Voice recognition software

Eric S. Johansson esj at harvee.org
Sat Jul 2 21:11:28 UTC 2005

Brian Astill wrote:
> After another RSI eposode :-(  it seems to me that the ultimate anti-RSI tool 
> is voice recognition software.
> Anyone had experience with the more recent Linux implementations and can offer 
> some guidance?

as someone who has lived with speech recognition for the past 10 years, 
I can say the only way to get usable, large vocabulary continuous speech 
recognition is to move back to Windows.  The open source speech 
recognition initiative (ossri.org of which I am co-cto and one of the 
founders) is trying to build speech recognition support on Linux. 
Unfortunately, it is a multimillion dollar effort to build a native 
speech recognition and there isn't a large enough user population yet to 
support that kind of investment.

Therefore we're looking at other options such as NaturallySpeaking on 
wine and vm based solutions with speech recognition running on Windows 
and a bridge to the Linux environment.

Like most of open source projects, we are funded out of our own pockets 
and as a result, people can only put as much effort in as they can 
afford to spend if their own money.

personally, for a first generation I believe the virtual machine and 
bridge solution is the most viable.  It's doesn't get rid of Windows but 
it does let us partition the problem and solve the 
accessibility/interfacing problems first.


More information about the ubuntu-users mailing list