Timing info in the result

Home › Forums › OpenEars › Timing info in the result

Viewing 6 posts - 1 through 6 (of 6 total)
Advertisement: “Did you know OpenEars™ can use rules-based grammars to recognize fixed phrases? And RuleORama lets you use them with RapidEars!”

  • Author
    Posts
  • April 1, 2013 at 7:12 am #1016833
    OT
    Participant

    Hello,

    Do you have plans to include the timing info in the recognition results (e.g. start/end times at the word, or even phoneme level)

    April 1, 2013 at 9:33 am #1016837
    Halle Winkler
    Politepix

    You can receive more timing information by turning on OpenEarsLogging, but there is no plan to return that info as delegate methods.

    April 2, 2013 at 10:09 pm #1016853
    wenhuizhao
    Participant

    Pocketsphinx has api to get start/end frame, acoustic score and language score for each word: ps_seg_frames, ps_seg_prob.
    This two api return timing information as well as score for each word.
    Is there anyway to get this api work in OpenEars?

    April 2, 2013 at 10:13 pm #1016854
    Halle Winkler
    Politepix

    Certainly, you can make any modifications to the Pocketsphinx code you like and recompile the framework. OpenEars doesn’t interfere with that, it just doesn’t return this data along with hypotheses through its own API.

    April 2, 2013 at 11:00 pm #1016855
    wenhuizhao
    Participant

    Thanks for the quick response. Is there any instruction for recompile the framework from source code?

    April 3, 2013 at 8:27 am #1016857
    Halle Winkler
    Politepix

    You’re actually just opening up OpenEars.xcodeproj after making your changes and pressing “run”.

  • Author
    Posts
Viewing 6 posts - 1 through 6 (of 6 total)
  • You must be logged in to reply to this topic.