1. The Problem Stated

When we talk about our musical experience, by which I mean our experience with the sounds that we perceive as music, what is "real"? There is an assumption that pervades much computer music research, not to mention more popular accounts of music, that acoustic instrumental sounds are real whereas synthesized sounds are unreal or artificial, even when people are not complaining that the synthesized sounds are not expressive or interesting to listen to. Beyond this, there may be implicit understandings that synthetic sounds should be beyond the bounds of our consideration, or that we do not need to talk about instrumental sounds because we already know all about them. In this discussion, I intend to focus upon these problems with a view to clarifying some of the issues in electronic music and musical perception, and with the goal of encouraging people to open up their minds to new and unusual musical experiences.

2. Basic Issues of Musical Perception

Music is not our first experience with sound; that is the province of language. But in studying some of the ways we acquire language, we can understand better some of the ways that we perceive music.

Linguists have explained how language is processed by breaking it down into three basic elements: phonetics (the speech sounds of the language), syntax (the way the sounds are put together to form words, phrases, clauses and sentences), and semantics (the meanings of the words). To understand and employ the language, the listener must sufficiently process all three components.

In some ways, the manner in which we acquire musical understanding mirrors these experiences. I would also point out that, while many of us are exposed to music as children, it is unlikely that music forms more than a small part of our early experience with sounds. There is a third category which must be mentioned here, namely our ability to process and understand the sounds of nature, animals, machines, and other things that have sonic structure but which are not employed to communicate ideas or aesthetic experiences (where I am lumping music for the present). These sounds can communicate quite specific properties of our environment to us. Children (and adults) can try to imitate them, sometimes quite successfully. If we try to imagine that we process speech as communication and music as art, or perhaps entertainment, this third category of sound fits somewhere else. Nevertheless, we can try to perceive these sounds as representing a language of communication, as people have attempted to explain the sounds of whales and dolphins as a form of communication, for example. We can also try to listen to them as music, and composers, particularly but not solely of electronic music, have used them both as material and as the source of inspiration. I would also throw open the possibility that someone could try to listen to music as language, although this wouldn't necessarily be a fruitful way of looking at music in general; it is a level that lends a different depth of meaning.

When we first encounter music and begin to try to make sense of it, we do some of the same things that we do when we process language. First, we must discern the sonic elements that go into the music, usually instruments and voices, and place these into separate categories. This is analogous to the phonetic aspect of language. Second, we begin to discern how these elements can be put together to form individual events, motives, phrases, and complete statements of a theme. This aspect of musical perception is analogous to the syntactical aspect of language, and it corresponds to our ability to internalize rules of harmony and other types of musical logic. Finally, we recognize the ability of complete musical compositions to portray more abstract ideas or feelings. This can be compared to the semantic aspect of language, although of course it is not the same thing as language semantics.

3. Cultural Biases

Many people, particularly those who have not been exposed to the music of other cultures, regard tonality as an innate quality of music. They do not realize, or appreciate, that it is partly our culture that has conditioned us to look at these qualities in a specific way. Let me clarify this by means of an analogy.

In the English language, the quality of pitch is used in very specific ways to communicate emphasis and indicate phrase and sentence structure, and indirectly to change meaning. While there may be a great deal of variation with a specific range, most speech is delivered within a narrow pitch range and with changes used for specific purposes. Stress is also achieved in other ways, by employing loudness and duration. Consider the aspects of pitch, loudness, and duration that would be used to convey different interpretations of the same sentence (from Peter Westergaard's Sung Language):

I like Jane's dress.

I like Jane's dress.

I like Jane's dress.

I like Jane's dress.

I like Jane's dress???

If you consider how pitch is used in these examples, you can get some idea of what I am talking about here.

Everybody who speaks English understands these points, and I would argue further that this has a subtle influence on the importance we place on pitch in music. Now consider a different language, such as Chinese, one in which pitch is used for completely different purposes. (I should make it clear that I do not understand Chinese, but I have heard it spoken.) Should it not be the case that listeners to music who speak this language may place some different importance or have a different understanding of music? The music that originated in different cultures organizes the elements of music differently, and in order to appreciate it, we have to acquire new modes of listening.

4. Creative Listening

I believe that listening to music is a creative experience, one in which the listener constructs his own interpretation and understanding of the music, and that this can be different for each listener and each piece. I find myself completely at odds with psychologists or musicians who have studied musical perception and have concluded that there are certain basic processes that apply to all music. Sometimes this is stated in such a way that a piece, when played to a group of musicians, produces one kind of response, and when it is played before a group of untrained listeners it produces something else. While listeners learn that certain strategies may apply to most tonal works, or to the works of a certain composer but not to others, it is a mistake to look for universal principles. The highest form of listening we can experience is one in which we understand and take for granted all of the elements corresponding to the phonetics (processing of sounds) and syntax (harmony, counterpoint and phrasing) of language and look for deeper meanings, perhaps comparable to the "deep structure" of language recognized by Chomsky. In my view, the Schenkerian analysts of tonal music have come closest to this ideal in their interpretations of individual compositions. So have certain theorists of twentieth-century music like Allen Forte in their analyses of seminal works.

Nevertheless, in order to make any interpretation of a piece of music at all, we have to have a point of view from which to investigate it. A point of view is an idea that we take to be relevant and successful in understanding the piece. There is an infinite number of points of view that we can adopt in attempting to interpret a piece of music, and we cannot hope to enumerate them all even in our lifetimes. A point of view gives you a perspective from which you can try to interpret the piece, which may or may not be successful. If a point of view is unsuccessful with respect to a given piece, then all we can conclude is that this is not a fruitful viewpoint from which to examine this piece. It does not mean that it is a bad piece, or that the viewpoint is a bad one for other music. Nevertheless, a point of view brings something of ourselves to the experience, which is unique and possibly different from the way that anyone else views the music. As an example of this process, consider the problems that listeners experienced during the time that atonal music was first being written (mainly by Schoenberg, Webern and Berg). Listeners attempted to hear these new works as tonal, and they didn't make much sense. Rather than concluding that something was wrong with the way in which they were trying to hear it, most listeners simply concluded that it was bad music, thereby depriving themselves of the opportunity to experience the music in more appropriate ways. To an extent, unless they have some helpful guidance, listeners still fall into this same trap.

In case there are any music critics listening to this, I would like to use this opportunity to make an aside about the difference between an evaluation and a judgment about a piece of music. An evaluation is a carefully-considered investigation of a piece of music from a viewpoint that is understood and acknowledged at the outset. Once we agree on the premises, the analysis can be completely objective. A judgment is a statement about our relationship to a viewpoint we may take toward a piece of music, for example, "I like loud music." If we agree on this premise, it is an easy matter to listen to a piece and decide whether or not it measures up to this criterion. The evaluation is objective; the judgment is personal; it does not have to be justified.

5. Active versus Passive Listening

In saying that listening is a creative experience, it is clear that I am asserting that it takes involvement, sometimes even work, to arrive at a well-formed interpretation of a piece of music. Unfortunately, in our culture, where music is thrust upon us in all kinds of unwanted circumstances that discourage active listening, it is difficult to achieve this. Passive listening is everywhere -- in elevators, shopping malls, at the office, in restaurants, and as the background of much broadcasting on both radio and television. Have you ever noticed that the music that is broadcast in these situations (with the exception of advertising, which I will return to shortly) is intentionally boring -- in fact, it degrades the whole idea of music as an art. Most of the pieces do not carry tunes that you are familiar with, or if you are, you recognize that they are from an earlier era. I once had a student who worked for Muzak. He lost his job because bank presidents called the company to complain that people were listening to the music!

Listening while you work is destructive to an active listening experience. Some plant managers play music in the background and deliberately speed it up so that their employees will work faster. I have a student who drives me up the wall in this manner. He listens to music while he works on his music at the computer -- usually a different piece from what he is working on!

Music in advertising is another matter. While few people have studied the music-theoretical aspects of advertising, there is a potentially fruitful field of investigation awaiting for some lucky researchers. Music in advertising works subliminally to try to sell us products which we rarely need and often do not want. The real meaning of functional harmony, prolongation and resolution of dissonance, and the use of music to create an atmosphere either in conflict with or consonant with the visual images, can all be found there. Even this use of music as a more active participant in the listening process does not necessarily advance the cause of creative listening, and actually creates an uncomfortable association for us with certain melodies, harmonies and other musical contexts. It is discouraging to find that we are continually being manipulated by the media to engage in behavior that has the ultimate goal of enriching someone else's pocketbook. Since this now happens to most of us since birth, I doubt if we actually realize how important these images and messages are. The ultimate effect of modern mass culture is to break down the distinctions between illusions and reality.

6. The Basic Properties of Sound

What are the basic elements of music? In elementary textbooks we are taught that they are melody, harmony, rhythm, and timbre, which is explained as those elements that enable us to recognize the instruments that play the sounds. This always struck me as a rather different thing than the others, but if you think about it, the idea of timbre is intimately connected to our perception of each of the other qualities. It is only when we sense a certain sameness about the tones in a melody that we have a reason to conceive of them as a single melody rather than as a succession of unrelated sounds. The same is true of harmonies; sounds that don't blend well are not necessarily heard as a harmony. Rhythm is also a complicated quality that has to be broken down into attack patterns and durations of tones.

Nevertheless, the problem with the traditional definition of timbre is that it is a "bushel-basket" quality that lumps many different characteristics under a single heading. Many people think that timbre is just what we call the spectrum or overtone series of a sound. This in itself raises many other questions, but consider some of the other things that you also have to rely on to be able to recognize the instrument that plays a sound. One of them is envelope, the attack and decay qualities of the sound. Within the two basic categories of sustaining and decaying sounds we can perceive many gradations of difference. Another is modulation, or the periodic variation of the frequency, amplitude, or some other quality of the sound.

The first thing that users of electronic musical instruments learn is that each of these qualities is a separate and distinct element which has to be programmed individually, and they can all be perceived distinctly from one another. A proper explanation of these qualities not only shows them to be distinct, but determined by different acoustic properties as well.

The term timbre is now pretty much used by electronic musicians to be the psychological equivalent of the physical quality of spectrum, which is defined as a list of overtones that accompany the sound, each with its own envelope. This is one of the most complicated yet basic qualities of musical perception. First, overtones have to be broken down into harmonic or < b>non-harmonic partials. The frequencies of harmonic partials are integral multiples of the fundamental, which is (usually) the perceived pitch. In fact, it is the presence of harmonic partials that creates the sensation of pitch. Sounds with non-harmonic partials, including the portions of complex sounds that contain both harmonic and non-harmonic partials, do not create the sensation of a single pitch, but sound more like a "cluster". Noise is the ultimate cluster, which does not create the sensation of pitch at all, and which is explained sometimes as possessing all frequencies simultaneously. Between noise and the sine tone, which is a single frequency with no overtones, there is a smooth continuum which varies from cluster sounds to single pitches with many overtones to the single sine tone. Until the present, most music dealt just with single pitches, and the concept of pitch was elevated to the highest level of perceptual qualities in the music. Twentieth-century, particularly electronic music, has explored these other sounds.

Since psychological timbre is supposed to be what enables us to recognize different sounds on the same instrument, it is obvious that the concept has to deal with different pitches. I sometimes wonder if it is even meaningful to compare very high to very low pitches at all; with high pitches, we have to account for the fact that fewer harmonic partials may lie within the limits that can be perceived by human beings. Nevertheless, when we consider different pitches over any range, we have to take account of the fact that there can be different spectral relations that can be taken as the basis of timbral similarities. The two most important are waveform transposition, by which the same spectrum is used on all notes being compared, and < b>formants, which use the same resonances as the basis of comparison. Waveform transposition is the basic similarity possessed by members of the clarinet family, and formants are what give double-reed instruments their distinct timbres. Formants are also the qualities that enable listeners of speech identify different vowels, so this is obviously a very basic similarity.

Our understanding of the properties of sounds has evolved over the years. It was the great nineteenth-century scientist Hermann Helmholz who developed our basic views of the properties of sound. He probably would have gone much farther if he had possessed the means of measuring these qualities that are now available with computers. But nevertheless, recognizing a property present in musical sound, like creative listening, also requires a concept from which you must conduct your investigation, and one of the problems in contemporary research is that few people bother to look deeply into this problem. Instead, they simply assume that the traditional properties are all that matter and proceed from the viewpoint that our understanding of them is incomplete, not measured properly, or requires special training to recognize. The modern explanation of the separation of envelope, spectrum or timbre, and modulation has only developed in the last thirty or so years. Other properties have also been investigated recently. What is now lumped under the heading of reverberation and effects may more properly be understood as a number of different qualities having to do with delays, comb filters, and modulation of these properties. Within the last ten years, new synthesizers have been developed which allow these properties to be included with and structured along with the generation of the tones. While reverberation has been used for many years to create room simulations and different acoustic environments in which to play music, only now can different reverberant qualities be used simultaneously within the same music, and I don't think that most listeners or composers yet know what to make of all these possibilities, which have not yet been widely used in compositions.

7. The Recognition of Instruments

How important is it that we recognize instruments when we listen to music? If you watch the musicians as they play the music in a concert situation, I think you would be convinced that this is an important skill to possess; but on the other hand, it is less likely that you would have difficulty in recognizing them under these conditions. If you are watching a live synthesizer concert, you may experience confusion. The main reason why we even ask this question is that most of us now hear the overwhelming majority of all the music that we hear in our lifetimes through recordings. The mental reconstruction of the instrumental qualities of the sounds has thus become an important aspect of listening to the music itself.

When we do not have the visual cues to help identify the instruments, we become aware of many problems and subtleties that are present in nearly all music. Many instruments are hard to tell apart, particularly on certain tones. The instrumental families (strings, brasses and woodwinds) all possess certain similarities and often blend together in such ways that make it difficult to isolate a single event. Some music itself encourages this merging of sounds in order to create certain aspects of its structure. Some people would argue that aspects like melody and harmony are in fact the structure of the music and the instrumental identifications meaningless, or at least not important.

I argue that we need to replace the idea of the recognition of instruments by the recognition of the basic properties I mentioned above -- spectrum, envelope, modulation, etc. -- as unique qualities in themselves. It does not matter whether you perceive that a viola played a particular tune, but rather that you perceive that tones were played which had certain spectral, envelope, and modulatory qualities that you could identify, both on these tones and on others elsewhere in the music. What matters is the qualities themselves, not that they add up to some preconceived notion of an instrument.

Electronic music, even if limited to the imitations of acoustic instruments that dominate commercial synthesizers, is capable of remarkable ingenuity and creativity with these elements. Acoustic instruments always produce a number of qualities together as an entire package. Each family of instruments has a basic envelope and is not capable of creating the qualities of other families. A woodwind instrument can't create a bowing sound or a pizzicato. All instruments produce noises in the form of scratchings, key clicks, or breathing, which we do not consider part of the musical qualities of their sounds. Since they are always produced, however, they are necessarily part of the instrumental sound, and much synthesizer programming is devoted to (wasted on, I would say) simulating these qualities. Much twentieth-century music contains explicit directions to performers to try to produce qualities that were formerly left entirely within the performer's discretion. While I am not advocating that composers shouldn't do this, these instructions nevertheless make much of the music extremely difficult to perform, and also seem to rob the performer of some of his options for expressiveness.

If we can accept the viewpoint that spectrum, envelope and modulation are all separate qualities, we can create music which presents a truly multi-dimensional structure, in which each of these qualities can be used to create associations with other tones in the same context. Imagine a sound as simple as a clarinet timbre with a pizzicato envelope. It is impossible to produce such a sound on an acoustic instrument, but trivial on a synthesizer. In a specific musical context, however, some of the "clarinet-like" sounds could have a sustaining envelope and others pizzicato, and the same context could have other instrumental sounds with the same modes of distinction. Our ears can hear some sounds connecting with others on the basis of the envelope, and the same sounds connecting with different ones on the basis of the spectrum.

8. Expressiveness

To return to my basic premise, all music already is multi-dimensional in the manner described above, but our traditional concentration on trying to recognize which instruments play which tones prevents us from appreciating the true depth of the music. What is the role of expressiveness in musical performance, and how does a performer achieve it?

Let me answer this question first by giving my own traditional view. The most convincing performances we hear of music are what rightly should be called interpretations, where the performer communicates not only his or her complete technical mastery of the performance difficulties but also knows this well enough to do it in a uniquely personal way. We can only rarely say what specific qualities a performer uses to emphasize certain notes in a melodic line or to deliver a phrase with a particular twist to it, although we can study them in recordings. One problem with this practice, however, is that many performers do not actually make their own interpretations, but instead learn them from their teachers, who in turn are often acting under the impression that they are conveying a particular tradition or school of thought. When some young performers try to be creative in their performances, their teachers criticize them. Sometimes these performances, such as with the late Glenn Gould's interpretations of Bach, become quirky and different. While we may find his performances unusual, I doubt that most experienced listeners would argue that Gould didn't know what he was doing.

It is a given of competent musical performance that all the notes, rhythms, and written-out aspects of the score be played correctly, and many performances, even those of some great musicians, fail to measure up to this standard. I am not quibbling about a missed note here or there, but about far more basic issues. Even when all the notes are played correctly, there are often mistakes of tempo, dynamics or phrasing. We overlook these things when we nevertheless consider than most of a performance was good, or that the performer at least made a serious effort to play the piece right; but if we are going to try to measure expressive qualities accurately through recordings, many attempts will not get beyond this stage. (As an aside, I want to encourage all young performers to make use of recordings to evaluate their performances objectively. This is a tool that was not available through most of the past, and even now many people are reluctant to use it.)

To continue my traditional response to the question of expressiveness: if we assume that the written aspects of the music must be played correctly, then it is the unwritten aspects on which the performer must rely for his or her means of individual expression. Vibrato, envelope qualities, dynamic differentiations, and timbral changes are all used regularly by good performers. Each performer learns about all the qualities that are available on his or her instrument for these purposes. As I mentioned before, the more that composers structure these qualities in their music, the more that they take away options for individual expressiveness from the performers.

What is the meaning of expressiveness in electronic music, where there is usually only one performance of a work? Before answering this question, I want to emphasize my own viewpoint that there is a performance aspect to electronic music, only the composer himself usually is solely responsible for it. Some composers, myself included, have made several different performances of the same work, usually releasing only the one we think is the best.

Early electronic music was criticized mostly because it was perceived as being mechanical and unexpressive. One problem was that early music synthesizers made it difficult to control all the aspects of sounds that we were used to hearing in live music. But a more important problem was that composers who had been accustomed to writing notes on the page suddenly found that they were responsible for handling all those aspects that they had been used to relying on performers to bring to the music: they had to think about envelope, timbre, and dynamics in ways that they actually understood but were not accustomed to specifying. The mechanical-sounding quality of their music was a result of simply playing the notes that they had written without the other qualities.

There are two important ways in which electronic music is expressive. First, there must be the recognition that all qualities of sound must be specified in the music. Second, there is the possibility that each of these dimensions can be structured independently, and that the kinds of structures that are possible are not limited to the qualities that come from live instruments. When both of these aspects are understood by the composers, truly interesting and expressive music is possible, and I don't believe that most listeners would deny that the music is expressive.

I would make one other point about expressiveness: interesting sounds are complex. Boring sounds are sounds that do not change. Similarly, interesting music is complex. Much music that may strike us as boring on a structural level can be quite interesting to listen to because its performances use interesting and complex sounds. I think that this generalization applies to so-called "new age" music, for example, but I would make yet another point about new age music. This is a style that has emerged at the same time as new synthesizers and effects devices have become available, and many people are intrigued by these new devices but are uncertain how they might be used in music. By providing a minimalist environment for the sounds, new age music allows us to concentrate on the sound qualities themselves without having to come to grips with a complicated musical structure as well. We can even get the idea that the musical structure can imitate or can even be the sound structure.

9. Valid Reasons for Studying Instrumental Sounds

While I am arguing against the notion that instrumental sounds are "real" whereas electronic sounds are "unreal", I would like to point out nevertheless that there are many valid reasons for studying instrumental sounds, apart from the mindless commercialism which has dominated the development of synthesizers. Since instrumental and vocal sounds have been used throughout the history of music, it is very important to understand and measure their qualities. For example, what exactly are the limitations of vibrato that are produced by different instruments? Can the vibrato of a violin be compared meaningfully to that of a flute? Can, or do, people hear these qualities accurately? Answering questions like these will help us to refine our own perception of all music, not just old works.

Another reason for studying instrumental sounds, which is explicitly the most important reason why this has been done, is to know how to reproduce these sounds. The goal of synthesizer manufacturers is not just to be able to do this, but to find the cheapest and simplest method and to patent it so that everybody else will have to pay them to use their method. In the seventeenth century, instrument manufacturers experimented with many different ideas and methods, and we know of many instruments which did not survive until the present time. The problem with acoustic experimentation of instrumental qualities is that it runs up against performers who spend their lives mastering the techniques of instruments which might now be seen to be highly flawed. In this respect, performers are a profoundly conservative force. Since their technique is demonstrated in a more impressive and more easily understood way by older music which does not challenge the listeners in new ways, many of them prefer to play this music and ignore most twentieth-century music. If any of you has had any experience with performing on electronic keyboards, one thing you would probably sense is that they respond much more easily than their acoustic counterparts. Performing on an electronic piano versus an acoustic piano is exactly like comparing touch typing to an old manual typewriter. I do not doubt that the time will come when people study performing technique on electronic instruments (which are much more than simply keyboards) much in the same way as they presently do on acoustic instruments. When that comes, all of the means of expression that are available on these instruments is likely to be exploited to the hilt.

Finally, another reason for studying instrumental sounds is to assist composers in deriving their own far-fetched variations of these sounds. Composers have always been interested in extending the range of qualities they can deal with; this is part of the meaning of originality. Just as a composition can develop variations on a melodic theme, so can it include variations on a sound. The important issue is maintaining enough of the original sound so that the variations can be perceived as being related to it.

So while there are many valid reasons for studying how musical instruments work, we must nevertheless not be misled into thinking that those sounds represent some kind of standard of excellence, or that sounds which are different are somehow less valid or less "real".

10. A Cynical View of the History of Synthesizers

The history of the development of synthesizers illustrates the biases I have been describing fully. The first complete synthesizers, different, incompatible and independently-developed systems invented by Robert A. Moog and Donald Buchla, were introduced in 1964. These were not commercial products, but were largely brought about by the encouragement of a few avant-garde composers who were able to raise the seed funding needed to carry their development to this level. Although they were used by many composers, some of whom, like Morton Subotnick, became quite well-known, it was not until the success of a single recording called Switched-On Bach that real interest in these machines grew.

The listeners to Switched-On Bach conjured up a fantasy of a single musician sitting in front of a vast machine capable of playing any and all instruments simultaneously. It meant no difference that the composer, or rather, performer, of that work gave detailed testimony to the painstaking and fanatical procedure that he employed to produce it, playing each melodic line of the works individually on a multi-track tape recorder and mixing the results carefully. The makers of the new synthesizers that began to appear at that time had in mind the singular goal of reproducing the sounds of acoustic instruments. You probably don't remember the dozens of imitations of Switched-On Bach that came out at that time. None of them has survived, and none of them was worthy of having survived, for they were not its equal in terms of performance.

The technology represented by the earliest electronic musical instruments is what is now called analog synthesis, so-called because its electronics were analog in the same manner of an analog computer, and this technology survives today in the instruments of several leading manufacturers. The concept behind analog synthesis is that you start with a complex wave, which is subjected to filtering to create the primary timbral characteristics. Both musical instruments and the human voice use comparable methods to create their sounds, so there was historical precedent for this method. Nevertheless, the engineers who were seriously interested in duplicating the qualities of acoustic instruments were disappointed with the results. The best imitations that could be obtained did not have the depth and realism of the instruments, and they believed that the fault lay in the technology rather than in their implementation of it.

The next advancement in synthesizer technology came out of computer music, specifically from the work of John Chowning and his colleagues at the Center for Computer Research in Music and Acoustics (CCRMA) at Stanford University. They had been able to produce quite realistic imitations of nearly all acoustic instruments through a process of frequency modulation now known as FM digital synthesis. The Yamaha Corporation of Japan, the largest manufacturer of musical instruments in the world, acquired and patented this method of synthesis in the late 1970s, refined it and incorporated it into their DX-7 Synthesizer (actually called an FM digital algorithm synthesizer), which was introduced to the U.S. market in 1983. This remains the most successful electronic musical instrument ever produced in history. Thousands of instruments were produced, which led Yamaha to pour increasing amounts of money into the development of newer instruments which have never realized the level of success of the DX-7.

The success of the DX-7 was really not due to the method of synthesis, but rather to the careful quantization that was allowed for structuring each of the different qualities of sound that it produced. The envelope generators were the most advanced of its time, and are still as versatile as nearly anything yet produced. A carefully-controlled low-frequency oscillator could be applied to amplitude, frequency, and other qualities with rates and amounts that could both be predetermined and changed by a real-time controller. The method of spectral production allowed the timbre to be changed in useful and varied ways, even if these did not match the same types of changes that filters produced. There was a well thought-out control for glissando and portamento, allowing the performer to control both the time lag and the interval of change both in steps and as a smooth glide. The second version of the DX-7 contained micro-tuning capabilities that still equal or exceed those of any other instruments on the market, and which have hardly been used.

Even more important, much of the research Yamaha poured into the DX-7 became incorporated into the MIDI Specification, which was adopted by the manufacturers of electronic musical instruments in the same year. Many of the characteristics of MIDI, including its bias toward keyboard instruments and its relegation of other real-time performance devices to the category of controllers, were first introduced with the DX-7.

While the DX-7 represented a significant advancement over the previous electronic instruments, and it did allow more realistic imitations of acoustic instruments, it was still incapable of duplicating all instruments, and commercial interests drove the market in still further directions. After using FM digital synthesis, licensed from Yamaha, in its first version of the Synclavier, the New England Digital Corporation, among others, finally got the idea that, if people wanted true reproductions of acoustic instruments, why not just record those sounds and allow them to be played back under the control of the keyboard just like a synthesizer? This is the origin of the third major method of electronic tone generation known as sampling. Sampling now has a history as long and varied as the other methods of synthesis, and in fact present technology tends to combine all of these methods into a complex process of sound production whereby a digitally sampled wave can be subjected to analog filtering and sometimes combined with FM tones to produce the final result. Sampling introduces problems and limitations of its own, which I do not want to get into here.

The purchasers of today's electronic musical instruments now can choose multi-voice multi-timbral instruments whose sounds cannot be distinguished from those of acoustic instruments. Many musicians fought against this possibility, by such methods as withholding their instruments from sampling and passing laws forbidding the use of such instruments in live performance situations such as Broadway theaters, but their efforts did little to slow the march of progress. Now that the brave new world is here, how are things different, and what has been lost?

11. Conclusions

Following the outline of history as I have presented it, we find that the original researchers' efforts to try to discover the properties of acoustic sounds, in order to enhance our understanding of music as it was then known and performed, have been overtaken by the relentless commercial drive to provide the sounds of a full orchestra at your fingertips. Whereas composers wanted to know the limitations of acoustic instruments so that they could explore extensions or variations of these qualities, the marketplace has assumed that these were simply the goals. One consequence is that you do not find electronic instruments used very extensively in live performance, or even in any music presented as "classical", other than in electronic music itself, but you do find them used ubiquitously in popular and commercial music, where they most often explicitly replace live performers.

Nowadays, performers have to justify their own work in terms of what they can bring to the music that is different from what some keyboard specialist can provide for less cost. Frankly, in much commercial music, it hardly matters. The designers of commercials are interested in selling soap, not in advancing the cause of art.

To my mind, what has been partly responsible for this series of developments has been the tendency, which was visible as long as people have been doing this kind of research, to categorize sounds in the "real" and "unreal" terms that I explained at the outset. If we hadn't experienced the past in the same way, with its known instrumental classifications and musical literature supported by them, perhaps we could have been able to maintain a more open mind.

Even though most of the presets of synthesizers simply provide rough imitations of acoustic instruments, most machines can be programmed over a much wider range of values, and there is no reason to limit ourselves to those categories. Even if we limit ourselves to instrumental qualities of envelope, timbre and modulation, we can achieve a much wider scope by intermixing these qualities and extending their ranges, as with my example of the clarinet tone with the pizzicato envelope. The proper role of electronic music is to extend the creative aspect to the design of the sounds as well as to the musical structures that composers have created throughout history.

While history may be a guide to the present, in the sense that those who don't know history are likely to repeat its mistakes, there is no reason why the future has to follow the same path as the past. Because music has used the sounds of acoustic instruments in the past does not mean that it must be limited to them in the future. But in order to appreciate the difference, you have to have an open mind; you may have to hear some of the same sounds in different ways to be able to understand the point. When you have an open mind, you will appreciate that those unusual, different, interesting, creative and challenging sounds of electronic music are as real as those of the instruments of the past.