Voice Recognition and Text to Speech in Python • Hacker News Board

Re: Voice Recognition and Text to Speech in Python

#51

Post by turnip1979 » Sun, Feb 28, 2016, 11:57 AM UTC

luke-stanley wrote: post #48

afsina wrote: post #7

They should move from Sphinx to Kaldi and from GMM to DNN acoustic models. Instant 30% improvement.

http://kaldi.sourceforge.net/about.html

Does Kaldi need Windows? I only saw installation instructions for Windows. Also .. I just tried Pocket Sphinx ... says it works on Windows and Linux. So .. no non-apple or cross platform speech rec for us mac devs?

Re: Voice Recognition and Text to Speech in Python

#52

Post by afsina » Sun, Feb 28, 2016, 7:22 PM UTC

turnip1979 wrote: post #51

luke-stanley wrote: post #48

Earlier quoted context omitted.

http://kaldi.sourceforge.net/about.html

Does Kaldi need Windows? I only saw installation instructions for Windows. Also .. I just tried Pocket Sphinx ... says it works on Windows and Linux. So .. no non-apple or cross platform speech rec for us mac devs?

AFAIK, On the contrary. They officially support only linux. But community provides Windows support. I do not know Mac support. AFAIK again, Kaldi is mainly targeted server and desktop applications

Re: Voice Recognition and Text to Speech in Python

#53

Post by afsina » Sun, Feb 28, 2016, 7:23 PM UTC

luke-stanley wrote: post #48

afsina wrote: post #7

They should move from Sphinx to Kaldi and from GMM to DNN acoustic models. Instant 30% improvement.

http://kaldi.sourceforge.net/about.html

They use Github now.