I can understand that you're having some frustration with your project. The folks here have pointed you to stuff they thought would help. It didn't. This happens. Most of the time, if you're asking a question which has some real meat to it (and yours does), the answers you get from just asking people at relatively speaking random don't help much other than to point you toward some possibilities and to confirm that you have a lot of work to do. If you can pull off realtime FFTs in Python and get something musically interesting out of it, you've done something that would probably qualify as graduate-level computer music work. I have to say I'm not surprised that someone else couldn't just hand you a solution! It does seem sometimes as if the Web is a cornucopia of instant gratification - so many things *are* available just for the asking, The fact that this one isn't should make you very happy! You have an opportunity to be "that guy" who's the one people think of when the subject of realtime music processing in Python comes up ... but only if you're willing to pursue the problem yourself, including all the hard bits and "aw crap I have to learn this/write this/figure this out from first principles". I'd *love* to hear what you finally come up with when you get there -- but I'm afraid it's probably going to be your own road. As an afterthought, have you seen the Snack package (http:// www.speech.kth.se/snack/)? It seems to have capabilities that you might be able to use with sufficient work, but Im pretty sure it's not an out-of-the-box fit. --- Joe M.