I have the following graph:
[audio capture]->[Frame grabber]->[Audio compressor]->[File writer]
The Frame Grabber just looks at the timestamps and store the timestamp
of the first frame so that streams can be synchronized later.
When I record audio and video from a web camera and use an MP3
compressor the timestamp received in the frame grabber can be used to
synchronize the streams.
However, when I replaced my MP3 compressor with the "WMAudio Encoder
DMO" the timestamps are off and can't be used for synchronization,
sometimes they're off by 250 ms, sometimes a bit more. Since the frame
grabber is first in the graph I don't think they should be affected if
the compressor is changed?
I have now tried with the Microsoft ADPCM encoder which works like the
MP3 compressor, this leads me to believe that's the WMA compressor
does something wrong...
I could use the ASF writer to compress the audio stream BUT I also
need to send the stream over network using a home built RTP sender
filter.