Subject: Re: [linux-audio-dev] query
From: Alexander Ehlert (ehlert_AT_gris.uni-tuebingen.de)
Date: Wed Feb 13 2002 - 20:02:42 EET
On Wed, 2002-02-13 at 17:39, Taybin Rutkin wrote:
> > bigger than the overlap you used before. So far the basic idea. Happy
> > implementing :-)
>
> Is that how the non-pitch-changing one works?
That's how sox does it, I suppose. But that only works for a limited
range. If you try to stretch it to much this way, you probably get some
strange artefacts. Another option are FFT or wavelet based algorithms.
Actually, this time stretching question would be something for a FAQ.
I'm subscribed to the music dsp list as well, and people frequently ask
how time stretching is done. There is a good explanation at
http://www.dspdimension.com/
Cheers, Alex
This archive was generated by hypermail 2b28 : Wed Feb 13 2002 - 19:53:54 EET