Abstract:

Sound synthesis algorithms which radically depart from acoustical equations, and seek out numerical quirks at audio rate, can still have a part to play in the art-science investigations of computer music. This paper describes a host of ideas in alternative sound synthesis, from dilation equations and nonlinear dynamical equations, through probabilistic sieves, to oscillators based on geometrical formulae. We close with some new ideas in concatenative sound synthesis, using sparse approximation as the analysis method for matching, and driving synthesis through an EEG interface.