Newsgroups: comp.speech
From: Lee@lowlevel.demon.co.uk (Lee Brown)
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!hookup!swrinde!howland.reston.ans.net!pipex!peernews.demon.co.uk!lowlevel.demon.co.uk!Lee
Subject: Re: Converting raw signals to frequencies
References:  <790779103snz@quinion.demon.co.uk>
X-Mailer: cppnews $Revision: 1.42 $
Organization: Home
Lines: 14
X-Posting-Host: lowlevel.demon.co.uk
Date: Mon, 23 Jan 1995 18:51:36 +0000
Message-ID: <790912296snx@lowlevel.demon.co.uk>
Sender: usenet@demon.co.uk

In article <790779103snz@quinion.demon.co.uk>
   Brian@quinion.demon.co.uk (Brian Quinion) writes:
> I am trying to write a very simple speech recognition program (6 or 7 words)
> and have fallen at the first hurdle.
> How do can the frequency and amplitudes that combine to produce the raw
> (amplitude against time) signal be determined from the raw signal?
You need something known as a FFT (Fast Fourier Transform) which time *time*
to compute, but will give you the amplitute through time for a range of
frequencies as I understand it.

I'm not a maths/physics person, but do know this much.  Hope it helps.

Regards,
||  Lee Brown  ||  Internet: lee@lowlevel.demon.co.uk ||  CI$: 100325,3304  ||
