Works for live electronics made through a no-input feedback system modulated by multi-scale feature extracted signals, i.e. both audio and sub-audio signals from sound system components, and a machine listening arrangement for timbral analysis consisting of spectral shape descriptors (‘centroid, spread, skewness, kurtosis, rolloff, flatness and crest’), mel-frequency cepstrum coefficients, and pitch detection (FluCoMa), recursively applied as modulation signals through an output combinatorial matrix.