| Summary: | mplayer a52 CRC and resampling error patch | ||
|---|---|---|---|
| Product: | Gentoo Linux | Reporter: | Matthew Schultz <mattsch> |
| Component: | New packages | Assignee: | Gentoo Media-video project <media-video> |
| Status: | VERIFIED UPSTREAM | ||
| Severity: | normal | CC: | Reimar.Doeffinger |
| Priority: | High | ||
| Version: | unspecified | ||
| Hardware: | x86 | ||
| OS: | Linux | ||
| URL: | http://www.pchdtv.com/forum/viewtopic.php?t=1048&postdays=0&postorder=asc&start=0&sid=16b6f3083f0351e50d3f0dedc9d5e2e2 | ||
| Whiteboard: | |||
| Package list: | Runtime testing required: | --- | |
| Attachments: |
Patches demux_ts.c to get rid of the a52 sound errors
Patches demux_ts.c to get rid of the a52 CRC sound errors Patch for 20060415 New ebuild for 20060415 which includes the sound patch |
||
I'd also like to add that I card I am using is the Dvico Fusion Gold 5 card (ATSC) and us-cable (QAM_256). mplayer development is actually moving quite rapidly, they just haven't made any releases in awhile.
It would be much more convienent for us if you could provide a patch ("diff -u original revised" is the standard patch generation command).
Created attachment 79298 [details, diff]
Patches demux_ts.c to get rid of the a52 sound errors
It seems to me like they need to make a new release given the large amount of bugs needing to be fixed with the current version. The users are patching the software faster than the developers.
I just emerged 1.0.20060217 and it _still_ has the a52 CRC errors. I still suggest using this patch if you want to be able to watch any dvb streams. Discussion and hopefully better patch in http://thread.gmane.org/gmane.comp.video.mplayer.devel/32140 I can confirm that the patch posted at gmane.org works and it is a patch posted by the author of that section of the code in mplayer so someone needs to patch the mplayer in portage with this patch ASAP. Created attachment 80798 [details, diff]
Patches demux_ts.c to get rid of the a52 CRC sound errors
This patch was written by the author and it looks like it's been submitted to mplayer's cvs. So I urge gentoo developers to grab the latest cvs from mplayer again and update.
I just tested 1.0.20060302 and unfortunately it looks like the patch hasn't made it in yet but I suppose it should pretty soon? I do hope so I don't know what's taking upstream so long to patch this. In the mean time, why don't we patch the ebuild with this patch? Created attachment 87763 [details, diff]
Patch for 20060415
This patch can be used to patch the 20060415 ebuild.
Created attachment 87764 [details]
New ebuild for 20060415 which includes the sound patch
I have created a revision to 20060415 to patch this sound issue since upstream can't seem to apply their own patches in a timely manner.
This patch is no longer needed as upstream has _finally_ checked this patch into source. Everyone should upgrade to Mplayer 1.0_pre8 or higher immediately. mplayer is already available, sorry for the delay Well it doesn't look like this patch was applied but it seems like the problem was fixed somewhere else as I don't have any a52 CRC errors on any of the streams anymore. |
I'd like to suggest a patch to the source for mplayer (given the glacial development pace of mplayer) that fixes (I've verified that it works on my x86-based system compiled from source) these errors when playing many dvb streams in mplayer: a52: error at resampling a52: CRC check failed! These errors pile up to such an extent that mplayer will halt after a minute or less of playing a dvb stream. Without this proposed patch, mplayer is wholly unusable to play dvb streams. In libmpdemux/demux_ts.c, in function parse_pes2() look for this code: else if ((p[0] & 0xF0) == 0x80) { mp_msg(MSGT_DEMUX, MSGL_DBG2, "A52 WITH HEADER\n"); es->start = p+4; es->size = packet_len - 4; es->type = AUDIO_A52; es->payload_size -= packet_len; return 1; } Quote: I've never seen an AC-3 packet with a header before but I think the header bytes are already gone at this point in the code so it's starting four bytes past the packet data. So take out the plus and minus fours: else if ((p[0] & 0xF0) == 0x80) { mp_msg(MSGT_DEMUX, MSGL_DBG2, "A52 WITH HEADER\n"); es->start = p; es->size = packet_len; es->type = AUDIO_A52; es->payload_size -= packet_len; return 1; }