Portfolio
In Suspended Harmony
A seven-minute composition for 20 Dynamic Stochastic Synthesis voices performing the overtone series of a single Bb fundamental. Xenakis's DSS algorithm runs at sample rate in a Web Audio API AudioWorklet: breakpoints evolve via Gaussian and Cauchy random walk with reflecting barriers, and the stochastic parameters themselves are subject to second-order random walk. The composition reveals partials 1 through 16 across nine harmonic regions with continuously interpolated trajectories, building from a bare octave through the fifth, major third, natural seventh, and ninth to a full spectral climax before dissolving back to the fundamental, altered by a ghost of the 15th partial. Four voice classes (4, 8, 14, and 6 breakpoints) are shaped by 12-second convolution reverb, shimmer delay, and tanh waveshaping saturation. No samples. No oscillator objects. Every sound is generated from arithmetic on breakpoint arrays.
Lemniscate
A seven-minute realization of Xenakis's Dynamic Stochastic Synthesis with no key, no scale, and no harmonic plan. Twenty DSS voices are defined only by breakpoint count and frequency barriers in raw Hz. The random walk determines every pitch. Consonance is a statistical event. Six stochastic fields trace continuous trajectories through parameter space: barrier widths open from 35% to full range and close again, Cauchy distribution blends from zero to 65% at the climax, and second-order walks let the stochastic parameters evolve autonomously within reflecting barriers. The master gain is flat for the entire duration. All dynamics emerge from the amplitude random walk alone. Every performance is unique at every time scale.
