Our course project needs API to catch video stream from webcam (as data packets/datagram) and we are responsible to to manipulate/schedule the video data at backend ourselves. Can openTok satisfy us? After browsing the documentation, I didin't see any separated interface to capture video; all seem integrated.
The audio stream in the downloaded videos is in the SPEEX codec. Many desktop audio players don't recognize the codec. You may need to use ffmpeg to transcode the audio before it is playable in programs like VLC.