Baslinux mailing list

[sindi keesan <keesan AT sdf.lonestar.org>]

 I also would really like to have the option
> of using speech-to-text software (like Dragon Naturally Speaking) since I
> need to write long papers at times and typing can create a bottle-neck.
> Ironically, IBM produced a very capable speech recognition package for
> Linux. But they stopped development and no longer make it available for
> download. Bootleg copies can still be gotten. So, if I want to have this
> speech-to-text software and remain within legal/ethical boundaries, I
> would have to use Windows. I wish this weren't the case. But for some
> fields, like higher level studies in the humanities, Linux just doesn't
> provide everything needed. Sad but true.
>
> James


See http://linux-sound.org/speech.htm for links to a few speech 
recognition (text to speech) projects including one that seems to be 
somewhat usable already called ISIP 
http://www.isip.msstate.edu/projects/speech/software/index.htm

The requirements for the previous prototype were 500MHz 128MB RAM 256MB 
swap and gcc 3.2 (later compilers have a bug).  The latest prototype now 
wants 256MB RAM and 512K swap space.

Works with linux and there is even a Windows port (which I did not manage 
to track down).

You need to train the program to recognize your speech.

There are also links to several text to speech programs (synthesizers).