[MPlayer-users] realmedia, num_of_packets and DATA chunks

Olivier Galibert galibert at pobox.com
Fri Feb 14 13:42:45 CET 2003


That format is interestingly crap.

In some cases you get:
Packet 17737 / 17739
00C8442F: packet v0 len=1015  id=1  pts=461035  rvd=0  flags=0
block: hdr=0x1, len=1143, offset=0, seqnum=2
Packet 17738 / 17739
00C84826: packet v0 len= 166  id=1  pts=461035  rvd=0  flags=0
block: hdr=0x81, len=1143, offset=147, seqnum=2
[chunks=1  subseq=2]
Packet 17739 / 17739
00C848CC: packet v0 len=1016  id=1  pts=461099  rvd=0  flags=0
block: hdr=0x1, len=1384, offset=0, seqnum=3
Packet 17740 / 17739
00C84CC4: packet v0 len= 406  id=1  pts=461099  rvd=0  flags=0
block: hdr=0x81, len=1384, offset=387, seqnum=3
[chunks=1  subseq=2]
Packet 17741 / 17739
00C84E5A: packet v0 len= 570  id=0  pts=461147  rvd=0  flags=0
Packet 17742 / 17739
[...]

Which shows you can't trust num_of_packets in the first place.  But
then you can have:

Packet 17686 / 17688
0063DEAC: packet v0 len= 603  id=15  pts=2543091  rvd=0  flags=2
unknown stream id (15)
Packet 17687 / 17688
0063E107: packet v0 len=  62  id=15  pts=2543091  rvd=0  flags=2
unknown stream id (15)
Packet 17688 / 17688
0063E145: PING, version=17473, len=5441

"PING" means stream_id=33 (I was going on a wrong hypothesis).

In fact there is:
0063e140: 80a8 9a0a 0044 4154 4100 211b 8a00 0000  .....DATA.!.....
0063e150: 0022 9800 84fc cf00 0100 f500 0000 0000  ."..............
0063e160: 0000 0003 4488 5771 7e97 c3a6 202c 8d9e  ....D.Wq~... ,..
0063e170: c0b0 f380 1c37 9127 2f9d 48b0 88c5 e62e  .....7.'/.H.....
0063e180: 396d f558 74d2 1eb8 095e 9d84 265c 75f2  9m.Xt....^..&\u.

i.e., a new DATA chunk.

Probably the best way is to try to check for a new tag when the packet
version number goes to hell, but the current code very much doesn't
like new chunks in the middle of the stream...

  OG.



More information about the MPlayer-users mailing list