[Ffmpeg-devel-irc] ffmpeg.log.20130402
burek
burek021 at gmail.com
Wed Apr 3 02:05:01 CEST 2013
[00:02] <Chat6237> Hi
[00:12] <craigruks> saste: Thanks again guys, you were right, upgrading did the trick.
[00:30] <No_Miaus> hello i want add a metadata but it contains commas " " and i'm getting an error.. some idea?
[00:34] <No_Miaus> please.. xD
[01:23] <No_Miaus> thank you llogan, i solved it, was a batch mistake
[01:23] <No_Miaus> :P
[01:23] <No_Miaus> a problem with double quotes
[04:19] <cloud1250000> hello :)
[04:23] <cloud1250000> dead :/
[04:23] <cloud1250000> with so many people connected
[04:23] <cloud1250000> xD
[04:23] <klaxa> not really
[04:23] <klaxa> you should just state your problem
[04:23] <klaxa> otherwise nobody will help you
[04:24] <cloud1250000> not really a problem
[04:24] <cloud1250000> I was looking for maxim
[04:24] <nownot> going form wmv to mp4 using avconv?
[04:24] <klaxa> nownot: for support on avconv, this is the wrong channel, try #libav
[04:25] <cloud1250000> ahah well I may have a problem then.. does anyone know how I can download rtsp with ffmpeg? using segmented download
[04:26] <cloud1250000> anyway, this is not the reason why I'm here, I'm looking for some news about the at3+ decoder ahah
[04:28] <cloud1250000> since vlc can do it, I was looking for a faster way to download the rtsp file... and I've found about the segemented way... but I can't make it works
[04:29] <klaxa> i don't think anyone ever implemented that?
[04:29] <klaxa> an at3+ decoder i mean
[04:29] <klaxa> nope, no at3+ in ffmpeg
[04:30] <cloud1250000> http://ffmpeg.org/pipermail/ffmpeg-devel/2013-February/139776.html
[04:30] <cloud1250000> yeah I know ahah xD
[04:30] <cloud1250000> it should be there soon... but I don't know what is the meaning of soon for maxim :x
[04:32] <klaxa> hmm well there is one for atrac3
[04:58] <cloud1250000> yeah :/
[04:59] <cloud1250000> and for rtsp, do you have an idea? xD
[10:34] <xintron> What format would be good for streamed data? Should one go for flv or is there something more suited (h264 + aac codecs)
[10:55] <SenGlar> Hi people
[10:57] <SenGlar> Is it possible to mux audio to a TS stream with sizes higher than 4096bytes if I'm using a configuration of 48kHz, stereo and S16 per audio frame at 25fps (so I got per frame 7680bytes)?
[11:30] <pagios> what is the video encoding done by this command: -f video4linux2 -s 320x240 -r 30 -i /dev/video0 -metadata title="my awesome video" -f flv rtmp://1234 ?
[11:31] <pagios> flv is the contained what is the encoding h264?
[12:05] <Keshl> I have a video with a segment that features the characters cast in very dim light with bright surroundings. Using the x264 algorithm, the details on them are just atrocious. I get that it's the point, since it's unlikely a dim area will have detail, but in this case it does. Is there a way to tell x264 when such footage is being fed to it so it doesn't cut the quality, oÉo?
[12:25] <zap0> constant bit rates do that
[13:20] <pagios> can i include in ffserver sections a stream to post directly to icecast?
[13:57] <xintron> I have "Strict -2" set on my output stream for ffserver but when running `ffmpeg -i <source> <ffserver feed>` I get "The encoder 'aac' is experimental but experimental codecs are not enabled, add '-strict -2' if you want to use it." Do I need to add it to the command line as well?
[14:40] <xintron> So, I need to run `-strict -2` even though it's specified in ffserver-config but why do I get the following error? It's like it doesn't use the fast preset: http://pastie.org/private/6euxa2ilif8pzmuwikhwma
[16:00] <petar> my local ffmpeg installation looks for presets in ~/.avconv. it also looks for *.avpreset instead of *.ffpreset files. is that expected? I'm pretty sure I'm running ffmpeg and not avconv (or libav or whatever else the fork is called)..
[16:17] <durandal_1707> petar: look at output of ffmpeg, it will show answer you seek
[16:18] <petar> durandal_1707: I strace'd ffmpeg and all it ever tries is ~/.avconv/*.avpreset even though AVCONV_DATADIR is not set
[16:19] <petar> also ffmpeg_opt.c has two .avpreset but no .ffpreset references..
[16:21] <petar> it also tries share/ffmpeg/*.avpreset but again: avpreset, not ffpreset..
[16:22] <petar> the documentation only mentions ~/.avconv in conjunction with AVCONV_DATADIR
[16:22] <durandal_1707> petar: i did not asked for strace and other jaga-maga
[16:23] <durandal_1707> when you run ffmpeg, just ffmpeg, whay you see?
[16:23] <durandal_1707> *what
[16:23] <petar> lots of stuff. whats your point?
[16:24] <viric> I don't think "avconv" is related to any ffmpeg fork, btw.
[16:24] <viric> I only know of libav, for a popular fork.
[16:24] <petar> isn't that how the libav executable is called?
[16:24] <durandal_1707> petar: ffmpeg output says if it is from fork or not
[16:25] <viric> ah I didn't know
[16:25] <petar> http://pastebin.com/YAvmXCaP
[16:25] <petar> durandal_1707: ^
[16:27] <durandal_1707> petar: than it is just how current ffmpeg behaves
[16:27] <petar> viric: http://packages.ubuntu.com/quantal/i386/libav-tools/filelist
[16:28] <petar> second line
[16:31] <viric> petar: aaah
[16:31] <viric> maybe there have been some crossed checkins :)
[16:49] <petar> durandal_1707: this behavour seems incoherent. the documentation is pretty clear about preset files, their names and paths. There is no reference to avconv besides AVCONV_DATADIR. presets in $HOME are supposed to be in $HOME/.ffmpeg. like it has been the case for years. also in the preset subdir of the source tree there are only *.ffpreset presets..
[16:50] <durandal_1707> petar: yes, it looks to be bug
[16:51] <durandal_1707> but i think it is there for compatibility
[16:51] <durandal_1707> so if presets in .ffmpeg still works there is no issue
[16:51] <petar> it doesn't
[16:51] <durandal_1707> presets in .ffmpeg does not work?
[16:51] <petar> no
[16:52] <petar> like i said: i strace'd ffmpeg and there was no open() of any file in ~/.ffmpeg
[16:52] <durandal_1707> hmm that seems to be serious regression, report bug immediately
[16:55] <petar> durandal_1707: it would be nice if someone with 1.2 or HEAD could reproduce the bug fist..
[18:54] <frogs> gooday fokes
[18:54] <frogs> can anybody demonstrate a working command to encode a dvd with lpcm? i am trying with -codec:a pcm_s16be and i get a great deal of errors (and dvd-player malfunctions on playing it)
[18:55] <frogs> i can do this successfully by producing file with sox and then joining bare video with mplex, but i would like to have a ffmpeg-only solution
[20:04] <Zylum> what's a good web player for playing m3u8 files on any mobile OS?
[20:05] <Mavrik> <video> tag? :)
[20:06] <Zylum> I've tried jwplayer, but it doesn't seem to support m3u8
[20:50] <llogan> relaxed: any progress on the compile instructions? or should i stop bugging you about that?
[22:19] <sergem> Hello. I just tried to upload a sample to ftp://upload.ffmpeg.org/, but I don't see it there. So I wanted to ask whether it works but hides everything, or it just does not work?
[22:26] <kippi> hey
[22:26] <kippi> is there any reason this would not work on a live stream? ffmpeg -v verbose -i rtmp://livestream -vf select='gt(scene\,0.9)' -vsync 0 -an keyframes%03d.jpg
[22:27] <sacarasc> sergem: I believe it hides stuff.
[23:18] <Jordan_> how can I be sure that the output will have near same duration?
[23:18] <Jordan_> within 1 second
[23:18] <Jordan_> preferably exactly the same
[23:27] <Zeranoe> Can FFmpeg encode large files without issues? I have a 5GB file that is giving an error "moov atom not found". Could it be due to the file size or the location of the moov atom?
[23:35] <Jordan_> and how often is this the case, is it most the time they won't match some of the time?
[23:36] <Jordan_> what is causing them not to have the same duration?
[23:38] <llogan> Zeranoe: i've generally only seen that with broken/corrupted inputs
[23:38] <llogan> does non-ffmpeg based stuff decode it normally?
[23:39] <iive> Zeranoe: are you running under windows? at least one other person had similar problem under windows.
[23:40] <Zeranoe> iive: Yes I am, and it seems to possibly be a bug on Windows. It is reporting it for a few files, all of which should not be corrupt. Was there any solution for the other person?
[23:41] <iive> Zeranoe: he said he would investigate.. don't remember talking with him again.
[23:42] <iive> are you building it yourself or using somebody's else build?
[23:42] <iive> ffmpeg uses largefile support under unix, and mingw should be able to provide it too.
[23:43] <Zeranoe> iive: I compiled FFmpeg myself, plus the toolchain
[23:45] <iive> the configure checks for _FILE_OFFSET_BITS and _LARGE_FILE_SOURCE, can you try to find the result of these checks, in the configure regular output and inside the configure log file (config.log or configure.log)
[23:47] <llogan> iive: most windows users probably use Zeranoe's builds now
[23:47] <iive> is there a link to him on the download page?
[23:47] <llogan> yes
[23:49] <iive> llogan: if he had same problem, he might have the solution as patch somewhere. patch that it should be merged :)
[23:50] <Zeranoe> iive: I just heard about this issue moments ago
[23:51] <Zeranoe> Looks like both FILE_OFFSET_BITS and LARGEFILE_SOURCE are defined in the builds
[23:51] <iive> Zeranoe: to the previous person I recommended using mingw64. it supports 32 bit compilation.
[23:51] <iive> Zeranoe: defined as 0 or 1 ?
[23:51] <iive> config.h i guess.
[23:53] <Zeranoe> config.h doesn't turn up either, but I see "-D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE" a lot in config.log
[23:54] <iive> so it is safe to assume it have been added to the cflags...
[23:54] <iive> maybe config.mak :)
[23:55] <iive> if it is -D ... if they are detected and present, then it might be mingw/windows bug.
[23:55] <iive> you can try mingw64 as alternative. it supports compilation of 32bit target, too.
[23:56] <iive> also... try the windows build llogan was talking about and see if it have same issue as the one you built.
[23:56] <woop> howdy...
[23:57] <woop> so I'm on a recent git checkout of ffmpeg
[23:57] <woop> how would I go about figuring out which h.264 profile a video is encoded with?
[23:58] <relaxed> woop: look for "Stream #0:0(eng): Video: h264 (High)," in the ffmpeg -i output
[23:59] <Jordan_> input file is a .mts Duration: 00:12:46.79, start: 1.000033, i encode to mp4 Duration: 00:12:46.93, start: 0.000000
[00:00] --- Wed Apr 3 2013
More information about the Ffmpeg-devel-irc
mailing list