Return to the archive index

Re: CMU Sphinx Speech Recognition

The thing that is really interesting about an
open-source voice recognition package is that you 
should be able able to start thinking about self-adaptive 
improvements using festival, the open-source speech synthesizer.

You should be able to fix one to speak with more articulation, 
or fix the other to recognize the idiocyncracies of the first. 
In any case, starting to share language data between to two 
projects might be a valuable research topic!

	-pehr

Lenny Foner wrote:
> 
>     Date: Wed, 02 Feb 2000 19:24:09 +0000
>     From: "R. Paul McCarty" <rpmc@troi.cc.rochester.edu>
> 
>     CMU opened their running speech recognition project as open source.
>     It's a relatively small buy powerful statistical speech recognition
>     program that will run on linux, freeBSD, SunOS and windows NT. For those
>     who don't know, there isn't really any speech recognition on linux yet.
>     This is the closest I've seen.
> 
> IBM announced ViaVoice for Linux on 26 Apr 99.
> 
>   http://www-4.ibm.com/software/speech/dev/
>   http://www6.software.ibm.com/dl/viavoice/linux-d?S_PKG=linux&S_CMP=&S_TACT=
>   http://www-4.ibm.com/software/speech/dev/faq_linux.html
>   http://viavoice.sparklist.com/
>   http://www-4.ibm.com/software/speech/dev/faq_tf.html
> 
> How does it compare to this?
> 
>     For those working on speech recognition in their wearable, have you
>     found something different, written your own, or just use windows?
> 
>     Almost forgot the link..
> 
>       http://www.speech.cs.cmu.edu/sphinx/
> 
>     It's available from source forge for download:
> 
>       http://sourceforge.net/project/?group_id=1904