Newsgroups: comp.speech
Path: pavo.csi.cam.ac.uk!pipex!uunet!panther!mothost!lmpsbbs!isu6!ronzu
From: ronzu@comm.mot.com (Ron Zuckerman)
Subject: Re: Vowel Formant Analysis
Reply-To: ronzu@comm.mot.com
Organization: Motorola
Date: Mon, 5 Oct 1992 15:14:32 GMT
Message-ID: <1992Oct5.151432.13922@lmpsbbs.comm.mot.com>
References: <1992Oct2.220813.1146@agog.com>
Sender: news@lmpsbbs.comm.mot.com (Net News)
Nntp-Posting-Host: 145.12.232.8
Lines: 30

In article 1146@agog.com, greg@agog (Greg Cockroft) writes:
>I'm working on vowel formant analysis of 8Khz mulaw sound.
>
>I have the complex roots of the Z transform, but I don't know how covert  
>those into frequencies.
>
[table deleted]

f = arg(real+j*imag) * fs / (2*pi)

where:

                       { arctan(imag/real)          real >= 0
    arg(real+j*imag) = { arctan(imag/real) + pi     real < 0, imag >= 0
                       { arctan(imag/real) - pi     real < 0, imag < 0

    fs = sampling frequency

However, since you're dealing with a real valued signal, you can take the absolute
value of the frequency from the above formula.          

---
------------------------------------------------------------------
| Ron Zuckerman, KA4RPD  E-mail: ronzu@isu.comm.mot.com          |
| Phone: (708)538-3134   FAX: (708)538-2818                      |
------------------------------------------------------------------
| "A madman got us into this, and it's beginning to look as if   |
|  only a madman can get us out." -- Chekov                      |
------------------------------------------------------------------

