Menu

How to output error rate with an adapted acoustic model?

Help
Freya
2016-04-19
2016-04-20
  • Freya

    Freya - 2016-04-19

    I am creating model_parameters using the tutorial "adpating the default acoustic model". I can tested it using microphone but it is diffcult to calculate the error rate.
    In an4 example, there are an4_test.fileids, an4_test.transcription and corresponding audio files for testing. I just tyed command

    ./scripts_pl/make_feats.pl -ctl etc/an10_test.fileids 
    perl scripts_pl/decode/slave.pl
    

    Is it possible to output sentence error rate and error rate with a given acoustic model, such as the adapted acoustic model? BTW, where is the output recognized words when I run pocketsphinx_continuous command? Thanks in advance.

     
    • Nickolay V. Shmyrev

      As covered here

      http://cmusphinx.sourceforge.net/wiki/tutorialtuning

      you have to create a fileids and transcription files yourself.

       
      • Freya

        Freya - 2016-04-20

        Hi, Nickolay, thank you very much. It worked!

         

Log in to post a comment.

Want the latest updates on software, tech news, and AI?
Get latest updates about software, tech news, and AI from SourceForge directly in your inbox once a month.