Re: Matlab/Octave ACF based pitch extraction

[Matt Flax <flatmax@xxxxxxxx>]

I have attatched the replies received in the past 24 hours for those who
are interested.

Matt

On Thu, Jun 27, 2002 at 05:47:37PM +1000, Matt Flax wrote:
> Hi,
>
> I was wondering wether anyone knew of a software base which implemented
> the Meddis/Hewitt fibre Autocorrelation pitch extraction algorithm ?
>
> thanks
> Matt
>
> --
> http://mffm.darktech.org
>
> WSOLA TimeScale Audio Mod  : http://mffmtimescale.sourceforge.net/
> FFTw C++                   : http://mffmfftwrapper.sourceforge.net/
> Vector Bass                : http://mffmvectorbass.sourceforge.net/
> Multimedia Time Code       : http://mffmtimecode.sourceforge.net/

--
http://mffm.darktech.org

WSOLA TimeScale Audio Mod  : http://mffmtimescale.sourceforge.net/
FFTw C++                   : http://mffmfftwrapper.sourceforge.net/
Vector Bass                : http://mffmvectorbass.sourceforge.net/
Multimedia Time Code       : http://mffmtimecode.sourceforge.net/
_______________ Begin replies ___________________________________
From: Massimo Grassi http://www.psy.unipd.it/~grassi

Michael Akeroyd implemented some of the Meddis stuff for Matlab.
You could give a look to his web page.

http://www.biols.susx.ac.uk/home/Michael_Akeroyd/

______________ Next reply _______________________________________

From: "Lowel P. O'Mard" http://www.essex.ac.uk/psychology/hearinglab/dsam

Take a look at

http://www.essex.ac.uk/psychology/hearinglab/dsam/

You will find the development system for auditory modelling software
there.  You can get the software library itself with which you can write
your own applications or the ready made AMS application which you can
use to implement the pitch extraction algorithm.


...Lowel.

______________ Next reply _______________________________________

From: Malcolm Slaney http://www.slaney.org/malcolm/pubs.html

I'm not sure what you are asking for.  There is Meddis/Hewitt hair cell
code in the Auditory Toolbox, along with various forms of cochlear filter
banks.  There is autocorrelation code in there too, and finally the pitch
summary code.

Isn't this what you want?

- Malcolm

_____________ My reply to Malcolm _______________________________

Yes - exactly ... I have already wrapped your functions like so :
function [p, sc]=slaneyPitch(signal, fs )
  %#  Apply Slanmey's pitch finding algorithm
  if !exist('CorrelogramPitch')
    LOADPATH=[LOADPATH '/home/flatmax/research/3rdParty/AuditoryToolbox:']
  end

  frameRate=50;
  width=2048;
  cor=CorrelogramArray(signal, fs, frameRate, width);
  %#size(cor)
  [p, salience, sc]=CorrelogramPitch(cor, width, fs);
end

______________ Next reply _______________________________________

From: "Dan Ellis" http://www.ee.columbia.edu/~dpwe/

There's something very like it in the 'vowelExplorer' module of
Sheffield's Matlab Auditory Demonstrations, I think,

  http://www.dcs.shef.ac.uk/~martin/MAD/docs/mad.htm

_________________ End Replies _____________________________