AudioDemo.py not Recognizing

Aug 8, 2012 at 4:13 AM
Edited Aug 8, 2012 at 4:14 AM

Is anyone able to get the AudioDemo.py example to work?

I can't get it to recognize any words.

After a short delay, it prints "Recognizing..." but never matches any of the 4 words in the grammar no matter how many times I say them.

The Speech Basics-D2D in the Kinect for Windows Developer Toolkit v1.5.1 works fine, so I don't think there are any issues with the Kinect or Speech components.

The other pykinect nui demos seem to work fine to adjust the camera, track skeletons etc.

I'd really like to add some speech recognition. There's no documentation on the pykinect.audio API.

Any help or tips would be appreciated.


Config:
Windows 7
ActiveState Python 2.7, 32 bit
Kinect SDK v1.5






Sep 8, 2012 at 6:36 PM

can you please help me? I cannot seem to find where the demos for pykinect are installed.

Sep 10, 2012 at 11:57 PM
Edited Sep 11, 2012 at 5:38 PM

Hi Rich,

Check out the response from Dinov at this link. That's where I found the zip file containing all the samples

http://pytools.codeplex.com/discussions/352762

Let me know if you get the audiodemo.py working. :)

Coordinator
Sep 11, 2012 at 12:08 AM

On the audio I just pushed a fix today which should correct that (and enables a few new features from the 1.5 SDK - http://pytools.codeplex.com/SourceControl/changeset/ac89a5a13147), but it requires a rebuild of the PyKinect C++ support for it to work.  But that'll be in the next release.

You might be able to workaround the issue it by first creating a Runtime object with RuntimeOptions.UsesAudio and then creating a KinectAudioSource but I didn't actually try that to make sure.  But the reason it's not currently working is that the global NuiGetAudioSource was failing - presumably because the device wasn't initialized yet.  The fix was to use a per-instance version of NuiGetAudioSource so I'm not quite certain if this workaround will work or not.