Newsgroups: comp.speech
Path: pavo.csi.cam.ac.uk!doc.ic.ac.uk!pipex!howland.reston.ans.net!math.ohio-state.edu!sdd.hp.com!saimiri.primate.wisc.edu!news.crd.ge.com!sunblossom!knight.vf.ge.com!sde8a12!patter_d
From: patter_d@sde8a12.vf.ge.com. (Patterson David M.)
Subject: Perceptual coding in voice coding		
Message-ID: <1994Mar3.153807.24874@knight.vf.ge.com>
Sender: news@knight.vf.ge.com
Nntp-Posting-Host: sde8a12.mdso.vf.ge.com
Reply-To: patter_d@sde8a12.vf.ge.com.
Organization: Sun Microsystems, Inc.
Date: Thu, 3 Mar 1994 15:38:07 GMT
Lines: 14

Hi Netters:


I have been reading through some (not much) of the literature on voice coding and noticed the main technique is a form of Linear Prediction.  I am curious to know what other techniques have been tried and how they compare to LP techniques such as the NSA - CELP standard.  

The latest techniques used in recording music employ some form of perceptual encoding.  Other than briefly reading about this in Stereo Review, I have not seen anything in the literature.  Does anyone have any references they could point me to?

Would perceptual encoding (pe for short) be worthwhile for voices?  My understanding of the benefits to music of pe is based on the harmonics associated with the music are masked by other harmonics.  Would the resonances of the human voice be similar. 

As you can see, these are some pretty basic questions.  Any help on these matters would be appreciated.


David Patterson
patter_d@mdso.vf.ge.com
