[Ffmpeg-devel-irc] ffmpeg-devel.log.20160731

burek burek021 at gmail.com
Mon Aug 1 03:05:03 EEST 2016


[04:52:52 CEST] <cone-460> ffmpeg 03Timothy Gu 07master:5522da9ee102: doxygen: Remove outdated EXPAND_AS_DEFINED
[04:52:52 CEST] <cone-460> ffmpeg 03Timothy Gu 07master:cebad64d4f06: Doxyfile: Reset unused options back to default
[07:51:58 CEST] <KGB> [13FFV1] 15TimothyGu pushed 1 new commit to 06master: 02https://git.io/v6eqX
[07:51:58 CEST] <KGB> 13FFV1/06master 14da40880 15Reto Kromer: Fix dead link of HuffYUV (#23)
[11:11:41 CEST] <kierank> atomnuker: ^
[16:39:47 CEST] <cone-091> ffmpeg 03Burt P 07master:999c6b1da85a: af_hdcd: fix a minor cosmetic annoyance
[16:39:47 CEST] <cone-091> ffmpeg 03Burt P 07master:390eea61be59: af_hdcd: give cdt expired counter a value for never set
[16:54:54 CEST] <bms20> Hello, I'm trying to determine if a video is interlaced or not.  I'm looking at the field_order entry in AVCodecParserContext and AVCodecContext - both are AV_FIELD_UNKNOWN on two test videos I've tried.  Is there something I'm doing wrong, or is this information just not coded in the files?
[16:55:33 CEST] <nevcairiel> that information is typically not available on a per-file basis
[16:57:41 CEST] <bms20> Is the information available on a per-frame basis?
[16:57:47 CEST] <bms20> Or is it just not stored?
[16:57:52 CEST] <bms20> (in any file)?
[16:58:01 CEST] <nevcairiel> per-frame yes
[16:58:11 CEST] <nevcairiel> AVFrame has it
[16:59:12 CEST] <bms20> Ok - cool.  Am looking now.
[17:00:08 CEST] <bms20> ok - think I've got it.  interlaced_frame and top_field_first.
[17:00:11 CEST] <bms20> Thanks for that!
[17:38:13 CEST] <bp0> michaelni, Alright, what do I need to do to assume 'maintainer of hdcd'?
[17:39:44 CEST] <michaelni> bp0, send a patch that adds you to the MAINTAINERs file for hdcd, if noone objects and the patch is applied then you are maintainer of it
[17:41:30 CEST] <bp0> ok
[17:48:19 CEST] <bp0> done.
[18:34:31 CEST] <cone-091> ffmpeg 03Michael Bradshaw 07master:c9ae8be5a87c: avformat/matroskaenc: fix Voids with size < 10
[18:35:28 CEST] <Timothy_Gu> michaelni: can you test this patch: http://sprunge.us/BUMU? I can't reproduce that bug here but the GCC documentation does show a few errors in my original patch
[18:41:51 CEST] <Bitterblue> Hi. I see that the PGS subtitle decoder doesn't return an AVSubtitle for every packet fed to avcodec_decode_subtitle2, even though the decoder doesn't have CODEC_CAP_DELAY. What is going on there?
[18:42:17 CEST] <nevcairiel> those things are not necessarily related
[18:42:33 CEST] <nevcairiel> sometimes you just need to assemble more then one packet to be able to produce output
[18:43:24 CEST] <nevcairiel> delay just means it needs to be flushed at the end, which might not apply as partial data may not be decodeable at all stand-lone
[18:43:27 CEST] <nevcairiel> +a
[18:43:57 CEST] <Bitterblue> I see.
[18:47:00 CEST] <michaelni> Timothy_Gu, make: *** [fate-lossless-shorten] Error 134
[18:52:13 CEST] <michaelni> seems not reproduceable under gdb with debug symbols
[19:00:05 CEST] <michaelni> Timothy_Gu, backtrace: http://pastebin.com/vccsQqFw
[19:02:26 CEST] <Timothy_Gu> michaelni: thanks, I'll just drop this patch. I can't repro on gcc (Debian 4.9.2-10) 4.9.2 and it's not worth the effort
[19:19:22 CEST] <cone-091> ffmpeg 03Timothy Gu 07master:04da20e58fd7: dynarray: Change AV_ to FF_ for AV_DYNARRAY_ADD
[19:21:37 CEST] <omerjerk> hey is this function supposed to be executed when encoding from wav to mp4 als? - https://github.com/omerjerk/FFmpeg/blob/alsenc/libavformat/alsenc.c#L83
[19:21:49 CEST] <omerjerk> the log line isn't coming though. 
[19:22:03 CEST] <omerjerk> The encoder gives seg fault as well. :|
[19:22:28 CEST] <omerjerk> it completes the encoding though. I receive the last frame and then after that somewhere is segfaults.
[19:22:36 CEST] <omerjerk> Any tips on how can I debug this?
[20:06:55 CEST] <Bitterblue> Huh, every other subtitle I get from the PGS decoder has num_rects=0.
[20:20:11 CEST] <nevcairiel> its probably to indicate that the display should be cleared
[20:21:50 CEST] <Bitterblue> Yeah, the packet pts does indicate that.
[20:22:47 CEST] <Bitterblue> Which is weird, because AVSubtitle has those start_display_time and end_display_time fields.
[20:23:06 CEST] <Bitterblue> Set to useless values, in this case.
[20:36:09 CEST] <Bitterblue> But if several packets are needed to produce a subtitle, and they all have different pts, which of those should be used as start time?
[20:38:53 CEST] <michaelni> Does anyone want FFmpeg to participate in the next Outreachy ?
[20:39:52 CEST] <michaelni> if so theres need to create an ideas page: https://trac.ffmpeg.org/wiki/SponsoringPrograms/Outreachy/2016-12
[20:40:48 CEST] <michaelni> similar to previous ones: https://trac.ffmpeg.org/wiki/SponsoringPrograms/Outreachy/2016-05 or https://trac.ffmpeg.org/wiki/SponsoringPrograms/Outreachy/2015-05 or https://trac.ffmpeg.org/wiki/SponsoringPrograms/GSoC/2016
[20:49:17 CEST] <bp0> patchwork.ffmpeg.org seems very good
[20:56:26 CEST] <kierank> michaelni: can we actually come up with new takmsks instead of the same crap each year
[20:58:05 CEST] <michaelni> kierank, new tasks would be ideal
[20:58:30 CEST] <kierank> Yes but don't put the same old junk in
[20:58:50 CEST] <kierank> Because it's so boring
[21:00:34 CEST] <michaelni> we need mentor task pairs, if we have mentors that want to mentor new tasks i would favor that, if we have nothing but old ideas and mentors who are happy mentoring them then thats the only thing we can put on the page
[21:05:12 CEST] <nevcairiel> if we only have those tasks that have lived on that page for 5 years and noone ever wanted to do them, maybe we should just not join
[21:08:57 CEST] <michaelni> what would you suggest for fresh tasks ?
[21:14:28 CEST] <jamrial> BBB mentioned some weeks ago an idea he had for outreachy i think
[21:14:36 CEST] <jamrial> but in general most ideas people had were considered too complex for this kind of thing
[21:14:46 CEST] <BBB> I forgot what it was already
[21:15:02 CEST] <jamrial> because they included designing a new API and such
[21:15:04 CEST] <BBB> but I agree that we should get rid of old tasks after N years, where N is small
[21:15:20 CEST] <BBB> if a list of tasks just contains stuff nobody ever wanted to do, its likely nobody actually wants to do it
[21:15:25 CEST] <BBB> and therefore it just pushes people away
[21:15:31 CEST] <BBB> theyll look for other, more appealing, task lists
[21:17:22 CEST] <BBB> was what I wanted to do the vp9 loopfilter? I dont think it was
[21:17:28 CEST] <BBB> does anyone have a log of what I proposed?
[21:17:53 CEST] <jamrial> BBB: i'm trying to find it
[21:18:16 CEST] <jamrial> should be somewhere in the irc logs
[21:19:34 CEST] <BBB> I hope so :-p
[21:21:19 CEST] <jamrial> https://ffmpeg.org/pipermail/ffmpeg-devel-irc/2016-July/003682.html
[21:21:30 CEST] <jamrial> it was vp8/9 alpha support on the native decoders
[21:42:36 CEST] <iive> well, that's 1 and it is from this month. Do you have any other good ideas?
[00:00:00 CEST] --- Mon Aug  1 2016


More information about the Ffmpeg-devel-irc mailing list