Newsgroups: comp.speech
Path: pavo.csi.cam.ac.uk!pipex!uunet!hub.cs.jmu.edu!pranav
From: pranav@hub.cs.jmu.edu (Pranav Singh)
Subject: Spectrogram interpretation.
Message-ID: <CJy6vG.44z@hub.cs.jmu.edu>
Organization: James Madison University
X-Newsreader: TIN [version 1.1 PL8]
Date: Thu, 20 Jan 1994 21:40:26 GMT
Lines: 11

Hello Speech experts.

I am writing my masters thesis on speech recognition using a neural network.
The sound aquisition is done by a software package that is called CSL; and
seems pretty powerful. I am studying spectrograms and LPC formant history
curves. Can anyone tell me what features in a spectrogram would be essential
in classifying phonemes ? And what metrics are used in the classification ?
Thanks a lot.
IneedallthehelpIcanget.
Pranav Singh

