Re: pitch extraction (beauchamp james w )


Subject: Re: pitch extraction
From:    beauchamp james w  <jwbeauch(at)UX1.CSO.UIUC.EDU>
Date:    Thu, 17 Jul 2003 12:43:31 -0500

Nicholas, Geoff, and Malcolm, For doing pitch detection with MAX/MSP, I hear that Miller Puckettes' "Fiddle" works well. I think it is intended for musical applications. Speaking of musical applications, I wonder if the ESPS pitch tracker that Malcolm mentions works well for various instruments as well as sung voice. We've tried Praat on music, and it seems to work well, although we haven't given it a thorough test. Another algorithm to consider is Alain deCheveigne's YIN, which was published in JASA last year. Both his and the Praat method are based on subtractive autocorrelation, I believe. It would be nice though to have the pitch come out (as a graph) in some sort of musical pitch units like semi-tones or cents -- i.e., in log freq units. I know this can be thrown together, but musicians and musicologists could use a standalone program, ready-to-go. Jim


This message came from the mail archive
http://www.auditory.org/postings/2003/
maintained by:
DAn Ellis <dpwe@ee.columbia.edu>
Electrical Engineering Dept., Columbia University