Ffmpeg-devel-irc

Download

ffmpeg-devel-irc@ffmpeg.org

December 2011

1 participants
67 discussions

ffmpeg.log.20111208
by burek 09 Dec '11

09 Dec '11

[00:11] <alyawn> so for map_channel to work, I need to specify -map_channel -1 for each channel I don't want, right? [00:20] <ubitux> alyawn: no no [00:20] <ubitux> -map_channel 0.0.0 -map_channel 0.0.1 [00:20] <ubitux> this will create a stream with 2 channels [00:20] <ubitux> if you add a few -map_channel -1, it will add muted channels [00:21] <alyawn> ah ok [00:21] <ubitux> as said in the doc, the number of -map_channel option you specify will be the number of output channels in the stream [00:21] <alyawn> ok... that makes more sense [00:21] <alyawn> thanks for the help [00:22] <ubitux> np :) [00:46] <freeminds> can I force ffmpeg to output the convert status in a different way? [00:48] <_klk_> hi all [00:48] <_klk_> i'm running into link errors when using statically bit libav* libraries from ffmpeg 0.8.5 [00:48] <_klk_> *statically built [00:49] <_klk_> here's the error: http://pastebin.com/wXE9A9wM [00:50] <_klk_> here's the configuration: configuration: --enable-static --disable-shared --enable-gpl --enable-libx264 --enable-nonfree --enable-postproc --enable-version3 --prefix=../.. [00:51] <_klk_> i am including -lz in my list of libraries along with the libav* [00:51] <_klk_> does anyone know why this is happening? [00:59] <osearth> _klk_: well, first your mom and dad met [01:00] <_klk_> osearth: what happened next??? [01:00] <osearth> i forget your mom got me drunk [01:01] <_klk_> :( did she take all your money [01:01] <osearth> no i was clear i needed cash up front [01:01] <osearth> ';] [01:01] <_klk_> oh lame i fixed this problem & just had to put the failing libraries AFTER the -lav* in my flags to gcc [01:02] <osearth> never seen that error [01:02] <osearth> nice i knew needed include [01:15] <walisser> hi [02:08] <StijnH> Hello. I see that FFmpeg is released as GPL and LGPL. If I make a program that uses FFmpeg, can the program itself be released under the MIT license, or does it need to have the same license? [02:17] <pasteeater> freeminds: what do you mean? [03:10] <freeminds> pasteeater, fixed already thx anyway [03:10] <freeminds> I meant: FFmpeg outputs the convert status by updating the last line [03:11] <freeminds> But for my application it would have been better if it would just print a new line [03:37] <jbwiv__> guys, I have a video (info here: http://pastie.org/2983851) from the way I read this output, the current bitrate is 147 kb/s. I'm trying to reduce the file size, so I've tried various things like "ffmpeg -b 24k -i UI_Overview.mp4 newvid.mp4", and I've varied the bitrate setting, but *anything* I do results in a video significantly larger. can anyone tell me what I should do to reduce it? [03:38] <sacarasc> You're doing it wrong! [03:38] <sacarasc> You put the bitrate as an input option, not an output option. [03:38] <sacarasc> ffmpeg [input options] -i input [output options] output [03:39] <jbwiv__> sacarasc: aha...that explains it. thanks ;-) [03:41] <jbwiv__> sacarasc: actually, even this creates a bigger file "ffmpeg -i UI_Overview.mp4 -b 16k UI.mp4" [03:42] <sacarasc> Pastebin the complete output. [03:42] <jbwiv__> in my example above, I typed it in wrong. I've really been putting it after the input [03:42] <jbwiv__> ok, one sec [03:44] <jbwiv__> sacarasc: I killed it when it got to 30M (the original file is 28M), but is this enough output? If not, I'll let it run all the way: http://pastie.org/2983876 [03:47] Action: sacarasc shrugs. [03:47] <sacarasc> Though you're using Libav, which is a fork and has a different channel. #libav I think. [03:48] <jbwiv__> sacarasc: should I be using something else? i just apt-get'd the ffmpeg package [03:49] <sacarasc> Ubuntu ships Libav rather than ffmpeg. There's a bunch of politics and I don't know which is best or even if there are many differences between them. [03:50] <jbwiv__> sacarasc: gotcha. ok, I'll go try libav. thanks much [03:51] <jbwiv__> guys, I have a video (info here: http://pastie.org/2983851) I'm trying to shrink it on ubuntu with ffmpeg, which appears to use libav. from the way I read this output, the current bitrate is 147 kb/s. I'm trying to reduce the file size, so I've tried various things like "ffmpeg -i UI_Overview.mp4 -b 16k newvid.mp4", and I've varied the bitrate setting, but *anything* I do results in a video significantly larger. can anyone tell me [03:51] <jbwiv__> what I should do to reduce it? [03:51] <jbwiv__> woops [03:52] <jbwiv__> sheesh, sorry, irc party foul [04:38] <jeremyS> Can someone give me a command line option for the current git master version ffmpeg which will use an audio filter from libavfilter? [04:40] <jeremyS> I have tried -af, -audiofilter and -filter:a [04:51] <relaxed> jeremyS: -vf [04:52] <jeremyS> for audio filters, or video filters? [04:52] <relaxed> I believe it's for all filters [04:52] <jeremyS> ohhhh. [04:52] <jeremyS> Mencoder uses -vf =video filters -af = audio filters, so I assumed vf would be only for video filters. [04:53] <jeremyS> hmmm... that doesn't work with the volume plugin [04:54] <jeremyS> It doesn't give an error, but it doesn't change the volume. [04:57] <jeremyS> Also, the documentation (doc/ffplay.html) says "A filtergraph can be represented using a textual representation, which is recognized by the -vf and -af options of the ff* tools, and by the avfilter_graph_parse() function defined in libavfilter/avfiltergraph.h. " [04:58] <jeremyS> but using -af volume=0.1 with ffmpeg give the error "Unrecognized option 'af' Failed to set value 'volume=0.1' for option 'af'" [05:07] <calc84> hello [05:08] <calc84> I can't figure this out, how can I force ffmpeg to not use the audio interleave preload duration in AVI output? [05:10] <relaxed> jeremyS: it's -vf [05:10] <calc84> actually, never mind, I don't even need to do that :/ [05:10] <calc84> I'm trying to figure out exactly what is causing some of my generated video files not to show up on my Nintendo 3DS [05:11] <calc84> and -vf isn't very specific at all anyway [05:13] <calc84> What bugs me is that I make 3 videos using the exact same options and only two show up on the 3DS [05:14] <calc84> and the one that didn't show up had an interleave preload duration, whereas the others did not [05:14] <relaxed> calc84: maybe it's the frame size or the par isn't 1:1 [05:14] <calc84> par? [05:14] <relaxed> pixel aspect ratio [05:14] <calc84> well, all 3 output videos were from the same source video [05:14] <relaxed> put ffmpeg -i $file on pastebin for all 3 [05:15] <calc84> I split it up into sections of 9:59 becasue that's the maximum size that shows up [05:17] <calc84> http://pastebin.com/GqmeL2fk [05:17] <calc84> VID_0001 is the one that doesn't work, VID_0002 and VID_0003 do work [05:18] <relaxed> does it not support h264/aac in mp4? [05:18] <calc84> nope, just mjpeg/adpcm [05:18] <calc84> because that's the output of the video recorder [05:19] <calc84> it wasn't meant for loading custom videos, but of course I'm going to do it anyway :P [05:21] <calc84> I don't think interleave preload duration actually has anything to do with it, because the videos I actually recorded on the 3DS have that as non-zero [05:21] <relaxed> can you stick up a sample that was recorded using the 3DS? [05:22] <burek> calc84, try setting the duration to 9:50 for the 1st video [05:22] <calc84> okay [05:27] <jeremyS> relaxed: so are you telling me that the volume filter is broken? [05:28] <calc84> the 9:50 version doesn't show up either [05:29] <burek> http://ffmpeg.org/ffmpeg.html#volume [05:29] <burek> calc84, I was thinking, maybe the file size is what makes your 3DS not to play the file, not the file duration [05:30] <calc84> I don't think so, I think I had one >100MB play before [05:31] <calc84> hmmm [05:31] <calc84> [avi @ 0000000001D3D280] parser not found for codec adpcm_ima_wav, packets or times may be invalid. [05:31] <burek> calc84, ffmpeg -codecs | grep adpcm_ima_wav [05:32] <calc84> I know I have it, it says DEA D [05:32] <burek> jeremyS [05:32] <burek> can you please use pastebin.com, to show your command line and its output? [05:33] <burek> calc84, ok :) then back to the lab :) [05:33] <calc84> that was from when I did the pastebin though [05:33] <calc84> I'll see if anything comes up when actually encoding [05:33] <relaxed> jeremyS: No. I don't get any errors when using it. [05:33] <burek> what cmd did you use to create those 3 files [05:34] <calc84> ffmpeg -i "tmp.avi" -async 1 -vcodec copy -acodec adpcm_ima_wav -ss 00:00:00 -t 00:09:50 "VID_0001.AVI" [05:34] <calc84> earlier I did this: [05:34] <calc84> ffmpeg -i %input% -vf scale=320:240 -aspect 4:3 -r 20 -vcodec mjpeg -qscale %quality% -acodec pcm_s16le -ar 44100 -ac 2 "tmp.avi" [05:34] <burek> I think you don't need -aspect along with -vf scale [05:34] <burek> it's redundant [05:35] <calc84> yeah, it was left over from somebody else's thing :P [05:36] <burek> calc84, can you show the cmd+output for the 1st file [05:36] <relaxed> I would use mkvmerge --split duration to cut the source into 9:50 pieces. Then have ffmpeg encode each one. [05:36] <burek> using pastebin of course [05:36] <jeremyS> burek: http://pastebin.com/cGi09z8B [05:37] <calc84> I'll show it for the whole batch if you want [05:37] <burek> jeremyS, instead of -vf put -af ? [05:37] <jeremyS> I already posted the output to that above in irc [05:38] <jeremyS> It gives an error instead of just silently not working [05:38] <burek> can you please use pastebin.com, to show your command line and its output? [05:39] <jeremyS> Okay [05:39] <jeremyS> http://pastebin.com/EbnTAGTi [05:40] <calc84> Here, I accidentally the whole thing: http://pastebin.com/FzcEJHxN [05:40] <calc84> First is tmp.avi, then are VID_0001 to VID_0003 [05:42] <calc84> it's also complaining about pcm_s16le, even though that is also DEA D [05:44] <burek> jeremyS, try af=volume=0.1:1 [05:45] <burek> wait sorry [05:45] <jeremyS> ? [05:46] <burek> calc84, Invalid and inefficient vfw-avi packed B frames detec [05:46] <burek> ted [05:46] <burek> you have an error in the input stream [05:46] <burek> try using -ss (seek) to skip a part of a video [05:46] <burek> ffmpeg -ss 00:00:05 -i ... [05:46] <calc84> it was a problem with the input video the entire time? >_< [05:46] <burek> i guess [05:46] <burek> jeremyS, just a sec [05:47] <calc84> yeah, I guess I should have tried some other videos too [05:48] <calc84> hmm, putting in another video gives the same message D: [05:48] <calc84> I'll try starting later in the video like you said [05:50] <burek> calc84, I think the error is at the very beginning of the file [05:50] <burek> that's why I suggested to use seek [05:50] <burek> to skip it [05:50] <burek> jeremyS, did you compile the ffmpeg from latest git? [05:50] <calc84> starting at 5 seconds gives the same thing [05:51] <jeremyS> yes [05:51] <jeremyS> I just git cloned a few seconds ago [05:51] <jeremyS> *minutes [05:51] <calc84> oh wait [05:52] <calc84> I was starting at 5 seconds in the wrong one [05:52] <burek> jeremyS, you might have just found a bug :) [05:52] <burek> can you try with 5 instead of 0.1 [05:53] <jeremyS> sure [05:53] <jeremyS> I think I already tried it with 10 with no effect [05:53] <jeremyS> but I'll check [05:53] <calc84> okay, even doing that in the right one doesn't work [05:54] <jeremyS> burk:yeah, no effect [05:54] <calc84> though, come to think of it [05:54] <burek> jeremyS, can you please create a bug report for that [05:54] <calc84> wouldn't it detect that frame when seeking, anyway? [05:54] <jeremyS> I was going to start development of LV2 plugin support for libavfilter...lol [05:55] <burek> calc84, i guess it would, but i was hoping it would also skip it and not encode it into the output [05:56] <calc84> well, that's possible too, I'll go through with the whole thing and try again [06:00] <calc84> ah, it shows up now (I used an offset of 10 seconds, but less probably would have been fine too) [06:03] <burek> :) [06:04] <jeremyS> burek: alright, bug reported: https://ffmpeg.org/trac/ffmpeg/ticket/720 [06:04] <burek> i think -ss 00:00:00.040 will also do :) [06:04] <burek> thanks a lot jeremyS :) [06:05] <burek> btw jeremyS, if you need to do get the job done, without waiting for the fix, you can use one of the older gits or use a stable release 0.8.5 or something [06:07] <burek> jeremyS, also, please add the pastebin output of using -af too [06:07] <burek> so that devels can see you tried the correct option [06:07] <jeremyS> burek: But that clearly says that the option is not supported [06:07] <burek> exactly, because it should be [06:07] <jeremyS> ooh [06:07] <burek> it's in the docs http://ffmpeg.org/ffmpeg.html#volume [06:08] <jeremyS> K [06:08] <jeremyS> What is a version which has working libavfilter audio plugins? [06:08] <relaxed> -filter:a doesn't work either [06:08] <jeremyS> https://ffmpeg.org/trac/ffmpeg/ticket/720 [06:08] <jeremyS> yeah. [06:09] <jeremyS> but it doesn't say it's not an option. [06:11] <calc84> yep, 00:00:00.040 does indeed do it :D [06:11] <burek> :beer: :) [06:11] <burek> hmh, jeremyS, you can try apt-get install ffmpeg maybe [06:11] <burek> just to get your job done [06:11] <calc84> I was wondering why nobody on neogaf seemed to be having the same problem as me [06:11] <burek> i mean, if you don't want to wait for the fix :) [06:12] <jeremyS> burek: I believe I get libav ffmpeg [06:12] <jeremyS> which has the plugins stripped [06:12] <jeremyS> it only has the anull audio plugin [06:13] <jeremyS> Well how long are we talking till it gets fixed? a few days, weeks, months? [06:14] <burek> I really can't tell, because frankly I don't know [06:14] <jeremyS> So *perhaps* the libav libavfilter is working, but I can't tell because the only plugin I can use is one which is supposed to do nothing to the audio. [06:14] <relaxed> jeremyS: libav has '-vol volume' [06:14] <jeremyS> Do you know a past version that should work? [06:14] <burek> hmh [06:14] <relaxed> or you could use sox [06:15] <jeremyS> relaxed: I'm trying to get a reference environment for developing a libavfilter plugin [06:15] <jeremyS> I don't actually care about changing the volume of a file [06:15] <relaxed> then what do you care about? [06:16] <jeremyS> Having an audio filter which works, so I can compare my new one to it. [06:16] <jeremyS> So I know any issues are on my end, and not a problem with ffmpeg [06:17] <jeremyS> If I could find any filter, that would be fine, there's nothing special about the volume one, except I assumed it would be the simplest, so it would be most likely to work [06:19] <burek> jeremyS, the volume filter has been added on 2011-11-05 [06:19] <burek> http://git.videolan.org/?p=ffmpeg.git&a=search&h=HEAD&st=commit&s=volume [06:20] <burek> and I don't see any updates for it, so I guess the bug exists since the beginning :/ [06:20] <jeremyS> bruek: lol... I assumed it would be the first one written. [06:20] <burek> :) [06:21] <jeremyS> So which one would you suggest I try? [06:21] <burek> http://ffmpeg.org/ffmpeg.html#toc-Audio-Filters [06:21] <burek> try the others [06:21] <burek> and see if they can help [06:21] <jeremyS> I tried earway [06:21] <jeremyS> *earwax [06:21] <jeremyS> didn't wokr [06:21] <jeremyS> *work [06:21] <calc84> bye, thanks for the help guys! [06:21] <jeremyS> and pan didn't work [06:21] <burek> calc84, :beer: o/ [06:21] <burek> none of them work? [06:21] <jeremyS> I haven't tried them *all* but the ones I listed, yeah [06:22] <jeremyS> ashowinfo did nothing [06:22] <relaxed> aresample does nothing [06:22] <jeremyS> hmmm... so I guess none of them are working? [06:24] <jeremyS> yeah, and aconvert does nothing either [06:24] <jeremyS> hmmm... scratch that! anull works! [06:24] <jeremyS> It does exactly what it's supposed to! [06:25] <burek> :)))) [06:25] <burek> jeremyS, how did you specify params for ashowinfo [06:26] <jeremyS> I didn't give any params, the documentation didn't list any [06:27] <burek> it might be that the cmd line parser is broken for -af [06:27] <burek> i just tried all those filters [06:27] <burek> all give the same error about unrecognized params [06:27] <burek> *option [06:27] <jeremyS> I would believe that [06:31] <relaxed> doesn't -filter:a = -af [06:31] <relaxed> they should get rid of -vf/-af and stick with -filter [06:32] <Skaag> how do I create mp41 instead of mp42 when I transcode to mp4? [06:33] <burek> relaxed, jeremyS this actually works: ffmpeg -i 2011-10-31.mp4 -filter:a ashowinfo -f mp4 /dev/null [06:34] <burek> or it doesn't :))) [06:34] <burek> i mean, the command works, but i dont see any output of ashowinfo :) [06:35] <jeremyS> oh yeah [06:35] <jeremyS> lol [06:35] <burek> Skaag, did you try -vcodec mpeg1video [06:35] <Skaag> hm. no [06:36] <burek> Skaag, try: ffmpeg -codecs | grep mpeg [06:37] <Skaag> right now I use -vcodec libx264 [06:37] <Skaag> I'm targeting old iPhones [06:38] <burek> i think there is a preset for iphones [06:38] <Skaag> for some reason it doesn't work for me& :-\ [06:38] <Skaag> this is what I use right now: http://pastebin.com/8i45L64d [06:41] <burek> you should really use -preset and -tune instead all those x264 specific options [06:41] <Skaag> I'm trying to find a guide on this [06:42] <burek> also, try -f ipod [06:42] <Skaag> I tried to google 'ffmpeg iPhone preset' but it gives me all kinds of weird pages [06:44] <Skaag> that gives me: [ipod @ 0x33f0ce0] Warning, extension is not .m4a nor .m4v Quicktime/Ipod might not play the file [06:44] <Skaag> I set the extension as .mp4 [06:44] <burek> hmh [06:46] <Skaag> I know that page, it does not use presets like you say [06:46] <Skaag> I read it earlier today [06:49] <burek> well [06:49] <burek> type x264 --help [06:49] <burek> and read about tune and preset [06:49] <burek> and then add it to ffmpeg like: ffmpeg -i ... -preset blah -tune blah -f mp4 ... [06:54] <jeremyS> burek: do you know a version+plugin combination that works? [06:55] <burek> plugin? [07:20] <jeremyS> burek: filter [07:24] <burek> jeremyS, no I don't, but you can try to search on git when was each filter added [07:25] <jeremyS> k, thanks [07:27] <Skaag> I think the source I'm given is broken to begin with [07:28] <Skaag> burek: with the -vpre iphone preset, the iPhone gives me a "This movie could not be played" error... [07:33] <burek> why -vpre [07:33] <burek> use -preset [07:33] <burek> vpre is deprecated [07:44] <Skaag> oh I see [07:45] <Skaag> it's what I find in the ffmpeg documentation& :-\ [08:22] <GordonFreeman> hi [08:53] <Skaag_> I'm trying to achieve h264 (Constrained Baseline) but no matter what params I pass to ffmpeg it remains on (high) [10:14] <bove> Do I need any extra arguments for multi threading a ProRes encode in 0.8.7? [10:16] <kollapse> Hi. I have 2 files here - a reference file called X and another file called Y. How can I convert Y to have exactly the same characteristics as X (same resolution, audio/video codec, fps etc.) ? [10:27] <impulze> convert but texactly the same characteristics? [10:27] <impulze> *but exactly [10:27] <impulze> i can't follow [10:27] <impulze> cp? [10:35] <kollapse> impulze: Well, I want to transcode file Y to have the same resolution, codec, fps etc. of file X [10:36] <kollapse> as file X* [10:38] <kollapse> Look here: http://pastie.org/2985072 I have two files, as can be seen: MVI_5999.AVI and video.mp4 - I want to transcode video.mp4 so that it will be the same type of file as MVI_5999.AVI (Video: mjpeg, yuvj422p, 640x480, 20 tbr, 20 tbn, 20 tbc, Audio: ... etc.) [10:43] <kollapse> Any ideas ? [10:53] <bove> a "get settings from existing file" would be nice, but I don't think it exists today [10:58] <kollapse> bove: Hmm, so I have to manually match the parameters ? [10:59] <kollapse> The reason I need this is that I want to upload a video file to a camera for playback. But understandably, the camera only supports a very specific type of file. [10:59] <kollapse> In this case, mjpeg / yuvj422p [11:04] <bove> upload the result of 'ffmpeg -i filefromcamera' to pastebin.com and I'll try to help matching it [11:05] <kollapse> bove: Well it's here, the first file (MVI_5999.AVI) - http://pastie.org/2985072 [11:37] <sybariten> bove: thats exactly a feature i've been thinking of, too! Surely this is a thing that must exist, somewhere in the vast video encoding business? [11:38] <Peace-> hi there , i would like undestand this error ffmpeg -s $(xrandr | awk '/, current /{print $8}')x$(xrandr | awk '/, current /{gsub(/\,/,"");print $10}') -f x11grab -r 10 -i :0.0 -vcodec gif output.gif [11:39] <Peace-> Could not write header for output file #0 (incorrect codec parameters ?) [11:39] <Peace-> but i have this [11:39] <Peace-> DEV D gif GIF (Graphics Interchange Format) [11:52] <Peace-> ok now it works [11:52] <Peace-> but black gif :D [11:53] <Peace-> ffmpeg -s $(xrandr | awk '/, current /{print $8}')x$(xrandr | awk '/, current /{gsub(/\,/,"");print $10}') -f x11grab -r 10 -i :0.0 -pix_fmt rgb24 -y output.gif [11:56] <bove> sybariten: We could get pretty far by just reading the Input and stream lines from the existing file [11:57] <bove> but a lot of files would also need very specific muxer options to be compatible [12:09] <Shimmy> Anyone knows why this project: http://code.google.com/p/amv-codec-tools isn't combined with the original FFmpeg development? [12:09] <Shimmy> It allows for FFmpeg to support AMV files [12:10] <Shimmy> the problem is I don't know how to compile it with the up-to-date FFmpeg version [12:10] <Shimmy> You can see my question here as well: http://stackoverflow.com/questions/8429733/a-windows-compatible-c-ide-that-… [12:16] <Peace-> hahahahahaha [12:16] <Peace-> it works! [12:16] <Peace-> :D ty all anyway :P [12:19] <kollapse__> bove: Any luck ? :) [12:53] <shevy> Hi. [12:53] <shevy> One can do a screencast with FFmpeg like so: [12:53] <shevy> ffmpeg -f x11grab -y -r 12 -s 800x600 -i :0.0+480,200 -vcodec ffv1 -sameq out.avi [12:53] <shevy> Is there a way to do "stop after 30 seconds" ? [12:53] <shevy> except hitting Ctrl-C. Reason is, I'd like to use this in a script. [12:58] <Mavrik> shevy, -t 00:30:00 [12:58] <shevy> oh [12:59] <shevy> hmm that does not seem to want to work [13:00] <Compn> shevy : you must be putting that command in the wrong order then [13:00] <Compn> also its -t 30 [13:00] <Compn> for 30 seconds [13:01] <Compn> -t 30:00 is 30 minutes i think [13:01] Action: Compn wonders what Mavrik is doing [13:01] <shevy> aaaaaah [13:01] <shevy> yes [13:01] <shevy> that was it :) [13:01] <shevy> it works now, thanks Mavrik and Compn [13:01] <Compn> np [13:03] <Mavrik> Compn, errr, yea [13:03] <Mavrik> my mistake :) [13:32] <Diogo> hi, this is not the right channel buy anyone work with erlyvideo? [13:33] <Diogo> *but [13:39] <raven> hi [13:39] <raven> maxrate obviously does not work - any suggestions? [14:23] <Bove> kollapse: What have you tried in making your file? [14:32] <kollapse> Bove: Does this look correct- ffmpeg -i video.mp4 -vcodec mjpeg -s vga -r 20 -acodec pcm_u8 -ac 1 -ar 11024 -ab 88 OUT.AVI Reference file is here - http://pastie.org/2985876 [14:33] <Bove> you probably need some bitrates in there [14:33] <Bove> ffmpeg -i Y -vcodec mjpeg -pix_fmt yuvj422p -s 640x480 -r 20 -b 16000k -acodec pcm_u8 -ar 11024 -ac 1 -ab 88000 [14:36] <kollapse> Ran ffmpeg -i on the resulting file. Audio is a match, but it shows me bitrate: 5727 kb/s [14:37] <kollapse> Even if -b 16000k [14:37] <Bove> you can try to add -minrate 16000k -maxrate 16000k [14:40] <kollapse> A constant bitrate of 16000k would probably increase file size, right ? [14:41] <Ave> unless you are bandwidth limited/constrained, its better to just use constant quality setting [14:42] <kollapse> Ave: I am limited by what the camera supports. [14:42] <Bove> I would try with the file you made from my line [14:46] <kollapse> Strange... camera says "Unrecognized format". The two files are pretty similar though - http://pastie.org/2985920 [14:48] <Bove> maybe you need a header? Or does the camera make thumbnail files for all videos? (Can't remember the extension) [14:49] <kollapse> The camera does make a thumbnail file for videos. I used a thumbnail from another video. (it uses .THM - but are actually jpeg) [14:50] <kollapse> Maybe it checks the metadata to see if the encoder is "CanonMVI06" ? And if not, it refuses to play ? [14:50] <sybariten> Bove: well... yes, i think theres still quite a bit of discrepancy between what you read in the ffmpeg output, and what you would supply as encoding instructions [14:51] <Bove> sybariten: Buffer sizes, rate limits etc. would probably be hard to read from an encoded file as well [14:52] <kollapse> Is there any way to force modify the metadata ? [14:52] <sybariten> Bove: yeah... but still [14:52] <sybariten> A lot of questions on mailinglists etc maybe, could be avoided :) [14:53] <Bove> kollapse: Check out http://www.videotoolshed.com/product/42/qtchange [14:54] <Bove> sybariten: It would probably be a lot easier for the devels [14:54] <kollapse> Well that costs (1) and there's no Linux version (2) [14:55] <Bove> kollapse: I was thinking more for the research phase. There is a demo version, but if you're linux only... [14:56] <kollapse> Well I don't know why it won't run this video. It should have been fooled. [14:57] <Bove> have you tried without audio? -an [14:59] <Bove> kollapse: You might also try qtstreamize (don't know if that's built into the regular ffmpeg branch yet) [15:00] <kollapse> Bove: That makes the video streamable ? [15:04] <Bove> puts the header in front [15:06] <kollapse> I tried using -metadata encoder="CanonMVI06" but it still sets Lavf53.5.0 [15:06] <kollapse> And -metadata creation_time="..." doesn't set. [15:08] <Bove> I might need an excact size for the header. What if you try and compare 3 files for the camera and see if you can find the header size [15:09] <kollapse> Bove: Oh, so using -metadata won't work? [15:10] <Bove> It could be the camera expects n bytes of header before the actual stream begins [15:10] <kollapse> How can I determine the header size ? [15:13] <Bove> If the bitrate is cbr, you can compare duration vs. file size on 3 files [15:16] <kollapse> I don't think it's cbr. This 6 second video has 12.5 mb. Calculation duration (seconds) * bitrate gives about 104 mb [16:16] <relaxed> kollapse: run mediainfo on both files [16:21] <OneFix_Work> Is there any hardware (besides a bigger CPU) that can be bought to speed up x264 encoding? [16:21] <Compn> ask in #x264 [16:22] <Compn> there are somethings, also there are hardware h264 encoders [16:22] <Compn> but better ask them for input [16:23] <OneFix_Work> Well, I know there is the acceleration that uses 3D cards, but I thought that was mainly for playback and not encoding... [16:24] <Compn> there is some work being done to speed up encoding, not sure how far it has gotten [16:24] <Compn> like i said, ask in the #x264 channel, they will know more [16:25] <Mavrik> OneFix_Work, faster processor with more cores [16:26] <OneFix_Work> Mavrik: Right, but that means at least 4 cores and faster than 2.5GHz which is usually really expensive [16:27] <OneFix_Work> At this point, I would rather spend $200 to buy something that can offload the encoding process rather than spend $300 (minimal) for yet another CPU upgrade [16:28] <Compn> there is a broadcom crystal hd h264 pcie card, but thats decoder only [16:28] <Compn> i cant remember any encoders [16:28] Action: Compn goes afk [16:28] <Mavrik> OneFix_Work, well, 3.3GHz 2500 quad is about 160¬ last I checked [16:28] <Mavrik> that'll help more than anything else [16:28] <Mavrik> since x264 doesn't support any of other hardware accelerators [16:29] <OneFix_Work> Mavrik: This is my home Linux box...it does ZoneMinder, Subsonic, and Plex [16:29] <Mavrik> yes? [16:29] <OneFix_Work> Mavrik: And occasionally I will rip a DVD or edit photos in DarkTable... [16:31] <OneFix_Work> I already have a fast FreeNAS box (quad core, 8GB RAM) for storage [16:33] <Mavrik> fine, but how is that connected to the x264 problem? [16:33] <OneFix_Work> Mavrik: Just saying that I've built a quad core system recently [16:37] <OneFix_Work> Mavrik: Actually, for half the price of the 2500, you can get an AMD Quad-Core Phenom II X4 @3.6GHz [16:42] <OneFix_Work> Mavrik: And I'm half thinking just trying to do transcoding using my FreeNAS box. [16:44] <OneFix_Work> Another question I had, has anyone done a comparison of the libx264 presets? [17:01] <OneFix_Work> I've seen the page that compares lossless, veryslow and veryfast ... but I haven't found anything that includes the newer ultra fast presetd [17:10] <raven> hi [17:10] <raven> ffmpeg/winff preset: how to limit mpeg4 super high quality preset to maximum 4mbit? [17:14] <OneFix_Work> raven: -br 4000k [17:16] <OneFix_Work> raven: Depneding on your version it might be just -b 4000k and you might need to add -bt 4000k [17:17] <raven> OneFix_Work, i do not understand, why maxrate 3000k does not take effect - any suggestions? [17:18] <OneFix_Work> raven: The -b or -br option is the average and -bt tells it how much it can change. [17:19] <raven> OneFix_Work, not here - when i put anything into b it still runs above to double or more [17:23] <calc84> hey, is it possible to change the sample rate in ffmpeg without resampling? I want to speed up the audio to double speed [17:23] <calc84> I already got the video to double speed [17:24] <gst-kaps> has anyone tried compiling ffmpeg.c on windows ? [17:25] <calc84> I was thinking that using -async would make the sound go as fast as the video, but apparently not [17:25] <raven> OneFix_Work, this is the winff preset for mpeg4 super high -crf 15.0 -vcodec libx264 -preset veryslow -acodec libfaac -ar 48000 -ab 192k -coder 1 -flags +loop -cmp +chroma -partitions +parti4x4+partp8x8+partb8x8 -me_method hex -subq 6 -me_range 16 -g 250 -keyint_min 25 -sc_threshold 40 -i_qfactor 0.71 -b_strategy 1 -threads 0 [17:27] <vcs> hi, I have a file with 4 video streams and 2 audio streams, is it possible using ffmpeg to split the files up into a file for each stream? [17:27] <vcs> or using some other utlity that uses ffmpeg [17:29] <alyawn> so the q=<num> on the output of ffmpeg, is that an average, or is it specific to the current frame [17:38] <qxt> Was wondering if there was a way to import video uncompressed from a video camera that uses DAT. [17:40] <qxt> once edited Ill toss it into a mkv container and compress it with x264 using ffmpeg like usual. [17:48] <kollapse> relaxed: Hello. You told me to run mediainfo on 2 files. [17:48] <relaxed> yes, to see if it reveals any differences you may have missed [17:50] <kollapse> relaxed: I see a difference. The camera file has "Writing application: CanonMVI06" and "Mastered date: Sun Aug 7 09:50:18 2011" while the ffmpeg file only has "Writing application: Lavf53.5.0" [17:52] <teratorn> how would I use ffmpeg to copy the first frame of a video in to a new file? [17:52] <teratorn> I just want a new video file with a single frame in it [17:52] <teratorn> ah, -vframes [17:54] <relaxed> kollapse: you could hexedit the video and replace Lavf53.5.0 with CanonMVI06. [17:55] <calc84> so can I make audio double speed? (pitch increase and all) [17:56] <kollapse> relaxed: Hmm, here's what I get - http://pastie.org/2986785 (could anything else in there be the cause?) [17:59] <relaxed> kollapse: -b 16M [18:01] <kollapse> relaxed: I've set that but ffmpeg still gives out a file with 5M bitrate [18:01] <kollapse> 5626 kbps [18:07] <qxt> How would I capture my dv1394 (firewire) video as raw video. I can do this in dvgrab but would like to use ffmpeg. This part I know "ffmpeg -f dv -i /dev/dv1394/0 " [18:09] <qxt> or just uncompressed is a better terminology. [18:12] <Mista_D> Why encoding to 4 output files (1st of 2 passes) utilizes 2.5 CPU cores? No matter what the number of outputs, `-threads` or `-thread_type` are, the utilization is ~2.5 cores of 16 available. http://pastebin.com/m5ctKziU [18:12] <calc84> I guess I'll be forced to use mencoder to speed up my videos :/ [18:15] <relaxed> Mista_D: before the input determines how many threads used for decoding. [18:17] <cryptopsy> i'm having a hard time converting 400+ kb/s flac to 320kb/s mp3 [18:17] <cryptopsy> ffmpeg -i 111.flac -ab 320k -f mp3 test.mp3 produces 160kb/s bitrate mp3 [18:17] <relaxed> Mista_D: place "-threads 0" after the input and consider using a preset. [18:17] <cryptopsy> and using -acodec copy produces a 400kb/s+ mp3 file that cannot be played [18:18] <Mista_D> relaxed: I tried that already - no changes. [18:18] <cryptopsy> i see [mp3 @ 0xc197e0] Unsupported sample rate. [18:18] <cryptopsy> don't get why it's doing this [18:18] <walisser> cryptopsy, what is the sample rate of the flac?, e.g. ffmpeg -i 111.flac [18:18] <relaxed> cryptopsy: -aq 0 will give you highest quality vbr mp3. [18:18] <cryptopsy> walisser: over 400 like i said [18:19] <walisser> thats bitrate, not sample rate [18:19] <cryptopsy> bitrate: 495 kb/s [18:19] <cryptopsy> this is the only thing ffmpeg -i is showing [18:20] <walisser> ok then specify the sample rate when transcoding, e.g. -ar 48k [18:20] <cryptopsy> is the sample rate in hertz? [18:20] <cryptopsy> 22050 Hz [18:20] <walisser> yes [18:20] <cryptopsy> what's the difference between the two? [18:21] <walisser> bit rate is the compressed rate (streaming rate) of the file [18:21] <walisser> sample rate is the playback rate of the audio, e.g. samples per second [18:22] <cryptopsy> thanks [18:37] <calc84> why DOESN'T ffmpeg support changing the sample rate without resampling? [18:38] <Tjoppen> try -ar before -i [20:34] <pasteeater> vcs: late reply, but the answer is "yes". [20:34] <pasteeater> ffmpeg -i VTS_02_1.VOB -c copy -map 0:0 out0.mpg -c copy -map 0:9 out9.ac3 -c copy -map 0:11 out11.ac3 [20:34] <pasteeater> there might be a better way. i'm not totally used to the new map structure [20:40] <vcs> hm ok, ill give it a shot thanks [20:46] <Skaag> can ffmpeg be used somehow to output just specific information about a media file? [20:46] <Skaag> such as its resolution? [20:46] <walisser> skagg, probably... why not grep ffmpeg -i output [20:47] <Skaag> I do that [20:47] <Skaag> I was just wondering if it has something simpler [20:48] <Skaag> I grep for "Stream" and "Video" first, then I locate (\d)x(\d) [20:48] <walisser> in libavformat... when I call avformat_find_stream_info() how can I get back the frames/packets read during the probe? [20:48] <Skaag> I just don't want it to break on me at some point [20:49] <walisser> skaag, that is probably ok [20:49] <Skaag> ok [20:49] <walisser> you would have to use the C library (libavformat for example) to avoid that I think [20:49] <Skaag> oh I see. [20:50] <cbsrobot> Skaag: use ffprobe [20:50] <cbsrobot> ffprobe -show_streams -output_format json input [20:50] <cbsrobot> ^ might have some errors [20:52] <Skaag> my version doesn't love the output_format option [20:53] <walisser> the docs say "examined packets may be buffered for later processing", does that mean they may not be buffered as well? [21:27] <stonie> hi [21:27] <stonie> how may i encode in lame and pass options to not encode in joint stereo? [21:28] <stonie> something like -lameopts ? [21:30] <pasteeater> stonie: ffmpeg -i input -f wav - | lame -m s - output.mp3 [21:30] <stonie> theres also a videostream inside [21:31] <pasteeater> and what do you want to do with this video stream? [22:05] <vcs> pasteeater: any clue when the map stuff was changed? [22:18] <alyawn> as ffmpeg runs, is the q= that's displayed an average or only the current frame? [22:19] <cek> Hi. What video decoder supports FLV8 video? how is it named/ [22:19] <Skaag> It is named Zod, kneel before Zod! [22:19] <cek> pSrZhb [22:19] <Skaag> just kidding, sorry, I don't know what it's called& :-) [22:21] <cbsrobot> cek: ffmpeg -codecs [22:22] <cbsrobot> check the vp* codecs [22:22] <cek> D V D vp8 On2 VP8 [22:22] <cek> but it doesn't decode the swf i have, only getting audio [22:22] <cek> and it used to work 2 years ago with old version [22:23] <cbsrobot> pastebin the command + output [22:23] <cek> hereyogo https://gist.github.com/92c4d4d00e5e3b18d11e [22:26] <cbsrobot> well I only see audio [22:26] <cbsrobot> but try with ffprobe [22:28] <cbsrobot> but it doesnt look like the video is recognized [22:31] <cek> thats what im saying [22:31] <cek> Hi. What video decoder supports FLV8 video? how is it named/ [22:31] <cek> the video plays in chrome [22:35] <gfto> mediainfo reports for the audio "Delay relative to video: -146ms" Is there a way for this delay to be 0 (or at lease less that 40 ms) without recompressing? The log is at : http://pastebin.com/rAabQtJM It is generated by ffmpeg but with avconv the result is the same. [22:36] <gfto> ...and the sentense should start with : I've asked this a couple of days ago at ffmpeg-user but nobody seems to know, so here it is again.... I'm remuxing mpegts files into mpegps (-f vob) and the problem I'm having is that mediainfo ... [22:57] <teratorn> anyone have a clue about this memory leak inside libavformat? valgrind output: http://codepad.org/wIgKypR6 [23:59] <gdoteof> can someone take a look here? i am trying to turn this very large (680M) 3:24 clip exported from finalcut pro to something manageble [23:59] <gdoteof> i am using some commands i got from a board http://pastebin.com/nMCzBg40 [23:59] <gdoteof> line one is the command with flags [23:59] <gdoteof> the rest is the output from ffmpeg and the resulting errors [00:00] --- Fri Dec 9 2011

1 0

ffmpeg-devel.log.20111207
by burek 08 Dec '11

08 Dec '11

[00:49] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rd9ced9fec6 10ffmpeg/libavcodec/h264.c: [00:49] <CIA-41> ffmpeg: h264: init prev_frame_num to -1 on init [00:49] <CIA-41> ffmpeg: Fixes Ticket711 [00:49] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [00:49] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r6287a356e3 10ffmpeg/libavcodec/h264.c: [00:49] <CIA-41> ffmpeg: h264: Fix memleaks on close, reset more fields [00:49] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [02:12] <CIA-41> ffmpeg: 03Janne Grunau 07master * rf5c05b9aa5 10ffmpeg/libavcodec/ (5 files in 2 dirs): [02:12] <CIA-41> ffmpeg: rv40: NEON optimised chroma MC [02:12] <CIA-41> ffmpeg: Signed-off-by: Mans Rullgard <mans(a)mansr.com> [02:12] <CIA-41> ffmpeg: 03Mans Rullgard 07master * rf054a82727 10ffmpeg/libavcodec/arm/ (Makefile h264cmc_neon.S h264dsp_neon.S): [02:12] <CIA-41> ffmpeg: ARM: move NEON H264 chroma mc to a separate file [02:12] <CIA-41> ffmpeg: This allows sharing code with the rv40 version of these functions. [02:12] <CIA-41> ffmpeg: Signed-off-by: Mans Rullgard <mans(a)mansr.com> [02:12] <CIA-41> ffmpeg: 03Janne Grunau 07master * r6c88988866 10ffmpeg/libavcodec/arm/ (Makefile rv40dsp_init_neon.c rv40dsp_neon.S): [02:12] <CIA-41> ffmpeg: rv40: NEON optimised weighted prediction [02:12] <CIA-41> ffmpeg: Signed-off-by: Mans Rullgard <mans(a)mansr.com> [02:12] <CIA-41> ffmpeg: 03Konstantin Todorov 07master * r4d9ec050a2 10ffmpeg/libavcodec/mpegvideo.c: [02:12] <CIA-41> ffmpeg: Code cleanup - mpegvideo.c - 500-1000line [02:12] <CIA-41> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje(a)gmail.com> [02:12] <CIA-41> ffmpeg: 03Janne Grunau 07master * r42d32cf53c 10ffmpeg/libavcodec/ (5 files in 2 dirs): [02:12] <CIA-41> ffmpeg: rv34: NEON optimised inverse transform functions [02:12] <CIA-41> ffmpeg: Signed-off-by: Mans Rullgard <mans(a)mansr.com> [02:12] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rff4800c16b 10ffmpeg/: (log message trimmed) [02:12] <CIA-41> ffmpeg: Merge remote-tracking branch 'qatar/master' [02:12] <CIA-41> ffmpeg: * qatar/master: [02:12] <CIA-41> ffmpeg: Code cleanup - mpegvideo.c - 500-1000line [02:12] <CIA-41> ffmpeg: rv40: NEON optimised weighted prediction [02:12] <CIA-41> ffmpeg: rv40: NEON optimised chroma MC [02:12] <CIA-41> ffmpeg: ARM: move NEON H264 chroma mc to a separate file [02:33] <CIA-41> ffmpeg: 03Carl Eugen Hoyos 07master * rc1e2ca1e31 10ffmpeg/libavformat/img2.c: Allow auto-detection of ljpg images. [03:21] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r76d7d1b0bd 10ffmpeg/libavformat/img2.c: [03:21] <CIA-41> ffmpeg: img2: 10l typo [03:21] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [03:28] <burek> pasteeater, I've updated the page https://ffmpeg.org/trac/ffmpeg/wiki/CompilationGuide with just a link to the tutorial, because it was easier for me to write it on phpBB board [03:29] <burek> I'll see tomorrow to convert it into an wiki article [04:57] <CIA-41> ffmpeg: 03Mashiat Sarker Shakkhar 07master * rbf8715719a 10ffmpeg/libavcodec/wmalosslessdec.c: [04:57] <CIA-41> ffmpeg: Fix two more int16_t vs. int confusion [04:57] <CIA-41> ffmpeg: If everything works fine for 16-bit streams, [04:57] <CIA-41> ffmpeg: the current code could be extended to do 24-bit. [04:57] <CIA-41> ffmpeg: 03Mashiat Sarker Shakkhar 07master * r6dd19c97c4 10ffmpeg/libavcodec/wmalosslessdec.c: Add a size argument to dump_int_buffer() [04:57] <CIA-41> ffmpeg: 03Mashiat Sarker Shakkhar 07master * r81a3c67169 10ffmpeg/libavcodec/wmalosslessdec.c: [04:57] <CIA-41> ffmpeg: Get rid of logging that are not required anymore [04:57] <CIA-41> ffmpeg: (Resolves some conflicts) [04:57] <CIA-41> ffmpeg: 03Mashiat Sarker Shakkhar 07master * rbe8a0d26db 10ffmpeg/libavcodec/wmalosslessdec.c: [04:57] <CIA-41> ffmpeg: Init s->cdlms[][].recent to order - 1 [04:57] <CIA-41> ffmpeg: Not sure if this is correct. But looks like [04:57] <CIA-41> ffmpeg: it was giving us off-by-one's when used to index [04:57] <CIA-41> ffmpeg: buffers. [04:57] <CIA-41> ffmpeg: 03Mashiat Sarker Shakkhar 07master * re6c66fcf68 10ffmpeg/libavcodec/wmalosslessdec.c: [04:57] <CIA-41> ffmpeg: Fix some int vs. int16_t confusion [04:57] <CIA-41> ffmpeg: Size should be determined from bits per sample. [04:57] <CIA-41> ffmpeg: Move num_lms reading out of a loop [04:57] <CIA-41> ffmpeg: Use correct value for range [04:57] <CIA-41> ffmpeg: Fix some int / int16_t / int32_t confusion [04:58] <CIA-41> ffmpeg: 03Mashiat Sarker Shakkhar 07master * r768261e50d 10ffmpeg/libavcodec/wmalosslessdec.c: [04:58] <CIA-41> ffmpeg: Fix some int / int16_t / int32_t confusion [04:58] <CIA-41> ffmpeg: For now force everything to int16_t, except [04:58] <CIA-41> ffmpeg: for variables used as index [04:58] <CIA-41> ffmpeg: 03Mashiat Sarker Shakkhar 07master * r29c2fcb677 10ffmpeg/libavcodec/wmalosslessdec.c: [04:58] <CIA-41> ffmpeg: Use correct value for range [04:58] <CIA-41> ffmpeg: The current range value causes an underflow [04:58] <CIA-41> ffmpeg: when negated and pushes anything less than zero [04:58] <CIA-41> ffmpeg: to the minimum. [05:33] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r901af94b69 10ffmpeg/ (19 files in 5 dirs): [05:33] <CIA-41> ffmpeg: movenc: Always write EDTS when we are able to. [05:33] <CIA-41> ffmpeg: Inspired by a patch from Michael Root [05:33] <CIA-41> ffmpeg: Idea-by: Baptiste Coudurier [05:33] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [06:31] <ods15> michaelni, can you bounce me the original patch email so i can reply to it? [07:01] <michaelni> ods15, bounced [07:02] Action: michaelni goes to bed, its already 7 in the morning it seems :) [07:38] <Sehns> Anyone here looking for work? i'm having a tough time finding ffmpeg experts on elance/odesk [07:38] <Sehns> we're just looking for a consultant who can help us out with how we're currently encoding for our user-contributed video site :) sorry for the spam [08:12] <ubitux> Sehns: did you check http://ffmpeg.org/consulting.html ? [08:35] <sam1> hi guys, does ffmpeg support ass filter? [09:33] <Sehns> ubitux: I emailed a few of the guys on there but they seem too busy to help me [09:34] <Sehns> i'd just like someone to go over the ffmpeg settings we're using and give us some solid advice and direction, more than happy to pay an hourly rate over paypal or something [10:33] <ubitux> Sehns: did you ask on ffmpeg-devel? [10:33] <ubitux> i remember someone asking for consulting a while ago, not sure it was you [10:33] <ubitux> (i mean the mailing list) [11:06] <ubitux> mmh fate-vsynth1-zmbv fails here [11:35] <CIA-41> ffmpeg: 03Carl Eugen Hoyos 07master * ra99dbe014d 10ffmpeg/libavcodec/pngenc.c: [11:35] <CIA-41> ffmpeg: Do not ignore full transparency when writing png palette. [11:35] <CIA-41> ffmpeg: This reverts r6689, "Make PNG produce correct 8-bit pictures". [11:53] <kriegerod> Is that correct that 0.8.7 git tag is diverged from master, not branch release/0.8? [14:52] <Tjoppen> reading up on how mxf indexing actually works. the upshot is that I can actually get to reviewing the opatom patches again [14:53] <Tjoppen> and maybe prettify that system_item hack [15:03] <Tjoppen> also, proper essence container support [15:13] <michaelni> kriegerod, yes [15:14] <michaelni> ubitux, zmbv fails probably because zlib doesnt produce the same compressed stream [15:15] <ubitux> it seems some tests also fail on fate.ffmpeg.org [15:15] <ubitux> but ok i see [15:16] <ubitux> wasn't there a WIP by Måns to replace the zlib? [15:17] <ubitux> anyway, is there any helper to display a complete filtergraph? [15:18] <ubitux> so we can see all the input, outputs, auto inserted filters, etc [15:30] <CIA-41> ffmpeg: 03Matthieu Bouron 07master * r070a40f1b7 10ffmpeg/ (libavformat/gxfenc.c tests/ref/lavf/gxf): [15:30] <CIA-41> ffmpeg: gxfenc: support timecode option [15:30] <CIA-41> ffmpeg: Reviewed-by: Baptiste Coudurier [15:40] <michaelni> last i looked that zlib replacement was somewhat buggy [15:41] <michaelni> out of array writes in valgrind IIRC [15:44] <michaelni> It would be nice to have a (graphical) ascii art drawing function for filtergraphs :) [15:49] <ubitux> yeah i agree with the ascii art :D [15:50] <ubitux> don't we have a libavformat/asciiartenc.c ? [15:52] <Tjoppen> would a generic dot2ascii tool be doable? [15:54] Action: Tjoppen shakes fist at Avid [15:55] <Tjoppen> they don't follow S377m wrt the key alignment grid (KAG) [17:20] <kriegerod> please somebody comment my question http://ffmpeg.org/pipermail/ffmpeg-devel/2011-December/117415.html [17:42] <Compn> kriegerod : does your patch apply to mpegvideo.c now ? [17:42] <Compn> that code was just reindented like hell [17:43] <michaelni> kriegerod, i cannot reproduce any crash in ffmpeg or ffplay [17:43] <michaelni> is this vlc specific ? [17:44] <kriegerod> michaelni, yes, it reproduces with vlc only [17:44] <michaelni> did you use latest ffmpeg libs ? (iam just asking because its easy to forget) [17:45] <kriegerod> git master HEAD a99dbe0 - Wed Dec 7 11:31:27 2011 +0100 Carl Eugen Hoyos <cehoyos(a)ag.or.at> Do not ignore full transparency when writing png palette. [17:46] <kriegerod> Compn, seems to me indentation is right - 4 spaces, as everywhere [17:46] <Compn> kriegerod : michaelni means the latest ffmpeg libs inside of vlc [17:46] <Compn> i think [17:47] <kriegerod> under linux vlc uses libavcodec installed into /usr/lib, via dynamic linkage [17:48] <Compn> just making sure you dont have old libavcodec.so hanging around [17:48] <kriegerod> sure [17:50] <ubitux> av_log(avctx, AV_LOG_ERROR, [17:50] <ubitux> "An with stream index %d was already specified\n", [17:50] <ubitux> mmmh [17:50] Action: ubitux doesn't understand the sentence [17:50] <Freakshow> yeah that doesn't make much sense [17:50] <ubitux> i'm trying to play a video while and play some audio [17:51] <ubitux> s/and play/playing/ [17:51] <ubitux> and get this error [17:51] <Compn> missing the first word it seems [17:51] <Compn> or something [17:51] <ubitux> trying that kind of stuff: ./ffplay -f lavfi 'movie=in.mp4:si=0[out]; amovie=in.mp4:si=1,volume=1.5[out]' [17:51] <ubitux> i wonder if it's even possible [17:51] <Compn> you mean video from one file and audio from another ? [17:52] <michaelni> kriegerod, no crash with "vlc ~/videos/test_precise.ts" here [17:53] <ubitux> Compn: same file atm, but could be another one [17:59] <Compn> kriegerod : anyways, does your patch change how ffmpeg works on it? ignoring the crash that only happens on vlc... [17:59] <Compn> does it play more video after your patch ? [18:01] <ubitux> maybe the missing word is "outpad" ? [18:01] <ubitux> or "output"... [18:01] <Compn> what file is it ubitux ? [18:01] <Compn> with the error i mean [18:01] <ubitux> just a random video [18:01] <Compn> in the source [18:01] <ubitux> with the cmd line above [18:01] <Compn> which filter [18:02] <Compn> c source file, so i can try to guess the context of the error :D [18:02] <ubitux> lavd/lavfi.c [18:03] <Compn> hmm, got me. [18:03] <ubitux> mmmh using [out0] and [out1] seems to work [18:03] <Compn> its under video [18:03] <Compn> soo "A video with stream index %d was already specified\n" ? [18:04] <Compn> or output video [18:04] <Compn> an output video [18:04] <ubitux> no it's not limited to video [18:04] <Compn> ubitux : btw, bonus points if you change error messages to include tips how to fix command lines, like 'error blah, try foobar instead' [18:04] <ubitux> "output" is the missing word imo [18:04] <Compn> oh ok [18:05] <ubitux> yes, i good idea the errors [18:05] <ubitux> i wonder if that couldn't be a gci task ;) [18:05] <ubitux> "improve error messages to give more hints to the users" [18:07] <Compn> hah yeah good idea [18:07] Action: Compn slow with ideas [18:10] <Compn> ubitux : how about "review error messages for incorrect english" ? ;P [18:11] <ubitux> "and get a free cookie to give more hints on how to solve the issue" [18:11] <Compn> two different gci tasks :P [18:11] <ubitux> kage bunshin no jutsu [18:27] <michaelni> kriegerod, ive just recompiled latest vlc, and still no crash :( [18:27] <michaelni> either of us is doing something wrong [18:40] <michaelni> kriegerod, i got it to crash with crash_5.ts, but its in vlc (picture_Release()) seems to be calling something close to NULL [18:40] <michaelni> #0 0x0000000000000025 in ?? () [18:40] <michaelni> #1 0x00007ffff7945025 in picture_Release (p_picture=0x7fffec267d30) at ../include/vlc_picture.h:182 [18:43] <michaelni> vlc_crash_example.ts and test_precise.ts dont crash for me with latest vlc + ffmpeg [19:52] <GordonFreeman> hello [19:52] <GordonFreeman> the latest snapshot's i.h files are in error [19:52] <Compn> hello gordon freeman. going to save the world again today? [19:52] <GordonFreeman> <; [19:53] <GordonFreeman> was trying to get ffmpeg working with http://www.lavrsen.dk/foswiki/bin/view/Motion/WebHome [19:53] <Compn> what error do you get ? [19:53] <GordonFreeman> looks like i have to download all versions, and go back in time to find a version that will work [19:54] <GordonFreeman> don't know, it already scrolled away [19:54] <GordonFreeman> and i have deleted the latest ffmpeg [19:55] <Compn> of course [19:56] <GordonFreeman> there are times when i feel like fixing things with a crowbar ;/ [19:56] <Compn> i'd pick the gravity gun if i had the chance [19:56] <Compn> but yeah, we recommend you use the git [19:57] <Compn> you should be able to download a git snapshot [19:57] <Compn> http://git.videolan.org/?p=ffmpeg.git;a=snapshot;h=070a40f1b7e7f8224108b646… [19:57] <Compn> which snapshot were you trying to build ? [19:58] <Compn> oh , wrong url, do this one http://git.videolan.org/?p=ffmpeg.git;a=snapshot;h=HEAD;sf=tgz [19:58] Action: Compn bets this is user error just to be a user-hater [19:58] <GordonFreeman> i have compiled this http://ffmpeg.org/releases/ffmpeg-snapshot.tar.bz2 [19:59] <GordonFreeman> then when compiling Motion it said there are errors in the .h files [19:59] <Compn> ah [19:59] <Compn> i wonder if motion requires an older version or api than ffmpeg head [19:59] <GordonFreeman> so this is not the latest working snapshot? [20:00] <Compn> thats built from last night, around midnight i think [20:01] <Compn> but your problem sounds like motion wants a different ffmpeg version, like ffmpeg 0.6.0 maybe [20:01] <GordonFreeman> btw i have linux and gcc 4.4.4 [20:01] <GordonFreeman> so you think 0.6 might work? [20:01] <Compn> you should check motion's website to see what they suggest :) [20:02] <Compn> i will look for it , cuz i'm bored [20:02] <GordonFreeman> tried 0.5.1 now and it fails badly [20:02] <GordonFreeman> /usr/local/lib/libavformat.a(file.o):(.data+0x0): multiple definition of `file_protocol' [20:02] <GordonFreeman> ffmpeg.o:(.data+0x60): first defined here [20:02] <GordonFreeman> /usr/local/lib/libavformat.a(matroskadec.o): In function `matroska_decode_buffer': [20:02] <GordonFreeman> /mnt/z/neo/linuxdvd/multimedia/webcam/ffmpeg/ffmpeg-0.5.1/libavformat/matroskadec.c:917: undefined reference to `BZ2_bzDecompressInit' [20:02] <GordonFreeman> /mnt/z/neo/linuxdvd/multimedia/webcam/ffmpeg/ffmpeg-0.5.1/libavformat/matroskadec.c:926: undefined reference to `BZ2_bzDecompress' [20:02] <GordonFreeman> /mnt/z/neo/linuxdvd/multimedia/webcam/ffmpeg/ffmpeg-0.5.1/libavformat/matroskadec.c:929: undefined reference to `BZ2_bzDecompressEnd' [20:02] <GordonFreeman> collect2: ld returned 1 exit status [20:02] <GordonFreeman> make: *** [motion] Error 1 [20:02] <GordonFreeman> ;/ [20:03] <Compn> are you missing bzip / bzip2 lib ? [20:03] <GordonFreeman> nope [20:03] <GordonFreeman> i have all [20:04] <Compn> is motion still in development? [20:04] <GordonFreeman> i'm a bug magnet btw [20:04] <Compn> "Motion must be installed with revision 0.4.8 or 0.4.9pre1 of ffmpeg." [20:04] <GordonFreeman> no, that is latest stable version [20:04] <Compn> maybe i am looking at old documentation [20:04] <GordonFreeman> hehe well then it states it should work with latest ffmpeg [20:05] <Compn> ok , can you build latest ffmpeg and paste the errors then ? [20:05] <Compn> also we should move to #ffmpeg , since this channel is for patches and stuff :) [20:05] <GordonFreeman> actually i wouldn't care which version it use, i just want to record webcam into a video. [20:06] <Compn> ffmpeg can record webcam to a video directly [20:06] <GordonFreeman> ok [20:06] <GordonFreeman> but i have to recompile again then [20:06] <GordonFreeman> should i git clone ? [20:06] <GordonFreeman> or the one i tried before? [20:06] <Compn> GordonFreeman : this one > http://git.videolan.org/?p=ffmpeg.git;a=snapshot;h=HEAD;sf=tgz [20:06] <Compn> its .targz , or you can git clone if you want [20:07] <GordonFreeman> no, i like tar.gz thanks [20:08] <GordonFreeman> well if it fails to work, and it's not ffmpeg's fault, then i will be coding [20:12] <GordonFreeman> ffmpeg-HEAD-070a40f]# ./configure | tee ./configure.log [20:12] <GordonFreeman> why isn't this working ? [20:12] <GordonFreeman> ;/ [20:13] <Compn> paste log somewhere [20:14] <GordonFreeman> this doesn't work right [20:14] <GordonFreeman> but i used to do this on other software [20:14] <GordonFreeman> it should write output, and tee it into a logfile [20:14] <Compn> maybe configure failed [20:15] <GordonFreeman> configure script does not write output [20:16] <GordonFreeman> only after a delay [20:16] <GordonFreeman> now running without pipe [20:18] <GordonFreeman> nahm its working i was just expecting output... [20:18] <GordonFreeman> now running make [20:21] <Sehns> Anyone here looking for consulting work? We're hiring a contractor with the mustard to go over our ffmpeg settings.. apologise in advance for the spam, we've had no luck on elance/odesk etc [20:22] <Compn> Sehns : its ok, feel free to ask on the ffmpeg-devel list as well [20:22] <GordonFreeman> Compn<< :) http://matrix.z-labor.com/tmp/C/sdltext_fps_d_asdf.c [20:23] <GordonFreeman> something agains a few minutes of boredness [20:23] <Sehns> ahh cool i'll do that as well, thanks Compn [20:24] <burek> pasteeater, https://ffmpeg.org/trac/ffmpeg/wiki/CompilationGuide :) if you find time, can you please just check if everything is ok [20:25] <GordonFreeman> uh this is a huge one libavcodec/h264.c [20:25] <Compn> Sehns : you might also want to clarify what kind of work you need done. e.g. frontend programming, scripting scripts, optimizing encoding command lines, etc [20:25] <Compn> GordonFreeman : should have make -j4 'd it :P [20:26] <GordonFreeman> i don't care actually, i let it do whatever [20:26] <GordonFreeman> CPU is to work [20:26] <Sehns> we're doing the frontend programming, we really just need an ffmpeg expert to help us with the scripts and the command line encoding options/settings [20:26] <GordonFreeman> Compn<< soo, how was this getting image/video from webcam using ffmpeg again? [20:27] <ubitux> are you sure this is not something you could even ask on ffmpeg-users? [20:27] <GordonFreeman> don't know [20:27] <ubitux> a lot of users are familiar with shell scripting and various options [20:27] <GordonFreeman> hehe [20:28] <ubitux> GordonFreeman: i'm talking to Sehns, and btw your question *definitely* belongs to ffmpeg-users :) [20:28] <GordonFreeman> so you have 2 in a row [20:29] <Sehns> ubitix: I'll give it a go, thanks [20:29] <Compn> ubitux : shhh, hes willing to pay [20:29] <Compn> :) [20:30] <ubitux> Compn: he already asked several times for this, i think it's time to get him some hints how to solve the issue :p [20:30] <ubitux> since no one seems to care unfortunately :) [20:31] <Compn> ah [20:31] <Compn> i'd just copy / paste out of the docs and charge em $50/hr for the privledge :P [20:31] <GordonFreeman> :) [20:31] <ubitux> he can still make a donation ;) [20:32] <ubitux> Sehns: isn't? :D [20:32] <Sehns> haha [20:32] <GordonFreeman> ahah still compiling [20:33] <GordonFreeman> only a 2.2GHz AMD64 [20:34] <GordonFreeman> nah done [20:34] <Sehns> np [20:34] <Sehns> i'll hit up the mailing list [20:34] <Sehns> need someone to basically do QA on our shit yesterday [20:35] <ubitux> will you use the name "Sehns" on the mailing list? [20:35] <Compn> hmmm [20:35] <ubitux> (just to see your post) [20:35] Action: Compn wonders if he could help [20:35] Action: Compn likes being lazy [20:35] <GordonFreeman> ok i have the error in log [20:36] <Compn> paste to pastebin [20:36] <GordonFreeman> ffmpeg.c:380: error: CODEC_TYPE_VIDEO undeclared (first use in this function) [20:36] <Compn> paaaaaaaasssssssttttteeeeeebbbbbbiiiiiiinnnnnnnnnnn [20:36] <GordonFreeman> ffmpeg.c:649: error: PKT_FLAG_KEY undeclared (first use in this function) [20:36] <GordonFreeman> make: *** [ffmpeg.o] Error 1 [20:36] <GordonFreeman> hehe [20:36] <Sehns> ubitux i'll post as pentae / swoonr.com [20:36] <GordonFreeman> ok [20:38] <ubitux> Sehns: okay :) [20:39] <GordonFreeman> heres motion configure http://pastebin.com/Wtc4MHHK [20:39] <pasteeater> burek: quick look seems good to me. thanks for adding it. [20:39] <burek> :beer: :) [20:40] <GordonFreeman> and here's the make with the error http://pastebin.com/fdAWYuJy [20:43] <GordonFreeman> looks ok until pasteline 43, where the warnings begin, and looks fail at line 48 [20:47] <GordonFreeman> but i see it is because the variable was removed from ffmpeg, and was used in motion [20:52] <GordonFreeman> Compn<< do you think it would be much work to make this work ? [20:53] <GordonFreeman> i don't know how incompatible if current ffmpeg with older versions [20:53] <Compn> ask in #motion [20:54] <GordonFreeman> ok i did [20:54] <GordonFreeman> :) [21:05] <cbsrobot> GordonFreeman: try to change CODEC_TYPE_VIDEO -> AVMEDIA_TYPE_VIDEO in motion/ffmpeg.c line 360 [21:05] <GordonFreeman> wait i have been told there is an svn version [21:06] <GordonFreeman> but that fails too [21:06] <GordonFreeman> ffmpeg-HEAD-070a40f/libavformat/matroskadec.c:1023: undefined reference to `BZ2_bzDecompressInit' [21:06] <GordonFreeman> in here [21:06] <GordonFreeman> after linking [21:06] <cbsrobot> then line 649 pkt_flag_key -> AV_PKT_FLAG_KEY [21:07] <GordonFreeman> thanks, please discard the errors i posted [21:07] <GordonFreeman> heres the motion svn version svn co http://www.lavrsen.dk/svn/motion/trunk/ . [21:08] <GordonFreeman> it shows now only errors after linking, in matroskadec.c [21:08] <cbsrobot> GordonFreeman: http://www.lavrsen.dk/foswiki/bin/view/Motion/BugReport2010x11x28x210835 [21:13] <CIA-41> ffmpeg: 03Alexis Ballier 07master * r1af91978db 10ffmpeg/ (libavcodec/qtrle.c tests/ref/fate/qtrle-1bit): [21:13] <CIA-41> ffmpeg: qtrle.c: Fix for ticket #226. [21:13] <CIA-41> ffmpeg: In 1bpp mode, interpret skip&0x80 as "start a new line" instead of "go to next [21:13] <CIA-41> ffmpeg: line", this is almost the same except for the first line which was always [21:13] <CIA-41> ffmpeg: skipped before and caused to try to write an extra line at the end of the frame [21:13] <CIA-41> ffmpeg: (ticket #226). [21:13] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [21:13] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r8b85c29463 10ffmpeg/libavcodec/qtrle.c: [21:13] <CIA-41> ffmpeg: qtrledec: Fix artifacts in tickets/226/qt_rle_bw.mov [21:13] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [21:13] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r0be95fe183 10ffmpeg/MAINTAINERS: [21:13] <CIA-41> ffmpeg: MAINTAINERS: ask maintainers to help keep their entries in this file upto date. [21:13] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [21:13] <GordonFreeman> cbsrobot<< cool thanks :) you have saved the day [21:13] <GordonFreeman> it was only a typo in makefile [21:13] <cbsrobot> GordonFreeman: np - enjoy [21:14] Action: cbsrobot has a degree in using google search [21:15] <cbsrobot> GordonFreeman: let me know if you get it running - and report it back to motion please [21:18] <GordonFreeman> what does this mean? [21:18] <GordonFreeman> [1] [ERR] [TRK] v4l2_select_input: Unable to query input 8. VIDIOC_ENUMINPUT, if you use a WEBCAM change input value in conf by -1: Invalid argument [21:18] <GordonFreeman> [1] [ERR] [TRK] v4l_start: ioctl (VIDIOCGCAP): Invalid argument [21:25] <burek> it means that camera can't read/write that property at the moment [21:26] <burek> maybe your cable is damaged or camera broken or something [21:29] <GordonFreeman> hmm [21:29] <GordonFreeman> well i probably need a reboot [21:29] <GordonFreeman> i was having segmentations faults when i tested things in C [21:47] <CIA-41> ffmpeg: 03Reimar Döffinger 07master * r7023fb81c7 10ffmpeg/tests/ (codec-regression.sh ref/vsynth1/zmbv ref/vsynth2/zmbv): [21:47] <CIA-41> ffmpeg: One last try to get zmbv test to work across all fate machines. [21:47] <CIA-41> ffmpeg: Change compression level to 0 this time. [21:47] <CIA-41> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger(a)gmx.de> [21:50] <GordonFreeman> Please use -b:a or -b:v, -b is ambiguous [21:50] <GordonFreeman> what is this all about? [21:51] <GordonFreeman> if i enter a : after -b then it will fail anyway [21:51] <GordonFreeman> ffmpeg -an -f video4linux2 -s 640x480 -b 1200k -r 15 -i /dev/video0 -vcodec mpeg4 -vtag DIVX ./out.avi [21:51] <GordonFreeman> did this [21:55] <Compn> oh bitrate [21:55] <Compn> use -b:v 1200k [21:55] <GordonFreeman> what if i just want best quality? or some autodetect ? [21:56] <Compn> use -vcodec libx264 -preset slowest then [21:56] <Compn> hehe [21:56] <GordonFreeman> :) ok thanks [21:56] <Compn> might be too slow for real-time ... [21:57] <michaelni> best quality is to use cp instead of ffmpeg, nothing beats the original ;) [21:57] <GordonFreeman> ahah yes [21:57] <GordonFreeman> actually i have a $10 webcam [21:57] <GordonFreeman> so [21:57] <GordonFreeman> ;>> [21:57] <GordonFreeman> was aiming for less noise only [21:58] <michaelni> there are various noise reduction filters, some postprocess filters like spp are also quite capable to remove noise [21:59] <GordonFreeman> that sounds cool [21:59] <GordonFreeman> so it can reduce video size without degrading quality? [22:09] <Compn> ubitux : when does volume normalization get committed ? hehe [22:10] <ubitux> hehe no idea [22:10] <ubitux> i won't have time for this soon [22:10] <GordonFreeman> you can do dynamic compression in real-time btw [00:00] --- Thu Dec 8 2011

1 0

ffmpeg.log.20111207
by burek 08 Dec '11

08 Dec '11

[01:00] <darkstarbyte> does dual pass encoding produce better results? [01:04] <cbsrobot> darkstarbyte: normally yes [01:04] <darkstarbyte> what about for a dvd, you know like a home movie for a class? [01:06] <cbsrobot> darkstarbyte: Your decision ... [01:06] <darkstarbyte> ??? [01:07] <darkstarbyte> How about mpeg2 in general? [01:08] <cbsrobot> darkstarbyte: sorry - whats the question ? [01:08] <darkstarbyte> multi-pass encoding on mpeg-2 size/quality [01:09] <darkstarbyte> in general [01:10] <darkstarbyte> high bitrate [01:11] <darkstarbyte> no quality cut [01:11] <cbsrobot> darkstarbyte: I cannot help you - just ask a proper question. [01:11] <cbsrobot> mpeg2 vs. what ? [01:12] <darkstarbyte> Well I have wondered if mpeg4 part2 is worth the extra time it takes to encode for a divx dvd [01:14] <darkstarbyte> vs an mpeg2 dvd [01:14] <cbsrobot> I'd say yes, but I never compared them [01:15] <darkstarbyte> Ok [01:15] <darkstarbyte> I am going to start encoding in mpeg2 until I can find something that says other wise [01:15] <cbsrobot> mpeg2 1 pass with ffmepg is pretty fast nice aswell [01:16] <darkstarbyte> ffmpeg it is [01:22] <TomCat2> I don't know what you mean by a divx dvd though... are you talking about a DIVX dvd (i.e. failed video rental format) or putting a DivX file on a DVD to play on a PC? [01:23] <TomCat2> actually technically DIVX disc != DVD - although similar [01:27] <TomCat2> if you're talking putting a file onto DVD media to play on a PC (or set top device that can recognize files on a DVD) use h.264 [01:27] <TomCat2> for quality [01:34] <darkstarbyte> I have noticed that my dvd player will play things encoded in xvid [01:35] <darkstarbyte> I found out when I forgot to re-convert a file off of my camera. [01:35] <TomCat2> it's likely a feature of your dvd player [01:35] <TomCat2> some will play .mp3 files too [01:36] <TomCat2> for compatibility - if you want to be able to give the dvd to others to play, stick with the standards [01:36] <TomCat2> which means authoring the dvd in the proper format - VOB files, etc., is what you'll see in the final print [01:38] <darkstarbyte> dvdauthor [01:40] <TomCat2> make an h.264 version of it anyway tho - if it's a keepsake, you never know, you might want to watch it on your blu-ray (or better) device in years to come [01:40] <TomCat2> h.264 will transcode directly to what goes on a blu-ray w/o any loss [01:41] <TomCat2> unless u can keep your master files - even better [01:43] <darkstarbyte> I wish movies could be kept on flash memory. [01:43] <darkstarbyte> instead [01:43] <darkstarbyte> of disks [01:45] <TomCat2> why can't they? [01:46] <darkstarbyte> Because a flash drive is easier to loose than a dvd I think [01:47] <TomCat2> you can put a movie on a flash drive... whether or not you lose it is irrelevant [01:47] <darkstarbyte> I am talking about after [01:47] <darkstarbyte> blu ray [01:47] <sacarasc> Flash drives are more expensive than DVDs. [01:48] <TomCat2> price doesn't determine whether it can or cannot get done either [01:48] <sacarasc> But whether it will or not. [01:49] <TomCat2> and actually I disagree with the notion of a flash drive is easier to lose than a disc :D [01:49] <TomCat2> I've never lost either of my 160GB SSD's but lost Blu Ray discs plenty of times [01:50] <TomCat2> yet I still manage to copy my blu ray backups onto either of the SSD's and play it back in my PCH when that happens [02:14] <teratorn> is it supported to encode a packet once, with avcodec_encode_video() and then write it out using av_interleaved_write_frame() to multiple different format contexts? assuming same format/stream layout of course [04:50] <Algo> Hi, how do I make my ffmpeg convert wmv to flv correctly? "ffmpeg -i <input>.wmv <output>.flv" seems to fail [04:54] <hi117> Algo both of those are containers, try ffmpeg -i input.wmv -acodec copy -vcodec copy output.flv [05:08] <kode54> copy probably won't work for that [05:08] <kode54> since flv probably doesn't support windows media codecs [05:10] <hi117> maybe it did work, its been 10 mins and he hasnt responded [05:17] <Algo> Thanks for the reply hi117: if output is "Motion9.wmv" - it works fine. if output is "Motion9.flv" - it doesn't work. Output for "Motion9.flv":http://pastebin.com/nNVvXyLE Output for "Motion9.wmv":http://pastebin.com/6AWswG5P [05:20] <hi117> my guess is what kode54 said, i forgot that it probably doesnt support the same stuff, flv seems to be able to support h264, use that [05:20] <kode54> flv seems to support h264, aac, and mp3 [05:21] <kode54> there's also some other primitive video crap that nobody should encode to anymore [05:21] <hi117> so it would be ffmpeg -i input.wmv -vcodec libx264 output.flv (your video doesnt have audio according to your pastebins) [05:21] <kode54> ah [05:23] <hi117> also, is there a way to pass some advanced stuff to x264 from ffmpeg, like the path to x264 and custom options? [05:24] <hi117> on my system i have 2 x264s, one for 8 bit and one for 10 bit and i would like to pass a lot of advanced stuff to it, not just profiles and presets [05:24] <hi117> would -vcodec x264:bla work maybe? [05:25] <Algo> hi117: http://pastebin.com/e6hC4CY6 [05:25] <hi117> ... did it fail? i dont see anything blatantly wrong [05:26] <kode54> ffmpeg uses the library directly, not the x264 binary [05:26] <kode54> the options are probably supported, but I'm not sure of the syntax [05:27] <hi117> ah, still can i specify which library to use? i can recompile 2 my 10 bit x264 with --enable-shared if needed [05:27] <Algo> hi117: Doesn't work... blank still ;p [05:27] <hi117> kode54 do you see anything wrong with it? [05:29] <kode54> weird frame rate issue there, I'm not sure what else [05:29] <hi117> it always says that the video framerate differs from container when i do it [05:30] <hi117> maybe we could try telling it what framerate to use? [05:31] <hi117> well, im going to bed, as a tip to help kode54, post the output of ffmpeg -i input.wmv [05:32] <kode54> I may not be able to help much, and I may not be staying around much longer for the night, but I leave my client idling [05:33] <kode54> and other people who may be more helpful will still be here [05:33] <Algo> kode54: Thanks! [07:04] <GordonFreeman> hi [07:05] <GordonFreeman> can someone help in ffmpeg? libavcodec.so and libavformat.so [07:05] <GordonFreeman> these files are not found after installing ffmpeg [07:05] <GordonFreeman> Package ffmpeg-0.6.1-1.el6.rf.x86_64 already installed and latest version [07:05] <GordonFreeman> but i have these /usr/lib64/libavcodec.so.52 [07:05] <GordonFreeman> /usr/lib64/libavcodec.so.52.72.2 [07:06] <GordonFreeman> /usr/lib64/libavformat.so.52 [07:06] <GordonFreeman> /usr/lib64/libavformat.so.52.64.2 [07:36] <sam1> hi guys, i want to use "ass filer", what's command? [07:37] <sam1> i use "ffmpeg -i *.avi -vf "ass=*.ass" -vcodec libx264 -acodec copy *.avi [07:38] <sam1> but the out avi without ass [07:39] <Sehns> Anyone have experience with embedding subtitles? [07:59] <sam1> anyone use ass filter?? [08:08] <ubitux> sam1: are the '*' really in the command? [08:08] <ubitux> any special output on ffmpeg while opening the ass file? [08:09] <ubitux> could you paste your full command line and output on pastebin? [08:31] <sam1> Hi ubitux, [08:33] <sam1> ffmpeg -i /home/video/test0.avi -vcodec libx264 -an -frames 50 -s 720x576 -vf "ass=/home/video/test0.ass" -y /home/output/test1.avi [08:34] <sam1> i use this command line [08:59] <ubitux> sam1: does ffmpeg output say anything? [09:02] <ubitux> could you paste it on pastebin.com? [09:05] <relaxed> ubitux: Is it possible to use an ass profile with srt subs? [09:06] <relaxed> Similar to mplayer's '-ass-force-style'? [09:08] <ubitux> you can not burn anything else than ass atm [09:08] <ubitux> the filter is just a temporary workaround [09:08] <ubitux> i'm working on something more mature for subtitles in my free time, but it might take a while [09:09] <ubitux> basically if you want to burn a srt, just convert it to ass with gaupol or similar, customize it, and burn it with vf ass [09:13] <relaxed> I applaud your effort [09:13] <sam1> hi ubitux , you mean ass could burn into a movie? [09:14] <ubitux> sam1: that's not what you wanted? [09:14] <relaxed> It's most commonly referred to as hardsubbing. [09:15] <sam1> yes i want that ,but i use the command "ffmpeg -i /home/video/test0.avi -vcodec libx264 -an -frames 50 -s 720x576 -vf "ass=/home/video/test0.ass" -y /home/output/test1.avi" [09:15] <sam1> the test1.avi does not have subtitle [09:15] <ubitux> sam1: i asked 2 times that you paste the ffmpeg output somewhere :p [09:16] <sam1> i paste here? [09:16] <sam1> i open your wedsite [09:16] <relaxed> ubitux: Do the subs have to be external? [09:17] <ubitux> relaxed: yes, it is limited to vf=file :p [09:17] <ubitux> sam1: just create a new paste copypasting the ffmpeg on pastebin.com and share the url here [09:18] <sam1> ok [09:18] <ubitux> the ffmpeg output* [09:18] <sam1> i do it ,ubitux [09:20] <relaxed> ubitux: I assume the more muture version will be able to specify an internal stream? [09:20] <relaxed> mature* [09:20] <ubitux> relaxed: yes we will certainly use the source audio/video filter [09:20] <ubitux> adding a support for subtitles [09:21] <relaxed> That's great! I appreciate your work on this. [09:21] <ubitux> relaxed: i started a dev thread here: http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/2011-November/117059.html [09:22] <ubitux> if you're interested in following it [09:22] <relaxed> I think this will appeal to a large number of people. [09:22] <ubitux> yes, but better get some work done first ;) [09:22] <ubitux> right now, we have the ass filter by Stefano which is great [09:22] <relaxed> Yes, I started following the thread early on but there wasn't much attention. [09:22] <sam1> hi ubitux ,there is website : http://pastebin.com/2QW59uav [09:24] <relaxed> ubitux: Of course, the current solution is better than no solution. :) [09:24] <ubitux> sam1: you added some debug? :) [09:24] <sam1> yes [09:24] <sam1> you see it ? [09:25] <sam1> does the command line is right ? [09:25] <ubitux> it looks right yes [09:25] <ubitux> do you mind sharing the ass file? [09:25] <sam1> how [09:26] <sam1> where i put [09:28] <ubitux> i don't know, upload it somewhere, or as a cheap solution you can use pastebin [09:29] <sam1> wait a moment [09:31] <sam1> http://pastebin.com/Rvmm0Kn6 [09:31] <sam1> there is ass [09:31] <sam1> ubitux, do you burn ass into a movie? [09:34] <ubitux> just a sec, i'm testing [09:34] <sam1> ok [09:34] <sam1> thanks [09:36] <ubitux> sam1: the duration or your subtitles is zero& [09:36] <ubitux> look at the timestamp [09:36] <ubitux> all the subtitles are from 00:00:00.00 to 00:00:00.00 [09:36] <sam1> o [09:37] <sam1> but it could work with vlc [09:38] <ubitux> could? [09:38] <ubitux> how so? [09:39] <ubitux> heh fun, mplayer displays all of them [09:39] <sam1> i play test0.avi ,and reload ass [09:39] <sam1> wait a sec ,i test it with kmplayer [09:39] <ubitux> anyway, the timestamps looks wrong to me [09:40] <ubitux> i'm not why mplayer display them though [09:41] <sam1> i also find it,but this ass could play with kmplay , [09:41] <ubitux> all the subtitles are displayed at the same time, right? [09:42] <ubitux> it's a pretty strange behaviour though [09:43] <sam1> only one sub [09:43] <ubitux> ok so we have 3 different behaviour [09:43] <sam1> only the first sub [09:44] <sam1> do you play my ass ? [09:44] <ubitux> yes [09:44] <ubitux> mplayer shows them all at once here [09:44] <ubitux> ffmpeg doesn't show them (which is likely the correct behaviour imho) [09:44] <ubitux> and you say your players display just the first one? [09:45] <sam1> kmplayer [09:45] <ubitux> yes ok, so well& just fix your subtitle ;) [09:45] <sam1> i fix it ,wait a sec [09:45] <sam1> thanks [09:46] <ubitux> where does this file comes from? did you "craft" it yourself? [09:46] <sam1> i use visualsubsync [09:47] <sam1> i have a sub,and convert to ass with this tool [09:47] <ubitux> and the timestamp where trashed? :p [09:47] <sam1> trash? [09:48] <ubitux> i mean, wrong [09:49] <sam1> i see the source sub, [09:49] <sam1> its timestamp all {0} [09:50] <sam1> maybe source is wrong? [09:51] <ubitux> the source has 0 as initial and final timestamp? [09:52] <sam1> yes [09:52] <sam1> {0}{0}Just gonna stand there and watch me burn [09:52] <sam1> {0}{0}But thats alright because I like the way it hurts [09:52] <sam1> {0}{0}Just gonna stand there and hear me cry [09:52] <sam1> {0}{0}But thats alright because I love the way you lie [09:52] <sam1> {0}{0}I love the way you lie [09:52] <sam1> {0}{0}I cant tell you what it really is [09:52] <sam1> {0}{0}I can only tell you what it feels like [09:52] <sam1> {0}{0}And right now its a still night in my wind pipe [09:52] <sam1> {0}{0}I cant breathe but I still fight while I can fight [09:52] <sam1> {0}{0}As long as the wrong feels right its like Im in flight [09:52] <sam1> this is the sub [09:56] <ubitux> sam1: microdvd right? [09:57] <ubitux> (.sub) [09:57] <ubitux> i don't how that's supposed to work with microdvd [09:57] <ubitux> the microdvd "specs" might says something about this like "display them all the time" [09:57] <ubitux> but it might not match the ASS specs (but it's likely undefined) [09:58] <ubitux> so well, better set the timestamp manually :) [09:58] <sam1> microdvd ?? [09:59] <ubitux> the format name is bad :) [09:59] <ubitux> iirc it was a software called with that name [09:59] <ubitux> https://en.wikipedia.org/wiki/MicroDVD [09:59] <sam1> o ,but this sub could play with kmplay,not vlc [10:00] <ubitux> it's not surprising [10:01] <ubitux> well anyway, you just have to fix the frame in case of the microdvd file (your .sub with {0}{0}), and/or set the correct timestamps in the ASS file [10:01] <sam1> you mean unstandard sub could play ? [10:02] <ubitux> ? [10:02] <sam1> i do it [10:02] <sam1> ubitux: it's not surprising [10:02] <sam1> you mean ? [10:02] <ubitux> i just don't understand what you are asking [10:05] <sam1> i mean wrong sub [10:05] <sam1> irregular sub [10:05] <sam1> kmplay could play irregular sub ? [10:05] <ubitux> well, the .sub is not really standard you know [10:06] <ubitux> it's just a software (that lives for 1 year) make such subtitles [10:06] <sam1> i konw [10:06] <ubitux> and support for those subtitles was added in various players [10:06] <ubitux> so "irregular" doesn't have much sense here [10:07] <ubitux> unspecified behaviour is just differently interpreted between players [10:07] <sam1> o [10:07] <sam1> i see [10:07] <ubitux> (the original software behaviour should be considered the standard though) [10:07] <sam1> ubitux, i see your RFC [10:08] <sam1> you are going to do the work? [10:09] <sam1> and you mean current ffmpeg could burn ass into a movie? you test it? [10:11] <ubitux> what work? [10:11] <ubitux> -vf ass=... is the way to burn ass [10:11] <ubitux> if you want muxed ass, it's also in ffmpeg since a while now [10:11] <ubitux> and sure i tested both [10:12] <sam1> http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/2011-November/117059.html [10:12] <sam1> [FFmpeg-devel] [RFC] Talk about subtitles [10:12] <sam1> i see it [10:13] <ubitux> ah, yes i'll take some time to improve the subtitles support [10:14] <sam1> add str filter like ass filter ? [10:14] <sam1> srt filter [10:14] <ubitux> something like that [10:14] <sam1> sorry [10:14] <ubitux> actually ffmpeg already has a srt and sub support [10:14] <ubitux> but not for burning [10:15] <ubitux> the aim is to make use of that support and add the subtitle burning [10:15] <sam1> only decode and encode ? [10:15] <ubitux> but it means some thinking [10:15] <ubitux> there is a partial support for decode/encode/mux/remux, but nothing for burning [10:16] <sam1> burning only for ass?? [10:16] <ubitux> yes [10:16] <ubitux> at the moment. [10:16] <sam1> does it support ass decode/encode/mux/remux ?? [10:17] <ubitux> yes afaik [10:19] <sam1> thanks [10:19] <sam1> i read ffmpeg source [10:20] <sam1> so bulky [12:18] <leristo> hey, is this the correct channel for questions on using ffmpeg with c++? [13:14] <impulze> is the video/audio synching chapter of the sdl/ffmpeg tutorial still up to date? seems like a lot of code is needed [14:51] <impulze> way too much code indeed, there must be some library i can use that basically provides me with an easier interface, no? [14:56] <impulze> what i basically want is to play a cutscene from within my 2d game [15:00] <Tjoppen> take ffplay.c, cut away all the crap -> done? [15:01] <impulze> yeah well i'd still have to maintain it and as i fail to understand most parts of the code i suppose this is a huge place for bugs :P [15:02] <Mavrik> well use a player library then [15:02] <impulze> can you recommend one/some? [15:05] <gst-kaps> is ffmpeg flac encoding faster then libflac one ? [15:06] <kriegerod> please dear libavcodec hackers, look at this bug, need help here: http://trac.videolan.org/vlc/ticket/5620#comment:3 [15:06] <Mavrik> impulze, not really, I've seen some references to SMPEG, but nothing portable :\ [15:06] <impulze> maybe libvlc? [15:23] <Kuukunen> impulze: ffmpegsource? [15:24] <gst-kaps> is ffmpeg flac encoding faster then libflac one ? [15:24] <impulze> maybe ask more often [15:25] <impulze> Kuukunen: looks interesting at first glance, thank you [15:28] <JEEB> Kuukunen, is it a playback lib tho? [15:28] <JEEB> although I guess you can use it as such [15:29] <JEEB> since aegisub uses it too [15:29] <Kuukunen> JEEB: well, now that I think about it, having to index stuff might cause a bit of an overhead [15:29] <Kuukunen> but... you can store the index in a file too [15:29] <JEEB> yeah [15:29] <JEEB> and you can use a specific lib version for it [15:30] <JEEB> so you can just use the pre-created index [16:11] <teratorn> does it work to encode a packet once, and write it out to multiple different AVFormatContext's ? [16:12] <teratorn> basically I want to stream some content to multiple clients, but only incur the encoding cost once.... presumably I can only start streaming to a new client on the next key frame? [16:18] <Mavrik> teratorn, there should be no problems with sending a packet to multiple avformatcontexts, just make sure the timestamps are ok [16:18] <Mavrik> and you can start streaming to client before keyframe, however the player will probably start playing when it reaches the first keyframe (depending on format/player) [16:20] <teratorn> Mavrik: thanks [16:42] <GordonFreeman> hi [16:42] <GordonFreeman> so? noone knows? [16:45] <Zeranoe> Do subtitles work like any other stream? If I wanted to add .srt subtitles to a video, would it be done like: ffmpeg -i "input-video.mp4" -i "input-subtitles.srt" [16:49] <ubitux> Zeranoe: yes [16:49] <ubitux> and just like -acodec and -vcodec you have -scodec [16:50] <ubitux> (or -c:a, -c:v, -c:s iirc) [16:53] <filo1234> hi all [16:53] <filo1234> hey guys does ffmpeg have some filter about brightness ? [16:55] <ubitux> mmh you may try the eq2 filter [16:55] <ubitux> through mplayer wrapper [16:56] <ubitux> http://ffmpeg.org/libavfilter.html#mp look for "brightness" [16:57] <Skaag> I made the mistake of pressing D during an ffmpeg run, how do I disable debug output? ;-) [16:57] <filo1234> ubitux: I'll look [16:57] <filo1234> thanks [17:00] <filo1234> ubitux: how can I see wich brighness value has my video? [17:01] <Zeranoe> ubitux: Thanks for confirming it [17:01] <ubitux> filo1234: i guess the default is 1.0 if that's your question [17:02] <filo1234> ubitux: nope I mean which value has my source video [17:02] <filo1234> my input video [17:03] <ubitux> i'm not sure how you're supposed to evaluate this [17:03] <ubitux> average Y maybe, there is nothing in ffmpeg afaik to extract such information [17:03] <filo1234> ok [17:04] <GordonFreeman> [065938] <GordonFreeman> can someone help in ffmpeg? "libavcodec.so and libavformat.so" these files are not found after installing ffmpeg Package "ffmpeg-0.6.1-1.el6.rf.x86_64 already installed and latest version" [17:05] <GordonFreeman> [070018] <GordonFreeman> but i have these /usr/lib64/libavcodec.so.52 /usr/lib64/libavcodec.so.52.72.2 /usr/lib64/libavformat.so.52 /usr/lib64/libavformat.so.52.64.2 [17:05] <filo1234> so I have to try some values [17:06] <ubitux> filo1234: as i said, the reference is 1.0 i guess [17:06] <ubitux> so 1.0 will likely not change the brightness [17:06] <ubitux> while 0.5 will darken it by half [17:06] <ubitux> you need to check mplayer man though [17:07] <ubitux> my bad, it's 0.0 by default, and you set a value from -1 to 1 [17:11] <filo1234> ubitux: yes I'm reading... but is that only a filter while I see videos with mplayer? or can I pass mp=eq2[parms] to ffmpeg conversion? [17:13] <ubitux> ffmpeg as a wrapper on the mplayer filters through the filter "mp" [17:13] <ubitux> so ffmpeg -i ... -vf mp=eq2... [17:14] <filo1234> oh -vf well thank as well [17:18] <filo1234> o.0 argh -vf means? video format?? [17:18] <filo1234> do you mean vcodec that I use? [17:18] <ubitux> video filter [17:19] <filo1234> ok but ffmeg say me that it doesn't recognise -vf option [17:19] <filo1234> ffmpeg* [17:19] <ubitux> paste your command [17:19] <ubitux> how old is your ffmpeg? [17:20] <filo1234> this is my version FFmpeg version SVN-r0.5.1-4:0.5.1-1ubuntu1.2 [17:20] <impulze> Kuukunen: from what i see, i have to do the whole syncing with ffmpegsource on my own aswell, don't i? [17:20] <impulze> at least i can't figure out where in the source they do that [17:21] <ubitux> filo1234: the current version is 0.8.x you know :p [17:21] <filo1234> and my command is ffmpeg -i myvideo -sameq -vf mp=eq2=1.0:2:0.5 output [17:21] <ubitux> ffmpeg 0.5 hasn't mp wrapper [17:21] <filo1234> ubitux: :/ yeah but I need to compile it if I want new version [17:21] <ubitux> it's too old [17:21] <Kuukunen> impulze: yea well usually it's meant less for playing and more for video editing etc... but at least it should still be easier than ffmpeg [17:21] <filo1234> ok :| [17:21] <Kuukunen> impulze: might want to check something else too :P [17:21] <ubitux> well i won't do it for you [17:22] <ubitux> filo1234: just upgrade your repositories [17:22] <ubitux> or yes, build it from source [17:22] <filo1234> yeah I think so [17:23] <filo1234> well going to work [17:23] <filo1234> ubitux: thanks a lot bye [17:23] <ubitux> 0.5 was released march 2009 btw& [18:44] <JodaZ> can i encode multiple ranges of video from a source video into a target video [18:44] <JodaZ> so like seconds 5 to 10 followed by seconds 20 to 25... [18:47] <leristo> is it possible to encode data from opencv iplimage with ffmpeg? [18:53] <relaxed> leristo: what formats can it output? [19:11] <leristo> relaxed: opencv's iplimage ? [19:14] <relaxed> yes [19:15] <leristo> its a basic c++ structure with a pointer to the image data [19:17] <leristo> http://opencv.willowgarage.com/documentation/basic_structures.html#iplimage [19:18] <leristo> question would be: is it possible to encode data from char pointer using ffmpeg [19:18] <leristo> i cant find guides on this, only stuff where files are handled [19:34] <GordonFreeman> latest snapshot fails. [19:35] <GordonFreeman> errors in include.h-s [19:42] <JodaZ> hmm, i want to make something like moving thumbnails... for that i need to extract like a few 2-5 second snipets of video from a movie and concat them, does anyone have any recommendations as for how to do that ? [19:46] <GordonFreeman> /usr/local/lib/libavformat.a(file.o):(.data+0x0): multiple definition of `file_protocol' [19:46] <GordonFreeman> ffmpeg.o:(.data+0x60): first defined here [19:46] <GordonFreeman> /usr/local/lib/libavformat.a(matroskadec.o): In function `matroska_decode_buffer': [19:46] <ubitux> JodaZ: heh, i'm working on a video filter for such thing [19:47] <GordonFreeman> so they put up these untested, how cool [19:47] <ubitux> i need to get something done on it, and it should be available upstream in a while [19:47] <JodaZ> :/ video filters are simple librarys ? [19:47] <JodaZ> or do they need to be integrated into ffmpeg ? [19:47] <ubitux> libavfilters is part of ffmpeg [19:48] <ubitux> you will do sth like: ffmpeg -i input.mp4 -vf thumbnail,scale=320:240 -vframes 5 out.mjpeg [19:48] <ubitux> or something like that :p [19:48] <GordonFreeman> looks like i must download them all and go from svn to back in time [19:49] <ubitux> GordonFreeman: svn? we're using git right now [19:49] <GordonFreeman> ok i mean latest version. [19:49] <ubitux> you might want to ask ffmpeg-users if you get no answer here for your issue [19:49] <GordonFreeman> since latest snapshot doesn't work with motion [19:49] <ubitux> you should report such issue :) [19:50] <GordonFreeman> well i did report it [19:50] <GordonFreeman> here now [19:50] <ubitux> the persons that can solve it might not be available now, or even reading this channel [19:50] <ubitux> the bug tracker exists for such purpose [19:51] <ubitux> also, you can try to point out the regression or find the incriminated code and discuss the issue on #ffmpeg-devel [19:51] <GordonFreeman> so http://www.lavrsen.dk/foswiki/bin/view/Motion/WebHome http://ffmpeg.org/releases/ [19:52] <GordonFreeman> and http://ffmpeg.org/ [19:52] <ubitux> JodaZ: if you need a workaround, i'd say you could use vf select [19:52] <ubitux> JodaZ: vf select has a eval that might help you picking 1 frame each N frames [19:53] <ubitux> something like -vf select=eq(n%100\,0) [19:53] <JodaZ> ubitux, thats not really what i want [19:53] <ubitux> not sure about the mod operator, you need to have a look to the documentation [19:53] <JodaZ> i want 5 seconds of video every 10 minutes of source video [19:54] <ubitux> oh, i misunderstood then sorry [19:54] <ubitux> mmh there is a segmenter, but i'm not sure it is mature enough (some ppl are working on it right now) [19:55] <ubitux> you can do it manually though [19:55] <ubitux> with a script, using -t and -ss [19:55] <JodaZ> that needs like two reencodes >_> [19:55] <JodaZ> i want to concat all the 5 second snipets [19:56] <ubitux> reencode? no [19:56] <ubitux> just using -c copy [19:57] <JodaZ> and how do i concat ? [19:58] <ubitux> mmh not all kind of video will support this, but you have the concat protocol [19:58] <ubitux> iirc you can even use ffmpeg -i ... -i ... [19:58] <ubitux> but i may be wrong [19:59] <pasteeater> GordonFreeman: how can your issue be duplicated? [20:00] <JodaZ> ubitux, doesn't seem to work, even tho i read somewhere that it ought [20:00] <GordonFreeman> download this, make install, http://ffmpeg.org/releases/ffmpeg-snapshot.tar.bz2 [20:00] <GordonFreeman> then try to compile, make this http://www.lavrsen.dk/foswiki/bin/view/Motion/WebHome [20:00] <GordonFreeman> it will say there are errors in the .h files [20:01] <GordonFreeman> (on linux and gcc) [20:01] <pasteeater> how did you configure ffmpeg? [20:02] <GordonFreeman> default [20:03] <GordonFreeman> ./configure && make [20:03] <pasteeater> oh, i see you started asking this in ffmpeg-devel. this is a more appropriate channel. [20:05] <pasteeater> by "this" i mean #ffmpeg [20:29] <Sehns> Anyone here looking for consulting work? We're hiring a contractor who has the mustard to go over our ffmpeg settings.. apologise in advance for the spam, we've had no luck on elance/odesk etc [20:30] <Sehns> mainly need someone to help with scripts/command line settings and thumbnails .. fyi we already have some bad code in place that kind of sucks [20:38] <burek> sucks in which way [20:47] <Compn> Sehns : i've only got an hour of free time right now, you want me to check it out ? i'll have more time tonight / tomorrow [20:48] <Sehns> Sure [21:52] <GordonFreeman> hi [21:52] <GordonFreeman> anybody used to grab video from webcam? [21:52] <GordonFreeman> what command line do you use? [23:20] <alyawn> I have an odd video with 8 channels of audio. I'm trying to transcode. any idea how to reduce it to the first 2? [23:27] <ubitux> -ac 2 [23:27] <ubitux> you mean just pick the first two, or downmix to two? [23:28] <alyawn> pick the first two [23:28] <ubitux> -map_channel 0.0.0 -map_channel 0.0.1 then [23:28] <ubitux> GordonFreeman: on linux you have v4l2 available [23:29] <ubitux> GordonFreeman: http://ffmpeg.org/ffmpeg.html#Video-and-Audio-grabbing [23:30] <ubitux> alyawn: http://ffmpeg.org/ffmpeg.html#Advanced-options just scroll down a little for everything about -map_channel [23:31] <alyawn> thanks, ubitux [00:00] --- Thu Dec 8 2011

1 0

ffmpeg-devel.log.20111206
by burek 07 Dec '11

07 Dec '11

[00:17] <CIA-41> ffmpeg: 03Carl Eugen Hoyos 07master * ra448a5d1c4 10ffmpeg/libavformat/mov.c: Do not fail fatally if chan atom is too short. [01:53] <CIA-41> ffmpeg: 03Diego Biurrun 07master * rdef5dc1260 10ffmpeg/ (doc/developer.texi tools/patcheck): doc: document preferred Doxygen syntax and make patcheck detect it [01:53] <CIA-41> ffmpeg: 03Alex Converse 07master * r6d23d19729 10ffmpeg/libavformat/mov.c: [01:53] <CIA-41> ffmpeg: mov: Allow empty stts atom. [01:53] <CIA-41> ffmpeg: Fixes regressions caused by 30c3d976 [01:53] <CIA-41> ffmpeg: 03Justin Ruggles 07master * rd1241ff3b2 10ffmpeg/avconv.c: avconv: use avcodec_decode_audio4() instead of avcodec_decode_audio3() [01:53] <CIA-41> ffmpeg: 03Justin Ruggles 07master * re2a2c49fcd 10ffmpeg/avplay.c: avplay: use a separate buffer for playing silence [01:53] <CIA-41> ffmpeg: 03Justin Ruggles 07master * rf199f38573 10ffmpeg/avplay.c: avplay: use avcodec_decode_audio4() [01:53] <CIA-41> ffmpeg: 03Justin Ruggles 07master * r0bafe04917 10ffmpeg/libavcodec/api-example.c: api-example: update to use avcodec_decode_audio4() [01:53] <CIA-41> ffmpeg: 03Diego Biurrun 07master * rc68fafe0d2 10ffmpeg/ (10 files in 4 dirs): doxygen: eliminate Qt-style doxygen syntax [01:53] <CIA-41> ffmpeg: 03Martin Storsjö 07master * r714cd7e758 10ffmpeg/tests/ (codec-regression.sh ref/acodec/g722): [01:53] <CIA-41> ffmpeg: g722: Add a regression test for muxing/demuxing in wav [01:53] <CIA-41> ffmpeg: Signed-off-by: Martin Storsjö <martin(a)martin.st> [01:53] <CIA-41> ffmpeg: 03Justin Ruggles 07master * rf08e54e83d 10ffmpeg/libavformat/utils.c: avformat: use avcodec_decode_audio4() in avformat_find_stream_info() [01:53] <CIA-41> ffmpeg: 03Justin Ruggles 07master * r9179f27c64 10ffmpeg/avconv.c: [01:53] <CIA-41> ffmpeg: avconv: only allocate 1 AVFrame per input stream [01:53] <CIA-41> ffmpeg: This avoids allocating/freeing an AVFrame for each packet decode. It also [01:53] <CIA-41> ffmpeg: simplifies error handling in transcode_audio() and transcode_video(). [01:53] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rb404ab9e74 10ffmpeg/: (log message trimmed) [01:53] <CIA-41> ffmpeg: Merge remote-tracking branch 'qatar/master' [01:53] <CIA-41> ffmpeg: * qatar/master: [01:53] <CIA-41> ffmpeg: mov: Don't av_malloc(0). [01:53] <CIA-41> ffmpeg: avconv: only allocate 1 AVFrame per input stream [01:54] <CIA-41> ffmpeg: avconv: fix memleaks due to not freeing the AVFrame for audio [01:54] <CIA-41> ffmpeg: h264-fate: remove -strict 1 except where necessary (mr4/5-tandberg). [01:54] <CIA-41> ffmpeg: 03Justin Ruggles 07master * r110d2af28e 10ffmpeg/avconv.c: avconv: fix memleaks due to not freeing the AVFrame for audio [01:54] <CIA-41> ffmpeg: 03Diego Biurrun 07master * re873c03ac7 10ffmpeg/ (8 files in 4 dirs): misc Doxygen markup improvements [01:54] <CIA-41> ffmpeg: 03Martin Storsjö 07master * r6fcbb0f553 10ffmpeg/libavcodec/ (g722.h g722dec.c): (log message trimmed) [01:54] <CIA-41> ffmpeg: g722dec: Signal skipping the lower bits via AVOptions instead of bits_per_coded_sample [01:54] <CIA-41> ffmpeg: This avoids using bits_per_coded_sample for this information. [01:54] <CIA-41> ffmpeg: bits_per_coded_sample should be 4 for this codec normally, [01:54] <CIA-41> ffmpeg: since two samples are encoded into one 8 bit codeword. [01:54] <CIA-41> ffmpeg: In principle, this might be info that needs to be passed from [01:54] <CIA-41> (15 lines omitted) [04:59] <sam2> hi all, i want to build ffmpeg invs 2010 ,is there a doc? [05:26] <Compn> i dont know of any visual studio docs for ffmpeg , sorry [05:26] <Compn> maybe someone else does... [08:10] <ubitux> michaelni: what are the issues raised by not having request/poll_frame() in the filter? how should i test it works fine with/without it? [11:26] <kriegerod> anybody knows in which timezone Carl Eugen Hoyos resides? [11:26] <kriegerod> cehoyos [11:30] <av500> AST [11:48] <cbsrobot> kriegerod: UTC+1 [11:53] <burek> one question :) I've managed to import ffmpeg git project into Eclipse and everything went ok, but when I do Build All I get errors like these: http://pastebin.com/zFNiypsz [11:54] <burek> something like Makefile:2: config.mak: No such file or directory [11:54] <burek> does it mean i have to manually type ./configure ... [11:54] <burek> or something? [11:54] <av500> yep [11:54] <av500> no configure, no compile [11:54] <burek> ok thanks :) [11:55] <burek> anyway, I wrote the tutorial about how to import ffmpeg from git into Eclipse [11:55] <burek> and I'll put it somewhere online [11:55] <av500> you ported ffmpeg to eclipse? [11:55] <burek> for the newbee devels :) [11:55] <burek> ported? [11:55] <av500> never mind [11:55] <burek> :) [11:55] <burek> :beer: :) [11:56] <av500> beer with burek? [11:56] <av500> svakako! [11:57] <burek> haha :) [11:57] <burek> to to :) [12:02] <burek> is this good enough: ./configure --enable-memalign-hack --disable-shared --enable-static --disable-mmx --disable-mmx2 --disable-optimizations --disable-stripping [12:02] <burek> i found it using google :) [12:05] <av500> good what for? [12:06] <av500> ./configure should be fine [12:08] <burek> woo hoo :))) it works :) thanks :) [12:09] <burek> btw, what IDE do you guys use for coding the ffmpeg [12:13] <av500> none [12:15] <cbsrobot> wow none - thats a decent IDE - can you give me a link ? [12:16] <cbsrobot> maybe I should switch from hexedit to none ... [14:51] <ubitux> btw, does anyone have a video sample with the frame id written for each frame? [14:51] <ubitux> could be generated without much pain but i feel lazy [15:06] <CIA-41> ffmpeg: 03Stefano Sabatini 07master * r3af5ddb24b 10ffmpeg/libavfilter/vf_scale.c: [15:06] <CIA-41> ffmpeg: vf_scale: give a clue in case of invalid expression self-reference [15:06] <CIA-41> ffmpeg: Address trac ticket #706. [15:06] <CIA-41> ffmpeg: 03Stefano Sabatini 07master * r7ecabc8505 10ffmpeg/libavfilter/vsrc_life.c: vsrc_life: fix comment in parse_rule() [15:29] <Compn> ubitux : oh you mean like each frame has a number on it ? [15:29] <Compn> but no, i do not [15:29] <ubitux> yes, burned on each pic [15:29] <ubitux> erf [15:29] <Compn> lol [15:29] <ubitux> ok i guess i'll have to do it [15:29] <Compn> we might have a cctv stream [15:29] <Compn> oh no, thats timestamp [15:30] Action: Compn tries to think up easy way to generate it [15:30] <ubitux> with imagemagick [15:30] <ubitux> generating N png [15:30] <ubitux> then ffmpeg -i ... [15:30] <ubitux> i already did it a while ago in another context [15:31] <ubitux> but well, lazyness when you have it& [15:34] <ubitux> convert -size 640x480 -gravity center -background black -fill white -pointsize 120 label:1 out.png [15:34] <ubitux> i guess i just have to loop with this [15:42] <michaelni> ubitux, ffmpeg should have problems with the wrong poll/request [15:42] <ubitux> what kind of problem? i didn't expect any with my (simple) tests [15:48] <ubitux> Compn: http://pastie.org/2975124 [15:48] <cbsrobot> ubitux: I have one [15:48] <ubitux> arh [15:48] <ubitux> i just made one with that script :p [15:50] <cbsrobot> ubitux: ok [15:50] <cbsrobot> next time [16:12] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rc89488588b 10ffmpeg/ffmpeg.c: [16:12] <CIA-41> ffmpeg: ffmpeg: Warn if filters poll & request frame behave wrongly. [16:12] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [16:23] <michaelni> ubitux, last commit should print a warning if the filters poll/request behave odd [16:24] <ubitux> ok, thanks :) [16:47] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rcf8ec019ab 10ffmpeg/libavformat/ (movenc.c movenc.h): [16:47] <CIA-41> ffmpeg: movenc: fragment support [16:47] <CIA-41> ffmpeg: Reviewed-by: Baptiste Coudurier [16:47] <CIA-41> ffmpeg: Tested-with-QT: Baptiste Coudurier [16:47] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [16:48] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r15130b94fc 10ffmpeg/libavformat/ (movenc.c movenc.h): [16:48] <CIA-41> ffmpeg: movenc: support spliting fragments based on bytesize instead of duration. [16:48] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [17:43] <ubitux> michaelni: "* Gets the number of immedeatly available frames." typo in immediately [17:44] <michaelni> wrong, its 2 typos :) [17:45] <ubitux> :] [17:49] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rf0d0ae3b1e 10ffmpeg/libavformat/mov.c: [17:49] <CIA-41> ffmpeg: mov: increase total_size only when it actually increases. [17:49] <CIA-41> ffmpeg: This is just cosmetic as the if() is always true. [17:49] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [17:49] <CIA-41> ffmpeg: 03Mihnea Balta 07master * rfef2da6a22 10ffmpeg/libavformat/mov.c: (log message trimmed) [17:49] <CIA-41> ffmpeg: movdec: Fix bad computed size for atoms with size 0 in MOV files [17:49] <CIA-41> ffmpeg: The computed size doesn't contain the header size because it's already [17:49] <CIA-41> ffmpeg: skipped by incrementing total_size, but then it's skipped again in the [17:49] <CIA-41> ffmpeg: last line. The atom comes out 8 bytes short and the function [17:49] <CIA-41> ffmpeg: mov_read_chan() aborts the whole parsing process. I think the computed [17:49] <CIA-41> ffmpeg: size should be atom.size - total_size + 8. [17:49] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r7d90bc9f2a 10ffmpeg/libavformat/mov.c: [17:49] <CIA-41> ffmpeg: movdec: Read extended 64bit size before the debug av_log() and only when there [17:49] <CIA-41> ffmpeg: was enough space to read the 32bit size. [17:49] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [19:50] <CIA-41> ffmpeg: 03Reimar Döffinger 07master * r6ecc78b707 10ffmpeg/tests/ (codec-regression.sh ref/vsynth1/zmbv ref/vsynth2/zmbv): [19:50] <CIA-41> ffmpeg: Choose compression level 6 for zmbv test. [19:50] <CIA-41> ffmpeg: The default of 9 gives different results on different FATE systems. [19:50] <CIA-41> ffmpeg: However the zlib test using compression level 6 works, so [19:50] <CIA-41> ffmpeg: try this instead. [19:50] <CIA-41> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger(a)gmx.de> [20:07] <pasteeater> burek: you can add compilation tutorials here: https://ffmpeg.org/trac/ffmpeg/wiki/CompilationGuide [23:01] <burek> pasteeater, ok :) I will :) [23:02] <burek> btw this Eclipse is so cool :) it shows me all the warnings as an exclamation marks to the left of the line where it is :) [23:02] <burek> I mean all the warnings which show during 'make' :) [23:04] <JEEB> Now think if they had implemented llvm/clang-based warning/error management system [23:04] <burek> http://i42.tinypic.com/18yjcg.png [23:04] <Daemon404> JEEB, you mean like xcode? [23:05] <JEEB> I mean, man [23:05] <JEEB> http://zi.fi/shots/clang.png [23:05] <JEEB> the difference is BIG [23:05] <Daemon404> JEEB, xcode has llvm built into the idea [23:06] <JEEB> yeah [23:06] <Daemon404> llvm parses as you type and gives feedback [23:06] <Daemon404> its amazing [23:06] <JEEB> llvm is much easier to implement [23:06] <JEEB> and much more robust for stuff like that [23:06] <Daemon404> thats why xcode is <3 atm [23:06] <Daemon404> even if i dont own a mac [23:06] <JEEB> now if only it could compile mpeg4videodec.c :D [23:06] <Daemon404> lol [23:06] <JEEB> (and yes, I know this thing only relates to the svn trunk, the 3.0 release should be fine) [23:09] <CIA-41> ffmpeg: 03Carl Eugen Hoyos 07master * rba3bb53bdc 10ffmpeg/libavcodec/8bps.c: [23:09] <CIA-41> ffmpeg: Fix 8BPS RGB32 output, the alpha channel was ignored. [23:09] <CIA-41> ffmpeg: Reviewed-by: Roberto Togni [23:10] <Daemon404> so uh [23:10] <Daemon404> also [23:10] <Daemon404> http://git.videolan.org/?p=ffmpeg.git;a=commit;h=6ecc78b707945458e5d126a19a… [23:10] <Daemon404> this [23:11] <Daemon404> why was the test changed instead of actually fixing teh decoder? [23:11] <Daemon404> er, encoder. [23:43] <CIA-41> ffmpeg: 03Michael Niedermayer 07release/0.7 * rba5bb0562b 10ffmpeg/libavcodec/h264.c: [23:43] <CIA-41> ffmpeg: h264: Use mismatching frame numbers in fields [23:43] <CIA-41> ffmpeg: to synchronize the first/second field state independant of them being reference or not. [23:43] <CIA-41> ffmpeg: Fixes Ticket354 [23:43] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [23:43] <CIA-41> ffmpeg: (cherry picked from commit 545ec935a4b4e0f032ebd975907b41f6fe4465c9) [23:43] <CIA-41> ffmpeg: 03Michael Niedermayer 07release/0.8 * r1550c0885d 10ffmpeg/libavcodec/h264.c: [23:43] <CIA-41> ffmpeg: h264: Use mismatching frame numbers in fields [23:43] <CIA-41> ffmpeg: to synchronize the first/second field state independant of them being reference or not. [23:43] <CIA-41> ffmpeg: Fixes Ticket354 [23:43] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [23:43] <CIA-41> ffmpeg: (cherry picked from commit 545ec935a4b4e0f032ebd975907b41f6fe4465c9) [00:00] --- Wed Dec 7 2011

1 0

ffmpeg.log.20111206
by burek 07 Dec '11

07 Dec '11

[01:01] <bbeattie> is there any way to extract an m4a's coverart with either ffmpeg or ffprobe? Google doesn't give something so I'm guessing no but before I give up or start digging through code I thought I'd see if anyone knew off the top of their head first. [01:02] <bbeattie> m4a/aac [02:42] <raven> ffmpeg where to find .ffpresets files? x264 presets (veryslow, ...) missing [02:55] <relaxed> raven: ffmpeg now uses libx264 internal presets with -preset. [02:56] <relaxed> So you don't need the *ffpresets. Run `ffmpeg -h | less` and scroll down to "libx264 AVOptions:" [02:57] <raven> relaxed, ok tnx [03:42] <rynop> what is the best (fastest in terms of speed/min resources) way to add a png overlay (watermark) when encoding a video from 1 format to another? [03:43] <rynop> im piping output of glc-play to mencoder, wondering if i can hook in-between those 2, and not have to do a 2nd pass just to add watermark [04:23] <relaxed> why are you using mencoder? [04:24] <relaxed> rynop: glc-play -> ffmpeg (adding watermark -> output [04:25] <rynop> relaxed, only reason i'm using mencoder is to get access to the gamma (eq2=) option. I'm a total ffmpeg noob, couldn't figure out how to do what i needed with straight ffmpeg. [04:26] <relaxed> I believe you can use any mplayer filter in ffmpeg with "-vf mp:$filtername=options". [04:27] <rynop> ok cool. is the benefit of going right to ffmpeg that i'll encode faster? [04:27] <relaxed> yes, see http://ffmpeg.org/libavfilter.html#mp [04:27] <relaxed> eq2 is listed [04:27] <rynop> coo [04:27] <rynop> thx [04:28] <relaxed> The benefit is mencoder is unmaintained, buggy and can't mux into any recent containers. [04:28] <rynop> okay thanks. [04:31] <relaxed> it's -vf mp=$filtername=options [04:32] <rynop> relaxed, ok thanks [04:34] <rynop> relaxed, would you mind taking a look at http://pastebin.com/Emk0dSw8 when you get a chance. its 2 lines, want to see if the ffmpeg cmd i created is equivalent to the mencoder one i had before... [04:35] <rynop> if ur not familiar with mencoder then nevermind [04:36] <relaxed> I am. What codec do you want for the output? [04:36] <hi117> for video editing, do i want to work with a lossless codec then transcode back to lossy or keep it lossy? the source file is already lossy [04:36] <rynop> hah so again im a freaking noob w/ this stuff. I just want to output an mp4 for youtube. Was thinking h264... [04:37] <rynop> relaxed, unless u have a better recommendation (I dont really care about quality that much). Its a 1280x720, replay of a video game. [04:38] <rynop> I'm gonna be encoding alot of these .glc files, so want to be able to encode them pretty fast. [04:38] <rynop> but not be effin huge cuz i have to upload to youtube (or another vid hosting site i find) [04:39] <relaxed> try glc-play "${filename}" -o - -y 1 | ffmpeg -i - -an -vcodec libx264 -preset fast -crf 18 -vf mp=eq2=1.3 -y "${OUT}" [04:39] <rynop> i'll give it a try, thank you so much. [04:40] <relaxed> slower presets give you better the compression. lower the crf value to increase quality [04:42] <relaxed> hi117: If the source is lossless then stick with lossless. If it's lossy it doesn't make much difference, right? [04:43] <rynop> cool. as for the watermark overlay, is doing it the -vf "movie=..." the right path to go down? [04:43] <relaxed> rynop: add -threads 0 after -vcodec libx264 [04:43] <hi117> ok ty, i was wondering if repeated encodes would mess with quality much [04:43] <rynop> relaxed, ok [04:44] <hi117> heh, im playing with watermarks too right now [04:44] <rynop> hi117, if i figure it out (i just want to add a single transparent .png to every frame), i'll post back here. [04:44] <relaxed> hi117: I mean, as long as you can import the lossy format. If not, then lossy -> lossless for the import makes sense. [04:45] <hi117> rynop i believe its the overlay filter [04:45] <hi117> http://ffmpeg.org/libavfilter.html#overlay-1 [04:45] <rynop> thanks [04:46] <hi117> np [04:46] <hi117> im using to to cover my name in a video i have :D [04:46] <rynop> haah im kinda doing the opposite :) [04:46] <hi117> such a useful filter :P [04:47] <hi117> also does anyone know if the 'transparent' color is alpha or 808080, i heard in a video it had to be 808080 gray [04:48] <rynop> relaxed, when i run that ffmpeg command i get "Unrecognized option 'preset'" - that cuz im on an old ffmpeg ver?(0.6-4) [04:49] <relaxed> rynop: yes. replace it with -vpre fast [04:50] <rynop> relaxed, thx. same error now with -vf . I should prob just upgrade ffmpeg. Was trying to just use pkgs in ubuntu and avoid compile. [04:51] <relaxed> rynop: 64bit linux? [04:51] <rynop> y [04:52] <relaxed> I have 64bit linux builds. https://sites.google.com/site/linuxencoding/builds [04:53] <relaxed> or there's https://sites.google.com/site/linuxencoding/install-script [04:54] <rynop> relaxed, awesome [04:58] <sam2> hi guys, is there #ffmpeg-devel? [04:59] <relaxed> indeed there is [04:59] <sam2> thanks relaxed [04:59] <hi117> ty, works like a charm [05:00] <sam2> hi all, i want to build ffmpeg invs 2010 ,is there a doc? [05:00] <relaxed> what is invs? [05:09] <hi117> how do i join videos? i tried the mpg method mentioned in the faq but the intermediate video is very pixelated.... [05:10] <relaxed> ffmpeg -i concat:1.mpg\|2.mpg -vcodec copy -acodec copy output [05:11] <hi117> oh ty they improved on the very bad method in the faq [05:11] <rynop> relaxed, thx for this static build. its great. if project im working on ever makes $ i'll be donating for sure. [05:12] <relaxed> you're welcome [05:13] <hi117> someone should probably change that, mpg isnt compatible with h264 and quality takes a real hit converting to and fro [05:14] <hi117> like 720p video turns into cheap cellphone quality [05:17] <hi117> er, nvm [05:17] <hi117> i forgot to add the -sameq [05:22] <rynop> hi117, that overlay worked like a charm. thx. [05:22] <hi117> np :3 [05:50] <alyawn> I feel like I'm missing something. Is there a way to determine the src video framerate from within a filter? [06:10] <relaxed> alyawn: you can get it using `ffprobe -show_streams INPUT` [06:10] <relaxed> example, r_frame_rate=24000/1001 [06:11] <alyawn> thanks relaxed, but I'm trying to get the rate from within a filter I'm writing [06:12] <alyawn> I have access to the time_base, but it's usually set to 1/1000000 [11:02] <kriegerod> hi. Anyone from libavcodec hackers, could you please help me investigate the problem i reported here http://trac.videolan.org/vlc/ticket/5620 [11:02] <kriegerod> The specific broken media file triggers abort() in libavcodec [11:02] <kriegerod> in last ffmpeg release, 0.8.7 [11:03] <kriegerod> when used through vlc [13:42] <zawapete> hi [13:47] <zawapete> pastebin links: [13:47] <zawapete> http://pastebin.com/cgTqSC91 and http://pastebin.com/K3FSNp37 [13:48] <zawapete> i've the macports build of 0.8.5.git-38a444e on a macpro [13:49] <zawapete> and i built the 0.8.5 tag checkout from git natively on an armv5te (synology 410j) [13:49] <zawapete> as you can see from the link the arm version miss -c/-codec cli param [13:50] <zawapete> i need "pure transmuxing", just to copy streams from a format to another, supported by my dlna client [13:53] <zawapete> i tried obtaining the same on the synology using -vcodec copy -acodec copy -scodec copy, but then i get a "Number of stream maps must match number of output streams" error [13:54] <zawapete> i'd link to start ffmpeg with autodetected full streaming mapping as transcoding script from mediatomb [13:54] <zawapete> *i'd like to [13:54] <kriegerod> zawapete: what about just '-vcodec copy -acodec copy', without '-scodec copy' [13:54] <zawapete> same error kriegerod [13:55] <kriegerod> could you give your full command with '-vcodec copy -acodec copy', sample file piece, and output of command? [13:56] <zawapete> with -map 0 all works fine [13:56] <zawapete> *without -map 0, i mean [13:56] <kriegerod> and you certainly need -map? [13:56] <zawapete> as far as i read from the man it's the only way to map all the input streams [13:57] <zawapete> if they're more the one for type (video/audio/subs) [13:57] <kriegerod> by default it'll save all streams [13:57] <kriegerod> ah [13:57] <kriegerod> not sure then [13:58] <zawapete> if i can't make -map 0 work i have to write a bash script to parse the input streams and then launch ffmpeg with the needed params [13:58] <zawapete> i'd like to avoid that [13:58] <zawapete> ;) [13:59] <kriegerod> strange. just tested -map0 -vcodec copy -acodec copy with file that has two audio streams. worked [13:59] <kriegerod> ffmpeg 0.8.7 from git [14:01] <zawapete> which platform? [14:04] <zawapete> on mac it works for me also [15:07] <chrome> michaelni: Hello! I am a GCI student.. Would it be possible for you to give me more details about this task? http://goo.gl/hV2JY Thank you! [15:13] <ubitux> oh that's a nice task :) [15:17] <chrome> ubitux: Is it so hard to find someone from ISO willing to provide specifications for free? (They are indeed expensive...) [15:18] <ubitux> i guess it requires some special authorizations from a high level [15:18] <ubitux> so the person you might contact will need to negociate for you with the person able to give the "ok" [15:19] <ubitux> i never tried myself though, so this is just some guessing [15:22] <chrome> ubitux: thanks for your suggestions. I might try to do this.. but it's rather hard to obtain an authorization from a high level.. [15:23] <ubitux> do you know someone working there? [15:23] <ubitux> or someone who know someone who ... [15:24] <chrome> it's something along someone who knows someone who.. [15:25] <chrome> would it hurt to just contact them directly (by phone, e-mail)? or that's just impossible.. [15:26] <ubitux> i can't say myself [15:26] <ubitux> btw, you should join #ffmpeg-devel [15:27] <ubitux> it's somewhat related to ffmpeg dev :) [15:27] <chrome> oh, ok.. i've tried #ffmpegdev, but it seems rather empty(?) [15:28] <ubitux> it was an old special channel [15:28] <ubitux> #ffmpeg-devel is the official one [15:36] <twid> ffplay example.avi gitve error Header Damaged.... what does it mean by that? thx... [15:54] <hi117> hello, im trying to add audio to a video, this is a pastebin of the command run and the output http://pastebin.com/w6XuxqBz [16:19] <kuberan> Got this error when I ran FFMPEG http://pastebin.com/h284fkTE [16:19] <kuberan> can somebody help [16:19] <kuberan> WARNING: The bitrate parameter is set too low. It takes bits/s as argument, not kbits/s [16:19] <kuberan> Unknown encoder 'libfaac' [16:22] <hi117> can you post the output of ffmpeg -codecs? [16:23] <kuberan> FFmpeg version SVN-r0.5.1-4:0.5.1-1ubuntu1.2, Copyright (c) 2000-2009 Fabrice Bellard, et al. [16:23] <kuberan> configuration: --extra-version=4:0.5.1-1ubuntu1.2 --prefix=/usr --enable-avfilter --enable-avfilter-lavf --enable-vdpau --enable-bzlib --enable-libgsm --enable-libschroedinger --enable-libspeex --enable-libtheora --enable-libvorbis --enable-pthreads --enable-zlib --disable-stripping --disable-vhook --enable-runtime-cpudetect --enable-gpl --enable-postproc --enable-swscale --enable-x11grab --enable-libdc1394 --enable-shared --disable-static [16:23] <kuberan> libavutil 49.15. 0 / 49.15. 0 [16:23] <kuberan> libavcodec 52.20. 1 / 52.20. 1 [16:23] <kuberan> libavformat 52.31. 0 / 52.31. 0 [16:23] <kuberan> libavdevice 52. 1. 0 / 52. 1. 0 [16:23] <kuberan> libavfilter 0. 4. 0 / 0. 4. 0 [16:23] <kuberan> libswscale 0. 7. 1 / 0. 7. 1 [16:23] <kuberan> libpostproc 51. 2. 0 / 51. 2. 0 [16:23] <kuberan> built on Sep 16 2011 17:08:44, gcc: 4.4.3 [16:23] <kuberan> ffmpeg: missing argument for option '-codecs' [16:23] <kuberan> sorry pasted it here. [16:24] <hi117> .... it should print something like this: http://pastebin.com/f21yTHGm [16:27] <kuberan> so whatz the issue ? [16:29] <hi117> i dont think you have faac installed [16:30] <kuberan> ok i ll check [16:34] <iive> just look at the year... soon it would be 3 years old :P [16:41] <kuberan> iive: are you talking to me ? [16:42] <kuberan> this is the package that is available as a part of bitnami EC2 AMI [16:44] <kuberan> iive: built on Sep 16 2011 17:08:44, gcc: 4.4.3 [16:46] <iive> check the copyright... you can rebuild the source every day and it won't make it any newer. [16:46] <iive> it won't fix a bug and won't add a single feature. [16:49] <kuberan> makes sense :) [16:49] <kuberan> FFmpeg version SVN-r0.5.1-4:0.5.1-1ubuntu1.2 [16:50] <kuberan> What is the latest version ? [16:50] <ubitux> 0.8.7 [16:51] <kuberan> great. [16:51] <kuberan> I will download and build it then. [16:55] <kuberan> To support conversion of avi,mov,wmv,flv,mpg,mpe,mpeg,3gp,mp4 files to FLV.. what is the ideal config script ? [17:09] <kuberan> anybody there to help ? [17:15] <JEEB> kuberan, note: flv is a container [17:16] <kuberan> Does that mean there are difference encodings for FLV itself ? [17:16] <JEEB> the contents of some of those things you've mentioned can be put into flv as-is [17:16] <JEEB> yes, you can put multiple different things into flv >_> [17:16] <hi117> JEEB why does ffmpeg -i videoM.mkv -i audio.aac -acodec copy -vcodec copy videoC.mkv fail? [17:17] <JEEB> no idea [17:17] <JEEB> I usually use mkvmerge for matroska muxing anyways [17:17] <JEEB> VP62, that old old H.263 sorenson thing, H.264 at the very least video wise > kuberan [17:17] <hi117> says Application provided invalid, non monotonically increasing dts to muxer in stream [17:17] <kuberan> But still to make those conversions, I can use FLV rite ? [17:17] <JEEB> ... [17:17] <kuberan> So for FFMPEG to support all those conversions what is my ideal config script [17:18] <kuberan> I mean to make those conversions I can use FFMPEG rite ? [17:18] <JEEB> you usually want the output to be of single type [17:18] <JEEB> you've set the container [17:18] <JEEB> now set video and audio tracks' formats [17:18] <hi117> aac.... [17:18] <hi117> and h264 [17:19] <JEEB> hi117, he _might_ have some dumb reason to use something older, not like we know :V And might be that even he doesn't know [17:19] <JEEB> thus, let's make him think a bit first shall we [17:19] <hi117> :3 [17:19] <hi117> forums says its an ancient issue [17:19] <kuberan> ?? [17:20] <hi117> ty JEEB, mkvmerge worked [17:20] <JEEB> unsurprisingly [17:20] <hi117> :3 [17:20] <JEEB> kuberan, you haven't limited yourself to one thing with just telling us you want to output stuff into the flv container [17:21] <JEEB> think with those little brain cells and seek information [17:21] <JEEB> what is it that you exactly need? [17:24] <kuberan> I want to make this command work [17:24] <kuberan> ffmpeg -i 263.mov -vcodec flv -r 28 -b 96000 -ab 64 -acodec libfaac -ac 1 -ar 44100 263.flv [17:24] <kuberan> sorry [17:24] <JEEB> ok [17:24] <kuberan> ffmpeg -i 263.mov -vcodec flv -r 28 -b 96k -ab 64 -acodec libfaac -ac 1 -ar 44100 263.flv [17:24] <JEEB> so you want to use that sorenson H.263 crap [17:24] <JEEB> and use faac for AAC [17:24] <kuberan> ok :) [17:25] <JEEB> interesting combination [17:25] <JEEB> is there any reason you'd not use H.264? [17:26] <kuberan> A website setup of my friend is breaking because of wrong FFMPEG installation. [17:26] <kuberan> Just needed to fix it. [17:26] <kuberan> to make that command work. [17:26] <JEEB> is there any reason you're not using libx264 instead of flv? [17:27] <JEEB> I'll stop asking if you just want that to work [17:27] <kuberan> We will be using it to stream [17:27] <kuberan> using Flash plugin. [17:27] <kuberan> stream it on a website. [17:27] <JEEB> (just saying that libx264 will get you generally much more quality per the same bitrate) [17:27] <JEEB> (and flash supports H.264 as well as Sorenson's H.263 look-a-like) [17:28] <kuberan> thats great :) [17:28] <kuberan> would love to try that out. [17:28] <JEEB> anyways [17:28] <JEEB> build and install current x264 and faac [17:28] <JEEB> first [17:29] <kuberan> ok. [17:29] <kuberan> then how do i build ffmpeg ? using what config options ? [17:29] <JEEB> tell me when you've got faac and x264 built and installed first [17:29] <kuberan> sure. thanks :) [17:31] <JEEB> ./configure --enable-gpl --enable-nonfree --enable-libx264 --enable-libfaac will be the command line then pretty much [17:33] <kuberan> great thanks :) [17:33] <kuberan> is it ok to build x264 with ASM disabled ? [17:34] <JEEB> no [17:34] <JEEB> install current yasm [17:34] <JEEB> http://yasm.tortall.net/Download.html [17:34] <JEEB> 1.2 is newest [17:34] <kuberan> fine thanks. [17:34] <JEEB> x264 with asm = LOL SLOW [17:34] <kuberan> oh. [17:35] <JEEB> I meant [17:35] <JEEB> without [17:35] <JEEB> >_> [17:35] <JEEB> with asm it gets fast [17:35] <JEEB> it's one of the fastest encoders around [17:36] <kuberan> you mean x264+yasm is one of the fastest rite ? [17:36] <JEEB> yes [17:36] <JEEB> ffmpeg uses yasm as well, but doesn't necessarily need as new version of it [17:37] <JEEB> so having the newest yasm is always a good idea :3 [17:39] <kuberan> great. [17:46] <kuberan> I am installing x264 with ./configure --enable-shared [17:46] <kuberan> mkae fails with error [17:47] <kuberan> common/x86/cabac-a.o: could not read symbols: Bad value [17:47] <JEEB> Works For Me [17:47] <kuberan> fails with error http://pastebin.com/fFjAfDhw [17:47] <JEEB> current git? [17:47] <kuberan> yes. [17:48] <JEEB> and uh [17:48] <JEEB> make distclean once? [17:48] <JEEB> then re-configure and rebuild? [17:48] <kuberan> ok :) [17:48] <kuberan> thanks a lot. [17:48] <JEEB> if it still does it you'll have to add -fPIC [17:48] <JEEB> just like the message says >_> [17:50] <TomCat2> hey guys - I'm using avcodec as a library to an app that needs h.264 decoding capabilities. Ideally, avcodec's functionality will be invoked via Java through JNI. Just as a test, I have a function in C that requires no parameters - it loads an .h264 (previously created with x264) off of disk and feeds it to avcodec_decode_video2(). If I call it just using C, all works fine. If I invoke my [17:50] <TomCat2> function via JNI, I get output to my console which looks like: [h264 @ 18db0a80] get_buffer() failed (-1 0 0 00000000) [h264 @ 18db0a80] get_buffer() failed (-1 0 0 00000000) [h264 @ 18db0a80] get_buffer() failed (-1 0 0 00000000) [17:52] <TomCat2> This is using Windows 7 as the OS - I'm suspecting that maybe a different flavor of h264 decoder is getting used while using JNI due to maybe some DLL is not being found under that context to deal with hardware acceleration, just a wild theory - but don't have any idea how to check that. Any ideas? [17:58] <kuberan> looks like youtube uses H.263 [17:58] <kuberan> JEEB: [17:59] <kuberan> thats what the wikipedia page says. [18:00] <JEEB> kuberan, no -- they use H.264 [18:00] <JEEB> they use libx264 actually :P [18:00] <kuberan> thats great :) [18:00] <JEEB> they used to use H.263 a long time ago [18:00] <kuberan> oh ok :) [18:00] <kuberan> thanks.. [18:00] <kuberan> Is there any advantage for H.263 than H.264.. [18:01] <kuberan> something like lower bitrate or something ? [18:01] <JEEB> H.263 is much older and worse off than H.264 [18:01] <JEEB> don't use H.263 if you only can help it [18:02] <JEEB> pretty much: if you need to support random bad phones from the beginning of the 2000s, then H.263 might get useful [18:02] <JEEB> but I don't think flv is used or that those phones could use Flash :P [18:03] <kuberan> ok :) [18:03] <kuberan> makes sense :) [18:03] <kuberan> Getting this error when I run ffmpeg -codecs now [18:03] <kuberan> http://pastebin.com/G83BzrX5 [18:04] <JEEB> it can't load the binary you built for x264 then >_> [18:04] <JEEB> s/binary/library/ [18:04] <JEEB> I would just recommend building everything as static [18:04] <kuberan> oh.. [18:04] <JEEB> and then there would be no such problem :P [18:04] <kuberan> ok :) [18:04] <JEEB> now you have to tell the system where to load the libs [18:05] <JEEB> which is a gigantic PITA [18:05] <kuberan> I know where the shared object is [18:05] <kuberan> can i provide the path to the binary ? [18:06] <JEEB> LD_LIBRARY_PATH was it? [18:06] <JEEB> I don't remember [18:06] <kuberan> I will better build it as static. [19:18] <alyawn> In my little test filter, I always seem to get "*** 1 dup!" on the very first frame no matter what. Is this because the pts of my first input is 0? [19:23] <DX099> hello , how do i copy input file name to output file name ? [19:26] <alyawn> ok, so I found out that ffmpeg.c:1270 reports a duration of 2 for the first frame. I was expecting to find a duration of 1 [20:12] <bf4648> Got a question...I need to start ffmpeg from php...how do I do that? I've tried <?php exec("ffmpeg -i myvid.avi");...anybody know? [20:13] <bf4648> Anybody there? [20:15] <pasteeater> kuberan: see the ubuntu ffmpeg compile guide if you want instrucitons [20:15] <pasteeater> http://ubuntuforums.org/showthread.php?t=786095 [20:16] <kuberan> thnx a lot :) [20:16] <kuberan> will check [20:22] <alyawn> I suspect this may be my problem: parser not found for codec huffyuv, packets or times may be invalid. [20:22] <alyawn> is there a better container for huffyuv other than AVI? [20:24] <relaxed> alyawn: matroska [20:25] <relaxed> also, ffvhuff may be faster than huffyuv [20:42] <bf4648> Alright, I've got several videos .mp4s that need to be combined together so that they can play on a DVD...do you all know how I could do this? [20:44] <relaxed> you mean as different titles on the dvd or one long stream of video? [20:45] <bf4648> no...I just got clips of a videos that need to be all combined together.. [20:47] <relaxed> 1) Encode the mp4s to mpeg2/ac3 in the vob containerwith the same settings. 2) ffmpeg -i concat:1.vob\|2.vob\|3.vob -vcodec copy -acodec copy combined.vob [20:48] <bf4648> ok...I see [20:48] <bf4648> do it's just piped [20:49] <relaxed> it's not piped [20:49] <bf4648> delimited by the | [20:49] <relaxed> right [20:49] <bf4648> or \| [20:50] <bf4648> or is the "\" part of the \| just the slash for the file location? [20:51] <relaxed> no, it's escaping the | so as not to be treated like a pipe by your shell [20:51] <relaxed> you could also use -i 'concat:1.vob|2.vob|3.vob' [20:52] <bf4648> Oh I'm glad I asked...this will be running inside php exec call... [20:52] <bf4648> so I will have to escape the | delimiter right? [20:52] <relaxed> yes [20:52] <bf4648> Alright...thank you very much for your help [20:55] <alyawn> thanks, relaxed [20:59] <relaxed> you're welcome [21:01] <bf4648> relaxed...I got another question... [21:03] <bf4648> I've got this web app that is storing the start and offsets for interesting clips...it's actually not storing the actually clips just the point @ which the videos are to play....the videos are stored via paths in the database...would I have to call ffmpeg @ the offsets as input values? [21:11] <bf4648> Does anyone here know how to load ffmpeg @ an offset? [21:11] <relaxed> so you need ffmpeg to encode certain sections of videos? [21:11] <bf4648> well, I need to have ffmpeg go to the server get the video...load up @ an offset to a certain length in the video and then combine all the clips into 1 big file that is playable on an DVD [21:12] <bf4648> sounds like a lot of work [21:12] <bf4648> lol [21:12] <blez> I think it can't load the video from offset [21:12] <relaxed> Look at -t and -ss in the man page [21:12] <blez> only seconds [21:13] <bf4648> no way...hold on [21:15] <bf4648> -itsoffset offset (input) [21:15] <DX099> hello , how do i copy input file name to output file name ? [21:16] <bf4648> DX099...huh? [21:17] <pasteeater> bf4648: you probably want -ss, as relaxed mentioned, not -itsoffset [21:17] <bf4648> what is -itsoffset then? .. [22:10] <DX099> I want the output file name to be the exact copy of the input file name, is there a way to do that ? [22:13] <bf4648> yes... ffmpeg -i output.avi output.mp4 [22:19] <TomCat2> can someone who's familiar with using JNI and avcodec take a look please? This is boggling me: http://pastebin.com/7S0uLmHL [22:22] <TomCat2> basically - 1 DLL, 1 Java app, 1 C app. The C app can call functionality in the DLL just fine, but the Java app calls the same thing via a pretty plain JNI function, and it doesn't work, as shown in the pastebin post [22:23] <TomCat2> we're getting other stuff (swscale, x264) working just fine with JNI otherwise [22:24] <TomCat2> here in this minimalist example, we've narrowed the problem down to it just being whether or not JNI fires it - so we've ruled out any issues with using ByteBuffer, etc., as being the problem [22:41] <DX099> you mean i've to enter it manually ? [22:57] <teratorn> TomCat2: seen xuggle? [22:58] <TomCat2> yes - but for our purposes, we'll have major distribution issues [22:59] <TomCat2> plus it just adds an extra layer that we don't really need [23:00] <TomCat2> applets... the fewer DLL's you're trying to ship with it, the better [23:04] <Zeranoe> If -threads 0 is used and no threads are found will it default to 1? I know it is supposed to use the max, what about only 1? [23:05] <Zeranoe> or will FFmpeg complain and fail? [23:09] <TomCat2> "Can I use Xuggler with Java Webstart or as an Applet?" [23:09] <TomCat2> A: "No, not today. We're interested in approaches to solving this problem, and if you have experience with making native libraries work in webstart (especially ones that have runtime dependencies you can't know in advance) we'd like to hear from you. " [23:11] <relaxed> Zeranoe: With libx264, yes (former). I think it will fail if you use any other codec. [23:24] <pasteeater> '-threads 0' is now aotumatically applied when using libx264, IIRC [23:28] <pasteeater> relaxed: i meant to say, "pkgconfig is _not_ installed on CentOS 6 minimal" in the email, but again, the "tip" should cover that [23:41] <JEEBsv> there's actually two -threads nowadays, no? First before the -i (decoder threads, zero tends to take the decoder to a moonlight cruise), and the second for the encoder, where at least with libx264 zero is both default as well as 'auto' [23:52] <relaxed> JEEBsv: Yes. I assumed he meant threads for encoding. [23:54] <relaxed> pasteeater: I have it checking for pkg-config. Need to make some more changes before uploading version 2. [00:00] --- Wed Dec 7 2011

1 0

ffmpeg-devel.log.20111205
by burek 06 Dec '11

06 Dec '11

[00:12] <ubitux> do we have any video filter to do some color substitution? [00:12] <ubitux> like replacing a certain color with another given a tolerance ratio [00:14] <michaelni> mp=geq could surely do that somehow [00:14] <michaelni> but its not going to be fast [00:15] <ubitux> ok [00:43] <CIA-41> ffmpeg: 03Nathan Adil Maxson 07master * r6eedeefe0b 10ffmpeg/libavcodec/xan.c: [00:43] <CIA-41> ffmpeg: xan.c: coding style clean-up. [00:43] <CIA-41> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje(a)gmail.com> [00:43] <CIA-41> ffmpeg: 03Tsvetelina Borisova 07master * r4a59eca97a 10ffmpeg/libavutil/pixdesc.c: [00:43] <CIA-41> ffmpeg: Cleaned pixdesc.c file in libavutil [00:43] <CIA-41> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje(a)gmail.com> [00:43] <CIA-41> ffmpeg: 03David Goldwich 07master * re96070074d 10ffmpeg/libavformat/oma.c: [00:43] <CIA-41> ffmpeg: oma: clearify ambiguous if condition [00:43] <CIA-41> ffmpeg: Signed-off-by: David Goldwich <david.goldwich(a)gmail.com> [00:43] <CIA-41> ffmpeg: Signed-off-by: Anton Khirnov <anton(a)khirnov.net> [00:43] <CIA-41> ffmpeg: 03Derek Buitenhuis 07master * ra548b6cbfc 10ffmpeg/libavcodec/wavpack.c: [00:43] <CIA-41> ffmpeg: wavpack: Properly clip samples during lossy decode [00:43] <CIA-41> ffmpeg: When decoding lossy WavPack samples, they are supposed [00:43] <CIA-41> ffmpeg: to be clipped, in order to be decoded correctly. [00:43] <CIA-41> ffmpeg: Signed-off-by: Derek Buitenhuis <derek.buitenhuis(a)gmail.com> [00:43] <CIA-41> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje(a)gmail.com> [00:43] <CIA-41> ffmpeg: 03David Goldwich 07master * r8ae5eb75df 10ffmpeg/libavformat/oma.c: [00:43] <CIA-41> ffmpeg: oma: better format detection with small probe buffer [00:43] <CIA-41> ffmpeg: Signed-off-by: David Goldwich <david.goldwich(a)gmail.com> [00:43] <CIA-41> ffmpeg: Signed-off-by: Anton Khirnov <anton(a)khirnov.net> [00:43] <CIA-41> ffmpeg: 03Reimar Döffinger 07master * r7679bd6e60 10ffmpeg/libavcodec/roqvideodec.c: [00:43] <CIA-41> ffmpeg: roqvideodec: set AVFrame reference before reget_buffer. [00:43] <CIA-41> ffmpeg: Otherwise it might return a write-only frame which would break [00:43] <CIA-41> ffmpeg: decoding completely. [00:43] <CIA-41> ffmpeg: Signed-off-by: Anton Khirnov <anton(a)khirnov.net> [00:43] <CIA-41> ffmpeg: 03Carl Eugen Hoyos 07master * r1e92d58e44 10ffmpeg/libavcodec/bmp.c: [00:43] <CIA-41> ffmpeg: bmp: fix some 1bit samples. [00:43] <CIA-41> ffmpeg: <= 8bpp BMP images always have palette. [00:43] <CIA-41> ffmpeg: Signed-off-by: Anton Khirnov <anton(a)khirnov.net> [00:44] <CIA-41> ffmpeg: 03Chris Berov 07master * ra4e21baa74 10ffmpeg/libavformat/yuv4mpeg.c: [00:44] <CIA-41> ffmpeg: yuv4mpeg: cosmetics: pretty-printing [00:44] <CIA-41> ffmpeg: Signed-off-by: Justin Ruggles <justin.ruggles(a)gmail.com> [00:44] <CIA-41> ffmpeg: 03Chris Berov 07master * rdc6d043050 10ffmpeg/libavcodec/adpcmenc.c: [00:44] <CIA-41> ffmpeg: adpcmenc: cosmetics: pretty-printing [00:44] <CIA-41> ffmpeg: Signed-off-by: Justin Ruggles <justin.ruggles(a)gmail.com> [00:44] <CIA-41> ffmpeg: 03Chris Berov 07master * r541d083a40 10ffmpeg/libavcodec/ac3dec.c: [00:44] <CIA-41> ffmpeg: ac3dec: cosmetics: pretty-printing [00:44] <CIA-41> ffmpeg: Signed-off-by: Justin Ruggles <justin.ruggles(a)gmail.com> [00:44] <CIA-41> ffmpeg: 03Konstantin Todorov 07master * r363114e813 10ffmpeg/libavcodec/mpegvideo.c: [00:44] <CIA-41> ffmpeg: mpegvideo.c: code cleanup - first 500 lines. [00:44] <CIA-41> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje(a)gmail.com> [00:44] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r707138593a 10ffmpeg/: (log message trimmed) [00:44] <CIA-41> ffmpeg: Merge remote-tracking branch 'qatar/master' [00:44] <CIA-41> ffmpeg: * qatar/master: [00:44] <CIA-41> ffmpeg: adpcmenc: cosmetics: pretty-printing [00:44] <CIA-41> ffmpeg: ac3dec: cosmetics: pretty-printing [00:44] <CIA-41> ffmpeg: yuv4mpeg: cosmetics: pretty-printing [00:44] <CIA-41> ffmpeg: shorten: remove dead initialization [00:45] <CIA-41> ffmpeg: 03Nathan Adil Maxson 07master * r8f8c785ffc 10ffmpeg/libavcodec/zmbv.c: [00:45] <CIA-41> ffmpeg: zmbv.c: coding style clean-up. [00:45] <CIA-41> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje(a)gmail.com> [00:45] <CIA-41> ffmpeg: 03Yordan Makariev 07master * r425b45d4b0 10ffmpeg/libavutil/ (crc.c lfg.c log.c random_seed.c rational.c tree.c): [00:45] <CIA-41> ffmpeg: Code clean-up for crc.c, lfg.c, log.c, random_see.d, rational.c and tree.c. [00:45] <CIA-41> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje(a)gmail.com> [05:04] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r738c603a22 10ffmpeg/ffmpeg.c: [05:04] <CIA-41> ffmpeg: ffmpeg: Avoid directly accessing AVFrame fields that differ between forks. [05:04] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [05:04] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rdff97be808 10ffmpeg/libavutil/ (opt.c opt.h): [05:04] <CIA-41> ffmpeg: av_opt: add av_opt_ptr() to return a pointer to a field of a object based on [05:04] <CIA-41> ffmpeg: an AVClass [05:04] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [05:04] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rfef411ef3b 10ffmpeg/libavcodec/ (avcodec.h options.c): [05:04] <CIA-41> ffmpeg: lavc: Add a AVClass for AVFrames. [05:04] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [05:05] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r618fb9cc06 10ffmpeg/ffmpeg.c: [05:05] <CIA-41> ffmpeg: ffmpeg: handle stream copy timebase selection for mp4 as if it had VFPS set. [05:05] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [05:05] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r406e9647a0 10ffmpeg/ffplay.c: [05:05] <CIA-41> ffmpeg: ffplay: Avoid directly accessing AVFrame fields that differ between forks. [05:05] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [10:36] <Node_140> while decodeing video from mjpeg to mpeg4 , i get mjpeg: unsupported coding type (ce)? thx.... [13:31] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rb9d2b45165 10ffmpeg/libavcodec/avcodec.h: [13:31] <CIA-41> ffmpeg: lavc: Document AVFrame av_opt_ptr() relation. [13:31] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [13:32] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r31f93fc5f8 10ffmpeg/libavcodec/avcodec.h: [13:32] <CIA-41> ffmpeg: lavc: Fix sizeof(AVFrame) useability docs [13:32] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [13:32] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r51177b1fe9 10ffmpeg/libavcodec/avcodec.h: [13:32] <CIA-41> ffmpeg: lavc: Mark fields that require av_opt_ptr() [13:32] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [16:21] <ubitux> michaelni: i did the changes you asked to the thumbnailer filter but it does not seem to output anything when calling avfilter_{start_frame,draw_slice,end_frame} in the end_frame() callback (when a thumbnail is selected) [16:21] <ubitux> basically, i have: [16:21] <ubitux> avfilter_start_frame(inlink, thumb->frames[best_frame].buf); [16:21] <ubitux> avfilter_draw_slice(inlink, 0, inlink->h, 1); [16:21] <ubitux> avfilter_end_frame(inlink); [16:21] <ubitux> thumb->frames[best_frame].buf being the AVFilterBufferRef for the best frame [16:22] <ubitux> (now draw_slice() callback just compute the histograms and doesn't call anything) [16:22] <ubitux> (and i use an empty function for the start_frame callback) [17:10] <ubitux> mmh i needed to use the outlink actually. [20:06] <CIA-41> ffmpeg: 03Reimar Döffinger 07master * r840ecc9e07 10ffmpeg/ (Makefile common.mak configure): [20:06] <CIA-41> ffmpeg: Add coverage support. [20:06] <CIA-41> ffmpeg: Adds --enable-coverage to configure and a "coverage-html" make target. [20:06] <CIA-41> ffmpeg: The dependency stuff in the Makefile is a bit questionable, but the [20:06] <CIA-41> ffmpeg: best I could think of so far. [20:06] <CIA-41> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger(a)gmx.de> [20:06] <CIA-41> ffmpeg: 03Reimar Döffinger 07master * rc66bd78e3d 10ffmpeg/ (7 files in 2 dirs): [20:06] <CIA-41> ffmpeg: vf_scale: support PAL8 output by producing BGR8. [20:06] <CIA-41> ffmpeg: While quality is bad, PAL8 support is needed to allow testing some [20:06] <CIA-41> ffmpeg: encoders that only support PAL8 input. [20:06] <CIA-41> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger(a)gmx.de> [20:06] <CIA-41> ffmpeg: 03Reimar Döffinger 07master * r61b1d85c1f 10ffmpeg/tests/ (9 files in 3 dirs): [20:06] <CIA-41> ffmpeg: Add tests for several encoders. [20:06] <CIA-41> ffmpeg: More specifically, PNG, v210, zlib and zmbv codecs. [20:06] <CIA-41> ffmpeg: zmbv needs vf_scale to be able to produce PAL8. [20:06] <CIA-41> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger(a)gmx.de> [20:06] <CIA-41> ffmpeg: 03Reimar Döffinger 07master * rd086c1203c 10ffmpeg/libavutil/ (adler32.c aes.c base64.c des.c eval.c): [20:07] <CIA-41> ffmpeg: Add coverage exclusions for test code. [20:07] <CIA-41> ffmpeg: For some of the code e.g. doing timing measurements there is no [20:07] <CIA-41> ffmpeg: real point in running regression testing on it, thus it should [20:07] <CIA-41> ffmpeg: not be counted against coverage. [20:07] <CIA-41> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger(a)gmx.de> [20:31] <Compn> JEEB : why does libass require fribidi ? [20:31] Action: Compn kicks JEEB [20:31] <Compn> who do i need to kick to get this fixeds [20:31] <JEEB> oi [20:31] <nevcairiel> it needs it for bi-directional font support, of course [20:32] <Compn> of course [20:32] <JEEB> zgreg is the person who develops libass, just to note [20:32] <Compn> ah, i've messaged him , waiting reply [20:32] <Compn> saw your name and libass thought i'd bug you :) [20:34] <Compn> wonder if mingw or other systems can handle fribidi [20:34] <Compn> guess mingw works [21:14] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rb27ac355b7 10ffmpeg/libavformat/mov.c: [21:14] <CIA-41> ffmpeg: movdec: Fix parsing of a very last empty atom of size 8. [21:14] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [21:14] <CIA-41> ffmpeg: 03Gavin Kinsey 07master * r4547d883d3 10ffmpeg/libavutil/opt.h: [21:14] <CIA-41> ffmpeg: Fix compilation for C++ applications [21:14] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [23:36] <durandal_1707> found yet another forgotten demuxer mthp [00:00] --- Tue Dec 6 2011

1 0

ffmpeg.log.20111205
by burek 06 Dec '11

06 Dec '11

[00:20] <pasteeater> relaxed: nice script. i didn't get much of a chance to look over and try it yet though, but i'll let you know when I do. [04:59] <relaxed> sacarasc: thanks [06:20] <DGMurdockIII> http://download.macromedia.com/pub/labs/cinemadng/cinemadng_spec_090211.pdf [09:50] <blackbetty> how can I combine mpegs? [09:50] <blackbetty> cat doesnt work [09:51] <relaxed> blackbetty: did you try -i concat: ? [09:51] <blackbetty> no, I will. I have been like beating my head into the wall for the past 4 hours [09:54] <blackbetty> how would I use that [09:56] <relaxed> lets see "ffmpeg -i input.mpg" output on pastebin. [09:57] <blackbetty> relaxed: http://pastebin.com/m1r9juHi [09:58] <relaxed> blackbetty: ffmpeg -i concat:1.mpg\|2.mpg -vcodec copy -acodec copy -f mpeg combined.mpg [10:00] <blackbetty> relaxed: you are a god [10:00] <blackbetty> i have literly been searching for hours [10:00] <relaxed> glad I could help. [10:01] <blackbetty> everywhere points me to cat file1.mpg file2.mpg > output.mpg [10:01] <blackbetty> and that just doesn't work [10:53] <twid> while decodeing video from mjpeg to mpeg4 , i get mjpeg: unsupported coding type (ce)? thx.... [15:20] <wpl> Is it safe to call audio decoding functions from several threads? I'm using avcodec-53.5. [15:23] <gfto> Guys I'm remuxing mpegts files into mpegps (-f vob) and the problem I'm having is that mediainfo reports for the audio "Delay relative to video: -146ms" - http://pastebin.com/7VPuTJvx Is there a way for this delay to be 0 /without recompressing/? [15:29] <gfto> the same as above but done with ffmpeg: http://pastebin.com/88yPtBKC (result is the same). [15:53] <relaxed> gfto: Is the audio in sync with the video? [15:54] <relaxed> gfto: the correct way would be, ffmpeg -i concat:file1.ts\|file2.ts\|file3.ts -vcodec copy -acodec copy -f vob combined.mpg [15:57] <relaxed> or avconv * [16:27] <gfto> relaxed: concat or not it doesn't matter in this case, yes the audio and video are in sync when played back or transcoded but transcoding duplicates 3-4-5 times the first frame. [18:29] <elderg> I'm trying to letterbox a 2:1 video to fit in a 16:9. I would used to use -padtop and -padbottom but now it says I need to use the pad filter. I don't know how to use libavfilter. How do I do this? [20:46] <burek> elderg [20:47] <pasteeater> he's gone, jim. [21:06] <Nagy> I keep getting this global network initialization warning, what is it and how do I get rid of it? [21:11] <pasteeater> Nagy: why didn't you show your command and the complete console output on a pastebin service? [21:11] <pasteeater> we can make guesses, but we aren't psychic [21:12] <pasteeater> help us help you by providing all relevant and useful information [22:13] <soda> hey guys, im trying to cut out a peace from a vmw file, but getting some not so good output -> http://pastebin.com/zGLGg98J [22:14] <soda> messing around with that stuff for over an hour now, but not really an idea where the problem is located... [22:16] <pasteeater> soda: does it work as expected if you move -ss so it's applied as an output option? [22:17] <pasteeater> (after -i input) [22:17] <soda> Invalid duration specification for ss: -acodec [22:18] <pasteeater> ffmpeg -t 00:01:00 -i foo.wmv -ss 00:03:00 -acodec copy -vcodec copy bar.wmv [22:19] <soda> okay, this seems to work [22:19] <soda> let me check the man page... [22:21] <pasteeater> still shouldn't crash though. do you have a sample that i can attempt to duplicate the issue? [22:23] <pasteeater> -ss after -i is different than -ss before -i [22:23] <pasteeater> http://ffmpeg.org/pipermail/ffmpeg-user/2011-November/003221.html [22:35] <soda> pasteeater, i cant give you the file, sorry, and i cant find any comparable file on google neither :/ [22:39] <soda> thank you for your help :) [00:00] --- Tue Dec 6 2011

1 0

ffmpeg-devel.log.20111204
by burek 05 Dec '11

05 Dec '11

[00:25] <ubitux> Compn: it did the trick: shift+c on each patch mail, appending them to the new created mbox (when saving the first one), and then git am /path/to/mail/mboxname did the trick [00:25] <ubitux> thanks [00:26] <ubitux> it's certainly possible to do better, but sounds efficient enough to me ATM :) [00:26] <Compn> whatever works i guess :) [00:27] <ubitux> the sbg patches are huge : [00:27] <ubitux> :o [00:28] <ubitux> libavformat/sbgdec.c | 1518 ++++++++++++++++++++++++++++++++++++++++++++++ [00:28] <ubitux> libavcodec/ffwavesynth.c | 421 ++++++++++++++++++++++++++++++++++++++++++++++ [00:28] <ubitux> libavutil/noise.c | 273 ++++++++++++++++++++++++++++++++++++++++++++++++++++ [00:28] <ubitux> @_@ [00:32] <Compn> did you see what it is ? [00:32] <Compn> would you rather a wrapper to that sbg lib ? [00:33] <saste> ubitux: http://thread.gmane.org/gmane.comp.video.ffmpeg.devel/129833 [00:33] <ubitux> oh hi saste, didn't see you :) [00:33] <ubitux> saste: nice, thx [00:34] <ubitux> saste: btw, do you plan to commit lifesrc soon? :] [00:34] <ubitux> Compn: no idea, i'm testing right now [00:34] <ubitux> the sbg sample in the documentation sounds a bit creepy to me [00:34] <Compn> hahaha [00:34] <Compn> are you listening to it over headphones ? [00:35] <saste> ubitux: ah, i was working on it right now [00:35] <saste> it still misses one feature and docs update [00:35] <ubitux> Compn: yep :p [00:35] <ubitux> sounds like "woOooOoOoooooOo" [00:35] <ubitux> i feel i'm going crazy :p [00:35] <Compn> haha [00:36] <Compn> he did say 'try to relax' [00:36] <Compn> you sure its the right speed ? [00:36] <ubitux> saste: ok nice :) [00:36] <ubitux> Compn: yeah sure [00:36] <saste> ubitux: do you want me to send and updated WIP patch? [00:36] <ubitux> i reminds me creepy ambiant in horror films [00:37] <ubitux> saste: no hurry :) [00:37] <ubitux> saste: i think you can commit directly when you think it's ok [00:37] <Compn> ubitux : post an mp3 for us to listen to :P [00:37] <saste> ubitux: no it's no fun if no one says "ok" [00:37] <ubitux> saste: haha ok ;) [00:38] <ubitux> Compn: i'll upload a flac, give me a few minutes [00:39] <ubitux> oh it's 30minutes long [00:40] <ubitux> there is no chance i'd feel relaxed with this :p [00:45] <ubitux> Compn: http://gao.pkh.me/sbg.flac [00:46] <Compn> i hear nothing! [00:46] <ubitux> just raise the volume :p [00:46] <Compn> oh [00:47] <Compn> sounds like some fan missing a ball bearing [00:47] <ubitux> or a grenade just exploded next to you [00:47] <Compn> lol [00:47] <Compn> half life 2 ftw [00:47] <ubitux> yeah ;) [00:49] <Daemon404> half life 1 ftw [00:51] <ubitux> there is no after-grenade effect :p [01:05] <ubitux> Compn: so, how do you feel? [01:05] <Compn> i turned it off after that second noise came in [01:06] <Compn> but also i didnt use headphones [01:06] <Compn> just using speakers [01:06] <Compn> so its an incomplete test :) [01:08] <Daemon404> ubitux, yet but you get to go to Xen! [01:09] <CIA-41> ffmpeg: 03Luca Barbato 07master * ra2fb4bcb01 10ffmpeg/libavfilter/vf_drawtext.c: [01:09] <CIA-41> ffmpeg: drawtext: refactor draw_text [01:09] <CIA-41> ffmpeg: Split the memory allocation from the actual drawing. [01:09] <CIA-41> ffmpeg: 03Janne Grunau 07master * r84994421bc 10ffmpeg/doc/developer.texi: doc: remove space between variable and post increment in example code [01:09] <CIA-41> ffmpeg: 03Luca Barbato 07master * ra0338b598b 10ffmpeg/doc/developer.texi: doc: break some long lines in developer.texi [01:09] <CIA-41> ffmpeg: 03Luca Barbato 07master * r2cf74eca70 10ffmpeg/ (doc/filters.texi libavfilter/vf_drawtext.c): [01:09] <CIA-41> ffmpeg: drawtext: make x and y parametric [01:09] <CIA-41> ffmpeg: Introduce variables "E", "PHI", "PI", "main_w"/"W", "main_h"/"H", [01:09] <CIA-41> ffmpeg: "text_w"/"w", "text_h"/"h", "x", "y", "n" and "t" in line with [01:09] <CIA-41> ffmpeg: vf_overlay and refactor the code accordingly. [01:09] <CIA-41> ffmpeg: 03Derek Buitenhuis 07master * r55a280569f 10ffmpeg/tests/ (fate2.mak ref/fate/vble): [01:09] <CIA-41> ffmpeg: fate: Add a test for the VBLE decoder [01:09] <CIA-41> ffmpeg: Signed-off-by: Derek Buitenhuis <derek.buitenhuis(a)gmail.com> [01:09] <CIA-41> ffmpeg: Signed-off-by: Janne Grunau <janne-libav(a)jannau.net> [01:09] <CIA-41> ffmpeg: 03Ronald S. Bultje 07master * radedd840e2 10ffmpeg/libavcodec/ (h264.c h264.h h264_refs.c): [01:09] <CIA-41> ffmpeg: h264: fix frame reordering code. [01:09] <CIA-41> ffmpeg: Fixes fate-h264-conformance-{mr2_tandberg_e,mr3_tandberg_b} without [01:09] <CIA-41> ffmpeg: requiring -strict 1. [01:09] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r7f92f3d812 10ffmpeg/: (log message trimmed) [01:09] <CIA-41> ffmpeg: Merge remote-tracking branch 'qatar/master' [01:09] <CIA-41> ffmpeg: * qatar/master: [01:09] <CIA-41> ffmpeg: h264: fix frame reordering code. [01:09] <CIA-41> ffmpeg: fate: Add a test for the VBLE decoder [01:09] <CIA-41> ffmpeg: doc: break some long lines in developer.texi [01:09] <CIA-41> ffmpeg: drawtext: make x and y parametric [01:09] <CIA-41> ffmpeg: 03Luca Barbato 07master * rec11ff8407 10ffmpeg/libavfilter/vf_drawtext.c: [01:09] <CIA-41> ffmpeg: drawtext: manage memory allocation better [01:09] <ubitux> Daemon404: the first half life was really great, but hl² is awesome too :) [01:10] <Daemon404> i love all of valves games [01:10] <Daemon404> (except CS) [01:10] <JEEBsv> oh boy, I remember when I played H-L for the first time [01:10] <Daemon404> i didnt play hl when it came out [01:10] <Daemon404> i was too busy with OOT [01:10] <Daemon404> :3 [01:10] <Compn> oot ? [01:10] <Daemon404> ocarina of time [01:10] <Compn> ah [01:11] <Compn> i just got hl for dreamcast , its got some nasty controller learning curve to deal with [01:11] <Compn> i've never played the first one. only seen speed runs and such [01:11] Action: Compn is a speed run watcher [01:12] <Daemon404> the only speed runs i ever watched were uh [01:12] <Daemon404> ninja gaiden black. [01:12] <Daemon404> and MN [01:12] <Daemon404> on MN* [01:12] <ubitux> the 45 min HL speedrun is insane :) [01:12] <Compn> i like it more when they dont do all bunny hops , its more like watching a movie [01:13] <Compn> since you cant skip cutscenes/ whatever those are called when the people talk to you [01:13] <Compn> really entertaining movie too :) [01:19] <ubitux> (hey the sbg sample also sounds like when you carry the railgun in q3) [01:19] <Compn> i'm more of a q2 kind of guy [02:47] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rfb266f5973 10ffmpeg/libavcodec/h264.c: [02:47] <CIA-41> ffmpeg: h264: show when has_b_frames is being increased. [02:47] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [02:47] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r297422b338 10ffmpeg/libavfilter/vf_drawtext.c: [02:47] <CIA-41> ffmpeg: vf_drawtext: Add main_w/h to improve compatibility with the modified UI [02:47] <CIA-41> ffmpeg: from the fork. [02:47] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [02:47] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rf88d5df383 10ffmpeg/libavfilter/vf_drawtext.c: [02:47] <CIA-41> ffmpeg: vf_drawtext: cosmetics to reduce diff to qatar [02:47] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [14:07] <Hadus> Hello [15:45] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r594090c6c2 10ffmpeg/ffmpeg.c: [15:45] <CIA-41> ffmpeg: ffmpeg: select best pixel format for the encoder if no exact match. [15:45] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [15:45] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rb4425f26bb 10ffmpeg/doc/filters.texi: [15:45] <CIA-41> ffmpeg: doc/filters.texi: Update drawtext docs so that the documented input width/height variables are compatible with qatar. [15:45] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [15:45] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rdda3d716d6 10ffmpeg/libavfilter/vf_drawtext.c: [15:45] <CIA-41> ffmpeg: vf_drawtext: support W/H parameters to reduce UI differences to qatar [15:45] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [16:58] <CIA-41> ffmpeg: 03Peter Ross 07master * rba8410cb44 10ffmpeg/ (5 files in 3 dirs): [16:58] <CIA-41> ffmpeg: Microsoft Windows ICO demuxer [16:58] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [21:20] <CIA-41> ffmpeg: 03Stefano Sabatini 07master * r2f8b6e909d 10ffmpeg/ (6 files in 3 dirs): lavfi: add life source [21:20] <CIA-41> ffmpeg: 03Stefano Sabatini 07master * r6c26fe8b5a 10ffmpeg/ (3 files in 2 dirs): aevalsrc: add duration option [21:22] <ubitux> \o/ [21:57] <rahulr92> Hi. I am new to open-source programming. I am interested in contributing to ffmpeg. Can anyone please tell me where to start? [21:59] <ubitux> is there something you don't like in ffmpeg? :) [21:59] <ubitux> (bug, lack of feature, &) [22:03] <rahulr92> Actually, I am new to ffmpeg also. I am a college student and found ffmpeg in last years GSoC accepted orgs list. Just now checked out the latest version from github, and installation is under way. :) [22:05] <ubitux> rahulr92: http://lists.mplayerhq.hu/pipermail/ffmpeg-devel/2011-October/116200.html [22:05] <ubitux> this might interest you [22:05] <Daemon404> http://wiki.multimedia.cx/index.php?title=Small_FFmpeg_Tasks [22:05] <Daemon404> or this. [22:06] <rahulr92> Thanks. I will go through both links. [22:40] <ubitux> saste: do you know a way to scale exactly? [22:41] <ubitux> i'm not able to get something neat scaling the life [22:41] <saste> ubitux: command? [22:42] <ubitux> ./ffplay -sws_flags bitexact -f lavfi life=s=30x20,scale=600:400 [22:42] <ubitux> doesn't seem to do the trick :( [22:46] <saste> ubitux: iirc that's a bug, ffplay doesn't accept swscale flags [22:46] <saste> or to say it better, it ignores them [22:46] <saste> (and it's not showing swscale opts in -h) [22:47] <saste> did you try with ffmpeg? [22:51] <ubitux> not yet [22:52] <ubitux> mmh how do you avoid the need of an input? [23:04] <michaelni> "ffplay -vf crop=32:32,scale=640:480:flags=16" works fine here :) [23:04] <michaelni> #define SWS_BITEXACT 0x80000 [23:06] <saste> yep, so the problem is that -sws_flags is ignored by scale [23:46] <ubitux> ah indeed the flags=16 as scale parameters does the trick [23:46] <ubitux> thx [00:00] --- Mon Dec 5 2011

1 0

ffmpeg.log.20111204
by burek 05 Dec '11

05 Dec '11

[00:46] <arpu> hjello [00:47] <arpu> is it possible to use swfscale -vf scale=320:240,format=rgb24 fo rthe output ? [00:50] <burek> arpu, did you try [00:51] <arpu> burek, yes i get this [00:51] <arpu> [buffer @ 0x22435e0] w:640 h:360 pixfmt:yuv420p [00:51] <arpu> [ffsink @ 0x22479a0] auto-inserting filter 'auto-inserted scaler 0' between the filter 'Parsed filter 1 format' and the filter 'out' [00:51] <arpu> [scale @ 0x2240100] w:640 h:360 fmt:yuv420p -> w:320 h:240 fmt:rgb24 flags:0x4 [00:51] <arpu> [scale @ 0x225cc40] w:320 h:240 fmt:rgb24 -> w:320 h:240 fmt:yuv420p flags:0x4 [00:51] <burek> can you please use pastebin.com [00:52] <arpu> first convert to rgb24 but than a nother convert to original yuv420p [00:52] <arpu> http://pastebin.com/N9UHmDE4 [00:53] <arpu> i know i can use -pix_fmt rgb24 but i need to use swfscale for converting [00:58] <burek> can you please provide complete command + output? [01:00] <arpu> http://pastebin.com/aXgL2dmC [01:08] <burek> i see [01:08] <burek> and what would -pix_fmt after -f rawvideo do? [01:09] <burek> also, did you try 320x180 [01:12] <arpu> http://paste.railsbox.eu/show/515/ [01:12] <arpu> looks good [01:12] <arpu> but the result on the /dev/video2 device is jerkling :-/ [01:12] <arpu> scampled [01:12] <arpu> works with mencoder [01:13] <arpu> http://paste.railsbox.eu/show/516/ [01:13] <arpu> this is the mencoder command that works fine [01:13] <pietr101> hello, does ffmpeg works with nvidia/vdpau? [01:17] <burek> arpu, did you check cpu [01:17] <burek> can it handle that task? [01:17] <arpu> burek, ? yes [01:18] <burek> also, you might use latest ffmpeg from git [01:18] <arpu> why ? [01:18] <burek> i can imagine a lot has changed in ffmpeg since 0.7.x [01:18] <arpu> hmm [01:18] <arpu> ok i will try [01:19] <arpu> but strange mencoder works but ffmpeg not [01:19] <burek> its not that hard really [01:19] <burek> and that way you are sure if something works incorrectly, that it's a bug :) [01:19] <arpu> ok i will built the source from svn [01:19] <burek> anyway, what is /dev/video2 [01:20] <burek> no, not svn [01:20] <arpu> v4l2loopback device [01:20] <burek> svn is deprecated i thing.. use git [01:20] <arpu> ah i see yes git [01:20] <burek> i see.. did you have to initialize it somehow? [01:21] <arpu> yes [01:21] <arpu> mom [01:21] <burek> maybe you should set -r 25 or such [01:21] <burek> before /dev/video2 [01:22] <arpu> http://paste.railsbox.eu/show/517/ [01:24] <burek> oh i see [01:24] <burek> it requires input to be 640x360 [01:24] <burek> in rgb24 [01:25] <burek> that's why it might be jerkling :) [01:25] <arpu> yes this is hardcoded i changed it to 320 240 [01:25] <burek> change it to 320x180 [01:26] <arpu> same result [01:29] <burek> what exactly is "jerkling" [01:31] <arpu> scrampled [01:35] <pietr101> hello, how do I generate N screenshorts evenly spaced from a movie file? with original resolution. [01:37] <arpu> burek, http://image-upload.de/uploaded/136676/kXt7UE/ [01:38] <burek> hmh [01:40] <cbsrobot> pietr101: ffmpeg -i input -vf "select=not(mod(n\,1000))" -qscale 1 output-%04d.jpg [01:41] <cbsrobot> will output every 1000th frame [01:45] <cbsrobot> ah too late [01:46] <arpu> burek, same with git HEAD [01:46] <arpu> does mencoder use the same swscale lib ? [01:49] <arpu> hmm no mencoder use the libswscale.so ffmpeg the built in swscale [02:39] <alex87> is there a setting that will allow me to play incomplete x264 files? i'd like to watch an encoded video before it's finished transferring from my server [03:26] <grepper> could anyone help with filters? What is wrong with this filter chain for current git ? (it was working in ffmpeg 0.8.7): -vf movie=bg.png [wm];[in][wm] overlay=0:0:1 [wm2]; [wm2] setdar=4:3 [out] [03:27] <grepper> the error I get is: [ffmpeg]: [overlay @ 0x9c73cc0] Missing key or no key/value separator found after key '1' [03:49] <burek> grepper, http://ffmpeg.org/ffmpeg.html#overlay [03:49] <burek> you have 3 arguments, only 2 are needed [04:05] <grepper> burek: ah. The third had to do with transparency before I believe ? [04:07] <grepper> burek: thanks a bunch, working again [04:08] <grepper> now to figure out how to make this backwards compatible in my script for older ffmpegs [04:10] <burek> :beer: :) [04:12] <grepper> gladly, if you are in Toronto :) [04:13] <grepper> unless you meant that as: beer is a vital ingredient in coding [04:14] <burek> well, it has vitamins ^^ [05:44] <Xgates> hey guys [05:44] <Xgates> say, all around, for needing ffmpeg for many different apps, is there one particular version that is all around a safer bet to use for better compatibility? [05:45] <Xgates> so far I ran across not being to compile against 0.8x so I compiled 0.7.8 and that worked out better... [06:20] <rahulr92> Hi, I am a linux newbie, trying to read ffmpeg docs. Can anyone please tell me how to view *.texi files (without all the tags). I tried installing texinfo and tex2html packages, but can't figure out how to use them. Thanks. [06:37] <Xgates> hey if we check out from git then what version/revision are we dealing with? I need to know a version/revision so I can package this and I'm not sure... [07:08] <ircdearia> Using `fffmpeg -an -s 1920x1080 -f x11grab -i :0.0 -s 1920x1080 -r 25 -vcodec libxvid -sameq v.avi the output video appears low quality. The pixels are not exactly identical to those that appear on my screen. I believe it's related to libxvid codec maybe? What codec can I use to preserve the exact color of all the pixels? [07:17] <ircdearia> There are a lot of `ffmpeg -codecs` to choose from. I'm not sure which one seems most appropriate. [07:18] <ircdearia> Ah, I shall output to png images [07:18] <sacarasc> ircdearia: It's often suggested that at first, one encodes to lossless h264 using libx264 and from there into lossy video, again using libx264. [07:20] <ircdearia> mm, thanks for tip. I will see if I can output to images though, because I'm taking each individual frame and making a gif animation anyhow. [07:33] <burek> ircdearia, use -f image2 bla%02d.gif [07:33] <burek> as an output format [07:33] <burek> hmm, can ffmpeg create animated gifs? [07:34] <ircdearia> burek, mm, interesting idea, but I would want to crop all the frames too. How about, does ffmpeg support xcf (the gimp)? [07:34] <burek> it has -vcodec gif [07:34] <burek> but im not sure if it knows how to write animated gif [07:35] <burek> oh yeah baby [07:35] <burek> it can! [07:35] <burek> :d [07:35] <burek> -f gif -vcodec gif output.gif [07:35] <burek> try that [07:35] <burek> :D [07:35] <ircdearia> mm, nice! In that case, I must lastly see if I can record only certain pixels of my desktop instead of the whole desktop [07:35] <burek> yes [07:35] <burek> correct your -i [07:35] <ircdearia> ah [07:36] <burek> it's a left-upper point [07:36] <burek> -s gives width x height [07:36] <burek> oh wait [07:36] <burek> -i :0.0+10,20 [07:37] <burek> 10=x, 20=y [07:37] <burek> http://ffmpeg.org/ffmpeg.html#X11-grabbing [07:37] <burek> there you go [07:38] <ircdearia> yay! ffmpeg is surely amazing! [07:39] <burek> :) [07:40] <burek> btw, you'll get huge gif that way [07:40] <burek> maybe the better would be to encode a video [07:42] <ircdearia> yeah, and then I will use `mplayer -vo png v.avi` to extract all the frames [07:42] <burek> try for example: ffmpeg -codecs | grep -i lossless [07:42] <burek> and try, i dunno.. i would try -f avi -vcodec camstudio [07:42] <burek> output.avi [07:43] <burek> thats also lossless codec, used for screen grabbing [07:43] <burek> so it can't be bad [07:43] <burek> wait, you need all the frames? [07:43] <burek> as images, i mean [07:45] <burek> anyway, need to go [07:45] <burek> syl :) o/ [10:08] <XeonBloomfield> Hello [10:09] <XeonBloomfield> When I am typing "-b 500" bitrate is set to "0 kb/s", why? [10:10] <relaxed> 500k [10:10] <XeonBloomfield> thx [10:10] <XeonBloomfield> now it works great [10:10] <XeonBloomfield> relaxed: thank you [10:10] <XeonBloomfield> have a nice day [10:10] <relaxed> you're welcome [10:12] <XeonBloomfield> How can I convert FLV to AVI lossless? [10:34] <relaxed> XeonBloomfield: why do you need avi? [10:34] <XeonBloomfield> relaxed: hmm... I wonder if PiTiVi will work at FLV directly [10:35] <XeonBloomfield> relaxed: I get 1:1 copy to AVI by "ffmpeg -i a.flv -vcodec copy a.avi" [10:36] <XeonBloomfield> relaxed: It works directly - AVI is unnecessary [10:37] <relaxed> Anyone want to be my guinea pig? https://sites.google.com/site/linuxencoding/install-script [10:38] <relaxed> sacarasc pasteeater ^^^ when you get a chance [11:02] <XeonBloomfield> relaxed: latest version of ffmpeg supports multi-core rendering? [11:06] <relaxed> multithreaded decoding, yes [11:06] <relaxed> ffmpeg has had support for a while now. [11:35] <ircdearia> booo [11:36] <ircdearia> It appears not a single video codec is available to capture the exact pixels of X desktop: http://privatepaste.com/75ada99a74 [11:36] <ircdearia> I captured 3 frames of video using all the video codecs, extracted png image and compared to png image produced with `scrot` and not a single image matched, not even closely. [11:42] <JEEB> ircdearia, use something capable of RGB if you want exactness [11:42] <JEEB> also lol sameq [11:43] <JEEB> most abused command ever [11:43] <JEEB> because it doesn't do what the documentation used to say, now at least the documentation is changed :P [11:44] <ircdearia> mm, yeah, I didn't bother to remove it. [11:44] <ircdearia> I see that using vcodec png accomplishes what I wanted. [11:44] <JEEB> which is RGB [11:44] <JEEB> unsurprisingly [11:44] <JEEB> I think that at least one or two of the lossless encoders could do RGB as well [11:45] <JEEB> (lossless video that is) [11:45] <JEEB> x264 can do RGB-in-4:4:4 but I'm not sure if it's fast enough for live capture [11:46] <JEEB> anyways, you'd have to see which lossless video codecs support RGB [11:46] <JEEB> if you want 1-to-1 [11:50] <relaxed> -vcodec ffvhuff [11:51] <JEEB> yeah [11:51] <JEEB> I thought ffvhuff could do RGB [11:51] <JEEB> relaxed, also do you remmeber if there's a way to make ffmpeg/avconv convert YUV<->RGB with the BT.709 matrix [11:51] <JEEB> and not the BT.601 one [11:51] <relaxed> Here it says input and output are bgra. [11:52] <JEEB> great [11:56] <relaxed> maybe -color_primaries bt709 [11:57] <JEEB> I'll try later [11:58] <JEEB> so far I've just made people output to RGB from whatever they used for capture, and then had them use avisynth for the RGB->YUV conversion [12:34] <ircdearia> `ffmpeg -pix_fmt rgb24 -an -s 780x600 -f x11grab -i :0.0+1140,15 -s 780x600 -pix_fmt rgb24 -r 25 -vcodec gif -vframes 10 v.gif` shows [gif @ 0x1005300] ERROR: gif only handles the rgb24 pixel format. Use -pix_fmt rgb24. [12:34] <JEEB> > gif [12:34] <JEEB> u mad? [12:35] <JEEB> also, you were already given ffvhuff as an example of a lossless video codec that can do RGB :P [12:36] <ircdearia> JEEB, I am experimenting. I previously discovered that using -vcodec png was exactly what I wanted. [12:36] <JEEB> well, I think png in video only works in mov or so? At least it's only usually supported like that. ffvhuff in avi tends to be more supported. I could be wrong of course. [12:37] <ircdearia> Regarding if I'm mad or not. Currently I recorded 1000 frames, converted them to png, loaded the 1000 png images as layers in the gimp, and then though that maybe I can prepare gif animations [12:40] <Tjoppen> in gimp you can by naming the layers "frame 1", "frame 2" etc, but that's hardly practical [13:55] <nguard> hello, does anyone know why thumbnail generation does not work for some of my .flv videos? this is an example: http://pastebin.com/TpLwtmvG [13:55] <nguard> it works for some, for others it just freezes and doesn't do anything [13:56] <nguard> I think it might have to do with "vp6f -> mjpeg" vp6f codec. [13:57] <nguard> (h264 -> mjpeg works fine) [14:04] <Hadus> Hallo [14:05] <Hadus> Gibt es die Möglichkeit über ffmpeg HD zustreamen? [14:06] <Hadus> Is there a possibility to stream HD to ffmpeg? [14:44] <nguard> Hadus: ich kenne mich nicht wirklich gut aus, aber ich glaube du musst es einfach in x264 mit der passenden auflösung konvertieren. pack es in einen mp4 container und dann sollte das zu einem player wie flowplayer oder jwplayer kompatibel sein. [14:45] <Hadus> Okay, weil den Stream selbst möchte ich auf einem Flashserver schicken per RTMP, die Streamquelle ist beispielsweise eine Dreambox [15:43] <Nagy> do I need to call avfilter_inout_free? [15:43] <Nagy> all the examples seems to use avfilter_inout_alloc without corresponding avfilter_inout_free, is that correct? [16:25] <sacarasc> relaxed: So far this script is pretty cool... [16:35] <sacarasc> relaxed: The script seems to work fine. [17:16] <coony> hi [17:16] <coony> ho [17:16] <coony> hi [17:18] <coony> i'm runnig this wiki http://ubuntuforums.org/showthread.php?t=786095 to compile ffmpeg on unbuntu 10.10 my question is on fine i have a encodind to mpeg2video or i add something on its confiritation? [17:55] <defaultro> hey folks, can ffmpeg do a slow motion like twixtor's effect? [18:17] <blez> can I compare flv/mp4 by header? [18:17] <blez> if they're the same [18:19] <coony> how to have ecoding to mpeg2video non ffmpeg [19:50] <rynop> pasteeater, not sure if u remember me from few days ago - was having "dark video" issues w/ ffmpeg (from a .glc source). I found out the solution: menocder has a gamma (eq2) option where you can bump up the gamma a bit. Installing newest ffmpeg did not help... [19:51] <rynop> not sure why the vid is dark in the 1st place, but bumping up the gamma a bit with mencoder does the trick [19:54] <JEEB> is it as if the most bright and most dark colors had been cut off? [19:58] <rynop> JEEB, na - the entire video is just darker. If i open the .glc file in glc-play looks fine. [19:59] <JEEB> well, video gets darker if the darkest colors get cut off :P [19:59] <JEEB> although derp [19:59] <JEEB> getting tired [19:59] <rynop> oh, ok. yea then maybe that is what is happening. u run into this before? [19:59] <JEEB> nope, but I kind of feel that you're not doing the RGB->YUV conversion properly [20:00] <JEEB> I'd check with Avisynth and AvsP [20:00] <JEEB> first load the source into avisynth in some way [20:00] <JEEB> load as RGB [20:00] <JEEB> then convert to YUV with avisynth and check result [20:00] <rynop> ok, not familiar with either of those tools but I'll look into it. this is what im doing: glc-play 1092-3.glc -o - -y 1 | mencoder -nosound -demuxer y4m - -ovc x264 -x264encopts qp=18:pass=1 -vf eq2=1.3 -of lavf -lavfopts format=mp4 -o test.mp4 [20:01] <JEEB> > qp [20:01] <rynop> ? [20:01] <JEEB> > using qp with x264 [20:02] <JEEB> qp is a mode practically only useful for testing of features etc., normal users are supposed to use crf or bitrate [20:02] <rynop> sry, im a total encoding noob. so just get rid of the qp option? [20:02] <JEEB> -.- [20:03] <JEEB> crf, bitrate and qp are different rate control modes in libx264 [20:03] <JEEB> of which only the first two should be used by normal users [20:04] <rynop> ok. u like crf or bitrate more? this vid is a capture of a video game... [20:04] <JEEB> they have different uses [20:05] <JEEB> bitrate needs two passes for best possible quality for that file size, while crf is meant for 'I just want this quality, now compress it.' kind of encoding [20:05] <JEEB> and yeah, crf only needs one pass [20:05] <rynop> ah ok cool, crf it is. just uploading this to youtube, dont need it be great. thx so much for ur help [20:06] <JEEB> so if you want 'constant quality' crf is for you, just find the highest crf value that works for you [20:06] <JEEB> basically the end result in both cases for similar bitrate is the same [20:06] <rynop> kk, gonna start w/ crf=18 [20:06] <JEEB> start around 24 [20:06] <JEEB> it's easier to go down than up [20:06] <JEEB> also, with crf you don't need to set passes [20:06] <rynop> is pass=1 needed with crf? so like crf=18:pass=1? or is 1 pass implied? [20:06] <rynop> :) [20:07] <rynop> 1 step ahead of me u are [20:07] <JEEB> nah, I just saw it on your command line and I know that if you set pass=1 it will use faster ways of doing things which aren't exactly a good thing for crf or other kind of encoding that only has one pass to do [20:07] <JEEB> in libx264 that is [20:08] <JEEB> anyways, if you're encoding for youtube you might as well set crf to zero and get lossless [20:09] <rynop> why's that? cuz youbue will downsample (not sure if thats the right word) from my source? so if i dont upload a good src to begin with it will look even shittier? [20:09] <rynop> shitter in youtube that is [20:09] <JEEB> downsampling means usually resizing with video, which youtube might do as well depending on the resolution of your clip [20:09] <JEEB> but no [20:09] <JEEB> that's not the problem [20:10] <JEEB> the problem is that youtube will re-encode your video [20:10] <JEEB> so if you want to minimize the rape factor that youtube does [20:10] <JEEB> you'll want to upload as good of a sample as possible [20:10] <rynop> gotcha, yea thats what i was trying to say :). i'm capturing in 1280x720 so youtube wont touch aspect [20:10] <rynop> or reszie [20:11] <JEEB> there are some services that let you upload stuff untouched as long as you go within some limits but youtube isn't one of those :) [20:11] <rynop> yea. I'm looking for a vid hosting service that lets me upload unlimited # of videos, and is cheap. [20:12] <rynop> doing a little weekend sideproject thing, want to keep costs low. [20:13] <rynop> crf=0 will encode the mp4 faster too correct? i'm gonna be encoding lots of these things. [20:13] <JEEB> well, yeah -- lossless H.264 has no b-frames, so one part of encoding drops completely out [20:14] <rynop> coo [20:14] <JEEB> so it should be somewhat faster as long as your HDDs can keep up [20:14] <JEEB> if it's still too slow you can set the preset parameter [20:14] <JEEB> http://mewiki.project357.com/wiki/X264_Settings#preset [20:15] <JEEB> presets are basically speed VS compression ratio [20:15] <rynop> oh, cool. thx. I'll be reading .glc from one HDD and writing to another. so hopefully it will help a bit. [20:16] <rynop> u know if its possible to leverage your GPU when doing encoding. I know this is totally different subject, but sounds like u know wtf ur talking about. [20:17] <rynop> i'm gonna be doing this encoding on a ghetto old box of mine. dual core p3. would like to leverage the nvidia GPU if possible. [20:17] <JEEB> it is possible, and some encoders do it: the key is that you shouldn't leave the whole encoding chain to the GPU unless you have made your format specifically for that use [20:18] <JEEB> some encoders have motion estimation for H.264 put onto the GPU [20:18] <JEEB> not that it is of higher quality or otherwise more efficient, but it's just using a piece of hardware that normally doesn't get used, thus (possibly) being more fast [20:18] <JEEB> but no, ffmpeg doesn't support these as far as I know :) [20:18] <JEEB> mostly because libx264 doesn't support these [20:19] <rynop> yea. ok. do you know the name off the top of your head of an encoder that does support GPU? just so i know where to focus my research [20:19] <JEEB> the thing's generally slower than x264 [20:19] <rynop> oh [20:19] <JEEB> so let's just say "It's not worth it" [20:19] <rynop> cool, well that saved me some time. thx [20:20] <JEEB> There seems to be an actual project to use OpenCL for the motion estimation in x264 [20:21] <JEEB> but I'm not expecting much, if at all [20:26] <rynop> ok [21:15] <coony> is it normal that fflay give me this kind of problem whe i want to see a file mov [21:15] <coony> [mpeg2video @ 0x9262a00]warning: first frame is no keyframe 0B f=0/0 [00:00] --- Mon Dec 5 2011

1 0

ffmpeg-devel.log.20111203
by burek 04 Dec '11

04 Dec '11

[00:24] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rb552e3f6d8 10ffmpeg/libswscale/swscale.c: [00:24] <CIA-41> ffmpeg: swscale: Fix all cast qualifier warnings: [00:24] <CIA-41> ffmpeg: libswscale/swscale.c:2744:40: warning: to be safe all intermediate pointers in cast from int16_t ** to const int16_t ** must be const qualified [-Wcast-qual] [00:24] <CIA-41> ffmpeg: libswscale/swscale.c:2745:41: warning: to be safe all intermediate pointers in cast from int16_t ** to const int16_t ** must be const qualified [-Wcast-qual] [00:24] <CIA-41> ffmpeg: libswscale/swscale.c:2746:41: warning: to be safe all intermediate pointers in cast from int16_t ** to const int16_t ** must be const qualified [-Wcast-qual] [00:24] <CIA-41> ffmpeg: libswscale/swscale.c:2747:78: warning: to be safe all intermediate pointers in cast from int16_t ** to const int16_t ** must be const qualified [-Wcast-qual] [00:24] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [00:24] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r3fd4fe0e03 10ffmpeg/libswscale/x86/swscale_mmx.c: [00:24] <CIA-41> ffmpeg: swscale_mmx: Fix cast qualifier warnings: [00:24] <CIA-41> ffmpeg: libswscale/x86/swscale_mmx.c:131:36: warning: to be safe all intermediate pointers in cast from int16_t ** to const int16_t ** must be const qualified [-Wcast-qual] [00:24] <CIA-41> ffmpeg: libswscale/x86/swscale_mmx.c:132:37: warning: to be safe all intermediate pointers in cast from int16_t ** to const int16_t ** must be const qualified [-Wcast-qual] [00:24] <CIA-41> ffmpeg: libswscale/x86/swscale_mmx.c:133:74: warning: to be safe all intermediate pointers in cast from int16_t ** to const int16_t ** must be const qualified [-Wcast-qual] [00:24] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [00:24] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * ra24c4758d5 10ffmpeg/libswscale/x86/swscale_template.c: [00:25] <CIA-41> ffmpeg: sws: remove unused variable [00:25] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [00:36] <CIA-41> ffmpeg: 03Giorgio Vazzana 07master * r208bcc6b5b 10ffmpeg/libavutil/samplefmt.c: [00:36] <CIA-41> ffmpeg: libavutil: increase struct SampleFmtInfo member 'name' length to 8 [00:36] <CIA-41> ffmpeg: This is necessary since some sample format names are longer than 3 characters. [00:36] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [00:55] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rebe4d5adc5 10ffmpeg/libswresample/swresample_test.c: [00:55] <CIA-41> ffmpeg: swr-test_: Fix mixed declaration and statements: [00:55] <CIA-41> ffmpeg: swresample_test.c:123:21: warning: ISO C90 forbids mixed declarations and code [-Wdeclaration-after-statement] [00:55] <CIA-41> ffmpeg: swresample_test.c:127:25: warning: ISO C90 forbids mixed declarations and code [-Wdeclaration-after-statement] [00:55] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [00:55] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r12804348f5 10ffmpeg/libswresample/swresample_test.c: [00:55] <CIA-41> ffmpeg: swr-test: drop one const, silence most const qualifer warnings. [00:55] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [03:13] <CIA-41> ffmpeg: 03John Stebbins 07master * rb88eb87630 10ffmpeg/libavformat/asfdec.c: [03:13] <CIA-41> ffmpeg: asf: use packet dts as approximation of pts [03:13] <CIA-41> ffmpeg: Having a somehow off seeking is better than having none at all. [03:13] <CIA-41> ffmpeg: Signed-off-by: Luca Barbato <lu_zero(a)gentoo.org> [03:13] <CIA-41> ffmpeg: 03Luca Barbato 07master * r73f027c17b 10ffmpeg/libavformat/asfdec.c: [03:13] <CIA-41> ffmpeg: asf: do not call av_read_frame [03:13] <CIA-41> ffmpeg: The asf_read_pts should read the bitstream directly. [03:13] <CIA-41> ffmpeg: 03Nathan Adil Maxson 07master * rd0fd6fc201 10ffmpeg/libavcodec/alacenc.c: [03:13] <CIA-41> ffmpeg: Cleaned up alacenc.c [03:13] <CIA-41> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje(a)gmail.com> [03:13] <CIA-41> ffmpeg: 03Martin Storsjö 07master * r30266038bd 10ffmpeg/libavformat/rtsp.c: (log message trimmed) [03:13] <CIA-41> ffmpeg: rtsp: Initialize the media_type_mask in the rtp guessing demuxer [03:13] <CIA-41> ffmpeg: The media_type_mask is initialized via AVOptions for the [03:13] <CIA-41> ffmpeg: rtsp and sdp demuxers, but it isn't available as an option [03:13] <CIA-41> ffmpeg: for the rtp guessing demuxer (since it doesn't really make [03:13] <CIA-41> ffmpeg: sense there). Therefore, it must be manually initialized [03:13] <CIA-41> ffmpeg: instead, since a zero value means no media types at all [03:13] <CIA-41> ffmpeg: 03Aaron Colwell 07master * re02dec25ab 10ffmpeg/libavcodec/vp8.c: vp8: flush buffers on size changes. [03:13] <CIA-41> ffmpeg: 03Mans Rullgard 07master * r150ddbc148 10ffmpeg/ (avconv.c libavformat/nullenc.c): (log message trimmed) [03:13] <CIA-41> ffmpeg: Do not memcpy raw video frames when using null muxer [03:13] <CIA-41> ffmpeg: Commit 035af99 made avconv always call an encoder when using the [03:13] <CIA-41> ffmpeg: null muxer. While useful for 2-pass encodes, it inadvertently [03:13] <CIA-41> ffmpeg: caused an extra memcpy of raw frames when decoding only. [03:13] <CIA-41> ffmpeg: This hack restores the old behaviour when only decoding while [03:13] <CIA-41> ffmpeg: allowing use of the null muxer with encoded streams as well. [03:13] <CIA-41> ffmpeg: 03Mans Rullgard 07master * r59807fee6d 10ffmpeg/libavcodec/arm/h264dsp_neon.S: [03:13] <CIA-41> ffmpeg: ARM: h264dsp_neon cosmetics [03:13] <CIA-41> ffmpeg: - Replace 'ip' with 'r12'. [03:13] <CIA-41> ffmpeg: - Use correct size designators for vld1/vst1. [03:13] <CIA-41> ffmpeg: - Whitespace fixes. [03:13] <CIA-41> ffmpeg: Signed-off-by: Mans Rullgard <mans(a)mansr.com> [03:13] <CIA-41> ffmpeg: 03Janne Grunau 07master * ra760f530bb 10ffmpeg/libavcodec/arm/ (h264dsp_neon.S neon.S vp8dsp_neon.S): [03:13] <CIA-41> (43 lines omitted) [03:27] <CIA-41> ffmpeg: 03Clément BSsch 07master * r215b7724e7 10ffmpeg/ (18 files in 2 dirs): lavf: rename remaining av_set_pts_info() to avpriv_set_pts_info(). [03:35] <CIA-41> ffmpeg: 03Carl Eugen Hoyos 07master * r2d6a45c12a 10ffmpeg/libavcodec/ (h264.c h264_parser.c): [03:35] <CIA-41> ffmpeg: Silence warnings when decoding QNAP Systems H264 codec. [03:35] <CIA-41> ffmpeg: Fixes ticket #571. [03:47] <CIA-41> ffmpeg: 03Peter Ross 07master * r56f6628bca 10ffmpeg/libavcodec/pictordec.c: [03:47] <CIA-41> ffmpeg: pictordec: decode 8bpp images when extra header marker is missing [03:47] <CIA-41> ffmpeg: Fixes ticket #696. [03:47] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [03:47] <CIA-41> ffmpeg: 03Peter Ross 07master * r52b7ed4f6e 10ffmpeg/libavcodec/iff.c: [03:47] <CIA-41> ffmpeg: iff: support ExtraHalfBrite (EHB) palette mode [03:47] <CIA-41> ffmpeg: Fixes ticket #663. [03:47] <CIA-41> ffmpeg: Based on patch by ami_stuff [03:47] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [03:47] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * racf22ec69e 10ffmpeg/doc/developer.texi: [03:47] <CIA-41> ffmpeg: developer.texi: Fix naming convention [03:47] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [04:50] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r3eb6e14618 10ffmpeg/libavcodec/zmbv.c: [04:50] <CIA-41> ffmpeg: zmbv: Fix keyframe fault protextion. [04:50] <CIA-41> ffmpeg: Fixes zzuf -s467 -r0.001 cat samples/zmbv/wc2_001-partial.avi [04:50] <CIA-41> ffmpeg: Bug-found-by: darkshikari [04:50] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [04:51] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r7a5eb40ea1 10ffmpeg/libavcodec/utils.c: [04:51] <CIA-41> ffmpeg: audeo_get_buffer: fix pkt_pos [04:51] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [05:35] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r8bd2b27f6b 10ffmpeg/libavcodec/zmbv.c: [05:35] <CIA-41> ffmpeg: zmbv: fix 2 memleaks [05:35] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [05:35] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * ra930cd0d19 10ffmpeg/libavformat/oma.c: [05:35] <CIA-41> ffmpeg: oma: Fix out of array read. [05:35] <CIA-41> ffmpeg: Input: 01-Untitled-partial.oma [05:35] <CIA-41> ffmpeg: ZZUF params: zzuf[s=7157,r=0.001] [05:35] <CIA-41> ffmpeg: Bug-found-by: darkshikari [05:35] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [05:35] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * raedd30b63a 10ffmpeg/libavformat/id3v2.c: [05:35] <CIA-41> ffmpeg: id3v2: Fix null ptr crash in get_extra_meta_func() [05:35] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [11:24] <ago> Hi folks! [11:25] <ago> Usually, every ffmpeg vulnerabilities, there are CVE assigned to them, so for the vulnerabilities reported in < of 0.7.7 ( https://secunia.com/advisories/46736/ ) there are no specified vulnerabilities. [11:26] <ago> So, we can't ask a CVE, can you please explain the issue so we can ask every CVE for every issue? [11:26] <ago> thanks in advance [11:28] <ago> If them were already assigned, please point me them, because I didn't see anything on secunia and oss-security [12:05] <CIA-41> ffmpeg: 03Nicolas George 07master * r90bfd511c4 10ffmpeg/libavcodec/libcelt_dec.c: [12:05] <CIA-41> ffmpeg: celtdec: adapt to avcodec_decode_audio4. [12:05] <CIA-41> ffmpeg: Original fix by Michael Niedermayer. [12:05] <CIA-41> ffmpeg: Slightly modified to avoid the now redundant multiplications [12:05] <CIA-41> ffmpeg: and divisions in the main code path. [12:38] <CIA-41> ffmpeg: 03Carl Eugen Hoyos 07master * r88bbabcc00 10ffmpeg/libavcodec/targa.c: [12:38] <CIA-41> ffmpeg: Support decoding right-to-left targa files. [12:38] <CIA-41> ffmpeg: Fixes ticket #698. [12:38] <CIA-41> ffmpeg: 03Carl Eugen Hoyos 07master * r8dcd2a41ec 10ffmpeg/libavformat/mov.c: [12:38] <CIA-41> ffmpeg: Allow last mov chunk to have an arbitrary number of samples. [12:38] <CIA-41> ffmpeg: Fixes ticket #673. [12:38] <CIA-41> ffmpeg: 03ami_stuff 07master * r28a11a67d2 10ffmpeg/libavcodec/bmp.c: [12:38] <CIA-41> ffmpeg: BMP: Support BMP OS/2 v2 with RLE compression. [12:38] <CIA-41> ffmpeg: Fixes ticket #699. [12:38] <CIA-41> ffmpeg: Patch also provided by Peter Ross. [15:15] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * r730945d941 10ffmpeg/ffmpeg.c: [15:15] <CIA-41> ffmpeg: ffmpeg: use standard duration calculation code in the stream copy case. [15:15] <CIA-41> ffmpeg: Fixes a.mkv non monotonic timestamps [15:15] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [15:33] <ubitux> the SBG decoder looks very fun :) [15:34] <ubitux> but i still have no clue how to save & git am the whole patchset quickly with mutt :( [15:34] <Compn> git should tie into it [15:35] <Compn> you should be able to hit one key to auto patch it [15:35] <Compn> how to set that up i have no idea [15:36] <ubitux> :( [15:37] <Compn> lesset [15:37] <Compn> er [15:38] <Compn> http://blog.ndrix.com/2007/05/mutt-and-git-format-patch.html [15:39] <Compn> or this maybe http://flavioleitner.blogspot.com/2011/03/patch-workflow-with-mutt-and-git.… [15:44] <ubitux> mmh thx i'll check this out [15:56] <CIA-41> ffmpeg: 03Nicolas George 07master * r757cb3b46a 10ffmpeg/libavutil/avutil.h: (log message trimmed) [15:56] <CIA-41> ffmpeg: avutil: av_x_if_null: remove const for the return value. [15:56] <CIA-41> ffmpeg: This is the usual weirdness with const/non-const pointers through a function. [15:56] <CIA-41> ffmpeg: See the prototype of strchr for a standard example. [15:56] <CIA-41> ffmpeg: The function had to be moved down, below the point where stdint.h is [15:56] <CIA-41> ffmpeg: indirectly included. [15:56] <CIA-41> ffmpeg: Signed-off-by: Nicolas George <nicolas.george(a)normalesup.org> [15:56] <CIA-41> ffmpeg: 03Nicolas George 07master * r976b095939 10ffmpeg/ (4 files in 2 dirs): [15:56] <CIA-41> ffmpeg: lavu/error: define AVERROR_BUG. [15:56] <CIA-41> ffmpeg: Signed-off-by: Nicolas George <nicolas.george(a)normalesup.org> [15:56] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [17:16] <ubitux> i think i'm happy not to have much GCI students messing with the code on ffmpeg-devel =) [17:17] <ubitux> it looks exhausting to review there re-styling patches :p [17:34] <Compn> i hope they are testing benchmarks :P [17:34] <Compn> mpegvideo.c review [17:34] <Compn> lol [17:54] <michaelni> mpegvideo.c "review", ROTFL [17:54] <michaelni> more correctly, human indent program [18:19] <Daemon404> nothing wrong with having a consistent coding style <.< [18:20] <Daemon404> not too sure why it's a gci task tho [18:34] <JEEB> Because no-one really wants to clean up stuff >_> [18:34] <JEEB> it's not a "rewarding" task [18:34] <JEEB> you can see it in a certain media player's renderer fixage branch as well [18:35] <JEEB> the developer likes to do new stuff, but nothing gets finished [18:35] <JEEB> because the reward for actually getting something to "work" for the first time is bigger than to fix it completely [18:39] <JEEB> oh, and not to mention that in most cases the debugging to find the cause(s) for problems adds the need for effort [18:39] <j-b> Daemon404: because it is a good way to have people not being afraid of opening the code and compiling it [18:39] <j-b> which is what GCi is about [18:40] <Daemon404> true enough [18:40] <JEEB> yeah, and that too [18:41] <Daemon404> j-b, mind you, tha majority of work can be aided by a couple nifty regex [18:41] <Daemon404> if you spot patterns [18:41] <JEEB> I wish there was GCI back when I was in high school :) [18:41] <j-b> Daemon404: sure. If people had brains [18:41] <Daemon404> lol [18:42] <j-b> Daemon404: to the 'remove tabs' task in VLC, i have seen 7 students so far [18:42] <j-b> only one succeeded [18:42] <Daemon404> O_o [18:42] <Daemon404> thats crazy [18:42] <j-b> one sent a 540MB patch [18:42] <Daemon404> O______O [18:42] <j-b> 3 sent a 50MB tarball [18:42] <j-b> while the source is 200MB with whole git, and tarball is 20MB [18:43] <j-b> you were saying? [18:43] <Daemon404> i ... cannot comprehend/ [18:43] <j-b> :) [18:43] <Daemon404> also, from what i can tell [18:44] <Daemon404> 90% of time spent in gci is "learn to use git" [18:45] <j-b> possibly [19:37] <gnafu> Daemon404: Considering many of the leading open source projects are using or moving to git, and that some of these students may go on to contribute to things other than ffmpeg/libav, learning git is probably more important for them than learning ffmpeg/libav :-). [19:39] <Daemon404> i agree [19:39] <Daemon404> :P [19:39] <gnafu> )^_^( [19:46] <JEEB> indeed [20:48] <CIA-41> ffmpeg: 03Reimar Döffinger 07master * r85372f9d57 10ffmpeg/libavcodec/x86/dsputil_yasm.asm: [20:48] <CIA-41> ffmpeg: Add forgotten %ifdef HAVE_AVX. [20:48] <CIA-41> ffmpeg: Fixes compilation with older YASM/NASM versions. [20:48] <CIA-41> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger(a)gmx.de> [22:47] <CIA-41> ffmpeg: 03Michael Niedermayer 07master * rff53c79d0a 10ffmpeg/libavformat/flvdec.c: [22:47] <CIA-41> ffmpeg: flvdec: Stop searching for streams once a audio & a video stream has been found [22:47] <CIA-41> ffmpeg: instead of when the 2nd stream has been found. [22:47] <CIA-41> ffmpeg: This isnt ideal as we will likely still like before miss a data stream. [22:47] <CIA-41> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [22:52] <ubitux> gnafu: not only learning git [22:52] <ubitux> it seems they need to learn how to use the mail too [22:53] <ubitux> (i'm not saying these are simple things) [22:53] <ubitux> i think google is wrong with operation like that though [22:54] <ubitux> <+Daemon404> nothing wrong with having a consistent coding style <.< // if the style is consistent within the file, i think it's okay; but anyway that wasn't my point: using 13-17 years old student for this sounds wrong to me& but well, whatever :) [22:55] <Daemon404> it seems ot be painful for all parties involved [22:55] <Daemon404> to* [23:13] <Compn> i think we spend too much time on what code is indented how [23:13] <Compn> and not enough time on actually fixing bugs or speeding things up [23:13] <Compn> but thats just my stupid opinion [23:13] <Compn> and 'we' being ... people who want to reindent code all day for 6000 commits [23:17] <ubitux> if the coding style is overall consistent, i think we can agree with some inconsistency between files [23:17] Action: Daemon404 just submitted a bugfix yesterday :3 [00:00] --- Sun Dec 4 2011

1 0