Re: multi-pitches of quasi-periodic musical tone (Matt Flax )

Subject: Re: multi-pitches of quasi-periodic musical tone
From:    Matt Flax  <flatmax(at)IEEE.ORG>
Date:    Tue, 3 Jun 2003 18:00:25 +1000

Hi Chen-gia, I am not 100% sure what is going on in that transcription ... it does look like you have captured some form of tonal reference in them. Assuming you have used temporal correlative techniques to generate these then you may also want to check out the place models. For references you can check the paper which is distributed with an algorithm of my own which is still in development ... Matt On Tue, Jun 03, 2003 at 08:08:47AM +0800, Chen-gia TSAI wrote: > Dear list, > > I am analyzing multi-pitches of musical tones which are predominated by 5th, 7th, 9th, 11th, and/or 13th harmonics. Multi-pitches of these quasi-periodic complex tones have been transcribed with the help of autocorrelation. However, it seems that stream segregations due to multi-pitches is noticeable, but the perceived pitches are ambiguous. > > An example of Chinese flute music with multi-pitches is at > The transcription is at > > Small notes above the prominent F0 pitches are determined according to the "streams" in the autocorrelogram. I found that it is difficult to compare the perceived pitches to the musical notes in the transcription. > > Any suggestions and/or references are much appreciated. > > Best, > Chen-Gia Tsai > gia(at) -- WSOLA TimeScale Audio Mod : FFTw C++ : Vector Bass : Multimedia Time Code :

This message came from the mail archive
maintained by:
DAn Ellis <>
Electrical Engineering Dept., Columbia University