in article tridndhBF-K9GdDfRVn-gQ@giganews.com, bouyang at
bing_research@yahoo.com wrote on 04/01/2005 12:25:

> I have looked at some literature on using autocorrelation or AMDF for picth
> detection/extraction.
> 
> If I understand correctly, all these methods have two parts -- first
> construct the functional that can be used to extrac the period/pitch
> (AMDF, autocorrelation etc.), second, find peak/null and the associated
> lag index to obtain the pitch.
> 
> What I need help is how the extraction part is generally implemented (i.e.
> how to pick the peak/null).

i guess, think about how you would pick those peaks manually, try to
describe explicitly every decision you make in pick the peak, and then code
it.
 
> Thanks in advance.

FWIW.

-- 

r b-j                  rbj@audioimagination.com

"Imagination is more important than knowledge."

I have looked at some literature on using autocorrelation or AMDF for picth
detection/extraction. 

If I understand correctly, all these methods have two parts -- first
construct the functional that can be used to extrac the period/pitch
(AMDF, autocorrelation etc.), second, find peak/null and the associated
lag index to obtain the pitch.

What I need help is how the extraction part is generally implemented (i.e.
how to pick the peak/null). 

Thanks in advance.
 


		
This message was sent using the Comp.DSP web interface on
www.DSPRelated.com