Towards the Creation of a new Musical Machine

by Barry Drogin

Back to guide - Next item in guide - Back to index - Next item in index

THE FIRST FIFTY YEARS - Electrical music synthesis before the age of transistors and computers.*

THE RCA ... AND BEYOND - Avant-garde composers' need for a new machine, and developments of the sixties.

A DECADE OF TRENDS - Popular machines of today, and classical music's return to live performance.

IS MUSIC ANALOG OR DIGITAL? - The two major design trends, and the need for a third.

A NEW MACHINE FOR NEW MUSIC - A proposal for a non-performance oriented, composers' machine.

* Most of the information contained in the section "The First Fifty Years" was culled from a column by Tom Rhea entitled "Electronic Perspectives" and appearing in Contemporary Keyboard magazine. Further factual information was obtained from other articles and advertisements in that magazine, and from private research and readings.

THE FIRST FIFTY YEARSoscillator circuit

The early history of electrical music synthesis design is a quirky one dominated by inquisitive composers, wide-eyed theorists, and amateur inventors. The desire to apply new technologies to music is an age-old one, for what are drums, flutes, violins and pianos but machines, resulting from developed crafts and skills? With electricity, however, the process of setting air into motion to produce sound is a coincidental outcome divorced from the device's operation. Whereas the sounding of a conventional instrument cannot be separated from the process of playing, electrical instruments are all, in the final analysis, devices that manipulate speakers. Disconnect the speakers, and the device will continue to function but cease to produce music.

Serious electrical music generation can be said to have started with the invention of the triode tube (called the "Audion") by Lee de Forest in 1906. By combining the triode with an LC tank circuit in a regenerative feedback configuration, de Forest created (and patented, as shown above) the first tube oscillator. In 1915 de Forest dreamed of an "Audion Piano" with an oscillator devoted to each key of a keyboard, but the tube itself needed further refinement before such a grand application could become possible.

In the '20's, interest in tube oscillators flourished, and the two most famous electrical instruments, the "Theremin" and the "Ondes Martenot," were invented. Both instruments produced audio pitches by heterodyning two high frequency oscillators to produce a low frequency difference tone. The Theremin, invented by Leo Theremin in 1927, must have seemed quite magical at the time, for the oscillators were controlled by producing disturbances in an electromagnetic field --- by waving your hands in the air! The Martenot, invented by Maurice Martenot in 1928, was easier to control --- a wire on pulleys, affecting a variable capacitor, was slid back and forth over a dummy keyboard.

Also in 1928, Friedrich Trautwein introduced the "Trautonium," a sawtooth generator using neon tubes and incorporating tuned filters to simulate musical formants, and Bruno Hellberger and Peter Lertes built the "Hellertion," in which a metal band was slid over a long resistor.

All of the devices considered so far exploited the portamento quality of a sliding oscillator. Capacitors, resistors and inductors are electrical components which, at the time, were varied mechanically. For a number of years, developments in electrical music synthesis were due more to mechanical inventions that to radical improvements in the de Forest design.

In 1929, Edouard Coupleux and Joseph Girelet decided to emulate the Player Piano and use paper tape to control an oscillator. Historically, this marks the first time that a symbolic representation of musical parameters (pitch, duration, volume) could be translated directly into music by an electrical machine without the participation of a human being.

The huge field of electronic music also encompasses the electrical distortion and amplification of conventional instruments, so it is appropriate to note that the first electrostatic pickup was invented in 1930 by Benjamin Miessner and used to create an "Electronic Piano."

In the early '30's, two similar but unique electromechanical machines were built. The "Rhythmicon," built by Leo Theremin at the request of composer/theorist Henry Cowell in 1932, used spinning perforated disks, a light source, and a photoelectric cell to produce simultaneous rhythms. In 1934, Ivan Eremeeff, encouraged by conductor Leopold Stokowski, invented the "Syntronic Organ," which also used disks, light, and a photoelectric cell, but operating in the audio range to produce pitches, not rhythms. Of course, simultaneous pitches (polyphony) were possible without devoting a separate oscillator to each pitch.

Throughout the '30's, electronic organs appeared and disappeared. Many used the one tube/one note concept suggested by de Forest and made possible by the development of stabler tubes. It became quickly apparent, however, that this approach was power-consuming, space-consuming, expensive, and maddening. The first "assignment" keyboard, devoting four oscillators to the highest four keys pressed, was developed by Harald Bode in 1937 for the "Warbo Formant Organ." Mr. Bode also introduced the first "touch sensitive"keyboard in 1938 (the "Melodium").

Finally, in 1939, someone at Hammond Organ Company realized that, by frequency division of high notes, less oscillators would have to be built. The "Novachord" and "Solovox" used this concept, but since three tubes were needed for each octave division, more tubes were needed than before! Nevertheless, throughout the '40's and '50's, the popularization of electronic organs by companies like Hammond became evident, and much stress was placed on producing interesting tone colors by the use of high and low pass filters, vibrato, and resonating formant filters.

This is not to say that experimental machines didn't crop up now and then. Composer Percy Grainger got Burnett Cross to build him some "Free Music Machines," one using moving "kangaroo pouches" to control potentiometers (circa 1944), another using painted cellophane and photoelectric cells. John Hanert developed for Hammond an "Electrical Orchestra" in 1945, using square-foot sized cards marked with graphite, and controlling frequency, intensity, duration, decay, timbre, and vibrato. In 1951, Dr. Earle Kent's "Electronic Music Box," using a perforated tape reader, was a possible influence on one of the most important electronic music machines ever built: the RCA Electronic Music Synthesizer.

THE RCA ... AND BEYONDRCA and friends

The Theremin and the Martenot notwithstanding, electronic music did not really break into the classical music world until after 1960, when the Radio Corporation of America, fearing financial disasters for its new monolith, the RCA Electronic Music Synthesizer, donated their Mark II model to the Columbia-Princeton Tape Center and gave up on trying to market the thing. Introduced five years earlier, the RCA Synthesizer was damned by the musician's unions for purporting to eliminate the performer from the picture. RCA's engineers, in fact, made the mistake of bragging about this.

The fortunate coincidence is that performers were also refusing to play the music of avant-garde composers of the time like Vladimir Ussachevsky, Milton Babbitt, Charles Wuorinen, and Otto Luening, who found a friend in the RCA Synthesizer (the cheery bunch are pictured above). Yes, it took a long time to learn how to program this early musical computer that occupied more than a room, but it didn't object to playing new music. Combined with the development of modern tape techniques, the RCA Synthesizer was used to create a large important body of new music, and in 1970, Charles Wuorinen was awarded the Pulitzer Prize for "Time's Encomium," an RCA Synthesizer masterpiece.

While RCA let composers in the door with the donation of its computer-like synthesizer, starting in the early '60's, three men have served as the modern innovators in synthesizer design, making the electronic music world a place of infinite capabilities for both composers and performers. These three are Donald Buchla, Robert Moog, and Dr. Max Matthews of Bell Labs: the dreamer, the populist, and the scientist. In terms of engineering design, these three are the inventors of, respectively, modular synthesizers, keyboard synthesizers, and digital synthesizers. Of course, all three took advantage of the new world of transistor technology to produce small, relatively inexpensive devices.

In the history of synthesizers, Donald Buchla is frequently ignored, perhaps due to his eccentricities and theories. Mr. Buchla has always stressed freedom in synthesizer design, with the composer as a subjective controlling force. Buchla synthesizers use potentiometer-controlled, voltage-input and output modules that can be combined in any order to produce complex sounds. Oscillators tune from the rhythm range past the audio range, and so are multiple-purpose. Modules are linked by a "patch bay" of banana plug-tipped cords, producing a web covering the surface of the synthesizer when in use. The user is encouraged to randomly experiment with the many possible combinations of the machine.

The Buchla synthesizer is a true "analog" synthesizer. There is nothing "discrete" about its components or its use. The composer twists knobs, or uses more sophisticated controls like body capacitance keypads or joy sticks, to produce sounds that feel right.

The Moog synthesizer, on the other hand, is a performance-oriented machine, with a keyboard the standard interface. The modules are voltage-controlled rather than potentiometer controlled and their uses are clear: a voltage-controlled oscillator (VCO) is used for pitch; a voltage-controlled amplifier (VCA) is used for volume; the envelope generator controls the VCA, the filters modify the outputs. The Moog synthesizer is easier to use and quicker to control. Of course, one cannot reinvent the entire machine by the use of patch cords, though some freedom is possible (a patch bay is used in early models).

The accessibility of the Moog synthesizer, not only in use but in sound, became apparent when Walter Carlos and Rachel Elkind produced the enormously popular "Switched-On Bach" album with their Moog. Due to this album and those that followed, the words "Moog" and "synthesizer" were as one word in the minds of the public for many years. Though Robert Moog has left the company that bears his name, he is still respected as an authority on commercial synthesizers.

While Donald Buchla and Robert Moog were busy dealing with the problems of manufacturing and customizing their analog/mechanical synthesizers, a group of research engineers at the Acoustic and Behavioral Research Center at Bell Labs, under the direction of Dr. Max Matthews, was busy coming up with a scheme to produce conversation-like electronic signals to test phone equipment. Thus was audio digital-to-analog conversion born, and Dr. Matthews was quick to realize that adding a fun aspect to the project wouldn't hurt. Experimental digital synthesizers were built, composers were snuck onto the premises at odd hours and encouraged to play, and Bell Labs became a leader in the field of music synthesis without ever having marketed a single device.

In term of capabilities, the Bell Labs synthesizers have remained, for two decades, at the forefront of electronic music design. They also, however, would cost an extraordinary amount, given that you could buy one. Up until now, the people at Bell Labs have been the leaders in showing what can be done if the sky is the limit. Each brand new prototype, though, is a single non-reproducible entity; it has been up to other companies to steal the design concepts involved, decide what is necessary and marketable, and come up with their own versions.


Whereas the previous two sections have presented an historical overview of electronic music synthesis up to the early seventies, it is difficult to speak confidently of the developments of the past decade; only with the advantage of hindsight can the significant devices be separated from those that will prove to serve no purpose. Also, electronic devices (like computers) are being used to investigate semi-musical phenomena like spatial placement (Stanford) and digital waveform generation (M.I.T.), and it is too early to write of their results or their bearing upon music, electronic or otherwise. This is not to say that distinguishable trends cannot be spoken of; what is at question is whether some trends will prove to be dead ends in the world of music.

For example, a mass of analog synthesizer components directed at the pop market, some by original innovators like Moog and Arp, others by newer but major entrants like Emu (one of their products is pictured), Oberheim, Korg and Roland, may lose their usefulness (and presence) if keyboard rock dies out. These companies are primarily interested in permutations of the keyboard-based systems started by Robert Moog, though behind the switches and knobs, some are turning to digital components to replace non-discrete analog devices, if only because anything labeled digital nowadays is "hot."

Two companies that started digital, New England Digital and its Synclavier, and Fairlight Instruments with its Computer Musical Instrument (CMI), have a strong secure lead over all other digital designs, and are the state-of-the-art in digital synthesizer design. Including a CRT and devoted processor with the standard keyboard, the Synclavier and CMI, which both cost well over $20,000, make available features like harmonics summing, real-sound sampling and manipulation, multi-voice sequencing and editing, etc.

In the classical music world, where the instant gratification of touch-sensitive keyboards controlling pre-programmed sound patches appears to be of less importance, and where the cost of glorified oscillators is an insurmountable obstacle, reel-to-reel tape techniques and the modular simplicity of Donald Buchla's inventions are dominant, when present at all. Surprisingly, major musical streams that began on electronic components are now written for and performed on non-electronic classical instrument. Serialism, shunned at the time of the RCA Synthesizer, is now old-hat for pianists, chamber ensembles, and contemporary-minded orchestras. Minimalism, its early theoretical stages based on such electronic devices as the very stable oscillator (days?!) And electronic techniques like tape phasing, is now accepted as performable by humans --- in fact, the human performance is an integral part of the music, if only to prove that the repetition is not mechanical but subtly (if not carefully) changing. Even concepts of spatial placement, first investigated musically by composers like Karlheinz Stockhausen (first, that is, to a great extent --- one can find examples dating back to 1550), and today a major concern of IRCAM in Paris as well as other academic research centers, is not permanently wedded to loudspeaker manipulation; musicians are being placed offstage, in the audience, in unusual stage configurations, by composers like Luciano Berio and Elliott Carter. There is, at present, a small backlash against electronics, though not against new techniques or technology as a whole. The revolutionary fervor of twenty years ago has been replaced with a concern for theory and expression, and since understanding and interpretation are tangibles for human performers and not machines, the presence of an intrusive and limiting electronic device runs counter to the intentions of all involved in the generation of music. Is there, then, a future for electronics in the music world?

IS MUSIC ANALOG OR DIGITAL?Rite of Spring opening

By attempting to imitate and replace other instruments, many researchers into electronic music synthesis have discovered just how complex music really is. Nonharmonic tones, irrational amplitude envelopes, imprecise tunings and rhythms are the rule, not the exception. And yet, there is much order in music, a pervasive rationality that led the synthesists to believe that the machines were possible in the first place.

These contradictory forces are epitomized by the two major design concepts of synthesizer design: the analog belief that a musician's "feel" in the manipulation and construction of sound events is paramount, and the digital belief that a sufficiently complex definition of the intricacies of music can result in the creation of any sound event. Of course, neither is entirely correct since music combines both discrete and continuous elements, both rational and irrational concepts.

Perhaps it is not surprising that Robert Moog is currently investigating the design of unusual multi-directional control devices, and Donald Buchla is building a computer-controlled synthesizer in which the software defines the machine. Both are, in effect, stepping into the world of the other in order to find solutions to their self-created problems.

What is missing from all discussions of music synthesis is the realization that while musical elements are not entirely discrete, each shares specific definable relationships to other musical elements in a piece. Thus, a duration is best defined by its relationship to other durations in a piece; a pitch is important due to its presence amongst other pitches; intensity is relative, and timbre provides continuity, emphasis, and interest. Could you imagine a score that defined sound events as individualized quanta of musical information, or one that showed nothing but the general flow of continuous musical parameters? Music is a combination of short-term and long-term ideas, of exact and imprecise components.

It should be clear, then, that a hybrid system, an analog/digital device, is necessary for the successful creation of quality music. At present, any synthesizer that uses D/A converters for anything other than direct sound generation is labeled a hybrid device. This is akin to having an analog synthesizer whose knobs are turned in jumpy, discrete ways. But what of an integration of analog and digital hardware, one in which it is hard to tell just where the digital ends and the analog begins?


Behind all of the discussions considered so far is the general question of why a new machine is needed. The difficulty of the mathematical harmonic-randomness of serialism is a challenge that performers are now willing to meet. The agility and endurance needed for the performance of a minimalist piece are now facts of life. What can machines do that humans cannot?

As pertains to musical performance, there is probably no need for a machine that operates without the participation of a human. Any parameter of music that is worth performing is worth conquering by human performers. This is not to say that electronic sound generation itself is doomed to disappear --- the vibraphone, the Martenot, electrostatic pickups (the works of George Crumb) even the Farfisa organ and sound mixers (the works of Philip Glass) are here to stay. But of what use is a machine that, by definition, produces sound on its own?

As a performance instrument, such a device is useless. But the contemporary music world has not exhausted music's possibilities, and an inexpensive device devoted to exploring new areas of musical theory would, at least until the new ideas are absorbed by performers, serve an inquisitive composer well.

In the area of rhythm, for example, are two battling theories. One, "analog notation," places the occurrence of sound events outside of any relation to a constant pulse --- instead, the performer plays a note according to its horizontal placement on a time graph. The other, "fractional beats and meters" (pictured), stresses rational proportional relationships between duration, as in the relationship between a quarter note and a triplet, and so on. Whereas the first theory produces music that is easily performed, music that takes full advantage of the second theory can quickly get quite complex, and a machine could help a composer and performer come together on the proper meaning of particular rhythms.

In the area of pitch, similar battles rage. Percy Grainger's graphical "Free Music" deprives tones of all but coincidental relationships to other tones. The invention of new "well-tempered scales," on the other hand, stresses the exact tuning of fundamentals in their relationship to other fundamentals. Again, a machine which could produce these well-tempered tuning would be turning theory into sound immediately, and would encourage experimentation rather than limit it.

It should be clear that a machine that operated in fractional meters and well-tempered scales would be quite useful, and that such a machine should be an analog/digital hybrid.

Back to guide - Next item in guide - Back to index - Next item in index

A Musical Contrarian 1999-2007

Last Updated: August 4, 2007