Four Voice Canon #8 ("Nerve Canon")

Ray Guillette and Larry Polansky


Brief Notes

FVC#8 is based on Nick Didkovsky’s "nerve events," from his Dr. Nerve CD Beta 14 ok. The nerve events are over 40 short sounds which are meant to be used by other musicians as "samples." FVC#8 was composed using only the 12-tone chromatic scale of events (8-19). Ray and I each made one tranformation of each pitch, using programs like TurboSynth, Tom Erbe’s Soundhack (especially the time stretching algorithms), and Eric Smith’s Mutator (on the NeXT computer). These 36 sounds were then distributed on a virtual MIDI keyboard, with the middle octave consisting of the unchanged events. Next, I wrote a complex HMSL program to compose the piece which implemented an enhanced version of my general Four Voice Canon permutation/mensuration algorithm (see the notes for my Artifact CD The Theory of Impossible Melody ), which generalized the algorithms I used in the other FVC’s (#3-7).

This new version of the software extends the permutation and canon process into several new areas. One significant new feature is that the permutation process gradually "contracts" in all parameters. For example, pitches are moved to the middle of a virtual keyboard from the extremes as each voice progesses. Each of the six (!) voices (even though I still call it a "four voice canon") uses only 2 of the 12 pitches (a tritone apart), so the piece gradually moves to the middle register and fills in the chromatic gamut (getting louder, denser, and higher in pitch, using pitch bend, as it does so).

In addition, the HMSL four voice canon software controls a spatialization technique of Ray’s design, using MIDI control of two commercial processors, precise EQ and mixing, and a multiplexing of the MIDI "keyboard" out to 16 different audio channels, placing each sound in one of 16 locations (also part of the FVC algorithm).

The "permuation list" used in this canon is, like some of the other canons, a permuation of five elements. In generally, all of the musical parameters under control in this work have four possible values and one "wild card" value, selectable at any time by the software. This is especially evident in the spatial location of the piece, which divides up the stereo space into a 4 x 4 grid (right-to-left and front-to-back). Each voice gradually moves towards the "front" of that space as it progresses (so the higher voices move up more quickly), but at any time, the fifth "wild card" value can place an event anywhere in that grid, combining a steady location progression with some unpredictable occurrences. A similar kind of thing happens in the pitch, duration, and timbral processes.