Newsgroups: comp.speech
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!news.mathworks.com!news.ultranet.com!news.sprintlink.net!howland.reston.ans.net!math.ohio-state.edu!jussieu.fr!univ-lyon1.fr!swidir.switch.ch!news.unige.ch!arolla!wassner
From: wassner@idiap.ch (Hubert Wassner)
Subject: Question :Wavelets for Speech analysis
Message-ID: <1995Apr7.184159.6292@news.unige.ch>
Keywords: Scale,frequency
Sender: usenet@news.unige.ch
Reply-To: wassner@idiap.ch
Organization: Institut Dalle Molle d'Intelligence Artificielle Perceptive
Date: Fri, 7 Apr 1995 18:41:59 GMT
Lines: 30

I'm making a Scalogram (=Spectrogram with wavelets instead of 
fourrier transform).

I used a wavelets decomposition with integer Scale, (not with a power of 2),
to have a large choice of different scales to analyse the speech signal.

But it's not enough:

The big problem is that the relation between scale and frequency is:


  Frequency = lambda/scale   where lambda is a constant


Then it's hard to reach the high frequency with an integer scale.
while seeing a Spectrogram and my scalogram I noticed that I can only reach
the frequency below 1000 Hz and the speech signal goes over 1000 Hz.

So I miss a large part of the frequency of the signal.

--> I'm loocking for any information (address,bibilography,..)
    to access 'high' frequency analysis with wavelets.


    Thank you. 

 



