Maxim Polijakowski <max_pole <at> gmx.de> writes: > In the current FFMpeg, PSMF will be > (mis)-identified as MPEG-TS stream.
I thought it is correctly identified as mpeg-ps. > Atrac Audio is carried out in PRIVATE_STREAM_1 > packets. Their format is proprietary but rather > simple to understand. > Atrac3+ audio streams are delivered inside of > packets of the same size, usually 2048 bytes. > They usually contain several audio frames. Frames > can be splitted across two packets and need to be > stitched together in the demuxer. That means you have to implement an atrac3p parser, see many examples in the libavcodec directory. > Now my questions: > 1) what's the proper code to be extended? I > attached a very draft patch that extends the MPEG > demuxer. Is it the right place or should I look > for any other? If the files can be identified as pmf, please add a new field similar to the existing imkh_cctv and sofdec fields and use that new field to decide that PRIVATE_STREAM_1 packets are atrac3p. (Additionally, allow users to force atrac3p for cut streams with ffmpeg -acodec atrac3p -i input.) Carl Eugen _______________________________________________ ffmpeg-devel mailing list ffmpeg-devel@ffmpeg.org http://ffmpeg.org/mailman/listinfo/ffmpeg-devel
