Understanding the Principles of Sound Reproduction--Preserving the Art

A scientific understanding of the physical world has allowed us to do many remarkable things. Among them is the ability to enjoy the emotions and aesthetics of music whenever or wherever the mood strikes us. Music is art, pure and simple. Composers, performers, and the creators of musical instruments are artists and craftsmen. Through their skills, we are the grateful recipients of sounds that can create and change moods, that can animate us to dance and sing, and that form an important component of our memories. Music is a part of all of us and of our lives.

However, in spite of its many capabilities, science cannot describe music.

There is nothing documentary beyond the crude notes on a sheet of music.

Science has no dimensions to measure the evocative elements of a good tune.

It cannot technically describe why a famous tenor's voice is so revered or why the sound of a Stradivarius violin is held as an example of how it should be done. Nor can science differentiate, by measurement, between the mellifluous qualities of trumpet intonations by Wynton Marsalis and those of a music student who simply hits the notes. Those are distinctions that must be made subjectively, by listening. A lot of scientific effort has gone into understanding musical instruments, and as a result, we are getting better at imitating the desirable aspects of superb instruments in less expensive ones. We’re also getting better at electronically synthesizing the sounds of acoustical instruments. However, the determination of what is aesthetically pleasing remains firmly based in subjectivity.

This is the point at which it’s essential to differentiate between the production of a musical event and the subsequent reproduction of that musical event.

Subjectivity-pure opinion-is the only measure of whether music is appealing, and it will necessarily vary among individuals. Analysis involves issues of melody, harmony, lyrics, rhythm, tonal quality of instruments, musicianship, and so on. In a recording studio, the recording engineer becomes a major contributor to the art by adjusting the contribution of each musician to the overall production, adjusting the tonal balance and timbre of each of the contributors, and adding reflected and reverberated sounds or other processed versions of captured sounds to the mix. This too is judged subjectively, on the basis of whether it reflects the artists' intent and, of course, how it might appeal to consumers.

The evaluation of reproduced sound should be a matter of judging the extent to which any and all of these elements are accurately replicated or attractively reproduced. It’s a matter of trying to describe the respects in which audio devices add to or subtract from the desired objective. A different vocabulary is needed. However, most music lovers and audiophiles lack this special capability in critical listening, and as a consequence, art is routinely mingled with technology. In subjective equipment reviews, technical audio devices are often imbued with musical capabilities. Some are described as being able to euphonically enhance recordings, and others to do the reverse. It’s true that characteristics of technical performance must be reflected in the musical performance, but it happens in a highly unpredictable manner, and such a commentary is of no direct assistance in our efforts to improve sound reproduction.

In the audio industry, progress hinges on the ability to identify and quantify technical defects in recording and playback equipment while listening to an infinitely variable signal: music. Add to this the popular notion that we all "hear differently," that one person's meat might be another person's poison, and we have a situation where a universally satisfying solution might not be possible.

Fortunately reality is not so complex, and although tastes in music are highly personal and infinitely variable, we discover that recognizing the most common deficiencies in reproduced sounds is a surprisingly universal skill. To a remark able extent we seem to be able to separate the evaluation of a reproduction technology from that of the program. It’s not necessary to enjoy the program to be able to recognize that it is, or is not, well reproduced.

How do listeners approach the problem of judging sound quality? Most likely the dimensions and criteria of subjective evaluation are traceable to experiences in live sound, even simple conversation. If we hear things in sound reproduction that could not occur in nature or that defy some kind of logic, we seem to be able to identify it. But, as in many other aspects of life, some of what we regard as "good" is governed by a cultivated taste. Factors contributing to the prevailing taste at different points in time are interesting to discuss, as they will ring bells of familiarity in the minds of many readers. Thompson (2002) states that "culture is much more than an interesting context in which to place technological accomplishments; it’s inseparable from technology itself".

1. BACK TO THE BEGINNING: CAPTURING SOUND QUALITY

In terms of basic sound quality, claims of accurate reproduction began early.

Edison, in 1901, claimed that the phonograph had no "tone" of its own. To prove it, he mounted a traveling show in which his phonograph was demonstrated in "tone tests" that consisted of presentations with a live performer.

Morton (2000) reports, " Edison carefully chose singers, usually women, who could imitate the sound of their recordings and only allowed musicians to use the limited group of instruments that recorded best for demonstrations". Of a 1916 demonstration in Carnegie Hall before a capacity audience of "musically cultured and musically critical" listeners, the New York Evening Mail reported that "the ear could not tell when it was listening to the phonograph alone, and when to actual voice and reproduction together. Only the eye could discover the truth by noting when the singer's mouth was open or closed" (quoted in Harvith and Harvith, 1985, p. 12).

FIG. 1 Singer Frieda Hempel stages a tone test at the Edison studios in New York City, 1918. Care was taken to ensure that the test was "blind," but it’s amusing to see that some of the blindfolds also cover the ears. Courtesy of Edison National Historic Site, National Park Service, U.S. Department of the Interior.

Singers had to be careful not to be louder than the machine, to learn to imitate the sound of the machine, and to sing without vibrato, which Edison (apparently a musically uncultured person) did not like. There were other con sequences of these tests on recordings. The low sensitivity of the mechanical recording device made it necessary for the performers to crowd around the mouth of the horn and find instruments that could play especially loud. Because the promotional "tone tests" were of solo voices and instruments, any acoustical cues from the recording venue would reveal the recording as being different from the live performer in the demonstration room. Consequently, in addition to employing what was probably the first "close microphone" recording technique, Edison's studios were acoustically dead (Read and Welsh, 1959, p. 205).

Live versus reproduced comparison demonstrations were also conducted by RCA in 1947 [using a full symphony orchestra (Olson, 1957, p. 606)], Wharfedale in the 1950s (Briggs, 1958, p. 302), Acoustic Research in the 1960s, and probably others. All were successful in persuading audiences that near perfection in sound reproduction had arrived. Based on these reports, one could conclude that there had been no consequential progress in loudspeaker design in over 50 years. Have we come much farther a century later? Are today's loud speakers significantly better sounding than those of decades past? The answer in technical terms is a resounding "Yes!" But would the person on the street or even a "musically cultured" listener be able to discern the improvement in such a demonstration? Are we now more wise, more aurally acute, and less likely to be taken in by a good demonstration? When the term high fidelity was coined in the 1930s, it was more a wishful objective than a description of things accomplished. Many years would pass before anything resembling it could be achieved. Although recreating a live performance was an early goal, and it remains one of the several options today, the bulk of recordings quickly drifted into areas of more artistic interpretation.

Morton (2000) states, "The essence of high fidelity, the notion of 'realism,' and the uncolored reproduction of music dominated almost every discussion of home audio equipment. However, commercial recordings themselves betrayed the growing divide between the ideals of high fidelity and the reality of what happened in the recording studio".

Whatever the musical content, high culture or low, there is still a reference sound that we must emulate in our listening spaces, and that sound is the sound of the final mix experienced in a recording control room. The sound of live per formers is a kind of hidden reference, buried in all of our subconscious minds.

It’s relevant but not in the "linear" manner implied in many heated arguments over the years. Many believe that only by memorizing the sound of live performances can one judge the success of a reproduction. The principal issue is that few of us have ever heard the sound of voices and instruments with our ears at the locations of microphones. These are normally placed much closer to mouths and instruments than is desirable or even prudent for ears in live performances.

The sounds captured by microphones are not the sounds we hear when in an audience. Microphones hanging above the string section of an orchestra inevitably pick up a spectrum that has a high-frequency bias; the sound is strident compared to what is heard in the audience due to the directional radiation behavior of violins (Meyer, 1972, 1978, 1993). The total sound output of a piano cannot be captured by the close placement of any single microphone, which therefore means that such recordings cannot accurately represent what is heard by listeners in unamplified performances in natural acoustical spaces. Yet, they have been good enough approximations to give plea sure to generations of listeners.

The practical reality is that all recordings end up in a control room of some sort, where decisions are made about the blending of multiple microphone inputs, sweetening with judicious equalization, and enrichment with electronically delayed sound. This is the second layer of art in recordings, added by some combination of recording engineers and performing artists. There are many written discussions of how to "monitor" the progress of recordings-some in books and many more in magazine articles. Opinions cover an enormous range.

Many mixers choose monitor loudspeakers that add a desired quality to the sound, instead of using a neutral monitor and achieving the same desired quality through signal processing. This attitude, which seems to be depressingly common, leads nowhere, because only those who are listening through the same loudspeakers will be able to hear that desirable sound quality. A good, recent perspective on the topic can be found in Owsinski (1999), where comments from the author and 20 other recording mixers are assembled. They all care about what their customers hear, but they differ enormously in how to estimate what that is. Using known "imperfect" loudspeakers, such as Yamaha NS-10Ms and Auratone 5Cs, is popular. Some also listen in their cars and through various renderings of inexpensive consumer products. The problem is that there are countless ways to be wrong (bad sound).

Choosing a single or even a small number of "bad" loudspeakers cannot guarantee anything. Nobody in this massive industry seems to have undertaken a statistical study of what might be an "average" loudspeaker. The author's experience suggests that the performance target for almost all consumer loud speakers is a more-or-less flat axial frequency response. Failure to achieve the target performance takes all possible forms: lack of bass, excessive bass, lack of mids, excessive mids, lack of highs, excessive highs, prominent resonances at arbitrary frequencies, and so on. The only common feature that distinguishes lesser products seems to be a lack of low bass. In short, an excellent approach to choosing a monitor loudspeaker would be to choose a state-of-the-art "neutral" device, adjust it to perform in its specific acoustical environment, and then electronically introduce varying degrees of high-pass filtering to simulate reproduction through anything from a clock radio to a minisystem. It’s very perplexing that no truly reliable technical standards for control room sound exist, making the reference a moving target.

2. BACK TO THE BEGINNING: DIRECTION AND SPACE

Sounds exist in acoustical contexts. In live performances we perceive sources at different locations, and at different distances, in rooms that can give us strong impressions of envelopment. A complete reproduction should convey the essence of these impressions. A moment's thought reveals that because our binaural perceptual mechanism is sensitive to sounds arriving from all angles, reproducing a persuasive illusion of realistic direction and space must entail multiple channels delivering sounds to the listener from many directions. The key questions are how many and where? This aspect of sound perception has been greatly influenced by both recording technology and also by culture. Blesser and Salter (2007) discuss this in terms of "aural architecture," defined as those properties of a space that can be experienced by listening. This begins with natural acoustical environments, but nowadays we can extend this definition to include those real and synthesized spatial sounds incorporated in recordings and those that are reproduced through loudspeakers in our listening rooms. In this sense, all of us involved with the audio industry are, to some extent, aural architects.

From the beginning, we have come to associate certain kinds of sounds with specific architectural structures; For example, a highly reverberant spacious illusion is anticipated when we see that we are in a large stone cathedral or a multistory glass and granite foyer of an office building. Rarely are our expectations not met as we make our way through the physical acoustical world.

However, in recordings, we now have the technology to deliver to a listener's ears some of the spatial sounds of a cathedral while seated in a car or living room. But are the illusions equally persuasive? Are they more persuasive if there is an image of the space on a large screen? Auditory spatial illusions are no longer attached to visual correlates; they exist in the abstract, conceivably a different one for every instrument in a multi miked studio composition. Traditionalists complain about such manipulations, but most listeners consider them just another form of sensory stimulation.

Blesser and Salter (2007) said, "Novelty now competes with refinement".

All of this stands in stark contrast to the spatially deprived decades that audio has endured. It began with the first sound reproduction technology, mono phonic sound, which stripped music of any semblance of soundstage, space, and envelopment. This was further aggravated by the need to place microphones close to sources; early microphones had limited dynamic range and high back ground noise. Adding further to the spatial deprivation was the use of relatively dead recording studios and film soundstages. Read and Welsh (1959) explained, "Reproduced in the home, where upholstered furniture, drapes and rugs quite often prevented such an acoustical development of ensemble through multiple reflections, the Edison orchestral recordings were often singularly unappealing" (p. 209). Recording engineers soon learned that multiple microphones could be used to simulate the effects of reflecting surfaces, so the natural acoustics of the recording studio were augmented, or even replaced, by the tools and techniques of the sound recording process.

With the passage of time, directional microphones gave further control of what natural acoustics were captured. With relatively "dead" source material, it became necessary to add reverberation, and the history of sound recording is significantly about how to use reverberation rooms and electronic or electromechanical simulation devices to add a sense of space. In the past, these effects were used sparingly and "the typical soundtrack of the early 1930s emphasized clarity and intelligibility, not spatial realism" (Thompson, 2002, p. 283).

A coincidental influence was the development of the acoustical materials industry. In the 1930s, dozens of companies were manufacturing versions of resistive absorbers-fibrous fluff and panels-to absorb reflected sound and to contribute to acoustical isolation for bothersome noises. Acoustical treatment became synonymous with adding absorption. Dead acoustics were the cultural norm-the "modern" sound-which aligned with recording simplicity, low cost, small studios and profitability (Blesser and Salter, 2007, p. 115). Thompson (2002) explains, "When reverberation was reconceived as noise, it lost its traditional meaning as the acoustic signature of a space, and the age-old connection between sound and space-a connection as old as architecture itself-was severed".

Read and Welsh (1959) recount the following statement, written in 1951 by popular audio commentator Edward Tatnall Canby in his "Saturday Review of Recordings":

"Liveness," the compound effect of multiple room reflections upon played music, is--if you wish--a distortion of "pure" music; but it happens to be a distortion essential to naturalness of sound. Without it, music is most graphically described as "dead." Liveness fertilizes musical performance, seasons and blends and rounds out the sound, assembles the raw materials of overtone and fundamental into that somewhat blurred and softened actuality that is normal, in its varying degrees, for all music. Disastrous experiments in "cleaning up" music by removing the all-essential blur long since proved to most recording engineering that musicians do like their music muddied up with itself, reflected.

Today recording companies go to extraordinary lengths to acquire studios, churches and auditoriums (not to mention an assortment of artificial, after-the-recording liveness makers) in order to package that illusively perfect liveness. This notion that reflections result in a corruption of "pure" music, and the apparent surprise in finding that musicians and ordinary listeners prefer "muddied up" versions, reappears in audio even today. We now have quite detailed explanations why this is so, but one can instinctively grasp the reality that, toward the rear of a concert hall, the direct sound (the "pure" music) is not the primary acoustic event. It may even be inaudible, masked by later acoustical events. Two ears and a brain comprise a powerful acoustical analysis tool, able to extract enormous resolution, detail, and pleasure from circumstances that, when subject to mere technical measurements, seem to be disastrous.

Something that in technical terms appears to be impossibly scrambled is perceived as a splendid musical performance.

When Sabine introduced the concept of reverberation time into acoustical discussions of rooms at the turn of the last century, he provided both clarification and a problem. The clarification had to do with adding a technical measure and a corresponding insight to the temporal blurring of musical patterns that occur in large live spaces. The problem appeared when recordings made in spaces that were good for live performances were often perceived to have too much reverberation. A single microphone sampling such a sound field that was then reproduced through a single loudspeaker simply did not work; it was excessively reverberant. Our two ears, which together allow us to localize sounds in three dimensions to separate individual conversations at a cocktail party and to discriminate against a background of random reverberation, were not being sup plied with the right kind of information. Multiple microphones that could convey information through multiple channels and deliver the appropriate sounds to our multiple (binaural) ears were necessary, but they were not avail able in the early years.

This disagreement between what is measured and what is heard has been the motivation for much scientific investigation of the acoustics of rooms, both large and small. In some ways, our problems with rooms, especially small rooms, began when we started to make measurements. Our eyes were offended by things seen in the measurements, but our ears and brain heard nothing wrong with the audible reality. As we will see, some of the resolution of the dilemma is in the ability of humans to adapt to, and make considerable sense of, a wide variety of acoustical circumstances. Separating sound sources from the spaces they are in is something humans do routinely.

Old habits die hard. The introduction of stereo in the 1950s gave us an improved left/right soundstage, but close microphone methods, multitracking, and pan-potting, did nothing for a sense of envelopment-of actually being there. The classical music repertoire generally set a higher standard, having the advantage of the reflectivity of a large performance space, but a pair of loudspeakers deployed at ±30° or less is not an optimum arrangement for generating strong perceptions of envelopment (as will be explained later, this needs additional sounds arriving from further to the sides). Perhaps that is why audiophiles have for decades experimented with different loudspeaker directivities (to excite more listening room reflections), with electronic add-ons and more loudspeakers (to generate delayed sounds arriving from the sides and rear), and with other trinkets that seem capable only of exciting the imagination. All have been intended to contribute more of "something that was missing" from the stereo reproduction experience. The solution to this is more channels.

3. A CIRCLE OF CONFUSION

When we listen to recorded music, we are listening to the cumulative influences of every artistic decision and every technical device in the audio chain. Many years ago, I created the cartoon in FIG. 2 to illustrate the principle and suggest how we may break the never-ending cycle of subjectivity.

The presumption implicit in this illustration is that it’s possible to create measurements that can describe or predict how listeners might react to sounds produced by the device being tested. There was a time when this presumption seemed improbable, but with research and the development of newer and better measurement tools, it has been possible to move the hands of the "doomsday clock" to the point where detonation is imminent. FIG. 3 expresses the ideas of the "circle of confusion" in a slightly different form, one that more accurately reflects the impact on the audio industry.

FIG. 2 The first version of the "circle of confusion," illustrating the key role of loudspeakers in determining how recordings sound and of recordings in determining our impressions of how loudspeakers sound. The central cartoon suggests how the "circle" can be broken, using the knowledge of psychoacoustics to advance the clock to the detonation time at which the "explosive" power of measurements will be released to break the circle.

===

FIG. 3 (a) The "circle of confusion" modified to more accurately reflect its effect on the audio industry. The true role of loudspeakers is shown here. (b) Unless the loudspeakers involved in the creation of the recordings are similar in performance to those used in reproduction, the "art" is not preserved.

monitor LOUDSPEAKERS that are evaluated using; microphones, equalization, ambience, and other EFFECTS evaluated using; RECORDINGS that are made using; that are enjoyed using consumer audio systems in homes, cars, etc; that are used as test signals to subjectively evaluate audio equipment

===

4. BREAKING THE CIRCLE: PROFESSIONALS HOLD THE KEY

The audio industry has developed and prospered until now without any meaningful standards relating to the sound quality of loudspeakers used by professionals or in homes. The few standards that have been written for broadcast control and music listening rooms applied measurements and criteria that had no real chance of ensuring good, or even consistent, sound quality. Many years ago, the author participated in the creation of certain of those standards and can report that the inadequacies were not malicious, only the result of not having better information to work with. The film industry has long had standards relating to the performance of loudspeakers used in sound-mixing stages and cinemas.

These too are deficient, but something is better than nothing.

FIG. 4 The result of 250 frequency response measurements made at the engineer's location in many recording control rooms, using functionally similar Genelec loudspeakers. The curves show the maximum/minimum variations exhibited by different percentages of the situations. As we will see later, such "room curves" are incomplete data when it comes to revealing overall sound quality, but they are excellent indicators of sound quality at low frequencies. With similar loudspeakers, as is the case here, they are also indicators of the influences of room, loudspeaker mounting, equalization, and so forth. The real point of this display is to show the range of variations that occur in these critically important professional audio venues. The curves are 1/3-octave smoothed. From Mäkivirta and Anet (2001b).

A consequence of this lack of standardization and control is that recordings vary in sound quality, spectral balances, and imaging. Proof of this is seen each time a person reaches for a CD to demonstrate the audio system they want to show off. The choice is not random. Only certain recordings are on the "demo" list, and each will have favorite tracks. This is because the excitement comes not in the music-the tune, lyrics, or musical interpretation-but in the ability to deliver a "wow" factor by exercising the positive attributes of the system.

A recent survey of recording control rooms revealed a disturbing amount of variation in spectral balance among them (Mäkivirta and Anet, 2001a, b). FIG. 4 shows that the differences were not subtle, especially at low frequencies.

FIG. 5 On-axis frequency responses for six entry-level consumer "minisystems," incorporating CD/cassette/AM & FM tuners/amplifiers and loudspeakers. Prices ranged from $150 to $400 for these plastic-encased, highly styled units. These were among the most popular low-cost integrated systems in the marketplace in the year 2000. The heavy curve is the average response, which falls within a tolerance of ±3 dB from 50 Hz to 20 kHz.

FIG. 6 (a) On- and off-axis frequency responses of an Auratone 5C, a five-inch (127 mm) full-range minimonitor. As can be seen in the photograph, it’s in a small closed box enclosure which contributes to seriously attenuated bass, and the single diaphragm shows evidence of resonances and strong directivity at high frequencies. (b) Comparable measurements for a UREI 811B large monitor loudspeaker of the same era, showing improved bass but similarly bad off-axis performance at middle and high frequencies.

Recording engineers who work in these circumstances, presumably approving of them, are doing the art no favor. This is an excellent example of the circle of confusion in action because members of this group of audio professionals cannot even exchange their own recordings with a reasonable certainty of how they will sound in one another's control rooms.

Many audio professionals insist on using their own recordings when setting up new or alternative mixing or monitoring sites. All this does is to perpetuate whatever distortions were built into the original site. There is no opportunity for improvement, of moving to more "neutral" territory. It’s disturbing to hear some such people argue that they attribute some of the success of their prior recordings to a monitoring situation that is clearly aberrant (one can only imagine that the composers and musicians feel insulted to hear that their contributions are subservient to a loudspeaker!). This is the kind of misguided argument that has led normally sensible people to promote the use of obviously "less than high-fidelity" loudspeakers for monitoring, on the basis that the majority of consumers will be listening through such loudspeakers.

It’s true that the majority of consumers live with mediocre, even downright bad, reproduction systems. The problem is that it’s possible to be "bad" in an infinite number of different ways, so any boom box or rotten little speaker that is chosen to represent "bad" is just one example of how to be bad, not a universal reference. The only aspect of their performance that is likely to be at all universal is the lack of low bass.

Looking at FIG. 6, it appears that all of the designers tried to make a "flat" system, but each of them failed in a different way. Plotting an average of these systems yields a respectably flat curve (50 Hz to 20 kHz, ±3 dB), although individual systems deviated greatly, but differently, from this specification. This suggests that using a monitor loudspeaker with a flat frequency response might be a good way to please a large percentage of entry-level listeners, as well as those with superb audio systems. However, there is one very important proviso: all of these small systems exhibit a serious lack of low bass, so if one uses a state-of-the-art monitor loudspeaker for all evaluations, it will be necessary to incorporate a high-pass filter in the signal paths to attenuate the low bass. This simple act will enable recording engineers to condition their recordings so they will sound good through average "bad" systems and remove more clutter from control rooms.

Discussing this topic from the perspective of mastering engineering, Katz (2002) basically agrees:

Mastering engineers confirm that accurate monitoring is essential to making a recording that will translate to the real world. The fallacy of depending on an inaccurate "real world-monitor" can only result in a recording that is bound to sound bad on a different "real-world monitor." Even the best master will sound different everywhere, but it will sound most correct on an accurate monitor system. Which leads us to this comment from a good client: "I listened to the master on half a dozen systems and took copious notes. All the notes cancelled out, so the master must be just right." (p. 82) FIG. 6 shows anechoic frequency-response measurements on two popular monitor loudspeakers of years past (some studios even today proudly advertise that either or both are still available). One is a small "near field" type, the Auratone, used to evaluate recordings as they might be heard out in the "real world." The UREI is a traditional large woofer, horn high-frequency configuration that typically would be built into a soffit or wall. The large loudspeaker has much better bass and can play at much higher sound levels, but in terms of the sound qualities of these two loudspeakers, there are more similarities than differences, and both have serious problems. Because of their distinctive imperfections, such loudspeakers are references only for themselves. The prime asset of the Auratone as a window into the world of bad loudspeakers is its lack of bass. As an indication of how far some of the guardians of our musical arts have strayed, it has been said that these minimonitors have "single-driver musicality." The mind is a marvelous instrument. We will see in SECTION 18 that the sound quality traditions of loudspeaker like these have been perpetuated in some contemporary products. Bad habits are hard to break.

Reflecting on all of this, it’s easy to be pessimistic about the integrity of the circle of confusion as it applies to sound quality. Things are improving, however, as we will see in SECTION 18, where it’s shown that, although the Auratone tonal personality can be seen in other newer products, there is another stream of superbly designed monitor loudspeakers that are remarkably neutral. In between, there is simply a boring collection of different versions of mediocrity.

Evidence exists that audio professionals are as susceptible to a good marketing story as are consumers, and, without double-blind listening tests, their opinions are just as susceptible to bias.

With large, professionally designed studio complexes now being replaced or supplemented by home studios, the challenges have multiplied. We need to master how to reproduce good sound in relatively ordinary rooms of different configurations.

Returning to the concept of being aural architects, Blesser and Salter (2007) usefully summarize the situation:

Acoustic engineers determine the physical properties of the recording environment; design engineers develop the recording and reproduction equipment; recording engineers place the microphones; mixing engineers prepare the final musical product for distribution; interior decorators select furnishings for the listeners' acoustic space; and listeners position themselves and the loudspeakers within that space. Often acting independently, these individuals are members of an informal and unrecognized committee of aural architects who don’t communicate with one another. With their divided responsibility for the outcome, they often create the spatial equivalent of a camel: a horse designed by a committee. So listeners are merely the last in a long line of aural architects but with no influence on, or connection with, what has happened before.

No matter how meticulously the playback equipment has been chosen and set up, and no matter how much money has been lavished on exotic acoustical treatments, what we hear in our homes and cars is, in spatial terms, a matter of chance. Blesser and Salter conclude that "spatial accuracy is not a significant criterion for much of our musical experience". They go on to explain, "The application of aural architecture to cinema is a good example of aesthetically pleasing spatial rules that never presume a space as a real environment.

Artistic space never represented itself as being a real space; it’s only the experience of space that is real; and achieving artistic impact often requires spatial contradictions".

5. MEASURING THE ABILITY TO REPRODUCE THE ART

The contradiction implicit in the title of this section will reverberate through this guide. How can we measure something that subjectively we react to as art.

Measurements are supposed to be precise, reproducible, and meaningful. Perceptions are inherently subjective, evanescent, and subject to various non-auditory influences within and surrounding the human organism. However, perceiving flaws in sound reproducing systems appears to be an activity that we can substantially separate from our critique of the art itself. We can detect flaws in the reproduction of music of which we have no prior knowledge and in which we find no pleasure.

The audio industry uses-indeed, needs-measurements to define bench marks of what is acceptable or not. Blesser and Salter (2007) contribute a simple, but not totally reassuring, perspective on the value of measurements. It begins with the recognition of a hierarchy in hearing:

¦ At the lowest level is sensation, an indication that the organism reacts to a sound-a detection threshold. This is probably quite well related to physical measurements of the sound.

¦ The next level is perception, which incorporates cognitive processes embracing cultural and personal experiences. Here we recognize what it’s that we heard, and perhaps initiate a process of adaptation. This means that some features in measurements may be neutralized by adaptation, and no longer be relevant.

¦ At the highest level of response to sound, we attribute meaning to the recognition, and this can range from irrelevant to highly relevant, from undesirable to good. Depending on the informational content of the sound, we may choose to pay attention or to ignore it. In the latter case, it matters not what measurements tell us.

As Blesser and Salter say, "Detectable attributes may not contribute to perceptible attributes, and perceptible attributes may not be emotionally or artistically meaningful. . . . Furthermore, affect can be at once meaningful and undesirable". What we, as individuals, consider to be meaningful and desirable is largely learned, although some of us show a more or less native ability to hear certain spatial and other attributes of sound. At this level of cognition, measurements are of dubious value.

Einstein's well-known quote is relevant: "Not everything that can be counted counts, and not everything that counts can be counted." The audio business hopes to convey much more than raw sensation; it aspires to perceptions and meanings as well. So, how do we quantify acoustical parameters in ways that correlate with the full panorama of subjective responses of individuals having wide-ranging personal and cultural characteristics? The hope is that we may be able to "connect" with some of the key underlying perceptual dimensions. The fact that several SECTIONs follow this one signals that there has been some success at doing so.