rick@hanauma.Stanford.EDU (Richard Ottolini) (09/05/90)
Prof. Grossman, one of the founders of the subject, spent this summer at Stanford's Center for Computer Music. They were using wavelets as a local spectral transform to analyse and synthesize music. The time-varying envolope of the musical signal is approximately the melody while the frequency modulation is approximately the instrument and tone. They weren't using the more recent Fast Wavelet Transforms in the current literature but the older Gaussian weighted Fourier Transforms. They plotted something that was a cross between a Witkin-like scale-space graph and a speech spectrum. Grossman and companysay to be writing a book on musical applications of wavelets.