Newsgroups: comp.speech
Path: pavo.csi.cam.ac.uk!doc.ic.ac.uk!uknet!pipex!howland.reston.ans.net!gatech!concert!unccsun.uncc.edu!usenet
From: cdgraham@uncc.edu (Christopher Darrell Graham)
Subject: Speech Recognition
Message-ID: <CGJG4D.4rq@unccsun.uncc.edu>
Sender: usenet@unccsun.uncc.edu
Nntp-Posting-Host: ws89.uncc.edu
Reply-To: cdgraham@uncc.edu
Organization: University of NC at Charlotte
Date: Mon, 15 Nov 1993 14:55:24 GMT
Lines: 15

Hello,
  I am working on a speech recognition software project.  I have read the FAQ and
a few books on the subject.  However, I am still a little confused on the details.
I am working on a Sun Sparc 2 station.  I have recorded about 8 words (recorded
in u-law then converted to linear) for the speech templates and processed them 
(DFT).  Then I run my program and speak one of the words into the mic and perform 
the DFT.  How do I compare the 2 DFT's to find a match?   Do I need to do 
time-warping or eliminate the silence (neither of which I really understand how 
to do)?  What about windowing?  I need to do all this with software.  Thanks 
for any help.

Christopher D. Graham
cdgraham@mosaic.uncc.edu


