Tristram Cary: Nonet: Computer Music In Four Tracks (1979)

Abstract

"Nonet 14 was composed early in 1979 at the Center for Computer Research in Music and Acoustics at Stanford University, using the large system at that time shared by the Artificial Intelligence Laboratory and the musicians. The sound output came from the 256-voice synthesiser, built by Peter Samson of Systems Concepts, San Francisco, and known as the Samson Box. Because of the large resources available, the entire score runs from the computer in real time, needing no further processing. Although I was aiming at an expressive, almost romantic sound, in the final result the music is tightly structured around the numbers 6, 9 and 18. There are nine 'instruments', the piece is exactly nine minutes long, and it is based on an octave divided into eighteen equal steps (giving 18-term series regarded as three hexachords ). The time structure is also controlled by the same 18-term series. The nine voices each have nine entries at nine different speeds, and all voices run for most of the time, producing a continuously changing texture. The score for Nonet is in two parts. One contains text and graphics, giving the material, manipulation details and all information needed for the piece program. The other is the piece program itself, the input to the computer, which is written in SCORE, a Fortran based music language by Leland Smith of Stanford. Figure 18 (on p.24) shows the graphic score from which pitch and time manipulations are derived. Two invertible 18-term rows, each divided into three 6-term motives, apply to both an 18-step pitch series (250 to 481 hertz in the middle octave), and an 18-step exponential time series (from 0.1 to 5 seconds). Multipliers are applied to these series to give nine time and pitch ranges (each of the eighty-one voice entries is unique in both respects). Figure 19 shows the pitch multiplier score (there is also a time multiplier score and a levels score). All these data were brought together into a final hand drawn input score, then typed into the program which runs the piece, a typical page of which I give in Figure 20. Parameters (P)2 and 3 are time and pitch data respectively, the@ symbol telling the computer to look for motivic data given at the head of the program, for example /@-W 3.28/ means: multiply the frequencies in motive W by 3.28 and retrograde the motive. The idea is to exploit the rich harmonic possibilities offered by an 18-step octave (further enriched by nonoctave multipliers), and the ever changing counterpoint of the nine voices (with time multipliers individual event durations vary from .04 to 12.6 seconds). Levels were originally serialised, but because subjective loudness is so dependent on timbre it is virtually impossible to predict the internal balance of nine voices, and in the end levels were fine tuned by listening and editing the program. In spite of the very controlled structural features None t is, I hope, anything but rigid in sound." -- Tristram Cary Show more