Re: An ASA cue for phase information

[Steve Beet <steve.beet@xxxxxxxx>]

Sorry to be slow in responding to this, but I think the most straightforward
(i.e. easiest to calculate and easiest to use) analysis of phase information
is the modified group delay function developed by Yegnanarayana. As I've
mentioned before on this list, my own "inverse phase velocity" auditory
parameter did this as well, and is probably more closely related to auditory
perception, being based on an early version of Dick Lyon's auditory model.
However, my parameter is relatively difficult to compute, and gives such
high "resolution" that it's also difficult to use in any standard ASR/DSP
framework. So basically I'd suggest using some form of group delay
parameter.

Good luck!

Steve Beet


-----Original Message-----
From: AUDITORY - Research in Auditory Perception
[mailto:AUDITORY@xxxxxxxxxxxxxxx] On Behalf Of azam rabiei
Sent: Tuesday, November 16, 2010 11:42 AM
To: AUDITORY@xxxxxxxxxxxxxxx
Subject: An ASA cue for phase information

Dear List,

 

I am working on CASA (Computational Auditory Scene Analysis). 

For feature extraction, I used an onset/offset analysis and a harmonic
analysis to keep the track of temporal and spectral information,
respectively. 

I need to have phase information, as well. But, spectrogram has no
information of phase. I'm looking for a cue for phase information.

 

I will appreciate anyone who could give me any point.

 

Regards

Azam