[Matroska-devel] Re: Status Quo of new features
paul at msn.com
Mon Dec 27 12:08:26 CET 2004
"Steve Lhomme" wrote...
>> - DV ( camcorder, lossy compression ) support, type 1 ( a +v = 1 stream )
>> and 2 ( a , a+v : 2 streams ); mainly some AVI parsing and defining how
>> to handle the embedded audio stream ( AVI has to repeat the PCM WAV audio
>> stream --> ENORMOUS overhead )
> As I've worked on MTX a bit, I think I could be able to add DV support
> too. Actually we would store the audio and video in separate tracks, not
> using the complex type, since the DV format is known we can demux it.
DV is strange because if you encode only video to DV, it still uses the same
space for the PCM, but it marks it as no audio. It would be possible to
store just the video frame data, but the splitter would need to recreate
Video + Null Audio stream on playback for anything to be able to use it. If
I recall correctly, it basically uses something like JPEG images for the
video, very similar to MJPEG video. I know we all have visions of using DV
video with Wavpack audio without having to store all of that silly PCM
"DV video information is carried in a nominal 25 megabit per second
(Mbits/sec) data stream. Once you add in audio, subcode (including
timecode), Insert and Track Information (ITI), and error correction, the
total data stream comes to about 29 Mbits/sec or 3.6 MBytes/sec."
That means you could shave off about 14% overhead for video only stuff by
storing only the video frames. Or about 9% for video + audio. Impressive,
but sounds like a good deal of work.
More information about the Matroska-devel