The old keyboard performance instruments began to grow into what we now term “synthesizers” when they abandoned reliance on the keyboard and developed sophisticated programmability. The first such device was built in 1945 by John Hanert for the Hammond Organ Company. Although the sound generator, which occupied a room filled with vacuum-tube oscillators, wave-shapers, and a white noise source, could be controlled by a keyboard, it was also coupled to a 60-foot long card scanner that determined the detailed continuous parameters of a sequenced composition. The scanner read graphite marks made on the cards by pencil that encoded the dynamic pitches, durations, timbre and volume profiles of the generated audio. Tempo was adjusted by varying the speed of the read head as it ran across the cards. Other composition-playing devices were built around that time, such as the Free Music Machines by Australian composer Percy Grainger and his colleague Burnett Cross, inspired by Grainger’s dream of scoring and realizing music that would not be humanly playable.
In 1951, Dr. Earle Kent, while working at C.G. Conn Ltd., an organ manufacturer in Indiana, developed a device that he termed the Electronic Music Box; a set of special heterodyned oscillators with pitch and many timbral and dynamic parameters all sequenced on a roll of perforated paper tape. This device was inspirational to Dr. Harry Olson of RCA, who used Kent’s concepts in the design of the famous RCA Synthesizer, introduced in 1955 and also employing paper-tape programming of all audio parameters. Because it was monophonic, an elaborate overdubbing method was developed that synchronized turntables and cutters to the paper tape (the audio storage medium was a 16″ lacquered record, replaced by magnetic tape recorders several years later). RCA upgraded and expanded the synthesizer, but because of several factors (possibly including brewing controversy with the American Federation of Musicians), RCA sold its Mark II to Columbia and Princeton for $10,000 in 1959, forming one of the world’s most productive music studios that inspired a generation of new-music luminaries. The RCA Mark II was in use until 1976, when vandals destroyed it.
Although other large synthesizer installations were built (such as the 1959 Siemens synthesizer in Munich), perhaps the most unusual step along this path was taken by New York bandmaster, jingle composer, electronic instrument designer and sequencer pioneer Raymond Scott, who finished his “Electronum” in 1970. This was a “music structure generator”, where an artist could enter a polyphonic piece and then parametrically adjust many of its parameters in real-time using various controls. Scott conceived it as a musical idea generator; after setting it in operation, one could rapidly explore a vast array of multidimensional musical and sonic variation, all of this without a computer. Another well-known device of this sort was the SalMar Construction, a massive real-time interactive composition machine composed of digital and analog circuitry that was finished in 1972 by Salvatore Martirano and colleagues at the University of Illinois. Also around 1970, a very unusual, low-cost commercial device of this sort was released called “The Muse”. Designed by MIT’s Marvin Minsky and Ed Fredkin, it housed a digital pseudorandom sequencer that produced audio output. Its many adjustments allowed realtime exploration of different musical variations.
Although elements of its design harkened back to previous work, such as Harald Bode’s instruments at the Cologne Studio in Germany, the modular analog synthesizer had its genesis in the USA. Donald Buchla’s synthesizers grew out of ideas that he developed while working at the San Francisco Tape Center in the early ’60’s. The Center then moved to Mills College, where Buchla’s first modular synthesizer is still located. Buchla’s modules dating back to 1963 are acknowledged to be the first use of “voltage control”, the protocol by which the different elements of a modular synthesizer communicate. Buchla, however, did not consider the diatonic keyboard to be the most appropriate controller for his system, therefore he developed other interfaces, like capacitive touch pads, for performer input. Likewise, Buchla’s early models used linear voltage control, which precluded harmonic multi-oscillator tracking. Robert Moog produced his first voltage-control modules in New York during 1964, but his systems generally included a keyboard and his exponential oscillators could be voiced in a stable, tracking chord. As a result, Moog’s modulars were embraced by popular music, beginning with Wendy Carlos’ classic 1968 Switched On Bach recording and forging into the rock mainstream when Keith Emerson adopted it in 1970 for touring with The Nice and ELP. The Moog became synonymous with synthesizer music at the time, spawning a new genre. Moog ensembles, such as David Borden’s Mother Mallard Portable Masterpiece Company, performed in the USA, while Germany witnessed the advent of Tangerine Dream and the Berlin sound. At the same time, the Buchla modular synthesizer (manufactured by CBS after 1969) was garnering respect and installation contracts in the academic community, receiving some notoriety through Morton Subotnick’s early Nonesuch releases. Although Buchla achieved much more commercial success with his Lightning and Thunder interfaces, he remained a synthesizer pioneer, designing and building a host of fascinating digitally-controlled analog synthesizers and all-digital devices throughout the following decades.
Following Moog’s success, other US modular manufacturers quickly appeared. Massachusetts-based ARP released the large 2500 system in 1970, E-Mu in California released their modular systems in 1973, and Serge Tcherepnin’s unique modular systems started in California during 1975, when Alan Pierce’s Polyfusion systems were also made in New York. These large modular conglomerates were both expensive and difficult for musicians to tour with, hence smaller, more integrated modulars started to appear, such as the ARP 2600 in 1971 and the ElectroComp 101 by Conneticut-based EML in 1972. America has a rich do-it-yourself tradition, and modular synthesizers were ripe for packaging as kits. The best known were John Simonton’s PAiA modules (released in 1972) and the Aries kits (started in 1975) designed by ARP veteran Denis Colin and colleagues, based in Salem, MA.
The first commercial tabletop synthesizer was the British VCS3 by EMS in London, released in 1969. Its pegboard signal routing system provided great configurability for such a small device. The MiniMoog, released in 1970, was very musician-friendly, since it was fairly small and accommodated few patchcords. Since the signal routing was hardwired, it was much easier to set up and play. Although this loss of flexibility severely limited its sonic possibilities, the MiniMoog was a major market success, and heralded the eventual end of the large modular systems. Other synthesizer manufacturers rapidly introduced similar low-end, preset-based products, such as the 1972 Arp Odyssey and the 1973 EML ElectroComp 500.
True polyphony hit analog synthesizers when Tom Oberheim licensed E-Mu’s digital scanning keyboard for incorporation into his famous 2,4, and 8-Voice series of synthesizers in 1974. These devices had no embedded computer however; the keyboard assignment and scanning logic was all done in discrete CMOS, designed for Oberheim by J.L. Cooper. Although E-Mu was working on its extremely high-end, computer-driven “Audity” hybrid synthesizer, Sequential Circuits was the first to drive a polyphonic analog synthesizer with a microprocessor system in the revolutionary Prophet 5, released in 1978. Even though E-Mu eventually dropped the development of the Audity to concentrate on the Emulator sampler, microprocessor-controlled analog synthesizers dominated the market well into the 80’s, where they were eventually overtaken by Yamaha’s all-digital DX7.
Because of his early experiments in computer-generated audio that date to the 1950’s, Max Matthews, who spent most of his career at Bell Laboratories, is widely acknowledged as being the father of computer music. His Music I package, finished in 1957, was the first sound-generating computer program. Music II followed in 1958, introducing the concept of programmable digital wavetables. Music III, in 1960, was much more complex, encompassing features such as scoring, timbral variation, modularity, and orchestration. Several composers and researchers, such as Jean-Claude Risset, James Tenney, and F. Richard Moore worked with Matthews and his Bell Labs systems, pushing the possibilities and technology still further. Matthews finished with Music V in 1968; written in FORTRAN, it could be distributed to offsite researchers who could compile it on different machines. Barry Vercoe, the founder of the MIT Experimental Studio and presently at the MIT Media Lab, was one of the researchers who extended Music V, eventually evolving it into the modern CSOUND environment, a powerful audio synthesis language that’s been written into the upcoming MPEG4 audio compression standard.
As the computers of that time were far too slow to make any kind of complicated, multi-timbral sound in real time, these programs ran in a batch environment, where audio data would be slowly spooled onto mass storage, then dumped to a set of analog-digital converters after the piece was finished. All of the digital academic studios of that time ran this way well into the 1980’s, since the complicated sonic textures that they were experimenting with couldn’t be efficiently rendered in realtime. The only way around this was to experiment with computer-controlled analog synthesis, as was being pursued starting in 1968 with a PDP-8 and what was to become the Synthi 100 at Peter Zinovieff’s EMS studios in London. Max Matthews, being very interested in the implications of computers in performance, built the first serious hybrid between computer and analog synthesizer, which he termed GROOVE (Generated Real-time Output Operations on Voltage-controlled Equipment) at Bell Labs in 1970. He developed his Conductor program on it, then went on to explore various conducting interfaces such as the Daton and Radio Baton. GROOVE was used by several composers, such as Emmanuel Ghent and Laurie Spiegel.
In the late 70’s, real-time digital oscillators were developed that could be controlled by microprocessors, ushering in the first all-digital synthesizers. They included the Systems Concepts Digital Synthesizer (better known as the SCDS or Samson Box), finished in 1977 at Stanford’s CCRMA, the Alles Synthesizer built by Hal Alles in 1977 at Bell Labs (which eventually became the commercial GDS synthesizer in 1979, then the Synergy in 1982), and the Dartmouth Digital Synthesizer, which became New England Digital’s Synclavier in 1977. As the initial hardware was limited by today’s standards, composers implemented algorithms on them such as FM synthesis, a compact and efficient means of controlling sonic spectra perfected by John Chowning of Stanford’s CCRMA in 1971. Yamaha licensed Chowning’s patent and developed low-cost hardware capable of running FM operators in real-time, producing a major commercial success in 1983 with the DX-7, the first affordable all-digital synthesizer.
Although other interesting digital synthesizers were developed during the early 1980’s, such as the Alpha Syntauri (an inexpensive wave engine hosted by an Apple II computer) and Casio’s popular CZ-101 (featuring dynamic phrase-distortion synthesis), the commercial market rapidly became dominated by wavetable synthesizers and samplers.
The emergence of the MIDI standard in 1983 returned modular functionality to the synthesizer world, and mandated that a microprocessor be embedded in every instrument. Huge new markets were opened up for software packages such as MIDI sequencers and librarians, and the personal computer rapidly became a central part of any electronic musician’s rig. MIDI also had implications for keyboard controllers, as they no longer needed to be part of the synthesizer that they were commanding. Although the MIDI specification can be severely limiting for high-bandwidth controllers, the benefits of commonality revolutionized the music industry.
One of the main current frontiers in music synthesis is physical modeling, an attempt to avoid the static nature of canned samples by generating complex, dynamic audio from computational models. As one might expect, when properly controlled, physical models of acoustic instruments are much richer and closer to their original sound. “Physical models” of artificial systems have no counterpart in nature, but can still exhibit the rich, dynamic sound qualities of real instruments (or in some cases, the old analog synthesizers, which are again prized in some quarters because of their gritty, dynamic texture). Although the first physically-modeled instruments, Yamaha’s VL series, were derived from the efficient Waveguide parameterizations developed by Julius Smith and colleagues at CCRMA, there are many other ways to model complex systems, leaving much room for exciting developments.
General-purpose personal computers are rapidly becoming powerful enough to subsume much of musical synthesis. Essentially all current computers arrive equipped with quality audio output capability, and capable software synthesizers are readily available that run on a PC or Mac without requiring additional hardware. Over time, the software synthesis capabilities of desktop computers will expand, and dedicated hardware synthesizers will be pushed further into niche applications.
From American Innovations in Electronic Musical Instruments
by Joseph A. Paradiso
© 1999 NewMusicBox