Hi list, "J_Zar" <romaninz(a)aruba.it> wrote: [..] Please don't forget that this is a very new site - it was opened just yesterday. Frank

Hi Erik, On Mon, 2004-04-05 at 22:25, Erik de Castro Lopo wrote: Could you reveal some details? :) Marek

Es geschah am Montag, 5. April 2004 23:33 als Erik de Castro Lopo schrieb: Which brings me to the question: how old are you? :P Just kidding, But I also planned to do some research on pitchshifting in conjunction with formant correction. If anybody has good material about that, don't hide it! Seems this field screams for getting elaborated. CU Christian

On Tue, Apr 06, 2004 at 08:38:09AM +1000, Erik de Castro Lopo wrote: well isn't that because you also spent about 10 years playing bass in a rock band? K.

Christian Schoenebeck wrote: I have already played with that stuff, only on the pure dsp side. For time scaling, one of the best approach, I think, is to use a phase vocoder with phase locking and adaptative frame size. The trick it to detect correctly the transients in the file: when there is a transient, choose a small frame (ie ~256 samples@44100), when there is no transient, choose a bigger frame. Check for example the article of dolson and Laroche on phase locked phase vocoder 4 *Improved phase vocoder time-scale modification of audio* /Laroche, J.; Dolson, M.;/ Speech and Audio Processing, IEEE Transactions on ,Volume: 7 , Issue: 3 , May 1999 Pages:323 - 332 About time transient detection, I happened to do my DSP master-like in England in the same lab than C. Duxburry, who di his PhD about that topic. Look for example the articles available at dafx.de. Look also to Axel Robel articles; Axel Robel worked at IRCAM, presently. One other idea I had recently, but havn't tried, is using Matching Pursuit to have a frequency/scale/time atoms decomposition, and play with it. In fact, basic time based time stretching techniques are kind of ganular synthesis ( see for example live! of ableton or variphrase), the idea of Matching Pursuit is to find an adaptative decomposition using dictionnary, which are redundant set of bases vectors, and split the signal in time/scale/frequency domain. The algorithms are quite difficult and very computive intensive. but the theory is beautiful. I used myself this technique to do multi pitch detection, following an article from Gribonval and Bacri, some results are impressive. Look for harmonic Matching Pursuit if you are interested. Cheeers, David

On Tue, Apr 06, 2004 at 11:28:28 +0200, Florian Schmidt wrote: Theres info about this on Stephan's website: http://www.dspdimension.com/start.html Its not that simple - but thats the basic idea. - Steve

[Florian Schmidt] please let me throw in that there's a high-quality time stretching GPL command line utility available at http://quitte.de/dsp/pvoc.html - based on a phase vocoder with fixed overlap (no transient detection) which easily beats soundtouch's results. also, there are phase-vocoding opcodes in csound which can, afaik, be used for the same purpose, and there's the sndan project which due to its obscure licensing ways i haven't looked into. tim

steve harris' plugins (plugins.org.uk) contains a LADSPA pitch shifting plugin, or maybe even 3. i think they all work in the time domain, perhaps 1 of them is frequency domain. [ ... ] yes. :)) seriously, the math is conceptually more complex than i think you may want to know. or i, for that matter. --p

On Wed, 7 Apr 2004 14:29:25 +1000 "Earle" <earle(a)ezi123.com> wrote: Most tracker style programs do not do real pitch shifting. Real pitch shifting (tm) modifies the pitch without affecting the playback duration. Tracker style pitch shifting (which is more accurately called vari-speed) uses a faster or slower playback speed to vary the pitch. Playing a sample back slower drops the pitch; faster raises the pitch. Varispeed, can be very easily implemented using a sample rate converter such as Secret Rabbit Code : http://www.mega-nerd.com/SRC/ In fact, in the examples directory of the source code tarball there is an example program called varispeed-play. Changing the pitch without changing the length is true pitch shifting. The code and concepts required to do this well is extremely complicated. Erik -- +-----------------------------------------------------------+ Erik de Castro Lopo nospam(a)mega-nerd.com (Yes it's valid) +-----------------------------------------------------------+ "If you are going to SCO to license superior driver technology, you need to reevaluate your product." -- Comment about SUN and SCO on Linuxtoday.com

On Wed, 7 Apr 2004 16:06:19 +1000 "Earle" <earle(a)ezi123.com> wrote: Sure, don't let me stop you. What you have just described is varispeed, not true pitch shifting. Erik -- +-----------------------------------------------------------+ Erik de Castro Lopo nospam(a)mega-nerd.com (Yes it's valid) +-----------------------------------------------------------+ "Hey, I've re-dorkulated." -- Prof. Frink (The Simpsons)

Erik de Castro Lopo wrote: To do it well is difficult, but the basic stuff is really simple, if you have a library to do sample-rate conversion: doing the time scale compression/expansion with the simplest method (overlapp add) is just a matter of buffering. It will sound awful, of course ! David

Yes! Thanks David and Dave! This was what I wanted to hear :) I hoped there was more than one way kill a sample. THanks very much! Earle

I forgot to add that there is a matlab script which does the basic phase vocoder stuff on musicdsp.org. As I am the one who writes this script, I can help you if you have any questions ;). As the script only uses the matlab language and fft, it should run on scilab and/or octave, but I haven't tried myself. cheers, David Cournapeau David wrote: