none
SAPI 5 and SR and TTS RRS feed

  • Question

  • hi,when i use SAPI 5 in my SR and TTS program,Should users of my program train their voice?thank u :)
    • Moved by Helen Zhao Monday, May 14, 2012 6:14 AM (From:Visual C++ Language)
    Saturday, May 12, 2012 2:10 PM

Answers

  • SAPI developer support can be found at: http://msdn.microsoft.com/en-us/library/ee431825(v=VS.85).aspx

    This forum is for Kinect for Windows, so it's only relevant here if your audio capture device is a Kinect.

    Kinect isn't designed to work w/ SAPI (it uses the more recent Speech Server 11).

    For SAPI, you will get much better recognition if you have your users do training.  If that's not feasible in your situation, I'd recommend considering Speech Server 11 instead, which uses a speaker independent recognition technology and achieves similar results as SAPI when it's been trained.  (And I recommend picking up a Kinect as your input device unless your users will be wearing a headset. :))

    Tuesday, May 15, 2012 5:43 AM

All replies

  • hi,when i use SAPI 5 in my SR and TTS program,Should users of my program train their voice?thank u :)
    • Merged by Barry Wang Monday, May 14, 2012 8:18 AM
    Friday, May 11, 2012 1:17 PM
  • Expect horrible results if you do not train and have complex grammar.


    The following is signature, not part of post
    Please mark the post answered your question as the answer, and mark other helpful posts as helpful, so they will appear differently to other users who are visiting your thread for the same problem.
    Visual C++ MVP

    Saturday, May 12, 2012 4:53 PM
  • Hi Samar21,

    According to your description, I'd like to move this thread to this forum for better support: Audio API and Programming—Kinect for Windows SDK.

    Thanks for your understanding.
    Best regards,


    Helen Zhao [MSFT]
    MSDN Community Support | Feedback to us

    Monday, May 14, 2012 6:13 AM
  • SAPI developer support can be found at: http://msdn.microsoft.com/en-us/library/ee431825(v=VS.85).aspx

    This forum is for Kinect for Windows, so it's only relevant here if your audio capture device is a Kinect.

    Kinect isn't designed to work w/ SAPI (it uses the more recent Speech Server 11).

    For SAPI, you will get much better recognition if you have your users do training.  If that's not feasible in your situation, I'd recommend considering Speech Server 11 instead, which uses a speaker independent recognition technology and achieves similar results as SAPI when it's been trained.  (And I recommend picking up a Kinect as your input device unless your users will be wearing a headset. :))

    Tuesday, May 15, 2012 5:43 AM