santosh.nath@ntlworld.com (santosh nath) wrote in message news:<6afd943a.0410150512.61393ce1@posting.google.com>...
> Speech is short stationary signals. Can anybody use "wavelet transform"
> for speech/voice or audio processing? Do they get any extra benefit out of it
>  e.g using variable window size - since speech is sparse in nature also?

Speech may be approximated as short stationary signals in a lot of
work, but I wouldn't call that a particular accurate description of
speech. That said, there should be a wealth of material to sort
through about wavelet coding of speech. In the early 90s there was a
lot of excitment about wavelets for audio coding, and especially for
speech. It was going to change the world. It was 3 months from mass
market deployment. It never really appeared in any product I know of.
 
> How about complexity of wavelet based analysis compared to conventional LPC 
> analysis - the question just came out of head. 

I'm not sure what you mean. LPC is fairly lightweight. Its the rest of
any decent codec, TTS synthesiser, or whatever which takes all the
heavy work. In early 90s technology one of the attractions in wavelets
for some people was the potential for well integrated part analogue,
part digital wavelet codecs for speech. Zilog made something along
those lines for digital answering machines. Never saw it deployed,
though.

Regards,
Steve

santosh nath wrote:
> 
> Speech is short stationary signals.

Sorry, I disagree. Speech is a really good example of a signal which is
inherently non-stationary.

> Can anybody use "wavelet transform" for speech/voice or audio processing? 

Sure, no reason why not.

> Do they get any extra benefit out of it
>  e.g using variable window size - since speech is sparse in nature also?

The wavelet transform doesn't have a window size, variable or otherwise.

Erik
-- 
+-----------------------------------------------------------+
  Erik de Castro Lopo  nospam@mega-nerd.com (Yes it's valid)
+-----------------------------------------------------------+
The idea that Bill Gates has appeared like a knight in shining armour to
lead all customers out of a mire of technological chaos neatly ignores
the fact that it was he who, by peddling second-rate technology, led them
into it in the first place. - Douglas Adams in Guardian, 25-Aug-95

Speech is short stationary signals. Can anybody use "wavelet transform"
for speech/voice or audio processing? Do they get any extra benefit out of it
 e.g using variable window size - since speech is sparse in nature also?

How about complexity of wavelet based analysis compared to conventional LPC 
analysis - the question just came out of head. 

Regards,
Santosh