Menu

Korean Speech Recognition

Anonymous
2001-11-08
2012-09-22
  • Anonymous

    Anonymous - 2001-11-08

    hello.
    I'm biginner.
    I want to make korean hmm model & language model.
    After downloading the SphinxTrain & sphinx2.0.3.
    How can I make model for speech recognition.

    Is SphinxTrain only execute in Linux OS?
    How can I use SphinxTrain in windows?
    How can I decode the speech waveform.

     
    • anselme dewavrin

      > How can I make model for speech recognition?

      There is a documentation describing the whole
      procedure. What you need is
      -audio files with korean people speaking
      -text version of what they say (=transcript)
      -choose a phonetic alphabet
      -for each word in the transcript, define phonetic
      translation yourself

      > Is SphinxTrain only execute in Linux OS?
      yes.

      > How can I use SphinxTrain in windows?
      recompile everything. I tried, but gave up...

      > How can I decode the speech waveform.
      Once you have the models for korean, use
      the sphinx2 decoder.

       

Log in to post a comment.