[PD] [soundtouch~] pitch shifter for Pure Data

katja katjavetter at gmail.com
Sun Aug 21 21:00:24 CEST 2011


Hello,

A few weeks ago I was reminded of my promise to share [soundtouch~], a
SoundTouch port for Pure Data. I use this class since quite some time now,
and had the code available on my website earlier on, but when I found that
it could not work for Linux 64 bit I withdrew the project for a while.
Reportedly, Olli Parviainen's newest version of SoundTouch is ready for 64
bit. I can not check this as I have no access to a 64 bit machine at the
moment. Anyway, I have built [soundtouch~] binaries for OSX(fat), Linux32
and Windows32, and uploaded it together with source and help patch:

http://www.katjaas.nl/pitchshift/soundtouch~.html

SoundTouch is a time domain pitch shift  / time stretch library, and I've
built it into [soundtouch~] as a real time pitch shifter. It is suitable for
the transposition of material with simple harmonic content, like voice,
flute etc. The SoundTouch library is not perfectly adapted to real time
processing, because you have to choose (at compile time!) between a build
which produces dropouts during pitch factor change, and a build which has
increased latency. I opted for the clicks instead of the increased latency
(which is by necessity already considerable) in my binaries distribution,
but if you want you can recompile according to your preference. To solve
this clicks-vs-latency issue, if possible at all, would require a
modification of the SoundTouch library itself.

Apart from this issue, I'm fairly content with [soundtouch~]. Let me know
what you think about it. If [soundtouch~] is considered a useful addition to
existing Pd methods for pitch shifting, it may be worth the effort to try
and solve the clicks-vs-latency issue. It is not a trivial task though.This
time I'm not promising anything.

Katja.
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.puredata.info/pipermail/pd-list/attachments/20110821/edd07947/attachment.htm>


More information about the Pd-list mailing list