Path: utzoo!utgpu!news-server.csri.toronto.edu!rpi!zaphod.mps.ohio-state.edu!sol.ctr.columbia.edu!emory!gatech!bloom-beacon!eru!hagbard!sunic!mcsun!ukc!slxsys!ibmpcug!hdrw
From: hdrw@ibmpcug.co.uk (Howard Winter)
Newsgroups: comp.periphs,connect.audit
Subject: Re: Looking for Speech Recognition
Message-ID: <1991Apr4.234022.5498@ibmpcug.co.uk>
Date: 4 Apr 91 23:40:22 GMT
Organization: The IBM PC User Group, UK.
Lines: 25

(Request for info about speech recognition)
The state of the art in speech recognition is pretty basic at present.
The only reasonably priced systems have to be 'trained' to the voice/
words of the speaker, and have very limited vocabulary.  They are 
unable to understand continuous speech and have a less-than-perfect
success rate on single words.  The reason for this is that speech
is very complex, and it is now recognised that Artificial Intelligence
is needed to get any further, since humans understand speech in the
most part by context - try listening to a conversation between people
with a strog foreign accent - until you have 'tuned' your brain to
their pronunciation, it is difficult to catch anything of what they
are saying.
Sorry to be negative about this, but your expectations outstrip the
capabilities currently achieved - language translation has also
come up against a 'wall' in technology.  They'll get there one day,
but don't expect it soon, cheap, or infallible.

Regards,
Howard.
-- 
Automatic Disclaimer:
The views expressed above are those of the author alone and may not
represent the views of the IBM PC User Group.
-- 
hdrw@ibmpcug.Co.UK     Howard Winter     0W21'  51N43'