbent as possible.⁵ The principle sought to drive the loudspeakers into something approximating an anechoic termination. With the monitors set flush into the front wall it would act as a baffle extension, but it could not reflect sound from the loudspeakers because they were radiating away from it. The wall would, however, in concert with the floor and equipment surfaces, provide life to the speech and actions of people within the room, thus relieving any tendency for the rooms to feel uncomfortably dead. The principles were not unlike those of the aforementioned Jensen rooms, but were taken to a greater extreme, approaching a hemi-anechoic chamber (not to be confused with a semi-anechoic chamber – see Glossary). It is also not unlike the Ishii-Mizutoni room, as described in Section 12.4, but with the loudspeakers flush mounted and the wideband reflectors removed.

By introducing more absorption into the room and reducing the quantities of reflexions, the ratio of the direct sound to the reflected sound is increased, and hence the levels of colouration are reduced. This is, however, achieved at the expense of any consideration whatsoever for mimicking domestic listening acoustic conditions. The consistency between Non-Environment type rooms is, perhaps, greater than that between most other types of rooms. One very famous recording engineer/producer who loves these rooms for mixing (George Massenburg) freely admits that for certain types of music recording, especially when the musicians are creating music in the control rooms (as is often the case these days), he must use rooms with a more inspiring life. Such things are very subjective, though, and he also produces people who love the rooms for the whole recording process. A typical ‘Non-Environment’ construction is shown in Figure 13.4., and the concept is discussed at length in Chapter 16.

Figure 13.4:

The Non-Environment concept.

13.1.5 Toyoshima Rooms

In 1981, Hirata et al, presented a paper on optimum reverberation times for control rooms and listening rooms.⁶ They made some interesting statements for the day, such as ‘It is easier for a recording engineer to assess the clarity of phantom images of reproduced sound in a dead room than in a live room, though this is not the purpose of listening to music’. They concluded that perhaps, for super-critical monitoring, a decay time of around 0.2 s would be optimum, with perhaps about 0.4 s being the optimum decay time for a listening room. The clear statement of different optimum decay times for quality control and listening for pleasure was important.

Around the same time, in Tokyo, Sam Toyoshima was designing studios for some of the major recording companies, which were also beginning to be noticed in Europe and the USA. In 1986, he presented a paper on control room acoustic design.⁷ The first of Tim Hidley’s Non-Environment rooms had also been built in Tokyo, so a momentum was beginning to build up from the east in the direction of dead rear walls.

Toyoshima stated unequivocally in his 1986 paper, ‘A control room should be designed for the live front wall and the dead rear end’. ‘To suppress … standing waves at low frequencies, the rear wall must be fully absorptive’. ‘If the rear wall is reflective, the wall must be designed to provide a high degree of diffusion. However, for 85 Hz, for example, the dimensions of the diffusing members must be as large as 4m, comparable to the wavelength of 85 Hz, which is practically impossible’.

Hidley was using absorber systems of about 1.2 m on the rear walls, and Toyoshima was claiming to have designs that would provide adequate absorption in a depth of approximately 60 cm. The decay time versus frequency for a typical Toyoshima room is shown in Figure 13.5. A typical wall construction is shown in Figure 13.6. By way of contrast, the innards of a large Hidley Non-Environment room are shown in Figure 13.7.

Thus, a clear division had arisen between people such as the Davises, who favoured live rear ends in control rooms, and Hidley, Toyoshima and others who opted for maximally absorptive rear walls. Effectively the option was becoming between ‘Live-End, Dead-End’ and ‘Dead-End, Live-End’. But what they all agreed upon was that, for best results, control rooms should have directional acoustics whose properties depended on source positions, and not a generally diffuse sound field of uniform decay time, irrespective of source position. Differences in opinion about which end should be live and which should be dead continue to the present day, and each philosophy has its followers. The underlying psychoacoustic philosophies will be discussed in detail in Chapters 16 and 17, when the concepts and constructions will be outlined by experienced practitioners specialising in one or other of the two points of view.

It must be appreciated, though, that the room concepts must be well understood if they are to be applied to their greatest effect, and that the principles involved are not necessarily interchangeable. Each one is a system in itself. It is the misapplication of many of the components of these philosophies which has led to some rather poor monitoring conditions. Nonetheless, some aspects of the designs are more or less common to all of the rooms described above, and it is worthwhile looking at some of these points.

13.2 Built-in Monitors

In the top-level control rooms it is general practice to build the monitor loudspeakers into the front walls. The flush mounting of the loudspeakers ensures that all the sound radiates in a forward direction. This avoids the low frequency problems associated with free-standing loudspeakers, where the rear-radiation strikes the wall behind the loudspeaker and returns to the listener with a delay, giving rise to response irregularities due to the relative phase differences of the direct and reflected waves. None of the current control room concepts has front walls that are sufficiently absorbent at low frequencies to avoid such effects from free-standing loudspeakers, although the Walker rooms of Figure 12.5 do go some considerable way towards limiting the damage.

Figure 13.5:

(a) Example of reverberation characteristic of a Toyashima control room; (b) Toyashima’s concept of the need for short decay times in control rooms. Relationship between impulse responses of control room and studio. The control room decay should be more rapid than that of the studio, and should generally be lower in level than the ambience of the recorded music to avoid masking caused by the superimposition of the two responses.

Incidentally, flush mounting is sometimes, wrongly, described as ‘soffit’ mounting. In many 1970s designs, the loudspeakers were flush mounted above soffits – ‘soffit’ being the architectural term for the underside of an overhang, or the ceiling of a recess. (The Italian word for ceiling is ‘soffitto’.) These overhangs were often built above windows, or to create recesses in which to house tape machines. This was never a good idea from an acoustic point of view because the recesses could create resonant cavities. They also allowed a certain amount of rearwards sound propagation below them, leading to the problems described in the previous paragraph. Nonetheless, the concept was widespread, and the loudspeakers were mounted above soffits. Somehow, by erroneous extension, the term ‘soffit mounting’ has come into rather common usage, especially in the USA, referring to flush mounting in general. Figure 13.8. shows a genuinely soffit mounted loudspeaker, whereas the photographs in this chapter all show flush mounted loudspeakers. Anyhow, loudspeaker positioning is a source of variability of room response, so it is a good idea for serious control rooms to standardise on a monitoring practice which has more in common from room to room, and flush mounting fulfils this purpose. The benefits of flush mounting are many.

Figure 13.6:

(a) Construction of ceiling, floor and wall in the style of Sam Toyashima; (b) Alternative ceiling construction; (c) The Sam Toyashima designed control room in Studio 4 at the Townhouse, London, UK.

Flush mounted loudspeakers drive a room from a boundary, as shown in Figure 4.6(c) so there is no rear radiation to add to the general confusion in the room as shown in Figure 4.6(d). Flush mounted loudspeakers, with well-designed, smooth, unobstructed front baffles, also significantly reduce diffraction problems because there are no edges to act as secondary diffraction sources. Flush mounted loudspeakers enjoy an extended baffle plane, which means that even when driven from the boundary of a hemi-anechoic chamber, the radiation impedance on the loudspeakers is increased because of the constrained angle of radiation, as explained in Chapters 11 and 14. This can mean a greater low frequency acoustic efficiency of between 3 and 6 dB compared to the same loudspeakers free-standing. Therefore, for the same SPL in the room, only a quarter to a half of the electrical drive power is needed. The reduced demand from the amplifiers results in a big cut in the working temperatures of the voice coils, which tends to both reduce power compression and distortion, and improves the transient response of the system at high SPLs. Essentially, for serious monitoring purposes, there is no real alternative to flush mounting if the flattest response is required.

Figure 13.7:

Extreme control: (a) Elevation; and (b) plan of a typical ‘Non-Environment’ absorbent control room (courtesy of Tom Hidley).

Figure 13.8:

Soffit-mounted loudspeakers. In the example shown, the loudspeaker is mounted above the soffit. The alcoves below are often used to house tape recorders. In general, the idea is not too good, both because the high mounting of the loudspeaker causes the high frequencies to arrive at the ear at a less than optimal angle, and because the alcove below can become resonant (the Italian word for ceiling is ‘soffitto’).

Flush mounting also allows the cabinets to be very solidly mounted in a massive front wall, of which they may form an integral part. This type of mounting, if correctly done in order to avoid structural transmission of the sound (essentially by using heavy, rigid boxes in a heavy, rigid wall – see also Chapter 20), can improve stereo imaging by reducing to a minimum any unwanted response artefacts due to the cabinets physically moving under high levels of low frequency drive. This can be of benefit to the overall system transparency and to the stereo imaging.

13.3 Directional Acoustics

All the leading philosophies of stereo control room design are directional inasmuch as the rear half of the room is acoustically different from the front half. It has been found, almost universally, that such a practice is necessary for rooms to exhibit the most detailed stereo images, smoothest pressure amplitude responses, and best low level perception (at least whilst maintaining a pleasant acoustic to work in), but this poses a dilemma for the design of surround-sound rooms. The two front halves of good stereo rooms cannot simply be built facing each other, as this would leave the loudspeakers facing inappropriate far boundaries; generally, the front walls are solid, which rear walls tend not to be. The ‘mirror imaged front-half approach’ led to many poor designs of quadrophonic rooms in the late 1970s. Good surround rooms are beginning to emerge,^8–11 but the design concepts are still quite contentious. Chapter 21 will deal with this subject in much greater depth.

13.4 Scaling Problems

Size differences affect room responses greatly. Wavelengths in air, at room temperature, are functions of frequency and the speed of sound in air, and the physical requirements of dealing with sound propagation remains fixed to the wavelengths which do not scale with room size. Wavelength-to-frequency ratios remain fixed because the speed of sound remains fixed. Therefore, it follows that a small room will exhibit a greater reflexion density than a larger one of similar nature, because the sound will encounter more boundaries per second in a smaller room. What is more, in smaller rooms, the first reflexions will return with more energy because they have travelled less distance. Therefore, when control rooms are designed, the physical dimensions of the room dictate some of the treatments needed. It thus becomes impossible simply to scale many of the room philosophies beyond certain limits. A LEDE room depends on the existence of reflexion free zones, and a suitable time lapse between the arrival of the direct signal and the first returning energy from any room boundaries. If the room is too small, the required time intervals cannot be achieved, and the philosophy will fail to function. The room shown in Figure 12.5 faces the same problems, whereby the reflexions would arrive sooner and higher in level if the room size were to be reduced; hence some of the prime design targets could not be met.

Non-Environment rooms, however, do not depend upon such carefully timed psychoacoustic phenomena, but they do require considerable space for their absorber systems. The absorber sizes are wavelength related, so in very small rooms, especially if the structural walls are heavy and rigid, and hence highly reflective, there may be insufficient space in the room to allow a useful working area if the necessary size of absorbent trapping is installed. In other words, they could become all absorber – no room! At best, a telephone kiosk size area may result from a 3 m cube.

Very small listening rooms of any design will produce problems due to effects such as the ones mentioned above, but they can also complicate issues by imposing additional loading on the fronts of the loudspeakers. The designs of most loudspeaker cabinets assume a relatively ‘free air’ loading on the outer face of the diaphragms. In small rooms, especially in cases where the low frequency absorption is less than optimal, reflexions can return from the room boundaries with considerable energy, as they have travelled much shorter distances than in larger rooms. Their superimposition on the loudspeakers can affect the performance of low frequency drivers and, more especially, any tuning ports. Pressure zone loading may also have implications here.

13.5 The Pressure Zone

This subject was briefly touched upon in Chapters 5, 6 and 11 (Sections 5.8.3, 6.10 and 11.2.3). Below the frequency of the lowest mode that any room can support, there lies a region known as the pressure zone. Transition into the pressure zone is gradual, and depends to some extent on the Q of the lowest mode. Within the pressure zone, travelling waves cannot exist, and the whole room is pumped up and down in pressure as the diaphragms of any loudspeakers move forwards and backwards respectively.

Myths and misconceptions abound with regard to loudspeaker response in the pressure zone. Some textbooks say that below the pressure zone frequency, also known as the room cut-off frequency, ‘… the environment loads any sources (such that often) the effect of this loading is to reduce the ability of the source to radiate sound into the rooms, and so results in reduced sound levels at these frequencies’.¹² Other texts state ‘In rooms that are both physically and acoustically small, the pressure zone may be useful to nearly 100 Hz’.¹³ Elsewhere, it can also be found in print that control rooms, for example, should have their longest dimension at least equal to half a wavelength of the lowest frequency to be reproduced in the room. For a ‘20 Hz room’ this would be around 8.6 m if the lowest useful axial mode was needed. The implication often seems to be that in the pressure zone there will be no significant loudspeaker response. Indeed, the concept is reinforced by the use of the term ‘room cut-off frequency’ to mark its upper boundary. The low frequency cut-off can be calculated simply from:

The origin of the concept of requiring the loudspeakers to face the longest dimension of the room, although it is not strictly true, is clearly shown in Figure 4.17. In this figure, the modal support of the combined loudspeaker/room response is considerable, showing almost 10 dB of ‘gain’ over the anechoic response. It can also be seen that the response below the lowest mode falls back to the level of the rolling-off anechoic response. Given that at 30 Hz the perception of a reduction in SPL of only 4 dB can be that of a halving of loudness (especially below around 85 dB SPL), then the typical 6 to 10 dB reduction, as shown in Figure 4.17 between the modally supported response and the pressure zone (dashed) response, would result in the pressure zone response being perhaps only about one-sixth as loud as the modally supported response. These frequencies would, in fact, tend to be inaudible in a modally supported room at sub-80 dB SPL listening levels.

Obviously though, in rooms in which the modes are substantially damped, the difference between the modally supported and pressure zone regions will be greatly reduced. What is more, half space (2π) loading, such as is approximated by flush mounting the loudspeakers in a heavily damped (absorbent) room, will often serve to reinforce the output in the dashed line roll-off region of Figure 4.17, thus reducing the difference even more.

As for the concept of the pressure zone loading the loudspeakers, there are many things to consider. Perhaps if the room was very small, hermetically sealed, and possessed infinitely rigid walls, the statement would be true. Taken to an extreme, if the room were to be the same size as the cabinet loading the rear of the loudspeaker, then the room loading effect could be readily considered, because it would be equal to the cabinet loading. The reality is that the loading effect of the room is inversely proportional to the room/cabinet volume ratio. This means that a room with a cubic volume only ten times that of the loudspeaker cabinet would have only one tenth of the loading effect of the cabinet, which is hardly significant.

The mechanisms at work, here, are very interdependent. Below the resonance of a sealed box loudspeaker, the velocity of the cone falls with decreasing frequency in such a way that exactly offsets the tendency for the cone excursion to increase with falling frequency for any given input power. The effect is that the displacement stays constant with frequency. In the pressure zone, the pressure is displacement dependent, so the response of a sealed box whose resonance was equal to or higher than the frequency of the room cut-off (pressure zone frequency) would exhibit a flat response in the pressure zone. If the loudspeaker resonance is significantly below the pressure zone frequency, then the response will tend to rise with falling frequency until the resonance is reached, when the response will flatten out. Ported loudspeakers will tend to exhibit a 12 dB/octave (second order) roll-off below resonance (as opposed to their normal 24 dB/octave [fourth order] roll-off) when the port resonance is below the pressure zone frequency, although loading effects on the ports can complicate matters.

However, all of this assumes that the room is sealed and highly rigid. Non-rigid (diaphragmatic) walled rooms, such as are found in many timber-framed houses and most purpose-designed sound control rooms, present a totally different set of circumstances. Most timber-framed structures are, to varying degrees, transparent to low frequencies. The pressure zone onset is then defined by whatever high rigidity containment shell exists without the inner structure. A flimsy garden shed may well equate to a 2π space at very low frequencies, being rigidly bounded only by the floor on which it stands. The classic pressure zone equation:

(which is merely another form of Equation 13.1) is therefore unlikely to yield an accurate representation of the upper limit of something sudden happening at the entry to the pressure zone; or at the lower limit of the room’s normal response – the room cut-off frequency – whichever way one chooses to view it.

Clearly, in a very small room where the pressure zone may exist up to 100 Hz (a 1.5 m cubed control cubicle, for example), the first modes would need to be very heavily suppressed if they were not to introduce severe colouration on their onset. The situation in rooms in general, however, is highly variable, and a very thorough understanding of the structural details would be needed before any accurate pressure zone response prediction could be made.

Nevertheless, one or two statements can be made definitively. Dipole radiators, such as many electrostatic loudspeakers, cannot radiate sound below the room cut-off frequency. Only at very small distances from the front or rear of the diaphragm would any sound be perceptible. A dipole source (see Chapter 11) would merely paddle back and forth, without creating any net change in the overall pressure within the room. The short circuit round the sides of the diaphragm would cancel out (except in the very close field) any local pressure changes due to the opposite polarity that would exist on the front- and rear-radiating surfaces – see Figure 11.3(a).

The very fact that dipole loudspeakers are an exception in itself implies that the room cut-off, as it is known, does not mean that conventional sealed box or ported loudspeakers cannot radiate useful output below that frequency. In fact, one characteristic of the response in the pressure zone is that the perceived response not only exists, but it is also extremely uniform, both in level and with respect to position, because no modes exist to create spectral or spacial variations. Nonetheless, a relatively flat extension of the low frequency response into the pressure zone can only be achieved in very damped (absorbent) rooms.

For these reasons, no loudspeaker system can be absolutely optimised for low frequency performance without knowledge of the room in which it will be used. Good monitor loudspeakers must perform in real control rooms, and fulfill the job for which they were designed. Good domestic loudspeakers must give pleasing results in less controlled circumstances. And as was also mentioned in Section 12.2.1, some loudspeakers have different bass alignments depending on the nature of the typical house constructions in their principal markets. There are so many things that a room can do to modify a loudspeaker response that no simple set of idealised test specifications can truly represent a loudspeaker in normal use.

13.6 One System

It should by now be clear that the design of the control rooms and their corresponding monitor systems cannot be done in isolation, because the performance of either one is dependent upon the performance of the other. This is why many studio designers require some degree of control over the loudspeakers to be used in rooms of their design if they are to put their name to the overall result. Directivity, for example, is a loudspeaker characteristic which may dictate how a particular loudspeaker suits a given control room design. Too wide a directivity pattern in a small control room may induce excessive early reflexions. Too narrow a directivity pattern in a large room may make the stereo listening area unworkably small. The vertically aligned ‘D’Appolito’ layouts of the loudspeaker shown in Figures 13.9 and 13.10 for use in the Non-Environment type rooms have strange vertical directivity. Nevertheless, in the horizontal plane the drivers remain reasonably time aligned wherever one walks or sits in the room, as long as one keeps one’s head at a relatively normal height. A highly absorbent ceiling is required for such loudspeakers because with the three drivers operating at the crossover frequency their reflexions from a hard ceiling (especially a low one) would be very much out of step, and could produce transient smearing. The floor reflexions are normally scrambled by equipment.

Laterally spaced drivers in the highly absorbent rooms can lead to a ‘phasiness’ whilst walking around the room, because the path lengths to the individual drivers change as one moves horizontally. If the upper frequency limit of the parallel drivers is low enough, and thus the wavelengths are long enough, this problem is less consequential, although some such horizontally mounted pairs of loudspeakers operate up to 800 Hz, which is not a good idea in a highly controlled room. In a more lively room, however, the reflected energy will tend to mask this problem. The concept is shown diagrammatically in Figures 13.11 and 13.12.

Figure 13.9:

Capri Digital, Capri, Italy. The control room, designed by Tom Hidley, with the vertically aligned Kinoshita monitor loudspeakers (1990).

Figure 13.10:

The control room at Area Master Studios, (now Planta Sónica 2) Vigo, Spain, designed by the author, showing the vertical alignment of the drivers in the Reflexion Arts loudspeakers (1997).

13.7 Aspects of Small Control Room Designs

Gilbert Briggs, in his book Loudspeakers, published in the 1950s, wrote that loudspeakers are somewhat like boxers … ‘In general, a good big one will always beat a good little one.’ Well, the same can also be said for control rooms; when the size comes down, several things happen which conspire against a neutral sounding room:

1.  Reflexions come back from wall surfaces earlier than in a larger room, because they have travelled less distance. This also means that they reflect more often, and hence small rooms produce a higher reflexion density than larger rooms.

2.  Reflexions come back from wall surfaces higher in level than from similar wall surfaces in larger rooms, again because they have travelled less distance. This gives them a greater ability to interfere with the direct signal.

3.  There is less space in the room to locate absorbent or diffusive materials or structures, so room effects (which are worse to begin with) will tend to be less able to be controlled than in larger rooms.

4.  The resonant room modes (often referred to as standing waves) begin to separate at higher frequencies in smaller rooms, so the reduced ability to absorb or damp the room effects are exacerbated by the greater intrusion of the irregular response range into the low frequency range of the audible spectrum, as shown in Figure 4.15.

5.  It becomes more difficult to work at a suitable distance from wide range, extended low frequency monitor loudspeakers, sufficient to avoid geometrical effects where the multiple loudspeaker drive units fail to gel into one integrated source … Highs can be heard coming from the tweeters, lows from the woofers, and small movements of the head cause changes in the perceived responses. (See Section 12.7.)

Figure 13.11:

Driver position considerations. At position A, all is well. The distance to each of the three loudspeakers is more or less equal. At position B, however, the distance to each loudspeaker is different. The path length to the left-hand bass driver is less than that to the horn, whereas the path length to the right-hand bass driver is longer. A sound emanating simultaneously from any two or more of the drivers will not arrive simultaneously at position B, so a flat frequency response would not be possible due to the phase shifts involved, unless, that is, the crossover frequency was so low that the wavelength at that frequency was significantly greater than the width of the box.

What should be clear from the above five points is that small rooms, in many cases even more so than large rooms, will tend to stamp their own individual ‘fingerprints’ on the sound of any given monitor system used in them. There is therefore a strong argument in favour of large control rooms, but economics and practical considerations have deemed that the majority of recordings are now undertaken in studios with relatively small control rooms … 24 to 40 m² being very typical. So, it would be useful to look at the above points in a little more detail.

Figure 13.12:

Vertical mounting of drivers. With the drivers mounted vertically, the arrival times from each driver to a listener at position ‘A’ are essentially identical. If the listener then walks to position ‘B’, the arrival time from the different drivers remains equal. A flat response is therefore possible at both points A and B, and indeed at all points in between.

Reflexions returning from walls (more quickly and with more energy in small rooms) will tend to colour the sound of the monitors and smear the stereo imaging. Specular reflexions are the worst offenders, so they must be dealt with by means of either diffusion or absorption. In small rooms absorption is the only solution, for whilst diffusion may in some circumstances maintain the stereo imaging over a larger ‘sweet’ area, it nonetheless returns considerable energy to the listening position, and this will colour the sound. Diffusers are statistically diffuse, which means that when area-averaged they are equally diffuse at all frequencies within their design range. However, this does not mean that at any given point there will never exist hotspots at certain frequencies. In small rooms, where the listeners are inevitably close to the diffuse surfaces, such hotspots can be problematical. Angus¹⁴ showed that the situation can, at times, be such that reflected energy from diffusers can be even more concentrated than from a flat reflective surface.

Adequate absorption, even if it does not remove all the reflected energy, will at least reduce the reflected energy and heavily damp the modes, making them broader in their frequency spread and lower in level, both of which tend towards producing less colouration. When the room is so small that the modal separation and entry into the pressure zone begins well into the lower octaves of the audible range, then heavy damping of modal activity is the only way to achieve a reasonably uniform frequency response.

The third point mentioned in the opening paragraph of this section was that in smaller rooms there was less space to install acoustic control contrivances. In the case of many absorber or diffuser systems, their lower-frequency cut-off points are a function of their depth. In other words, if a system needs, for example, 1 m of depth to be effective down to 70 Hz, then it requires that depth independently of the size of the room. Absorber and diffuser sizes cannot be scaled with the size of the room. These depths are fixed by the wavelengths of the frequencies, if 1m is needed to be effective to 70 Hz in a large room, then 1 m is still needed in a small room. Many studio owners complain when a designer asks to use 1 m depth of absorber on a rear wall when the room is only 4 m from front to back in the first place. Nevertheless, physics dictates this, not the designer, so a studio owner resisting the loss of floor space will pay the price of poorer response flatness in the room. Unfortunately, the reality is that to many of them, what they see is more important than what they hear, which is hardly an attitude that can be called professional, but it is rather widespread.

Point four is related to the scaling problem for absorbers and diffusers. Figure 4.15 clearly shows the identical modal pattern for two identically shaped rooms, but one having dimensions three times larger than the other. The modal patterns are a function of the relative room dimensions, i.e. the shape, and very little can be done to change this state of affairs. The perceived room response usually tends to follow the envelope of the modes and the spaces between them, so where the modes are closely spaced, a generally smooth response will be perceived. When the separation begins in the audible low frequency range, as is always the case in small rooms, the individual frequencies that coincide with the modes will be heard to ring on in the room, giving it a ‘boomy’ character on some notes, but not on others.

It is relatively futile to try to change these modal frequencies by angling the walls. Certainly, a small angle to avoid parallelism helps to reduce objectionable flutter echoes or their associated colouration, but as can be seen from Figures 5.7 to 5.10, even an angling of one wall by 1.5 m, between surfaces 10 m apart, has almost no effect on the pressure distribution at 70 Hz, and so will not change the modal frequencies significantly. The only viable solution is to damp the modes, by using the maximum practicable quantity of low frequency absorption on the side walls, the rear wall and the ceiling.

The last of the problems mentioned was the geometric consideration, relating to the spacial separation of the various drive units in a typical monitor loudspeaker. No studio for music recording can seriously call itself professional unless it has adequate full-range monitoring. Relying on small monitors that roll off rapidly below 60 Hz is not a professional approach. However, due to the physics of electroacoustics, loudspeakers which go down to 25 Hz, or thereabouts, and which can produce an adequate sound pressure level, tend to be relatively large. In a small room, large loudspeakers at short distances tend not to have flat responses, and the response tends to change with small movements of the head, which can be misleading.

In small rooms, it seems essential to flush mount the loudspeakers in the front wall of the room. This puts them at points of maximum distance from the listening position. It also avoids the low frequency response irregularities caused by the rear radiating waves reflecting from the front wall (from free-standing loudspeakers). These, as we have already seen, recombine with the direct signal with a phase relationship that causes peaks and dips in the combined response. In fact, flush mounting is the only realistic option for high-quality monitoring other than in free-field (anechoic) conditions. Otherwise, a maximally flat response cannot be achieved. Somewhat perversely, small control rooms tend to be the ones in which one is least likely to find flush mounted monitors, which is a pity, because well-designed small rooms can sound very even.

When looking for a site for a suitable control room, or when designing a building to house one, there are certain things which can be helpful towards later acoustic control. A ceiling height of at least 4 m is almost essential, because floors tend to be flat and hard. They are thus highly reflective to low frequencies, and must be opposed by a ceiling that is capable of dealing with the floor/ceiling modal activity. A floated isolation floor, which is usually required, may need around 20 or 30 cm, and the ceiling of the usual type of acoustic shell could require another 40 cm of height. If a finished ceiling height of 2.4 m was the goal, with a further 15 cm being needed for recessed lighting, then only 85 cm would remain from a 4 m original headroom for mounting acoustic devices to break up the floor/ceiling modes. Remember, when a room is already small, the ceiling height can be used to great effect to control the overall acoustics without losing too much floor space.

Rear walls, of necessity, do tend to need to take up a considerable area of floor space, because effective absorption (or diffusion) requires significant depth in order to deal with the lower frequencies. Somewhat like the floor/ceiling problem, the rear walls face front walls, which almost all tend to be reflective at low frequencies, so they must be adequately treated if a flat response is the objective.

13.7.1 Conflicting Requirements

One of the aspects of small control room construction that creates many problems is the need for sound isolation. If the need for sufficient isolation means the creation of a containment ‘bunker’ that is hard and heavy, then the bunker will be achieving its isolation by reflexion. External noises are kept out by reflexion, and internal noise is kept in by reflexion. The more reflective the containment shell, then the more the absorption that must be provided within the room to keep things under control, and space for this can be very hard to find in small rooms. In the rare cases where little isolation is required, then internal acoustic control tends to become much easier to achieve, because all of the sound that escapes from the room is effectively absorbed by the walls. Transmission through them or absorption within them leads to the same results – less energy reflected back into the room, hence better acoustic control.

These are some of the problems that seriously militate against the achievement of excellent monitoring conditions in small rooms. This is not to say that the objective cannot be achieved, but it can rarely be produced inexpensively, either in terms of money or of lost space. Again, somewhat perversely, money and space are often the two things that many people who want small control rooms do not have at their disposal, which leads to an enormous number of bad small control rooms.

13.7.2 Active Absorbers

Whether the future will see solutions to the small room problems is yet to be seen, but there is the possibility that we could soon see active absorber systems which will take up much less space than conventional passive absorption systems. Active absorbers use driven loudspeakers as energy sinks. The acoustic energy from the sound in the room is converted into heat in the voice coils of the absorber loudspeakers. At high frequencies, the technique is fraught with problems, but at low frequencies, practicable systems could be on the horizon. This sort of technology is rapidly advancing.

13.8 A Short Overview

Rooms form the final link in the sound chain from the loudspeaker to the ear. They are also the most variable links in the chain, and are the most difficult to standardise. In rooms with poor acoustics, no loudspeakers can be expected to perform well. Given rooms with reasonably controlled acoustics, the general tendencies are for rooms with some acoustic ‘life’ to sound more musical, and in some cases to support stereo images over a somewhat wider listening area. Very dead rooms allow perception of timbre and detail that is more accurate, but they are sometimes considered unsatisfactory for the enjoyment of acoustic music. They do, however, allow very precise stereo imaging and great clarity in the designated listening positions, and thus have their places in situations where quality control and attention to detail are important.

As highly damped rooms receive no ‘support’ from reflexions, they tend to require more power output capability from the amplifiers and loudspeakers as compared to more lively rooms of comparable size. On the other hand, because they contribute very little to the total perceived sound field, they are generally more consistent in their performance.

Considering the wide range of sound control room concepts and the great weight of experience that has been applied to their designs, the continuing existence of such variability of implementation suggest that there is no simple solution to the problem of room standardisation which is consistent with the provision of all the desired acoustics. Of course, the desired acoustic can be very personal – and most of the generally accepted control room philosophies have their partisan followers. Somewhat similarly, there is not one type of tennis racquet which is used to win all championships. Different weights, different string tensions and different designs suit different styles of play. No single room environment has been shown to be ‘correct’ for all purposes.

13.9 Summary

Control rooms should have a clarity of monitoring that is always one step ahead of the consumer systems.

Studio monitors are designed according to a different hierarchy of priorities compared to domestic hi-fi loudspeakers.

Most high-quality stereo control room designs rely on directional acoustics. The front and rear halves of the rooms are acoustically different.

Different designers have different opinions as to whether the front half or the rear half of control rooms should be reflective, and whether to use absorption or diffusion as the principal means of acoustic control.

For the best overall response, monitor loudspeakers should be flush mounted in the front wall. The increased reflexion density in small rooms makes adequate control all the more necessary. Flush mounting helps the low frequency control.

Because absorber and diffuser sizes are wavelength related, control can become difficult in small rooms which lack the available space for treatment.

Loudspeaker responses in the pressure zones are subject to many influencing factors.

For optimum performance, a loudspeaker and a room should be designed as one system.

At close distances, multi-way loudspeakers may not gel into a cohesive source.

At close distances, diffusers may produce reflective hot-spots.

References

1  Rodgers, C. A. (Puddie), ‘Pinna Transformation and Sound Reproduction’, Journal of the Audio Engineering Society, Vol. 29, No. 4, pp. 226–34 (April 1981)

2  Davis, Don and Davis, Chips, ‘The LEDE Concept for the Control of Acoustic and Psychoacoustic Parameters in Recording Control Rooms’, Journal of the Audio Engineering Society, Vol. 28, No. 9, pp. 585–95 (September 1980)

3  Haas, H., ‘The Influence of a Single Echo on the Audibility of Speech’, Acoustica, Vol. 1, p. 49 (1951). Reprinted in the Journal of the Audio Engineering Society, Vol. 20 (1972). This work was also investigated by Wallach, H. et al., American Journal of Psychology, Vol. 62, p. 315 (1949)

4  D’Antonio, P. and Konnert, J. H., ‘The RFZ/RPG Approach to Control Room Monitoring’, presented at the 76th Audio Engineering Society Convention, Preprint No. 2157, New York, (1984)

5  Newell, P. R., Holland, K. R. and Hidley, T., ‘Control Room Reverberation is Unwanted Noise’, Proceedings of the Institute of Acoustics, Vol. 16, Part 4, pp. 365–73 (1994)

6  Hirata, Y., Matsudaira, T. K. and Nakajima, H., ‘Optimum Reverberation Times of Monitor Rooms and Listening Rooms’, presented to the 68th Convention of the Audio Engineering Society, Hamburg, Preprint No. 1730 (March 1981)

7  Toyashima, S. and Suzuki, H., ‘Control Room Acoustic Design’, presented to the 80th Convention of the Audio Engineering Society, Montreux (Switzerland), Preprint No. 2325 (1986)

8  Walker, Robert, ‘A Controlled Reflection Listening Room for Multi-Channel Surround’, Proceedings of the Institute of Acoustics, Vol. 20, Part 5, pp. 25–36 (1998)

9  Holman, Tomlinson, 5.1 Surround Sound: Up and Running, Focal Press, Oxford, UK (1999)

10  Various authors, ‘Changing Rooms’, Studio Sound, Vol. 43, No. 2, pp. 54–60 (February 2001)

11  Various authors, The Proceedings of the AES 16th International Conference, Spatial Sound Reproduction (1999)

12  Howard, David M. and Angus, James, Acoustics and Psychoacoustics, 2nd Edn, p. 301, Focal Press, Oxford, UK (2001)

13  Davis, Don and Davis, Carolyne, Sound System Engineering, 2nd Edn, p. 209, Focal Press, Oxford, UK and Boston, USA (1997)

14  Angus, James A. S., ‘The Effects of Specular Versus Diffuse Reflections on the Frequency Response at the Listener’, presented at the 106th Audio Engineering Society Convention, Munich (1999). Preprint No. 4938

Bibliography

Adams, Glyn, ‘The Room Environment’ (revised by John Borwick), Chapter 7 in Loudspeaker and Headphone Handbook, 3rd Edn (John Borwick ed.) Focal Press, Oxford, UK (2001)

Davis, Don and Davis, Carolyn, Sound System Engineering, 2nd Edn, Focal Press, Oxford, UK and Boston, USA (1997)

Howard, David M. and Angus, Jamie, Acoustics and Psychoacoustics, 4th Edn, Focal Press, Oxford, UK (2009)

Kutruff, H., Room Acoustics, 4th Edn, E. & F. N. Spon Publishers, London, UK (2001)

Newell, Philip, Studio Monitoring Design, Focal Press, Oxford, UK (1995)

Toole, Floyd E., ‘Listening Tests – Turning Opinion into Fact’, Journal of the Audio Engineering Society, Vol. 30, No. 6, pp. 431–45 (June 1982)

Toole, Floyd E., ‘Subjective Measurements of Loudspeaker Sound Quality and Listening Performance’, Journal of the Audio Engineering Society, Vol. 33, No. 1–2, pp. 2–31 (1985)

Toole, Floyd E., ‘Loudspeaker Measurements and Their Relationship to Listener Preference: Part 1’, Journal of the Audio Engineering Society, Vol. 34, No. 4, pp. 227–35 (1986)

Toole, Floyd E., ‘Loudspeaker Measurements and Their Relationship to Listener Preference: Part 2’, Journal of the Audio Engineering Society, Vol. 34, No. 5, pp. 323–48 (1986)