Newsgroups: comp.speech
Path: pavo.csi.cam.ac.uk!doc.ic.ac.uk!warwick!zaphod.crihan.fr!jussieu.fr!math.ohio-state.edu!cs.utexas.edu!swrinde!news.dell.com!tadpole.com!uunet!hub.cs.jmu.edu!pranav
From: pranav@hub.cs.jmu.edu (Pranav Singh)
Subject: Speech Data
Message-ID: <CMB28H.Gnu@hub.cs.jmu.edu>
Organization: James Madison University
X-Newsreader: TIN [version 1.1 PL8]
Date: Mon, 7 Mar 1994 17:36:15 GMT
Lines: 24

Hello netters.
I have been working on my master's thesis this semester and have obtained
some data that some of you may be interested in, or may have some suggestions
about..
The software that I have been using is KAY Elemetrics' Computerized Speech
Lab (CSL).
From a set of sustatined vowels sung by both males and females, I have
extracted the following details:
1. The positions of the formants (Frequency)
2. The average power curve over the entire signal
3. The frequency response curve at a particular instance in time
The combined details that all these factors contributed have helped me to 
build up some sort of descriptive language about phonemes. The set is limited,
however and I will need to do the same procedure for all the phonemes in
English.
All this has been with the goal of characterizing a vowel and expressing it
as a unique vector to be used in a vector quantization neural net.
So far I have just completed the pre-processing part and will start on the
real network issues soon.
Any comments or questions will be appreciated. It took me almost 3 months
just to figure out what I was looking for!
Thanks
Pranav Singh

