none
Converting wav file to text RRS feed

  • Question

  • Good morning all,

    I developed an application which converts wav file to text using c#. Using SAPI TTS app tool i saved the wav file in microsoft voice itself. For accurate recognition only i saved it in microsoft voice. Though the result is not accurate. It is recognizing the words wrongly, such as meeting as needing and cute as dubed etc.

    Now what can i do for accurate result.

    Thanks in advance.

    Tuesday, February 7, 2012 4:30 AM

Answers

  • Hi,

    Kinect for Windows Acoustic Model does not officially support text to speech.

    We integrate with Speech Platform SDK V11, the "server" version, which does not have TTS.

    I assume the SAPI version you are referring to is the client version and not the speech platform V11 that we refer to in the documentation, correct?

    In any case, you can try the following:

    - Set the gain of the Kinect microphone to 3, like indicated in the readme file

    - Use the standard acoustic model rather than the Kinect one


    - hope this helps - Mauro.

    I don't think that's quite his problem. Basically, he's trying to recognize the TTS voice, and because the TTS voice has pitch and timing issues, it's not recognizing it correctly. If that's true psgviscom, I don't see a way to fix it. 

    Mauro, below is what I have for speech products installed. If I understand correctly, these are the server versions. I am using the Microsoft.Speech.Synthesis.SpeechSynthesizer class to do TTS in my Kinect app. Works perfectly.

    Tuesday, February 7, 2012 7:42 AM

All replies

  • Hi,

    Kinect for Windows Acoustic Model does not officially support text to speech.

    We integrate with Speech Platform SDK V11, the "server" version, which does not have TTS.

    I assume the SAPI version you are referring to is the client version and not the speech platform V11 that we refer to in the documentation, correct?

    In any case, you can try the following:

    - Set the gain of the Kinect microphone to 3, like indicated in the readme file

    - Use the standard acoustic model rather than the Kinect one


    - hope this helps - Mauro.

    Tuesday, February 7, 2012 7:05 AM
  • Hi,

    Kinect for Windows Acoustic Model does not officially support text to speech.

    We integrate with Speech Platform SDK V11, the "server" version, which does not have TTS.

    I assume the SAPI version you are referring to is the client version and not the speech platform V11 that we refer to in the documentation, correct?

    In any case, you can try the following:

    - Set the gain of the Kinect microphone to 3, like indicated in the readme file

    - Use the standard acoustic model rather than the Kinect one


    - hope this helps - Mauro.

    I don't think that's quite his problem. Basically, he's trying to recognize the TTS voice, and because the TTS voice has pitch and timing issues, it's not recognizing it correctly. If that's true psgviscom, I don't see a way to fix it. 

    Mauro, below is what I have for speech products installed. If I understand correctly, these are the server versions. I am using the Microsoft.Speech.Synthesis.SpeechSynthesizer class to do TTS in my Kinect app. Works perfectly.

    Tuesday, February 7, 2012 7:42 AM
  • I have an enterprise level medical app that runs on a Tabet PC. We use extensive handwriting, voice, handwriting and voice recognition. I want to add the abilities of the Kinect for handsfree app use. The recognition is a problem. MSP doesn't cut it, I need to use the desktop SAPI and recogniser. If I use the Kinect but the S.S desktop recogniser will I be able to get the motion detection to work for scrolling and mouse movement only?

    Wednesday, February 8, 2012 8:11 PM
  • ChrisCicc,

    Would you be willing to share with the rest of the class?  :)   I'd love to see TTS samples with the Kinect.

    Friday, March 9, 2012 11:27 PM