home *** CD-ROM | disk | FTP | other *** search
- Newsgroups: comp.programming
- Path: sparky!uunet!europa.asd.contel.com!darwin.sura.net!spool.mu.edu!sol.ctr.columbia.edu!sol.ctr.columbia.edu!mhall
- From: mhall@colossus.comp.programming (Matthew Hall)
- Subject: Power spectrum analysis help wanted.
- Sender: nobody@ctr.columbia.edu
- Organization: Oberlin College Computer Science
- Date: Wed, 11 Nov 1992 15:27:25 GMT
- Message-ID: <MHALL.92Nov11102725@colossus.comp.programming>
- Distribution: comp.programming
- X-Posted-From: colossus.cs.oberlin.edu
- NNTP-Posting-Host: sol.ctr.columbia.edu
- Lines: 25
-
- Hello-
- I want to find power spectra for small (fixed at 512 or 256 points)
- waveforms. I only want 16 or 32 different frequency gradations. My
- problem is, I have no idea how to do this.
- I do have the FFT algorithm from Numeric Recipies, however I
- can't seem to get it to work very well, and it's slow.
- The reason for wanting this is that I am trying to build a
- speech recognition system (very small, discrete, speaker dependant)
- for a Macintosh. I need a very quick way to determine spectra - using
- integer arithmetic if possible (The data points are merely signed
- bytes). I have seen some products (soundtracker, Sound edit) perform
- (if not real time) very quick spectral analysis so I know it's
- possible.
- Any pointers, or source code (I'm fluent in pascal, but I can
- read C pretty well.) that you can give would be greatly appreciated.
-
- Thank you
- -matt hall
- --
- -------------------------------------------------------------------------------
- Matt Hall. mhall@occs.oberlin.edu OR SMH9666@OBERLIN.BITNET
- (216)-775-6613 (That's a Cleveland Area code. Lucky Me)
-
- F(X)=M*X*(1-X)
-
-