Brain Custard

Inspired by a 1990 IEEE conference paper by S. Endo and Y. Kinouchi, Neural network with interacting oscillators to generate low frequency rhythm: multiple interacting chaotic audio-rate signals, in the hope of finding unexpected patterns, beat frequencies and rhythms in the outputs. In practice it is mostly noise, and lots of it, modulated by the inputs, with a great variety of tones from the knobs; fed to VCFs and phasers it cleans up into all sorts of harmonically rich waveforms. The chaotic oscillators are those of the Jerkoff chaos module, modified to run at audio rates (built otherwise it is a source of ridiculously complex CVs). The inputs take anything: audio, gates, triggers or CV. X, Y and Z are mixes of the same stages of different oscillators, minor sum neurons; CUSTARD is everything jammed together, the main output neuron; BLEND mixes X, Y and Z through the BLEND knobs, of ±1 gain: inverted at 0, off at the middle.

Key features

• IN HI, IN MID and IN LOW in; X, Y, Z, BLEND and CUSTARD out
• Two knobs under each of HIGH, MID and LOW, and three BLEND knobs
• Three LEDs
• 12 HP; Nonlinearcircuits publishes no current draw, depth or assembled price (it sells the DIY PCB and panel), and ModularGrid lists none