Skip to content

Latest commit

 

History

History
26 lines (19 loc) · 790 Bytes

asr.md

File metadata and controls

26 lines (19 loc) · 790 Bytes

ASR - Automatic Speech Recognition

New updates Fall 2015

The other direction is TTS

  • https://github.com/stormqloud/kazoo-tgbbok/blob/master/tts.md

  • Option 1: OpenSource

  • If you are an academic or other dreamer, this might be for you.

  • Anyone that can demonstrate a working system please contact me.

  • I've plaowed many hours into this without anything working.

  • Works in theory.

  • Option 2: ispeech.org

    • Kazoo to isppech.org for ASR has been fixed.
    • Voxter announced at Kazoocon 2015
    • I've not tested but I'm sure Voxter fixed the API
      • voice recognition is hard
      • Just because the voice recording hits ispeech API doesn;t mean they can recogniez it!
  • Option 3:

    • There is no option 3

ASR costs per word recognized.

http://www.stormqloud.ca