Early Computer Music: 1955-1975

"I can see coming ahead a time when the musician who is creator can create directly in tone, not on paper"
--Leopold Stokowski, 1930

1950s

With the advent of magnetic wire and tape it became possible to create sound not based on the abstract musical instruments, but work with "concrete" sounds of nature. By the 1950s there were artists whose complete work was defined through these tapes. At around the same time in Cologne, Germany, sound waves were being generated mathematically through sine wave functions. Because the waves were being created electronically, they called the work "elektronische Musik." Karlheinz Stockhausen, already an established composer, was immediately drawn to the possibilities.

Computers of the 1950s were not capable of being programmed to sense and create arbitrary sounds. Mathematics were all that was needed, and the computer was capable of complex work. Lejaren Hiller had a simple idea: have the computer generate random numbers representing anything we like about music, such as pitches, rhythms, dynamics, etc, then program "rules" by which one choice would be allowed to follow another. Such computational work was applied to existing numerical systems of music, with the entire scale of a piano having a 1-88 value.

John Cage and Lejaren Hillier, 1968.

Around 1955 Max Mathews, and others at Bell Labs, reasoned that because computers were capable of improving telephone speech analysis, processing and synthesis, they could be used to improve the same set of actions with musical sound. The digital signal, which reduces noise in the channel and reduces error, was recognized as a revolutionary step. The "on/off" command eliminated channel noise completely and was thought to be applicable to synthesizing sound. In 1957 Mathews wrote the first music synthesizing program capable of, in theory, producing any sound that could come from a loudspeaker. His was the first published recording of "computer music," called Music From Mathematics. In the liner notes he described the computer as an "instrument," but one which was capable of being involved in the creative process:

"In the words of three composer's whose works are heard on this recording, man's music has always been acoustically limited by the instruments on which he plays. These are mechanisms that have physical restrictions. We have made sounds and music directly from numbers, surmounting conventional limitations of instruments. Thus, the musical universe is circumscribed only by man's perceptions and creativity."

Much of the early work was done in labs, and most of it failed to be intriguing in any sonic capacity. Computations were done on computer tape, representing the digital form. These computations were then sent to a digital to analog converter to be turned into voltages. This signal was then fed into a loudspeaker. With the stage set for producing sounds from a digitized form in place, all that was needed was to speed the process up and work on the variables for producing more interesting sounds.

1960-1975

Between 1965 and 1975 the attention of those interested in sound created by computers was focused on synthesizing sounds with musical characteristics. It appears that one has two choices: make sounds like those made by traditional instruments, or synthesize new ones which would be musically satisfying. Expectations would play a huge role in what the computer would be applied to. "Musical" had become a questioned term in the 60s, with a great emphasis being placed on "noise," or unintentional sound. A computer could easily be applied to create random sounds. One would be studying both what the computer was capable of, but also what the human ear and perception would be capable of preferring.

Jean-Claude Risset, a physicist and composer, based his work around recreating musical sounds, and thus would learn about timbre, color and tone. It was during this time that the perceptual, psychological effects were discovered of the sound they were trying to manipulate. Pretty soon it was discovered that the psychoacoustics of sound involved having knowledge of the relationship between the physical and psychological aspects of sound. Jean-Claude Risset's "brass tone hypothesis" was based on figuring out the essential number of harmonics involved in a trumpet. This analysis of the sound amplitude enabled both knowledge of brass instruments in general and also the ability to synthesize similar sounds.

A continuum was forged between understanding human perception and digital computation. Physical models were for synthesizing sound were theoretically possible, but required too much computing power.

Music V, GROOVE and FM Modulation

John Chowning

Music V, developed at Bell Labs, was the first computer music program to be distributed. It existed in several boxes, each containing two thousand punch cards. Music V contained virtual instruments which could be given typed instructions that mimic traditional written music. This synthesis program was extremely static, operating in a performance free environment, with very little room for interpretation. Because it mimicked the traditional instruments well, it lacked any nuance or interesting performance attributes. In short, it was an accomplishment that had little to do with what one would expect from music.

The performance problem in computer music was solvable in two general approaches. One is to use its computational power to write complex scores which only the precise ability of a computer could replicate. The second would be to allow the computer to play on traditional instruments. GROOVE (Generated Real-time Operations On Voltage-controlled Equipment) was the first real-time synthesizer, utilizing computer processes in order to play input from traditional instruments.

John Chowning, a grad student at Stanford was using Music V to study the effects of extreme vibrato when he accidentally inputted a vibrato rate way beyond what a human voice would be capable of creating. Instead of sounding like a vibrato, it went carried a completely different tone quality. Though frequency modulation (FM) was well known in radio transmission, none had thought to listen to FM on its own. In essence FM theory was applied to the new realm of musical psychoacoustics.

Computation was simple: one needed to modulate the frequency of one sine wave with another. Before his discovery no one had any idea that this simple and efficient synthesis process could provide straightforward control over the perceptually important characteristics of musical sound. Yamaha licensed the discovery through Stanford and was the leader in computer music instrumentation for many years.

John Chowning's Turenas (Clip 1 | Clip 2)

1975-1985: Digital Computer

From the 1960s onward, integrated circuits became increasingly affordable and powerful. With synthesis and the creation of sound being the focus of most true "computer musicians," the digital revolution became both market and research orietned. Yamaha created cheap synthesizers that were extremely popular worldwide. Computer music itself developed out of these models of synthesis and desire for a performance. FM synthesis and its psychoacoustic malleability defines much of what is done with computers--instead of sampling and manipulation (music conrete) or creation of sine waves without computation (electronische Musik). Though computers have been used to do many things for music, "computer music" is privileged as being the first to use the processes for pure sound manipulation--spatially and temporally.