ok so here's what happened:
i took a 1 second clip of andy's track and tested out my new patch with it. so a wavelet is this thing with an envelope whose duration is based on the number of zerocrossings it contains. (traditionally this is not the case, but traditional wavelet analysis has some limitations in resynthesis). anyway i took the bitch and played around with quantization grids for the wavelets to fall into on the frequency time axes. the piece uses the following ratios: [1, 1:1.2, 1:1.5, 1:1.6, 1:2, 1:2.4, 1:3, 1:3.2] . this pitch constellation is a minor chord with a flatted sixth in just intonation. the time quantizations vary throughout the piece, and are affected by such things as number of zerocrossings (ranging from 1 to 128), density (or how much each wavelet overlaps onto the next), in addition to a variable rounding multiple and max value. at times there are as many as 128 particles overlaping each other. these dense figures are made even denser occasionally by increasing the overal particle size. notice how the quantization effects are much subtler than simply dealing in terms of pitch and timing. the true richness of these effects is exerted on the spectrum of the sound. these techniques are brand new to me; the patch is less than an hour old. i plan to explore the spectral implications of convolving these wavelets with bandlimited impulses. i also plan to fuck with andy a little more before the weekend...
