[Ffmpeg-devel-irc] ffmpeg-devel.log.20130926

burek burek021 at gmail.com
Fri Sep 27 02:05:02 CEST 2013


[00:00] <kolosos> michaelni: upload of 720p mp4 version here : http://www.sendspace.com/file/iuejix
[00:00] <kolosos> downloaded from Youtube
[00:03] <kolosos> oh nice, sendspace wants to download a .exe
[00:15] <kolosos>  <JEEBsv>                   http://johnvansickle.com/ffmpeg
[00:16] <cone-549> ffmpeg.git 03Luca Barbato 07release/0.10:75b1b13aff73: vc1: check mb_height validity.
[00:16] <cone-549> ffmpeg.git 03Luca Barbato 07release/0.10:15620c153a35: ogg: Always alloc the private context in vorbis_header
[00:16] <cone-549> ffmpeg.git 03Luca Barbato 07release/0.10:896baaaad85c: segafilm: Error out on impossible packet size
[00:16] <cone-549> ffmpeg.git 03Luca Barbato 07release/0.10:91355bec88ce: h261: check the mtype index
[00:16] <cone-549> ffmpeg.git 03Luca Barbato 07release/0.10:86c169c5b691: dxa: Make sure the reference frame exists
[00:16] <cone-549> ffmpeg.git 03Luca Barbato 07release/0.10:3dff283de11a: ac3: Do not clash with normal AVERROR
[00:16] <cone-549> ffmpeg.git 03Luca Barbato 07release/0.10:37e69e2dee7c: ac3: Clean up the error paths
[00:16] <cone-549> ffmpeg.git 03Michael Niedermayer 07release/0.10:c6a4397410ab: Merge commit '37e69e2dee7c5167083bb42d669f73f038111a79' into release/0.10
[00:17] <kolosos> JEEBsv: i tried ./ffmpeg-2.0.1-64bit-static/ffmpeg -i jump720.mp4 -target ntsc-dvd jump720m.mpg            same buffer issue
[00:26] <cone-549> ffmpeg.git 03Luca Barbato 07release/0.10:c225c620c664: ac3: Return proper error codes
[00:26] <cone-549> ffmpeg.git 03Anton Khirnov 07release/0.10:5773065a7105: pictordec: break out of both decoding loops when y drops below 0
[00:26] <cone-549> ffmpeg.git 03Anton Khirnov 07release/0.10:8d2a86a29055: lavf: avoid integer overflow when estimating bitrate
[00:26] <cone-549> ffmpeg.git 03Anton Khirnov 07release/0.10:23f73fc241da: ape demuxer: check for EOF in potentially long loops
[00:26] <cone-549> ffmpeg.git 03Martin Storsjö 07release/0.10:068c86728668: matroskadec: Check that .lang was allocated and set before reading it
[00:26] <cone-549> ffmpeg.git 03Michael Niedermayer 07release/0.10:7a9af1da3900: mjpegb: Detect changing number of planes in interlaced video
[00:26] <cone-549> ffmpeg.git 03Martin Storsjö 07release/0.10:5473d23ece9e: mpegvideo: Avoid 32-bit wrapping of linesize multiplications
[00:26] <cone-549> ffmpeg.git 03Luca Barbato 07release/0.10:b0ca5fef09d1: dv: Add a guard to not overread the ppcm array
[00:26] <cone-549> ffmpeg.git 03Michael Niedermayer 07release/0.10:c8ca38567669: Merge commit 'b0ca5fef09d1b1268ea0c8f89bf53cd38aaa85e7' into release/0.10
[00:36] <cone-549> ffmpeg.git 03Anton Khirnov 07release/0.10:9978c24abfbb: lavf: fix the comparison in an overflow check
[00:36] <cone-549> ffmpeg.git 03Michael Niedermayer 07release/0.10:d8af960e7eba: Merge remote-tracking branch 'qatar/release/0.8' into release/0.10
[00:46] <kolosos> <kolosos> Here's a video that gives the buffer error: http://www.youtube.com/watch?v=yqGcZ6WBZnY  and here is same one (720p flv version) that i downloaded that gives buffer error: http://shell2.reverse.net/~kaht/singlejumps.flv
[03:14] <kolosos> Result of   uname -a  ; wget shell2.reverse.net/~kaht/singlejumps.flv  ; ffmpeg -i singlejumps.flv -target ntsc-dvd singlejumps.mpg   Check http://http://pastebin.com/mgfKmD7M
[03:15] <kolosos> ^^^^^^ buffer underflow result of ffmpeg on certain youtube videos such as  http://www.youtube.com/watch?v=yqGcZ6WBZnY
[03:21] <kolosos> GAH the right buffer underflow link is  http://pastebin.com/mgfKmD7M
[03:27] <Compn> ok
[03:27] <Compn> kolosos : what ver ffmpeg ?
[03:27] <Compn> 1.2.2 eh
[03:28] <cone-549> ffmpeg.git 03Michael Niedermayer 07master:ddd0e23d587e: avformat/mpegenc: improve field names in underflow error message
[03:28] <cone-549> ffmpeg.git 03Michael Niedermayer 07master:b63711808619: avformat/mpegenc: change assert to av_assert0
[03:29] <cone-549> ffmpeg.git 03Michael Niedermayer 07master:930406325c29: avformat/mpegenc: Better default for max_delay
[03:29] <cone-549> ffmpeg.git 03Michael Niedermayer 07master:32cde9629693: avformat/mpegenc: Better heuristic for ordering packets
[03:29] <michaelni> kolosos, should be fixed
[03:34] <kolosos> michaelni: fixed where?
[03:34] <kolosos> git?
[03:35] <kolosos> release/0.10:d8af960e7eba: ??
[03:36] <kolosos> sheesh, this could easily make ffmpeg the most responsive software updates i have ever seen
[03:42] <michaelni> kolosos, ffmpeg master:HEAD
[03:45] <kolosos> oh i see, there's a separate git on ffmpeg.org
[10:13] <durandal_1707> anybody working on xll dts extension?
[10:14] <nevcairiel> not that I am aware, its apparently quite annoying to support
[10:18] <durandal_1707> apparently?
[10:42] <cone-843> ffmpeg.git 03Kostya Shishkov 07master:4370f65be8d7: g2meet: Respect cursor_stride properly everywhere
[10:42] <cone-843> ffmpeg.git 03Michael Niedermayer 07master:1906e00030b3: Merge remote-tracking branch 'qatar/master'
[11:21] <nevcairiel> Daemon404: ever since i updated my fate box to your make, i have one fate run which *always* segfaults during building, at different spots though, but only this one, and everytime
[11:32] <Daemon404> nevcairiel, no idea
[11:32] <Daemon404> sounds like a job server problem
[11:33] <Daemon404> go go make
[11:33] <nevcairiel> looks exactly like the old crash that happened rather randomly, except that now it seems to happen more often :p
[11:33] <Daemon404> nice.
[11:33] <durandal11707> faulty hardware?
[11:33] <Daemon404> nah i used to get it randomly too
[11:34] <nevcairiel> i reverted my fate box to the old make, lets see if it just passes agian
[12:27] <motowidlo> hi 
[12:27] <motowidlo> anyone can help regarding use of ffmplayer on multimonitor environment ?
[12:32] <durandal11707> motowidlo: what is ffmplayer?
[12:32] <BBB> j-b: what did I do?
[12:32] <BBB> j-b: *hug* back btw :)
[12:33] <motowidlo> i mean ffplay
[12:37] <Daemon404> probably vp9
[12:37] <Daemon404> motowidlo, afaik nobody actually uses ffplay to do anything other than basic testing
[12:39] <motowidlo> Daemon404, looking at the source code, too, so I think. :)
[13:05] <Compn> motowidlo : mplayer supports multi monitor stuff with -xineramascreen option
[13:14] <motowidlo> Compn, MS/win world ?
[13:27] <Compn> for windows, mplayer can use the -vo directx:display or :adapter to select the screen
[13:27] <Compn> or maybe can just stretch it if you extend desktop
[13:27] <Compn> i forgot which option. you should ask in #mplayer or read the manual
[13:28] <motowidlo> ok, thx, i'll check it
[14:20] <cone-63> ffmpeg.git 03Michael Niedermayer 07master:cd05c4060a2f: avformat/mpegenc: minor simplification / use AV_WB16()
[14:21] <Daemon404> i thought direct use of AV_RB/WB was discouraged
[14:24] <ubitux> aligned/unaligned?
[14:30] <Daemon404> ubitux, safe bytereader
[14:30] <Daemon404> probably 'too slow' or something
[14:50] <iive> huh?
[15:04] <durandal11707> shouldn't AV_WB do same what replaced code does?
[15:07] <cone-63> ffmpeg.git 03Michael Niedermayer 07master:e02495372240: avformat/wavdec: check smv_frames_per_jpeg
[15:12] <nevcairiel> Daemon404: well muxing seems to be less prone to such issues, since you control the buffer and the values you write, not sure if there really is much that uses a "safe byte writer"
[15:12] <Daemon404> true
[15:13] <nevcairiel> AV_WB is at least in many many places
[15:13] <nevcairiel> +used
[15:14] <ubitux> any idea where the dvb subtitles are standardized?
[15:15] <Daemon404> ... i assue at dvb
[15:15] <Daemon404> :V
[15:15] <Daemon404> http://www.dvb.org/standards
[15:16] <ubitux> ah cool it's open great
[15:16] <ubitux> thx Daemon404 
[15:18] <ubitux> the subtitles paper doesn't seem to cover much the muxing though
[15:19] <Daemon404> ofc
[15:19] <nevcairiel> the dvb blue book there has details on the PES format in mpegts, if thats what you need
[15:20] <ubitux> i'd like to understand wtf is wrong with subtitles muxing into mpeg-ts in ffmpeg
[15:20] <ubitux> (#2989)
[15:20] <ubitux> it might have work at some point actually
[15:20] <nevcairiel> look at the bluebook then, it might have the info
[15:20] <ubitux> ok
[15:20] <ubitux> thanks
[15:20] <nevcairiel> syntax tables of the structure in PES etc
[15:22] <ubitux> there are a bunch of them
[15:50] <ubitux> erm afaict ffmpeg was never really able to mux those subs
[15:50] <ubitux> at least not with that kind of input
[15:50] <ubitux> i wonder in what situation this was supposed to work
[16:59] <ubitux> what is the old equivalent of ffmpeg -i in.ts -c copy -map 0:0 -map 0:5 out.ts, assuming 0:0 video and 0:5 dvbsub?
[17:00] <ubitux> -c copy ’ -scodec copy -vcodec copy, but what about the rest? 
[17:00] <Compn> you mean the map isnt working correctly ?
[17:02] <ubitux> yes
[17:02] <ubitux> i probably need some -new* or sth
[17:04] <Compn> what ffmpeg ver ?
[17:04] <ubitux> 0.7.x
[17:04] <ubitux> Number of stream maps must match number of output streams
[17:04] <ubitux> http://pastie.org/8357515
[17:04] <Compn> i think you want just the one -map 5:5
[17:05] <ubitux> huh?
[17:06] <Compn> you want one ts file output with video at 0 and dvsub at 5 ?
[17:06] <ubitux> yes
[17:06] <ubitux> i want to pick 0:0 and 0:5, stream copy them
[17:06] <Compn> i dont even remember if ts muxer had subtitle copy back then
[17:07] <ubitux> ow..
[17:07] <Compn> ehe
[17:17] <ubitux> michaelni: 04704a7881d94f8fc4c0e9abe609c98cb573776f
[17:18] <ubitux> michaelni: who am i supposed to contact, except you?
[17:18] <ubitux> mail from the frontpage?
[17:24] <michaelni> ubitux, i dont know :( 
[17:24] <michaelni> just checke my inbox
[17:25] <michaelni> but i cant find anything so i dunno who sent me that link
[17:35] <ubitux> Compn: it seems old ffmpeg has stream copy, so it should work
[17:35] <ubitux> but that's the problem here afaict
[17:35] <ubitux> i just have no idea how the cmd line is supposed to look like :(
[17:37] <durandal11707> remove dead links ASAP
[17:40] <ubitux> anyway, when i try -newsubtitle/-newvideo, i get a "At least one output file must be specified"
[17:40] <ubitux> as if it wasn't detecting the output anymore
[17:40] Action: ubitux wonder what the hell is going on
[17:42] <ubitux> http://pastie.org/8357610
[17:42] <ubitux> any idea?
[17:45] <Armada> Is there anyone available to help us solve this regression? http://trac.ffmpeg.org/ticket/2962
[17:47] <durandal11707> ubitux: why you use so old software?
[17:47] <ubitux> durandal11707: i want to check if the problem i have is a regression
[17:47] <ubitux> it sounds really weird that no one ever tested dvb muxing in mpegts
[17:47] <ubitux> so at some point it might have work
[17:48] <ubitux> and here i'm just trying to achieve a remuxing with older versions
[17:49] <durandal11707> if it ever worked it was by accident
[17:49] <ubitux> well there are some muxing code
[17:50] <ubitux> dvb seems to be only available in mpegts, so it's really weird if it never worked
[17:50] <durandal11707> no coverage - no fate
[17:50] <ubitux> they would have wrote a dvb sub encoder if thye weren't able to mux them
[17:53] <cone-63> ffmpeg.git 03Michael Niedermayer 07master:5dc6c0ea9ee7: avformat/mpegts: check avio_tell() return code before using it
[17:53] <durandal11707> anyone need this: https://github.com/richardpl/FFmpeg/tree/rgain ?
[17:54] <ubitux> ah, an alternative to ebu? :)
[17:55] <Armada> michaelni: thanks! that fixes our blocker
[18:05] <ubitux> im not even able to ffmpeg -i in.ts -vcodec copy -map 0.0 out.ts
[18:06] <ubitux> -newvideo doesn't help ("Number of stream maps must match number of output streams" ’ "At least one output file must be specified" if i add it)
[18:06] <ubitux> archeology is hard :(
[18:06] <durandal11707> you lose more time than fixing issue itself
[18:07] <ubitux> probably
[18:07] <ubitux> but now i'm curious about that specific problem more than the original issue
[18:11] <ubitux> oh, it seems -sn -an actually helps
[18:14] <ubitux> ok so this was what i was looking for: ./ffmpeg -y -i in.ts -vcodec copy -scodec copy -an -map 0.0 -map 0.5 -y out.ts
[18:14] <ubitux> and... 0.7 has the proble
[18:14] <ubitux> problem
[18:15] <durandal11707> who is admin on ffmpeg-user ?
[18:35] <durandal11707> so anyone needs replaygain filter or not?
[18:36] <wm4> wut
[18:36] <wm4> yes here
[18:36] <wm4> no idea how to test it though
[19:47] <rcombs> is there any way to access all input file streams from an avfilter?
[19:50] <rcombs> I've got a video filter here, and I need to access the attachments on the input file as soon as they're available
[21:04] <TheRyuu> nevcairiel: does LAVFilter just use the standard ffmpeg stuff to open the files for playback?
[21:04] <Daemon404> what do you mean
[21:05] <Compn> Daemon404 : he means lavf demuxers
[21:06] <Compn> TheRyuu : are you talking about libavfilter or some other 3rd party project which has nothing to do with ffmpeg development ?
[21:06] <Daemon404> lav
[21:07] <Daemon404> as n nevcairiel's thing
[21:07] <Compn> annoying another related name to confuse everyone
[21:07] <TheRyuu> https://github.com/Nevcairiel/LAVFilters <---this
[22:01] <cone-63> ffmpeg.git 03Michael Niedermayer 07master:f31011e9abfb: avcodec/parser: reset indexes on realloc failure
[00:00] --- Fri Sep 27 2013


More information about the Ffmpeg-devel-irc mailing list