Newsgroups: comp.speech
Path: cantaloupe.srv.cs.cmu.edu!das-news2.harvard.edu!news2.near.net!howland.reston.ans.net!math.ohio-state.edu!jussieu.fr!oleane!pipex!uknet!info!atlantic!eeopensh
From: eeopensh@atlantic ()
Subject: Re: speech recogn.
Message-ID: <D0Fp2t.6B6@info.swan.ac.uk>
Sender: news@info.swan.ac.uk
Nntp-Posting-Host: atlantic.swan.ac.uk
Organization: Swansea Info Server
X-Newsreader: TIN [version 1.2 PL2]
References: <3b9l23$7li@newsbf01.news.aol.com> <3bl9s7$cft@jupiter.planet.net> <3bnrj2$sp3@netaxs.com> <3bo520$j48@agate.berkeley.edu>
Date: Wed, 7 Dec 1994 09:30:28 GMT
Lines: 27

John Lazzaro (lazzaro@boom.CS.Berkeley.EDU) wrote:
: >: e33029@aol.com (E33029) wrote:
: >: >
: >: > Could someone tell me the best speech recognition problem on the market.
: >

: Micropower, under $2 (US, OEM quantity) chips that do
: speaker-independent wordspotting of a few control commands in noisy
: environments, with acceptable performance for everyday applications.
: All the descriptive words above are needed, any less makes a tietack
: instead of a solution.  These chips could be to mechanical switches
: what light-emitting diodes were to filament bulbs.


This sounds an interesting development. Are these chips actually on the 
market and do you have anything more in the way of a technical spec.
Do they actually perform word-spotting in connected speech, or do they 
need discrete utterances? What are the control commands and what languages
are supported, is it just American English? What is the definition of 
acceptable performance for everyday applications? I know of applications
that could use this device as you've described it.


Cheers, 

John Openshaw

