To add to Brian's email: 1) The libfreenect API is sufficient to replicate the OpenNI ROS driver. As Brian mentioned, OpenNI itself supports more functionality (see openni.org). I actually don't know how well this extended functionality was supported in ROS, since I have never touched it. 2) I am willing to take ownership/maintenance of a libfreenect based driver. The Kinect is the primary sensor on our robots, and I don't want to find an alternative at this stage. If you plan on using this libfreenect based driver on a platform other than Ubuntu, I will need help with testing/patches. About the audio: 1) libfreenect mentions the audio, but it is not well documented. 2) This won't be very high on my priority list. Patches enabling support will be more than welcome. In either case, I'll move ahead with the driver as I find time later this week and work on a rosrelease. Thanks! Piyush On Wed, Oct 17, 2012 at 11:02 AM, Patrick Goebel wrote: > I hope this isn't too far off topic, but if the driver gurus out there do > some work on either the openni or freenect versions, it would be awesome to > get some audio support for the Kinect's microphones (ditto for the Xtion if > possible). This thread on ros-kinect seems to suggest that audio support is > still in a state of flux. I can think of three aspects of driver support > (not having a clue how to write a driver myself) that would be nice to have > in ROS: > > * basic audio (used as a regular audio input source) > * noise canceling (not sure if that takes place in the firmware or driver) > * sound localization > > --patrick > > > _______________________________________________ > ros-users mailing list > ros-users@code.ros.org > https://code.ros.org/mailman/listinfo/ros-users >