Abstract

In this paper we present an example of the use of the singing voice as a controller for digital music synthesis. The analysis of the voice with spectral processing techniques, derived from the Short-Time Fourier Transform, provides ways of determining a performer's vocal intentions. We demonstrate a prototype, in which the extracted vocal features drive the synthesis of a plucked bass guitar. The sound synthesis stage includes two di#erent synthesis techniques, Physical Models and Spectral Morph.

Citations