Skip to content
/ talk.js Public

JavaScript Speech Recognition Engine based on PocketSphinx 0.7

License

Notifications You must be signed in to change notification settings

mihai/talk.js

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

39 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

talk.js

talk.js is a direct port of PocketSphinx 0.7 lightweight speech recognition engine to JavaScript using Emscripten.

Demos are available at http://cirlanaru.com/srec-web, highlighting the JavaScript interface to capture the microphone audio stream for decoding, using WebRTC and WAMI-recorder (Flash-based plugin). Among the functionality highlighted in the demos the user can:

  • select model, recording sample rate, N-best hypotheses number (i.e. the top N most probable decoding results)
  • test PocketSphinx audio files ( numbers, goforward)
  • upload WAVE/RAW audio files for decoding
  • record own audio using WebRTC (HTML5) or WAMI (Flash)

The webRTC-based demo builds on top of Recorderjs (developed by Matt Diamond), which is modified to allow resampling of audio using resampler.js (developed by Grant Galitz).

About

JavaScript Speech Recognition Engine based on PocketSphinx 0.7

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages