so this new constant-q transform thing is working sorta. if i am to record buffer data i currently am limited to 16 bins/octave. which sucks for realism. but i like the squelch of it. different time / frequency distortions can happen because it writes arrays that are summed to produce the spectrum. in this recording the playrate and pitch are manipulated. there is a jitter parameter for playrate as well. this demonstrates the transform's timing resolution advantages over fft. also the frequency analysis happens at a logarithmic scale, unlike the linear scale of fft. this makes the transform more effective than an fft using the same number of bins.
