preface: I've only been working with max a few months; I'm using max 5 on mac os.
I've been working with sigmund~ and CNMAT's objects and tutorials to analyze and transform the spectrum of a live signal. I've incorporated resonances~ and am using list-interpolate to transition from resonance models.
it seems like I'm having the same issue brought up in this thread: http://cycling74.com/forums/topic.php?id=39530
so I'm just going to quote the relevant part with the last sentence framing the key issue:
"Using the CNMAT spectral tutorials as a base, I'm trying to make simple, smooth interpolations between frames using the list-interpolate object. But whereas the tutorials' example SDIF files produce smooth transitions where partials fade in and out by amplitude (or widen by frequency scaling and then adjust by amplitude) my partials produce sweeps across the audio range during interpolation. Almost always the lower partials cross the higher ones as they nearly switch places, and often new partials begin at the low end of the audio range and then spread out across the audio range "taking their positions" in the new model. Nothing in between the two models sounds like either sound source."
I'm hearing a lot of sweeping sines in the interpolation steps that seem too huge (like across the frequency spectrum). Is this unavoidable somehow with list-interpolate or otherwise?
here's part of the patch:
as a sidenote I'm really not sure if the way I've built the freq and amp pairs list in the subpatchers is really the best way or if this might be to blame for my interpolation issue. so I'd appreciate any comments on patch efficiency/construction too.