While I was debugging some dvd still frame issues in xbmc i finally got around to generating a sample. I've uploaded dvd_still_frame.vob to upload.ffmpeg.org/incoming. I think adding a fate regression test for this sample could be a good idea to make sure the SEQ_END check in mpeg parser works as it's supposed to. Just checking timestamps and order of output packets should show if that ever breaks again. I've also added a single frame mpeg too which is the other way still frames can work on dvd's. In that regard it would more be a test of probing to make sure it is still detected as a mpeg file with a video stream. /Joakim
On Sat, Apr 07, 2012 at 12:56:28AM +0200, Joakim Plate wrote:
While I was debugging some dvd still frame issues in xbmc i finally got around to generating a sample. I've uploaded dvd_still_frame.vob to upload.ffmpeg.org/incoming.
I think adding a fate regression test for this sample could be a good idea to make sure the SEQ_END check in mpeg parser works as it's supposed to. Just checking timestamps and order of output packets should show if that ever breaks again.
I've also added a single frame mpeg too which is the other way still frames can work on dvd's. In that regard it would more be a test of probing to make sure it is still detected as a mpeg file with a video stream.
i can upload these 2 files to our fate server, do you want to write the fate tests ? and in which directory should they go ? [...] -- Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB DNS cache poisoning attacks, popular search engine, Google internet authority dont be evil, please
i can upload these 2 files to our fate server, do you want to write the fate tests ? and in which directory should they go ?
mpeg2 I tried to provoke the parser to not output the still frame. But it seems it has gotten much more robust in this regard. I could drop all the SEQ_END checks in the parser and it still worked. I can't really be sure if this sample is good enough to show the complete problem, but i don't think it would hurt. /Joakim
On Sun, Apr 08, 2012 at 12:35:24PM +0200, Joakim Plate wrote:
i can upload these 2 files to our fate server, do you want to write the fate tests ? and in which directory should they go ?
mpeg2
uploaded [...]
@@ -0,0 +1,955 @@ +0, 0, 1244160, 0x53653c5b +1, 0, 6144, 0x341ec6f6 +1, 2880, 6144, 0x95ed19dc +1, 5760, 6144, 0xc3b466b0 +1, 8640, 6144, 0xae69b2e7 +1, 11520, 6144, 0x6906167f +1, 14400, 6144, 0xcb25418b +1, 17280, 6144, 0x224bfa39
this doesnt look right, how old is your ffmpeg ? [...] -- Michael GnuPG fingerprint: 9FF2128B147EF6730BADF133611EC787040B0FAB DNS cache poisoning attacks, popular search engine, Google internet authority dont be evil, please
Ah, seems I had the old 0.10.2 release checked out. Will resend later.
participants (2)
-
Joakim Plate -
Michael Niedermayer