home *** CD-ROM | disk | FTP | other *** search
- Path: sparky!uunet!charon.amdahl.com!pacbell.com!decwrl!concert!gatech!udel!bogus.sura.net!howland.reston.ans.net!spool.mu.edu!caen!batcomputer!munnari.oz.au!manuel.anu.edu.au!andosl!tridge
- From: tridge@andosl.anu.edu.au (Andrew Tridgell)
- Newsgroups: comp.speech
- Subject: Re: Non-realtime speech transcription?
- Date: 26 Jan 1993 23:08:40 GMT
- Organization: CSLab, Autralian National Uni.
- Lines: 26
- Distribution: world
- Message-ID: <1k4g9oINNmer@manuel.anu.edu.au>
- References: <TOM.93Jan25090338@amber.ssd.csd.harris.com>
- NNTP-Posting-Host: 150.203.15.21
-
- In article <TOM.93Jan25090338@amber.ssd.csd.harris.com>, tom@ssd.csd.harris.com (Tom Horsley) writes:
-
- > An awful lot of the stuff in this group seems to be concerned with realtime
- > speech recognition (which is understandable, since it would be so useful),
- > but I am curious to know if there has been any research or actual
- > development of systems which can do a *really* good job of generating
- > transcripts of multi-speaker conversations, even if it can't be done in real
- > time (for instance, taking overnight to transcribe a 1/2 hour recording
- > would not be too unreasonable).
-
- I think most research systems are NOT real-time. I know mine is far
- from it at the moment. The difference between a real-time and off-line
- recogniser is basically the grunt of the computer it runs on.
-
- I think its mostly the commercial systems that try for real-time
- performance. These start as non-real-time systems and then they put
- them on fancy hardware.
-
- Andrew
-
-
- --
- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
- Andrew Tridgell CSLab, Research School of Physical Sciences
- Andrew.Tridgell@anu.edu.au Australian National University (x3064)
- =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
-