h o m ed———o———w———nl—o—a——dc———i——t——a——t———i—o——n———
J. MacCallum, A. Schmeder, and D. Wessel. Timbral migration: Stochastic processes for the control of smooth spectral transformation. In Proceedings of the International Computer Music Conference, 2009.
We present an additive synthesis engine that can produce smooth transitions between timbres without glissandi or audible transients. With this algorithm, one can explore the musically rich area of hybrid timbres that arise as one timbre "migrates" into another. This method operates in real-time using live or pre-recorded sinusoidal analyses as input. Parameterization of the migration behavior enables interactive control. We present the history of this engine that begins with work by David Wessel at IRCAM in the 1970s. We discuss two implementations: a statistical method based on sampling from a discrete probability mass function derived from a sparse input spectrum, and a stochastic local-update method using parallel Markov Chain Monte Carlo processes, which enables control over spectral flux, spectral smearing and migration rate. We conclude with a summary of notable musical uses in contemporary electroacoustic music.