Newsgroups: comp.speech
Path: pavo.csi.cam.ac.uk!doc.ic.ac.uk!mrccrc!dcs.warwick.ac.uk!warwick!zaphod.crihan.fr!univ-lyon1.fr!scsing.switch.ch!ira.uka.de!howland.reston.ans.net!noc.near.net!uunet!mcsun!sunic!ericom!etlxd22h!etlmts
From: etlmts@etlxd20.ericsson.se (Martin A. Thomas)
Subject: Damaged vocal chords
Message-ID: <1993May13.100549.7382@ericsson.se>
Sender: news@ericsson.se
Nntp-Posting-Host: etlxd22h.ericsson.se
Reply-To: etlmts@etlxd20.ericsson.se
Organization: Ericsson Telecom, Stockholm, Sweden
Date: Thu, 13 May 1993 10:05:49 GMT
Lines: 16

How does a speech recognition program cope with the voice of someone
whose vocal chords are damaged (ie no longer sufficiently close to 
produce the normal speech frequencies but so they are still capable
of whispering, resulting in a lot more of the higher frequencies)?

What I really want to know is if there are any of the normal voice
frequencies in there that can be processed to produce a more normal
sounding voice. There must be some way that the voice can be processed
if what is said is still recogniseable as speech.  Any thoughts, ideas,
references etc gratefully received.

Thanks

Martin Thomas

No signature yet...
