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: Richard Dobson <email@hidden>
- Date: Sat, 20 Feb 2010 19:28:16 +0000
On 20/02/2010 18:42, tahome izwah wrote:
2010/2/20 Richard Dobson<email@hidden>:
On 20/02/2010 12:45, contact wrote:
..
The free dspdimension code is expensive on CPU, and a wonderful
misdirection on how to do pitch shifting (but that's another story!).
Optimised implementations (not least of the FFT itself) are ~very much~
faster, and of course that technique is targetted at arbitrary sounds
(polyphonic, inharmonic sounds, whole orchestras etc), maybe OTT for use on
monophonic pitched sounds and speech, where techniques such as PSOLA can
shine; but so can other f/d techniques such as Ceptstral analysis.
Just curious: faster than what exactly - faster than PSOLA or than
smbPitchShift()? FFT based implementations are not generally faster
than PSOLA as far as my measurements go.0
Sorry not to be clear - faster than smbPitchShift. I have not made any
comparisons with PSOLA as my ongoing interest is primarily in generic
pvoc-style tools, rather than anything specific to the voice or to mono
sources generally. Don't forget that in referring to PSOLA as a
comparison you do need to include the pitch detection stage which feeds
into the PSOLA algorithm itself. There are enough ways to perform that
detection step (including combinations of time and frequency-domain
methods) to make a comparison somewhat complex to define in general terms.
Richard Dobson
_______________________________________________
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