Re: Need to hire a software developer for a voice processing app!
Re: Need to hire a software developer for a voice processing app!
- Subject: Re: Need to hire a software developer for a voice processing app!
- From: "contact" <email@hidden>
- Date: Sat, 20 Feb 2010 12:45:39 -0000
On 20/02/2010 11:07, Paul Sanders wrote:
If you can't use Audio Units, there is some 'wide open license'
pitch-shifting source code here:
http://www.dspdimension.com/download/
There's also a link to an article which explains how it works.
Yes, it is a very good reference. Note however that being a demo it is not
optimised for speed; it would need to be considerably modified (e.g. to
use "in-place" computation) to run comfortably in real time on
lower-performance hardware. The iPhone (though no slouch) is a timely
reminder of the need to employ all those robust optimisation techniques
that modern multi-GHz machines have allowed us to forget!
As I noted earlier, all the dsp specified can be found in Csound, which
therefore offers an immediate (and free!) laboratory for designing and
testing things, with some efficient streaming phase vocoder tools for
pitch shifting, tracking, etc - the pvs framework, implemented by me :-).
Richard Dobson
The free dspdimension code is expensive on CPU, and a wonderful
misdirection on how to do pitch shifting (but that's another story!).
A PSOLA algorithm is probably going to be the answer, especially as the
fundamental frequency is going to be needed in the future.
It also solves the issue of formant preserving, which I presume is
automatically needed - and can be an expensive step for vocoders.
So a demonstation of real-time pitch marking would be the first task for the
phone app, me thinks.
Dave
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Coreaudio-api mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden