December 2011 - Ffmpeg-devel-irc

ffmpeg.log.20111228
by burek 29 Dec '11

29 Dec '11

[00:08] <JodaZ> wtf, why would ffprobe and ffmpeg be exclusive debian packages [00:09] <JodaZ> -_- nvm one is debian default, one is multimedia [03:29] <pythonirc1011> I've a file which is either an mp4 file or a flv file. But I dont know what type it is. Can ffmpeg help me rip out the audio from it -- in mp3 please? [03:35] <JodaZ> pythonirc1011, maybe ffmpeg -i FILE -vn -acodec libmp3lame mp3 OUTFILE.mp3 [03:38] <pythonirc1011> I'm getting ERROR: audio conversion failed: Unknown encoder 'libmp3lame' [03:38] <pythonirc1011> how do i fix this [03:38] <pythonirc1011> JodaZ: I installed ffmpeg using "apt-get install" on ubuntu [03:39] <JodaZ> try without -acodec libmp3lame [03:41] <JodaZ> pythonirc1011, alternatively add this repository to get a less broken ffmpeg http://medibuntu.org/repository.php [03:41] <JodaZ> or follow compilation guide in topic [03:41] <pythonirc1011> JodaZ: http://pbin.compgeom.com/show/2/ [03:42] <pythonirc1011> it seems its recognizing the filetype [03:42] <JodaZ> your ffmpeg is shit [03:42] <pythonirc1011> ffmpeg -i inpfile -vn -acodec libmp3lame mp3 OUTFILE.mp3 [03:43] <JodaZ> yes [03:43] <JodaZ> now get a proper ffmpeg using the comilation guide in topic or the repo i showed you [03:43] <pythonirc1011> JodaZ: I've a program that can use it to do what i want...so there is some trivial way that my crap ffmpeg should be able to do this [03:44] <JodaZ> does that work ? [03:44] <JodaZ> just check ps aux for the way ffmpeg is called and do it like that [03:44] <relaxed> ffmpeg -i input -vn -acodec libmp3lame -aq 0 output.mp3 [03:44] <pythonirc1011> relaxed: Thanks ! [03:45] <pythonirc1011> relaxed: also, how do i get good quality mp3 with VBR? [03:45] <relaxed> -aq 0 [03:45] <pythonirc1011> is this the best quality i can get? [03:45] <relaxed> yes [03:46] <relaxed> use flac if you want lossless [03:47] <pythonirc1011> whats flac and how can i use it? libmp3flac? [03:49] <relaxed> it's a lossless audio codec. ffmpeg -i input -vn -acodec flac output.flac [03:50] <pythonirc1011> ah ok...thanks [05:16] <kcm1700> can ffmpeg decode h.264 video encoded with forced intra refresh technology? [05:42] <pythonirc1011> how can i make ffmpeg not output anything except in case there is an error? [05:42] <pythonirc1011> I tried -loglevel quiet -- doesnt help [06:06] <patrakov> Hi. I have a mov file with mjpeg codec and 29.97 fps. How to extract a frame with a known number to a regular jpeg file? [06:58] <Aiwass> hi there, there a ffmpeg parameter that will turn off the configuration information when ffmpeg is used? [08:03] <defaultro> hey folks, can ffmpeg split odd and even frames then save each file as jpeg? [08:10] <defaultro> i got it [09:15] <shifter1> anyone know how to set input buffering in ffmpeg? [09:29] <heisenmink> can I do something like "ffmpeg -i IMG_1138.MOV -r 15 -f image2 -q 0 -s 480xANYTHING sharp\img-%d.jpg" [09:29] <heisenmink> to get dynamic scaling after the source video? [09:33] <heisenmink> "480x*" doesn't work, "480x" doesn't work [09:33] <heisenmink> and not 480xH [09:35] <sacarasc> -1 should scale it. [09:35] <Mavrik> heisenmink, try replacing -s with -vf scale=480:-1 [09:39] <heisenmink> thanks! works perfectly [09:40] <defaultro> what param specifies the quality of jpeg output? [09:41] <Mavrik> -qscale I think [09:41] <defaultro> k [09:41] <heisenmink> -q 0 gives max quality [09:41] <heisenmink> in my script [09:42] <defaultro> so -q 0 is similar to Photoshop's 12 when saving an image to the max as jpeg [09:46] <heisenmink> defaultro, I think so [09:46] <defaultro> cool [09:55] <verde> Hello, how do i extract subtitles from a .mkv file and save them as .srt? [12:54] <MindSpark> hello, sorry to ask this here, but does anyone work with red5? [13:32] <kollapse> Hi. What's the best way of cutting about 10 seconds out of a video clip ? [14:45] <Kartagis> hi [14:45] <Kartagis> how to stick in a MUXER at command line? [14:48] <Kartagis> basically I want to extract audio from a webm file to mp3 [14:50] <ubitux> the webm file has a mp3 track? [14:50] <ubitux> (is it even ok to have such file?) [14:50] <Kartagis> it's video [14:50] <ubitux> yes of course, but is the audio mp3 already? [14:51] <ubitux> or you need to re-encode the audio? [14:51] <Kartagis> let me see [14:51] <ubitux> run ffprobe on the file [14:51] <Kartagis> hmm [14:51] <Kartagis> audio is vorbis [14:51] <JEEB> unsurprisingly [14:51] <ubitux> you can do a ffmpeg -i in.webm -vn acodec copy out.ogg then [14:52] <ubitux> ffmpeg -i in.webm -vn -acodec copy out.ogg [14:52] <Kartagis> thanks [14:53] <ubitux> if you really want a mp3 (then you will have a lossylossy transcode), ffmpeg -i in.webm -acodec libmp3lame -aq 1 out.mp3 should do the trick [14:53] <Mavrik> ubitux, since webm is slightly changed matroska container it can carry MP3 without problems [14:53] <Mavrik> Kartagis, webm is not video -_- [14:53] <ubitux> Mavrik: yes, but is it permitted? [14:54] <JEEB> in the webm specs no [14:54] <JEEB> AFAIK [14:54] <Mavrik> I think it's defined only for vorbis as well, but I'm gonna go check [14:55] <Mavrik> spec claims "Video codec SHOULD be VP8, audio codec SHOULD be vorbis" [14:56] <Kartagis> webm is a container, right? [14:56] <JEEB> it's a mixture of limited specs [14:57] <JEEB> matroska, with limited features specified with certain video and audio codecs [14:58] <Kartagis> Unknown encoder 'libmp3lame' [14:58] <JEEB> then you have built ffmpeg without it [14:58] <Kartagis> I thought I could do mp3 [14:58] <JEEB> it's an external encoder [14:59] <JEEB> ffmpeg can decode mp3 just fine, as well as copy the stream itself just fine [14:59] <JEEB> for encoding you need LAME [14:59] <Kartagis> hmm [15:00] <Kartagis> the one I have another directory should have it [15:00] <Kartagis> I remember building ffmpeg with libmp3lame [15:01] <Kartagis> I don't :( [15:46] <Leoneof> JEEB: i've used 'ultrafast`, but it didn't work on psp, that's why you didn't recommend it? [16:00] <JEEB> Leoneof, I don't think ultrafast by itself makes it non-playable... [16:11] <Leoneof> JEEB: ouch [16:12] <Leoneof> i will try to convert some files later [16:17] <JEEB> but I think that superfast in most cases should be lolfast already :V [16:46] <Milos> I have a quad core. How do I make use of all the threads with this command: [16:46] <Milos> ffmpeg -i video_origine.avi -target pal-svcd video_finale.mpg [16:46] <Milos> When running that only uses 1 CPU core. [16:46] <Milos> I have ffmpeg compiled in with threading and 'cpudetection' enabled. [16:48] <Milos> Nvm, -threads worked. [17:05] <Slony> Hey [17:05] <Slony> I'm reading this, and I see they used "-bf 0", why? http://code.google.com/p/gource/wiki/Videos [17:14] <Grafica> I'm a newbie at programming. Is there something I can download so I can convert flv files to ogv, mp4, and webm? [17:19] <Slony> there are guis for ffmpeg if thats what youre asking [17:26] <Grafica> Yes, a GUI would be nice. I don't understand how to get to it. [17:27] <Slony> try google [17:28] <Slony> "ffmpeg gui" [17:29] <pom> Can I run FFMPEG without specifying an output-file AND have it return 0, as in no error occured, if I only want it's informational input? [17:30] <sacarasc> pom: ffprobe does that. [17:30] <Mavrik> pom, you should run ffprobe [17:37] <pom> If I've previously used ffmpeg to parse stream data, is ffprobe an inplace replacement as far as it's output goes? [17:38] <pom> I've only executed ffmpeg with ffmpeg -i [inputfile] [17:38] <JEEB> more or less [17:38] <JEEB> I haven't tested ffprobe myself yet tho :) [17:38] <ubitux> pom: ffprobe will give you all you need [17:39] <ubitux> it also has various output such as csv, json or xml [17:39] <ubitux> so you can extract the information in a simpler way [17:39] <ubitux> see the -show_streams, -show_format or -show_packets options [17:41] <pom> ubitux: Interesting. Just deployed ffprobe and using it pretty extensively thruout this project. Seem in-place-replaceable enough so far. =) Very happy to have this. ^^ Since when has this been around? [17:42] <ubitux> Date: Sun Feb 21 11:43:01 2010 +0000 [17:42] <pom> Time flies. [17:42] <ubitux> csv, json or xml outputs are pretty recent though [17:43] <ubitux> json being the first one, added in september 2011 [17:45] <pom> Nice! My proper regexp-fest is soon to meet it's maker, then. :) [17:46] <Slony> I'm recording a video from png frames, a guide recommends using -bf 0, now then way I do it is to first capture using -preset ultrafast and -crf 18, and then i do a two pass -preset slower recoding. At which stage(s) should I use -bf 0, the ultrafast one, the slower first or second, or all? [17:46] <Milos> I type -b 2000 [17:46] <Milos> and it says "2kb/s" [17:46] <Milos> wtf? [17:47] <Mavrik> Milos, of course. [17:47] <Mavrik> you told it to encode to 2000 b/s [17:47] <Milos> so I add four zeros and it says it's too high? [17:47] <Milos> I mean three zeros [17:47] <Mavrik> -b 2000k ;) [17:48] <Milos> oh.. I did 2000kb and it yelled at me [17:48] <Milos> ok, thanks [17:49] <Milos> Does it make any sense whatsoever if runnning ffmpeg to pal-dvd with -threads 8 gives a load average of 0.08? [17:49] <Mavrik> can't say with that info. :) [17:50] <Milos> it's like [17:51] <Milos> either my xeon server is amazing, or this is barely doing anything [17:51] <Mavrik> whats "pal-dvd"? [17:51] <Mavrik> mpeg2? 2-pass? [17:52] <Milos> mpeg2 apparently [17:52] <Mavrik> where did you put the -threads? [17:52] <Mavrik> etc etc :) [17:52] <Milos> at the beginning [17:52] <Milos> oh [17:52] <Milos> here's the whole thing [17:52] <Milos> ffmpeg -threads 8 -i derp.avi -target pal-dvd -b 2000k derp.mpg [17:52] <Mavrik> huh [17:53] <Mavrik> Milos, would you mind pasting the ffmpeg output for that somewhere? :) [17:53] <relaxed> you're trying to decode derp.avi using 8 threads [17:53] <Milos> sure [17:53] <relaxed> put it after the input [17:53] <Mavrik> Milos, and yes, you just told ffmpeg to decode input with 8 threads [17:53] <Milos> D: [17:53] <Mavrik> put it after -i, preferrably after -target as well [17:54] <Milos> hmm makes no difference to speed [17:54] <Milos> oddly [17:56] <relaxed> Milos: Works here. Use htop to see how many threads are used while it's encoding. [17:56] <Milos> relaxed, yeah 8 [17:56] <Milos> but my load average is 9 [17:56] <Milos> -_- [17:56] <Milos> I mean [17:56] <Milos> 0.09 [17:56] <relaxed> ffmpeg version? [17:56] <sacarasc> What is CPU usage? [17:56] <Milos> well 0.06 now [17:56] <Milos> all cores are on average 25% [17:56] <Mavrik> Milos, look at CPU usage, not load average [17:57] <Milos> still does the job in like 2 minutes [17:57] <Milos> but wtf [17:57] <Mavrik> there's also a possibility mpeg2 encode doesn't allocate more threads [17:57] <Mavrik> if you're using mpeg-2 encoder at all [17:57] <Milos> well, without -threads 8, only one thing is in use and that one's at 100% as far as I remember [17:57] <Milos> Mavrik, do you still want output? [17:58] <Milos> What is dup= btw? [17:58] <Milos> Sounds like "duplicate" [17:58] <sacarasc> It is. [17:58] <Milos> but it goes up a lot [17:58] <Milos> it's at 5,000 - that can't be good? [17:58] <Mavrik> yeah, dup= means number of frame duplicated [17:58] <Mavrik> nop [17:59] <Milos> nop what? [17:59] <Mavrik> 5000 duplicated frames isn't good :) [17:59] <Mavrik> are you sure your input isn't broken? [17:59] <Slony> anything I can do to get rid of what looks like posterization in gradient areas? [17:59] <Milos> does that mean... the source has duplicate frames, or the destination, or ... ? [17:59] <Mavrik> Milos, that means that encoder is duplicating frames to keep up with designated framerate [18:00] <Milos> oh.. [18:00] <Mavrik> 5000 duplicated frames means either you're increasing framerate by a rather large margin or your video is broken and the decoder isn't producing output [18:00] <Milos> I see. [18:01] <Milos> And a quick question, is it possible to estimate output size before doing something? [18:02] <Mavrik> (bitrate / 8) * (number_of_video_seconds) [18:02] <Mavrik> + same for audio of course [18:02] <Grafica> Thanks Slony. [18:03] <Milos> Wow perfect thanks Mavrik. [18:03] <Milos> Are there any flaws in that formula? [18:04] <Mavrik> yeah, the encoder doesn't really encode to the exact bitrate you set so it's only an estimation [18:04] <Milos> Ah, that's fine. [18:04] <Mavrik> (video encoding is VBR) [18:04] <Milos> Yeah. [18:04] <Milos> What if I use -min -max to set it [18:04] <Milos> (the bitrate) perfectly [18:04] <Mavrik> you'll just ruin your video quality :) [18:06] <Milos> I must do that then. [18:06] <Milos> I've got size limitations. [18:06] <Milos> It's going over my given bitrate. [18:08] <Mavrik> it's very unlikely the mpeg2 encode will listen to your -max setting [18:08] <Mavrik> use 2-pass encoding if you need to hit certain filesize instead [18:08] <Milos> Yeah it's jumped to 2400 again. [18:08] <Milos> I set it at 1945. [18:08] <Milos> (k) [18:08] <Milos> Well that's great. [18:09] <Mavrik> Milos, that's completely normal. [18:09] <Milos> Yeah I bet but just unfortunate for me, although I'm experimenting anyway. [18:09] <Mavrik> no, you don't understand: [18:09] <Mavrik> even if it jumps to 2400kbps at times, the encoder will most likely hit the rit filesize because it'll compensate with lower bitrate on other scenes [18:10] <Mavrik> you can make it even more accurate (and gain quality) with 2-pass encoding [18:10] <Milos> I see. [18:10] <Mavrik> if you try to force upper limit you'll just destroy quality on high-movement scenes and waste alot of bytes for still scenes [18:10] <Milos> That makes sense, actually. [18:10] <Milos> I've seen those crappy artifacts when stuff moves. [18:10] <Milos> I guess that's what that is. [18:10] <Mavrik> question is just how good the MPEG-2 encoder is (libx264 does that quite well, I have no idea how good the ffmpegs MPEG-2 encoder is) [18:11] <Mavrik> yeah, that's exactly what it is :) [18:11] <Milos> I've got x264, should probably find out how to use it. [18:11] <JEEB> well, when creating a DVD f.ex. there are some rules on how high you can go [18:11] <JEEB> and in what kind of a buffer [18:11] <Mavrik> JEEB, yeah, I hope -target pal-dvd sets those limits as well [18:12] <Milos> Essentially what I'm trying to do is fit three AVIs on a PAL DVD. Ignoring quality issues, is there a more automated way to do this with ffmpeg? [18:12] <Milos> Or do I have to do the research myself and just plug in the right params? [18:13] <Milos> Then use some DVD authoring tool to burn it. [18:13] <JEEB> Mavrik, yes it should [18:14] <Bertl> greetings! I'm wondering if there is a good way to combine two video streams (e.g. from two cameras) into one file suitable for stereoscopic display (e.g. side by side or similar)? [18:22] <relaxed> Bertl: use the pad filter to extend width*2, then use the overlay filter and place it over the padded area. [18:22] <Bertl> ah, good idea, let me try that [18:27] <Bertl> hmm, I see filter examples with -vf "..." but my ffmpeg says unknown option -vf ? is my ffmpeg outdated? [18:27] <Mavrik> Bertl, version? [18:27] <Bertl> FFmpeg version 0.6.3, Copyright (c) 2000-2010 the FFmpeg developers [18:27] <Mavrik> yeah, that's ancient [18:27] <Mavrik> current stable is 0.9 [18:28] <Bertl> ah, thought so, no problem, let's build our own then, thanks! [18:29] <Bertl> any big advantages/changes/new features in the snapshot/trunk vs. 0.9? [18:30] <Mavrik> bugs? :P [18:30] <Mavrik> Bertl, ffmpeg releases are pretty quick lately so I doublt there's anything really significant from 0.9 [18:30] <Mavrik> 0.9 is only like 2 weeks old or so [18:31] <Bertl> ah, great, so I stick with that then, tx! [18:31] <JEEB> you can check the changelog [19:03] <Bertl> okay, I'm using 'ffmpeg -i left.ts -vf "movie=right.ts [right_side]; [in] pad=3840:1080:0:0 [left_side]; [left_side][right_side] overlay=1920:0 [out]" output.mkv [19:04] <Bertl> this works, but gives really bad quality, despite the fact that the input streams are high quality, I guess I'm missing some options for the bandwidth or so [19:05] <Mavrik> Bertl, yeah, you're missing all the optioins :) [19:05] <Mavrik> Bertl, add -vcodec libx264 -crf 23 at least [19:05] <Bertl> thought so :) [19:05] <Mavrik> Bertl, change crf up (worsequality/smaller file) or down (better quality/larger file) [19:23] <Bertl> hmm, obviously I'm missing libx264, but I tried with mpeg4, but quality is still bad, will search for libx264 now [19:31] <relaxed> Bertl: for mpeg4 set -qscale 3 for high quality (1=highest - 31=lowest) [20:20] <defaultro> anyone around [20:22] <defaultro> i have a video at 30fps, each frame(odd/even) was recorded at different iso. I extract them and use the files as inputs to enfuse command line tool to generate an hdr image. Finally, I use ffmpeg again to reencode the hdr image. However, the video is faster. I slowed it down but I didn't like the playback [21:37] <mmarschall> Hi, I already transcoded some dv files to mov using the very same command on the very same box successfully. Now I encountered at least two file where ffmpeg simply hangs fully using one CPU without finishing the transcode. Please see http://pastebin.com/rz6tipT3 for details. Any help is highly appreciated! Thanks. [23:05] <Shimmy> Anyone please help me: https://ffmpeg.org/trac/ffmpeg/ticket/747 [23:06] Last message repeated 1 time(s). [23:23] <relaxed> Shimmy: maybe you should offer a bounty [23:29] <Guest51148> Hi all , i am using the libav* in an application on my linux-ubuntu pc and everything is working fine, however when i port the same code on the android platform i get an "signal 11 (SIGSEGV), fault addr" on avformat_write_header() or av_write_frame() call. Does anyone have an idea ? Or what exactly the 2 above calls do ? thanks [23:30] <Mavrik> Guest51148, basically, first saves the file header and the other one actually saves encoded frames to file [23:30] <Mavrik> there's something wrong with your file output... do you have permissions to write? [23:31] <Guest51148> yes i have [23:31] <Guest51148> basically i am writting on rtp output [23:31] <Guest51148> and i also have permission INTERNET [23:32] <Guest51148> which gives me the network permissions afaik [23:32] <Shimmy> @relaxed how do i offer a bounty [23:34] <Guest51148> i am using the latest ffmpeg-source [23:37] <Guest51148> i don't know how to debug it but i checked with some prints on logcat that these 2 functions are the problem. [23:41] <relaxed> Shimmy: Offer some money for a developer to add the support you need. [23:43] <JodaZ> so someone said that i need to output to - instead of a file to prevent ffmpeg from muxing to something if i use -fyuv4mpegpipe [23:43] <JodaZ> is that true ? [23:44] <JodaZ> because i'd really rather not use - [23:45] <Shimmy> relaxed: what means money? i.e. what's considered an entry point? [23:48] <relaxed> Shimmy: I have no idea how much work is required or what would be an entry point. Ask in #ffmpeg-devel if anyone would be interested and then go from there. [23:54] <huglester> hey everyone [23:56] <Nitsuga> How do yo use the audio filters? -af does not seem to work [00:00] --- Thu Dec 29 2011

1 0

ffmpeg-devel.log.20111227
by burek 28 Dec '11

28 Dec '11

[01:27] <CIA-49> ffmpeg: 03Diego Biurrun 07master * r4e616d0915 10ffmpeg/tests/ (Makefile fate.mak fate/microsoft.mak fate2.mak): fate: split off Microsoft codec FATE tests into their own file [01:27] <CIA-49> ffmpeg: 03Janne Grunau 07master * r6e9bb5aa3e 10ffmpeg/libswscale/swscale_unscaled.c: [01:27] <CIA-49> ffmpeg: swscale: prevent invalid writes in packed_16bpc_bswap [01:27] <CIA-49> ffmpeg: Writes past the end of the destination buffer were occuring when its [01:27] <CIA-49> ffmpeg: stride was smaller than the stride of the source. Fixes Bug #183. [01:27] <CIA-49> ffmpeg: 03Diego Biurrun 07master * rbebf9f923e 10ffmpeg/tests/ (Makefile fate.mak fate/vp8.mak fate2.mak): fate: group all VP* codec FATE tests together in one file [01:27] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r49b909b980 10ffmpeg/: (log message trimmed) [01:27] <CIA-49> ffmpeg: Merge remote-tracking branch 'qatar/master' [01:27] <CIA-49> ffmpeg: * qatar/master: [01:27] <CIA-49> ffmpeg: fate: split off vqf/twinvq FATE tests into their own file [01:27] <CIA-49> ffmpeg: fate: split off mpc FATE tests into their own file [01:27] <CIA-49> ffmpeg: fate: split off libavcodec FATE tests into their own file [01:27] <CIA-49> ffmpeg: fate: split off Microsoft codec FATE tests into their own file [01:27] <CIA-49> ffmpeg: 03Diego Biurrun 07master * r723c35feb0 10ffmpeg/tests/ (Makefile fate.mak fate/vqf.mak fate2.mak): fate: split off vqf/twinvq FATE tests into their own file [01:27] <CIA-49> ffmpeg: 03Diego Biurrun 07master * r913b6b4b98 10ffmpeg/tests/ (Makefile fate/libavcodec.mak fate2.mak): fate: split off libavcodec FATE tests into their own file [01:27] <CIA-49> ffmpeg: 03Diego Biurrun 07master * r9e5bd6cc94 10ffmpeg/tests/ (Makefile fate.mak fate/mpc.mak fate2.mak): fate: split off mpc FATE tests into their own file [06:07] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r80d2dddad1 10ffmpeg/ffmpeg.c: [06:07] <CIA-49> ffmpeg: ffmpeg: disable dr1 in case of resolution changes, as this does not work currently. [06:07] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [06:07] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * rf2f8632aa5 10ffmpeg/ffmpeg.c: [06:07] <CIA-49> ffmpeg: ffmpeg: Check that frames are user type before treating them as such [06:07] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [06:07] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r9971af3aa8 10ffmpeg/ffmpeg.c: [06:07] <CIA-49> ffmpeg: ffmpeg: add remaining hunks to enable dr1 support [06:07] <CIA-49> ffmpeg: Author: Anton Khirnov [06:07] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [06:07] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * rb72544370a 10ffmpeg/libavcodec/mpegvideo.c: [06:07] <CIA-49> ffmpeg: mpegvideo: fix visualization buffer size. [06:07] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [08:26] <shifter1> Do any know how ffmpeg reads from character files? Is this question better suited for #ffmpeg ? [08:59] <shifter1> ping [09:07] <burek> ffmpeg is missing docs info on "analyzeduration" [09:07] <burek> i dont want to file a bug report just because of that :) [09:10] <cbsrobot> burek: ffmpeg -h | grep analyzeduration [09:10] <cbsrobot> shows [09:10] <cbsrobot> -analyzeduration <int> .D... how many microseconds are analyzed to estimate duration [09:11] <shifter1> it doesn't seem to be in the man page [09:11] <cbsrobot> ah ok [09:11] <shifter1> I could be worng [09:12] <shifter1> it should also probably be in the ffprobe man page too [09:13] <burek> yes, its not in the html doc :) [09:13] <burek> crap, works only in 1 chan :) [09:13] <burek> http://ffmpeg.org/ffmpeg.html [09:13] <burek> :) [09:15] <ubitux> a lot of options are undocumented [09:15] <burek> really? [09:16] <ubitux> you might want to file a bug report for this [09:16] <burek> too bad :/ people usually use google to find help.. [09:16] <burek> maybe it would be a good idea to periodically do ffmpeg --fullhelp and make html file out of it and publish it? [09:16] <ubitux> check libavformat/options.c [09:16] <ubitux> then git grep each option [09:16] <ubitux> most of them don't appear in the documentation [09:17] <burek> i see [09:17] <ubitux> burek: --fullhelp is x264 specific [09:18] <ubitux> i'm not sure having a ffmpeg -h page is a really good idea though [09:18] <burek> why not [09:19] <burek> just setup a cron that does ffmpeg -h > /var/www/bla.html :) [09:19] <ubitux> because some options will be documented two times [09:19] <ubitux> otoh, pointing out ffmpeg -h in the main documentation page, saying it might gives a more complete list, might be better [09:20] <burek> im talking this because of google and its indexing power [09:20] <burek> so that people can search for a problem and get results fast [09:20] <burek> it doesn't have to be on the same page as the official documentation [09:20] <ubitux> where would you integrate a link to this? [09:20] <burek> it could be (and should) completeley separate page [09:21] <burek> wherever suitable :) [09:21] <burek> google bot will find it [09:22] <ubitux> if you're going to do it, i'd suggest you add anchors to the list [09:22] <burek> well i could, but i don't have enough privileges to access the machine and create a cron job :) [09:22] <shifter1> the man page is pretty full [09:23] <shifter1> why not dump the man page to html [09:23] <burek> well, it's a good idea [09:23] <burek> the only problem i see here [09:23] <ubitux> the manpage is the same as the online documentation [09:23] <ubitux> (the source is the same) [09:23] <shifter1> is it? how often is updated? [09:23] <burek> is that people have to update 1) html docs 2) ffmpeg -h options and 3) man ffmpeg [09:23] <ubitux> so if the online documentation is lacking documentation, the manpage too [09:24] <burek> i see [09:24] <ubitux> burek: so piping ffmpeg -h, and generates a page with each line like this: <a href="#opt_analyzeduration">-analyzeduration</a> [...] might be a solution [09:24] <ubitux> this will help user-helpers to point out an option [09:24] <ubitux> this could be scripted quite easily [09:24] <burek> yes, that would be great [09:25] <ubitux> also note afaik the -h option depends on how you build ffmpeg [09:25] <ubitux> (for codecs/formats specific options) [09:25] <shifter1> --help is what I use [09:25] <burek> yes, that's why i wanted to build a big fat static ffmpeg, with everything in it, but when we discussed the legal issues, I just gave up on it [09:25] <ubitux> :) [09:26] <ubitux> OTOH, you could also just add the documentation to the most common missing option docs [09:26] <ubitux> sending a patch for this isn't much trouble [09:26] <burek> the problem is to learn git :) [09:27] <burek> i mean, not for me now [09:27] <burek> because i alredy did it [09:27] <ubitux> simple usage is pretty straightforward [09:27] <burek> but for new people, which just use ffmpeg and notice something to report [09:27] <burek> do we have forum ready ? [09:27] <burek> or still not : [09:27] <burek> :( [09:28] <ubitux> no idea, still a WIP i guess [09:28] <burek> man.. we could build 5 forums so far.. :s [09:28] <burek> why does it go so slow [09:28] <ubitux> people who care seems not much available [09:32] <burek> well i have spare time [09:35] <ubitux> then contact zeranoe [09:35] <ubitux> :p [09:37] <burek> i will again :) [11:09] <CIA-49> ffmpeg: 03Clément BSsch 07master * r15a2a29ba3 10ffmpeg/ (5 files in 3 dirs): lavfi: add thumbnail video filter. [11:15] <CIA-49> ffmpeg: 03Clément BSsch 07master * rf1f87439e5 10ffmpeg/libavfilter/vf_thumbnail.c: thumbnail: fix error code in case of invalid args. [13:01] <CIA-49> ffmpeg: 03Stefano Sabatini 07master * rc405a22d85 10ffmpeg/Changelog: Changelog: move thunbnail filter entry from 0.9 to "next" section [13:01] <CIA-49> ffmpeg: 03Stefano Sabatini 07master * r93613338fe 10ffmpeg/ (ffmpeg.c ffplay.c ffprobe.c ffserver.c): ff* tools: move descriptions from copyright header to @file doxy [13:01] <CIA-49> ffmpeg: 03Stefano Sabatini 07master * rf1a4182e8e 10ffmpeg/ (doc/ffprobe.texi ffprobe.c): ffprobe: add show_private_data option [13:01] <CIA-49> ffmpeg: 03Stefano Sabatini 07master * r20ac5849f8 10ffmpeg/ (5 files in 2 dirs): ffprobe: add XML writer [13:01] <CIA-49> ffmpeg: 03Stefano Sabatini 07master * r5764301df3 10ffmpeg/libavfilter/avfilter.c: [13:01] <CIA-49> ffmpeg: lavfi: don't try to indent words in trace messages [13:01] <CIA-49> ffmpeg: The traces end up not indented and ugly anyway, but with the additional [13:01] <CIA-49> ffmpeg: spaces are much more painful to grep/parse. [15:28] <CIA-49> ffmpeg: 03Stefano Sabatini 07master * re9af732a1a 10ffmpeg/libavfilter/defaults.c: (log message trimmed) [15:28] <CIA-49> ffmpeg: lavfi: fix avfilter_default_get_audio_buffer() after changes in av_samples_alloc() [15:28] <CIA-49> ffmpeg: av_samples_alloc() behavior changed in bbb46f3ec, resulting in random [15:28] <CIA-49> ffmpeg: data filling the data[] and linesize[] arrays of the returned [15:28] <CIA-49> ffmpeg: AVFilterBufferRef, which was resulting in wrong behavior in case of code [15:28] <CIA-49> ffmpeg: checking on data[i] nullity. [15:28] <CIA-49> ffmpeg: In particular fixes crash in avfilter_filter_samples(): [15:28] <CIA-49> ffmpeg: 03Stefano Sabatini 07master * rae21776207 10ffmpeg/libavfilter/avfilter.c: [15:28] <CIA-49> ffmpeg: lavfi: add missing check in avfilter_filter_samples() [15:28] <CIA-49> ffmpeg: Avoid out-of-buffer data access when nb_channels is 8. [15:28] <CIA-49> ffmpeg: 03Stefano Sabatini 07master * r440e984b71 10ffmpeg/ (6 files in 3 dirs): lavfi: add asplit filter [18:32] <CIA-49> ffmpeg: 03Carl Eugen Hoyos 07master * r613cbf3168 10ffmpeg/libavcodec/pngdec.c: [18:32] <CIA-49> ffmpeg: Fix some 1bit png samples, they can be palettised. [18:32] <CIA-49> ffmpeg: Fixes ticket #767. [18:32] <CIA-49> ffmpeg: 03Carl Eugen Hoyos 07master * r387d86077f 10ffmpeg/libavcodec/ (mjpegdec.c mjpegdec.h): [18:32] <CIA-49> ffmpeg: Support decoding yuv442 and yuv421 jpeg images. [18:32] <CIA-49> ffmpeg: Fixes ticket #787. [18:32] <CIA-49> ffmpeg: 03Carl Eugen Hoyos 07master * rec0dc7e2e0 10ffmpeg/ (libavformat/ipmovie.c tests/ref/fate/interplay-mve-8bit): Make Interplay palette opaque and increase its dynamics. [20:03] <ubitux> ffplay -f lavfi 'amovie=a.mp3,asplit[a][out1],[a]showaudio[out0]' e e e [20:20] <ubitux> heh it seems to have trouble synching a/v :p [21:58] <CIA-49> ffmpeg: 03Michael Niedermayer 07release/0.8 * r680880c98d 10ffmpeg/libavcodec/h264.h: [21:58] <CIA-49> ffmpeg: h264: fix init of topleft ref/mv. [21:58] <CIA-49> ffmpeg: Fixes Ticket778 [21:58] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [22:23] <cbsrobot> any reason avfilter_copy_frame_props is defined in avcodec.h and avfilter.h ? [22:32] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r29c1b258ab 10ffmpeg/Makefile: [22:32] <CIA-49> ffmpeg: Makefile: fix path to ffprobe.xsd [22:32] <CIA-49> ffmpeg: Idea-by: beastd [22:32] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [22:35] <michaelni> cbsrobot, merge from libav [22:35] <michaelni> hmpf saste is not here [22:38] <michaelni> cbsrobot, ive sent mail to saste [22:44] <cbsrobot> michaelni: ok [23:58] <CIA-49> ffmpeg: 03Frank Vernaillen 07master * rad1c502557 10ffmpeg/libavcodec/ (bmpenc.c pcxenc.c): (log message trimmed) [23:58] <CIA-49> ffmpeg: Fixed crash in palette handling when converting certain .png images to .pcx or .bmp. [23:58] <CIA-49> ffmpeg: The existing code expected a palette buffer holding 256 uint32_t's allocated in [23:58] <CIA-49> ffmpeg: the data[1] field of the AVFrame structure, but data[1] was NULL. The bug is [23:58] <CIA-49> ffmpeg: fixed by using a fixed local array (palette256) to hold the palette instead. [23:58] <CIA-49> ffmpeg: This solves http://ffmpeg.org/trac/ffmpeg/ticket/833 [23:58] <CIA-49> ffmpeg: Signed-off-by: Frank Vernaillen <fr_ve(a)hotmail.com> [00:00] --- Wed Dec 28 2011

1 0

ffmpeg.log.20111227
by burek 28 Dec '11

28 Dec '11

[00:29] <tbruff13> can someone help me convert mp3 files to mp4 audio files using Winff [00:29] <tbruff13> what i need is the ffmpeg command [00:45] <tbruff13> can some one help me create a preset for mp4 audio [00:45] <tbruff13> is it -acodec libfacc or just acc [00:46] <Mavrik-> tbruff13, either libfaac or libvo_aac, depending on which library your ffmpeg is linked to [00:48] <tbruff13> Madtree: ok so the codec is -acodec libfacc -b 128 [00:48] <tbruff13> right [00:49] <Mavrik-> it's not libfacc. [00:53] <tbruff13> Mavrik-: what is it if it is not libfacc [00:53] <Mavrik-> read what I wrote. [00:54] <AnlamK> Hello, I've installed FAAC from https://sourceforge.net/projects/faac/files/faac-src/faac-1.28/faac-1.28.ta…. But when I do, "./configure --prefix='/opt/local/' --enable-libmp3lame --enable-libfaac --enable-nonfree" in my ffmpeg directory, I get an error message saying libfaac wasn't found. [00:54] <tbruff13> you wrote libfaac [00:54] <tbruff13> oh [00:54] <tbruff13> i see [00:54] <tbruff13> thanks [00:55] <AnlamK> When I do, 'which faac', I get the following: /opt/local/bin//faac -> so it is installed [00:57] <AnlamK> This is the output message from the config.log: [00:57] <AnlamK> gcc -D_ISOC99_SOURCE -D_FILE_OFFSET_BITS=64 -D_LARGEFILE_SOURCE -DPIC -std=c99 -mdynamic-no-pic -fomit-frame-pointer -fPIC -c -o /var/folders/__/_sdl9_3s25g2mx7jnq92rvp40000gn/T//ffconf.xeoUy3Zk.o /var/folders/__/_sdl9_3s25g2mx7jnq92rvp40000gn/T//ffconf.xgdp9WUJ.c /var/folders/__/_sdl9_3s25g2mx7jnq92rvp40000gn/T//ffconf.xgdp9WUJ.c:2:18: error: faac.h: No such file or directory /var/folders/__/_sdl9_3s25g2mx7jnq92rvp40000gn/T//ffconf.xg [00:58] <Mavrik-> AnlamK, hmm, your include path is wrong then [00:58] <AnlamK> I do have that faac.h here: /opt/local/include/faac.h [00:59] <Mavrik-> you need to set library/include paths [00:59] <AnlamK> When I do ./configure ? [00:59] <Mavrik-> yea [00:59] <AnlamK> I thought those were automatically set when you did --prefix=''? [00:59] <Mavrik-> well, the gcc lines show that they're not [00:59] <Mavrik-> so i guess --prefix only sets output [01:00] <AnlamK> Ugh I'm gonna give it a try but I mean this is the default: --incdir=DIR install includes in DIR [PREFIX/include] [01:01] <AnlamK> so /opt/local/incude should be there? [01:02] <AnlamK> I tried this: ./configure --prefix='/opt/local/' --enable-libmp3lame --enable-libfaac --enable-nonfree --incdir='/opt/local/include' [01:02] <AnlamK> Still getting the same error [01:02] <Mavrik-> AnlamK, that's an output setting [01:02] <Mavrik-> that means where the avcodec.h files will be copied to [01:02] <Mavrik-> you need to set -I and -L paramemters [01:02] <Mavrik-> for linking [01:04] <AnlamK> Is there an environment variable I can set in my ~/.bash_profile [01:04] <AnlamK> so that I can include /opt/local/include in my linking ? [01:04] <AnlamK> And how do I set the -I and -L parameters? Are you suggesting that I edit the Makefile? [01:05] <Mavrik-> no, there's a config parameter [01:05] <Mavrik-> I don't remember what the exact parameter is, something like LDFLAGS and CFLAGS [01:06] <AnlamK> --extra-cflags=ECFLAGS add ECFLAGS to CFLAGS [] --extra-cxxflags=ECFLAGS add ECFLAGS to CXXFLAGS [] --extra-ldflags=ELDFLAGS add ELDFLAGS to LDFLAGS [] --extra-libs=ELIBS add ELIBS [] [01:11] <AnlamK> Ok making progress [01:11] <AnlamK> gcc -L/opt/local/include -Wl,-dynamic,-search_paths_first -o /var/folders/__/_sdl9_3s25g2mx7jnq92rvp40000gn/T//ffconf.rzzfVeKC /var/folders/__/_sdl9_3s25g2mx7jnq92rvp40000gn/T//ffconf.yesBz2Ld.o -lfaac -lm -framework CoreFoundation -framework VideoDecodeAcceleration -framework QuartzCore -lbz2 -lz ld: library not found for -lfaac collect2: ld returned 1 exit status ERROR: libfaac not found [01:11] <Mavrik-> AnlamK, -L has to point to the .so/.a file directory, -I to .h file directory [01:12] <Mavrik-> usually those are .../lib/ and../include/ [01:12] <AnlamK> I see [01:13] <AnlamK> Thanks so much! [01:13] <AnlamK> I configured [01:13] <AnlamK> What exactly does -L do? [01:13] <AnlamK> I need to read upon how compilers work [01:14] <Mavrik-> AnlamK, -L tells the linker in which (additional) directory to find library (.so/.a) files to link to (the libraries to link to are specified with -l) [01:14] <Mavrik-> -I tells the compiler in which (additional) directory to look for .h files [01:15] <Mavrik-> that is files for#include directive [01:15] <AnlamK> Yeah I know about #include directives... Just trying to figure out what -L does [01:15] <AnlamK> These libraries... These are the apis that the programmer invokes? [01:15] <Mavrik-> yeah [01:16] <Mavrik-> libfaac has encoding APIs for AAC, and the compiler needs to know where to find the library file so it can correctly link function calls [01:16] <Mavrik-> and with -L you tell it where to look for the libfaac.so file [01:16] <AnlamK> Is this libfaac.so file just compiled code? [01:16] <AnlamK> Is it a binary? [01:17] <Mavrik-> it's a binary yea [01:17] <Mavrik-> you probably compiled that before compiling ffmpeg if you have it in that folder [01:18] <AnlamK> yeah i compiled faac separately [01:18] <AnlamK> i'm a little surprised because i thought /opt/local/include and /opt/local/lib were standard include/lib directories [01:18] <Mavrik-> no they're not [01:18] <AnlamK> what's the standard? /usr/local/include? [01:19] <Mavrik-> /usr/lib, /usr/include, /usr/share/lib, /usr/shrare/include are defaults on most distros [01:19] <Mavrik-> or /usr/local/lib,/usr/local/include [01:19] <Mavrik-> files in those are also automatically found [01:19] <AnlamK> i'm using a package installer called macports - it's adding the stuff at /opt/local [01:19] <Mavrik-> ah, you're on OS X [01:19] <Mavrik-> different there probably :) [01:20] <AnlamK> also if you feel for it - i have another question [01:20] <AnlamK> do you know the statistics package R? [01:21] <Mavrik-> heard of it, never used it :) [01:21] <AnlamK> is there a way to use ffmpeg to read .avi files in R? [01:21] <AnlamK> i'm working on this challenge at kaggle.com [01:21] <Mavrik-> huh... sorry, I don't know :\ [01:21] <AnlamK> ok ok thanks [01:21] <AnlamK> i mean now i'm using octave to read them and then load them to r [01:21] <AnlamK> it would save some hassle to read them directly to r [01:21] <AnlamK> but this is fine as well [01:23] <AnlamK> the native aac codec at ffmpeg won't play in ituens so i'm using libfaac to encode some audio into .aac files [01:24] <Mavrik-> yeah, the native aac is still experimental and broken [01:24] <Mavrik-> most people use libfaac [01:24] <Mavrik-> some are now switching to libvo_aac [01:28] <AnlamK> is libvo_aac better than lame? [01:28] <AnlamK> this also doesn't play on itunes [01:28] <AnlamK> i think I'll just stick with liblame [01:29] <Mavrik-> put the encoding command and output to a pastebin [01:29] <Mavrik-> it should play in iTunes just fine [01:31] <AnlamK> Here is the command and the output [01:31] <AnlamK> http://pastebin.com/TUC9tqGC [01:32] <Mavrik-> AnlamK, I think iTunes just doesn't like the container [01:32] <Mavrik-> rename the output file from .aac to .m4a [01:32] <Mavrik-> and it should play just fine [01:33] <cbreak> gcc -print-search-dirs, `gcc -print-prog-name=cc1` -v [01:33] <AnlamK> when I switch .aac to .m4a [01:33] <AnlamK> it's trying to encode the video as well [01:33] <AnlamK> this is a video off youtube [01:33] <AnlamK> whose audio i'm trying to get [01:34] <AnlamK> so inputs got two streams [01:34] <Mavrik-> AnlamK, add "-vn" to parameters to switch off video [01:34] <AnlamK> dude, you are soooo gooood [01:34] <AnlamK> yes it's working now [01:35] <AnlamK> thanks so much [01:47] <AnlamK> also... [01:47] <AnlamK> these youtube videos already have encoded audio streams [01:47] <AnlamK> i don't want to decode and then re-encode - like i'm doing now [01:48] <AnlamK> is there a way to capture the audio stream directly? [01:50] <Mavrik-> AnlamK, if the audio is already in aac [01:50] <AnlamK> yes the audio is already in acc [01:50] <Mavrik-> then you can just do ffmpeg -i <infile> -vn -acodec copy output.m4a [01:50] <Mavrik-> that will just copy and remux the stream without reencoding it [01:52] <AnlamK> yes!! sounds better [01:52] <AnlamK> Mavrik, thank you so much [02:40] <GorDonFreeMan> some relaxing animation after a long night coding :) http://matrix.z-labor.com/tmp/C/sdltext_fps_d_asdf2.c [07:00] <darkstarbyte> will ffmpeg repair damaged streams like how vlc will play damaged streams? [07:35] <Phazorx> hola, i'm trying to figure out a way to play some avc videostream (came from video registrator) it works in "default" player that came with hardware, which provides zero informaitonm about format and exports video stream as a group of 3 files (.avc, .idx. and .grp). The output i get on my box is - http://dpaste.com/677938/ plus a solid grey screen. Any ideas what options i might try to be able to see the video? [08:06] <darkstarbyte> Phazorx, Are you running windows? [08:08] <Phazorx> does it look like windows from the paste? [08:08] <Phazorx> nope, i have gentoo on my box [08:08] <darkstarbyte> You say the source is from a camera right? [08:08] <Phazorx> the original soft which came with hardware is windows and very proprietry [08:09] <Phazorx> it is from ip cam but i font no idea what runs on device which archives the foorage [08:09] <Phazorx> and that software is pretty much the only official way to connect to it [08:09] <Phazorx> so i kinda have to deal with what it gives as export [08:09] <darkstarbyte> See if you can download a clip from it [08:09] <darkstarbyte> then go into the terminal and use [08:09] <darkstarbyte> ffprobe [08:10] <darkstarbyte> like [08:10] <darkstarbyte> ffprobe file_name.ext [08:10] <Phazorx> i have the clip, which is how i got the output on that paste [08:10] <Phazorx> lemme try ffprobe [08:11] <Phazorx> http://dpaste.com/678208/ [08:12] <Phazorx> what i see via default player look more like 4:3 and probably 12-16 fps [08:12] <darkstarbyte> Did that help? [08:13] <Phazorx> same errors which make me think that codec cant handle the input [08:13] <Phazorx> so either encoder uses some weird profile or the codec is unable to deal with the stream for some other reason [08:14] <Phazorx> the questio is how do i figure out what exactly does it want [08:14] <Phazorx> accroding to specs in pdf that came with the device it is supposed to be baseline 3.0 profile [08:15] <Phazorx> if i run ff* with --help i get a whole bunch of info, but i'm not sure how to use it (specifically atributes for options which seems to be flags?) [08:16] <Phazorx> for example how do i force particular workaround for a bug or how do i specify specific options for some codec? [08:17] <darkstarbyte> there is a way to do that, but I have not faced anything like that before. [08:17] <Phazorx> also, considering i have .avc and .idx (so it looks similar how TS would) how do i specify auxilary input files? [08:18] <burek> it's h264 video stream [08:19] <Phazorx> i'm pretty sure it is [08:19] <burek> anyway [08:19] <burek> how do you connect to your camera [08:20] <Phazorx> it's an ip cam and i dont connect to it [08:20] <Phazorx> this is archived footage [08:20] <burek> why don't you connect to it and take a live stream [08:21] <Phazorx> specifically at this point i need the archoved footage for certain period [08:21] <Phazorx> and i'm pretty sure cameras are on different net due to security reasons so it might not be easy to connect [08:21] <burek> well what's the purpose of ip camera then [08:21] <Phazorx> security cameras [08:22] <burek> you could more conveniently buy a dvr [08:22] <burek> for less money for the same purpose [08:22] <Phazorx> it is a dvr like system [08:22] <burek> well, what is your overall goal [08:22] <burek> what are you trying to achieve [08:23] <Phazorx> i am trying to cut out particular segment from a captured stream [08:23] <Phazorx> and i want to be able to convert it to soemtihng that cab be consumed by regular media players (win, linux, whatever) [08:24] <shifter1> are there any devs in the channel? [08:24] <burek> shifter1, try #ffmpeg-devel [08:25] <shifter1> thanks [08:25] <Phazorx> i have like an hour worth of footage in form of avc bitstream, i need to cut like 2 minutes form it and wrap it in some conventional container [08:25] <Phazorx> i'm sure i'll be fine with converting and containing the video once i am able to see it [08:26] <burek> Phazorx, try running mediainfo on the .avc file [08:26] <Phazorx> i did [08:26] <Phazorx> it isays it is an avc [08:27] <Phazorx> if you want exact output - i can launch virtualbox [08:27] <burek> no need [08:27] <burek> just [08:27] <burek> are you able to play that stuff in ANYTHING [08:27] <Phazorx> nope [08:27] <burek> not even their prop. softw? [08:27] <Phazorx> only their soft on winbox [08:28] <Phazorx> but their soft is not really for playing it is for exporting data [08:29] <Phazorx> basiacly it connect to storage array and lets you select node and timeframe [08:29] <burek> well, i don't know then.. it seems you have a device that doesn't work.. [08:29] <Phazorx> and it is able to export the data (during that stage you see the preview) and it results in a group of 3 files (.avc .grp .idx) [08:30] <burek> return it, take money back and buy something more appropriate :) [08:30] <Phazorx> considering i can see the footage on the device - it seems to work [08:30] <burek> well yes, but think about it [08:30] <Phazorx> and considering player shows the preview it seems that footage is more or less valid [08:30] <burek> if they really want to keep you as their customer, by forcing you to use (and pay) their tools only [08:30] <burek> you shouldnt support them [08:30] <burek> so the least you can do is not to buy from them anything anymore [08:31] <Phazorx> this isnt really my choice i'm just helping a co worker [08:31] <burek> well, ok :) your choice :) [08:31] <Phazorx> i have nothing to do with the security setup or anythong relevant to how it is used [08:31] <burek> I've been working with people who intentionally switched bytes in video stream, just so that end users can't watch the video with standard tools [08:31] <burek> so they are practically forced to use their (shitty) player [08:32] <Phazorx> that could very well be the case [08:32] <shifter1> im having a problem when ffmpeg tries to read from a character device. it seems to fail with EOVERFLOW on a read() [08:32] <burek> the smartest choice would be to reject such companies [08:32] <burek> and provide no support for them [08:32] <burek> i.e. not buy anything from them [08:32] <shifter1> I looked up tha man pages for read, and EOVERFLOW doesn't seem to be in the possible erros [08:33] <shifter1> any ideas? [08:33] <Phazorx> again that is good for future [08:33] <Phazorx> in mean time the event i want to see has already ahppened [08:33] <Phazorx> and is captured using this particular hartdware [08:33] <burek> call the company that sold the camera to your friend [08:33] <burek> and ask them? [08:33] <Phazorx> not like i have options on re-selecting the company at that stage [08:34] <Phazorx> burek: it's an option but will probbaly result in lengthy email conversation [08:34] <burek> fortunately there is an option in your case [08:34] <Phazorx> i was hoping that there are just ways to play with ffmpeg to try different aproaches [08:34] <burek> in my case the company closed the doors couple of months after the big purchase happened [08:34] <Phazorx> that is not nice for sure [08:34] <burek> so the people didnt really have any kind of support [08:35] <burek> avoid emails, call them [08:35] <burek> use skype, cheap calls [08:35] <Phazorx> anyway, my original question was only on format in which ff* tools accept attrinbutes for their params [08:35] <Phazorx> i dont get at all what "EDVAS" flags mean that i see in help [08:36] <shifter1> E = encode [08:36] <shifter1> D = Decode [08:36] <shifter1> V = video [08:36] <burek> shifter1, how did you figure it was EOVERFLOW? [08:36] <shifter1> A = Audiso [08:36] <shifter1> strace said so [08:36] <shifter1> S = subtitles [08:36] <Phazorx> shifter1: thanks [08:37] <burek> its just a constant, you know, so try to see what the actual number is and compare it to the return codes that read() returns [08:37] <Phazorx> shifter1: -err_filter <flags> .DVA. [08:37] <shifter1> the read returns -1 [08:37] <Phazorx> and if i want to use "bitstream" option [08:37] <shifter1> errno is set to EOVERFLOW [08:38] <burek> i see [08:38] <shifter1> ffmpeg fails with "PES packet size mismatch. [08:38] <shifter1> [mpeg2video @ 0x9237720] invalid cbp at 14 38" [08:39] <shifter1> upon further inspection, its because the read() call fails [08:39] <shifter1> its reading from a character device, specifically a tvtuner card [08:39] <burek> http://linux.die.net/man/3/errno [08:39] <burek> (ENOTSUP and EOPNOTSUPP have the same value on Linux, but according to POSIX.1 these error values should be distinct.) [08:39] <burek> EOVERFLOW [08:39] <burek> oops sorry :D [08:39] <shifter1> the weird thing is [08:40] <shifter1> cattting the character device to a real file and then passing that file into ffmpeg works [08:40] <shifter1> but ffmpeg directly reading from the character device fails [08:40] <burek> use fifo? [08:41] <burek> if the dev is not ready, it's normal for read to fail, unless you specificaly told it to wait [08:41] <shifter1> tried, the error just appears on the cat into the fifo side [08:41] <burek> which could block [08:41] <shifter1> I was hoping it would block [08:41] <shifter1> rather than fail [08:42] <shifter1> maybe this is a problem with the driver for the device? [08:42] <burek> well if cat fails too, that's not an ffmpeg issue, is it [08:42] <burek> :) [08:42] <shifter1> in that case [08:42] <shifter1> I have an OT question [08:43] <burek> :) [08:43] <shifter1> is there a way to pipe stderr from one program into stdin of another? [08:43] <burek> sure [08:44] <shifter1> strace dumps to stderr [08:44] <burek> 2 > &1 [08:44] <shifter1> im doing: strace cat /dev/dvb/adapter0/dvr0 > out.ts [08:44] <shifter1> so stdout is being used already [08:45] <shifter1> im trying to grep the stderr output, so redirect stderr into stdout would not work [08:45] <burek> http://stackoverflow.com/questions/7431689/how-to-connect-stderr-to-stdin-u… [08:45] <burek> fifo is the way togo [08:45] <shifter1> I was hoping to avoid temporary files [08:46] <burek> or just redirect 2>&1 and then 1>/dev/null [08:46] <shifter1> even ephemeral fifos [08:46] <burek> make sure 2>&1 comes before 1>/dev/null [08:46] <burek> take a look at that link [08:46] <burek> you'll see there [08:47] <shifter1> works, thanks [08:48] <burek> :beer: [08:51] <Phazorx> may i interjecT? [08:52] <Phazorx> http://dpaste.com/678214/ for this specific option how do i pass "ms" attribute, for example, to -bug option (and further more - how do i combine it with some other) [08:54] <burek> -bug ms [08:56] <shifter1> burek, I think ffmpeg is still doing something funny. catting from the cfile to the fifo and then catting the fifo into a real file works [08:56] <shifter1> but if the second cat is replaced with ffmpeg, it fails [08:57] <Phazorx> hmm... what does ffprobe rely on - supplied fourcc or it actualy checks format signature? [08:58] <shifter1> http://pastebin.com/KfurdJxe [08:59] <burek> Phazorx, if you ommit -f before -i then it autodetects [08:59] <burek> shifter1, just a sec [09:00] <burek> oh.. [09:00] <burek> shifter1, that doesn't necessairly have to work [09:00] <burek> because, when you have fifo [09:00] <burek> it's a stream of bytes [09:00] <burek> which might not (and usually is not) seekable [09:00] <burek> file, on the other hand is [09:00] <burek> meaning, ffmpeg, when reading a file [09:00] <burek> can seek to the end of file, to read the metadata [09:01] <burek> like length, fps etc [09:01] <burek> with fifo it can't do that [09:01] <burek> anyway [09:01] <burek> why don't you do ffmpeg -i /dev/dvb... [09:01] <burek> and give us the pastebin [09:02] <shifter1> http://pastebin.com/rcgPUWFj [09:02] <shifter1> that's with the fifo, but it's the same effect [09:02] <shifter1> 3 is the fifo [09:02] <shifter1> 4 is out.mkv [09:02] <shifter1> when read returns 0, the PES warning shows up [09:03] <burek> please try ffmpeg -i /dev/dvb/adapter0/dvr0 [09:03] <burek> and show the pastebin [09:04] <shifter1> http://pastebin.com/00zJ0bJL [09:05] <shifter1> without strace you mean? [09:05] <Phazorx> burek: autodetects - yes, but what exactly the detection is based on ? [09:05] <burek> did i mention strace anywhere? :) [09:06] <burek> Phazorx, based on matching first N bytes of stream with known formats [09:06] <shifter1> http://pastebin.com/ymyqCY9k [09:06] <shifter1> the dvr0 device should be infinite [09:06] <shifter1> it should never have an end [09:08] <burek> Phazorx, try with -analyzeduration <microseconds> [09:09] <burek> shifter1, something is weird, there are too many errors [09:09] <shifter1> -analyzeduration <int> .D... how many microseconds are analyzed to estimate duration [09:09] <burek> concealing 1760 DC, 1760 AC, 1760 MV errors [09:09] <shifter1> burek, the postinit failures are normal [09:10] <shifter1> those errors don't occur when catting dvr0 to a file, and then passing that to ffmpeg [09:10] <shifter1> I can turn up log level if you would like [09:12] <Phazorx> burek: thanks... now i tihnk i understand... and previously you mentioned -i is that this option: -i avconv compatibility dummy option ? [09:13] <burek> Phazorx, ? [09:13] <Phazorx> 11:59 < burek> Phazorx, if you ommit -f before -i then it autodetects [09:14] <burek> ffmpeg -f <format> -i <input> ... [09:14] <burek> that tells ffmpeg to force the format to <format> [09:14] <burek> i.e. disable autodetect [09:14] <burek> if you know your video is mp4 but some idiot named the file .jpg [09:14] <burek> you can still say ffmpeg -f mp4 -i fool.jpg [09:15] <burek> it will read mp4 file without problem [09:15] <burek> shifter1, i don't know what to say.. something is wrong at the very beginning [09:15] <burek> so I don't know where to start off :) [09:15] <Phazorx> burek: yup got that [09:15] <Phazorx> now it also shows this: Seems stream 0 codec frame rate differs from container frame rate: 50.00 (50/1) -> 25.00 (50/2) [09:16] <Phazorx> which kinda makes sense since the data looks like half fields based on size [09:17] <burek> try adding -deinterlace [09:24] <Phazorx> err.. --deinterlace is for ffmpeg... but is there same option for ffplay [09:25] <Phazorx> since at this time i want to be able to see if the video actualy shows something rather than monotone gray before encoding it to sometihng [09:25] <Phazorx> i guess i can pipe it [09:26] <burek> :D [09:26] <burek> ffplay uses almost the same params as ffmpeg, as all of ff* binaries do :) [13:13] <JoeGibken> Hi! I tried installing the script I found here --> https://sites.google.com/site/linuxencoding/install-script -- but it failed and gave me this error --> ==> ERROR in build_libvpx(); git clone failed! -- Then the build stopped. Would it be alright to ask how to fix this? I'm using Linux Mint Debian 64bit. [13:16] <JoeGibken> Oh - hmmm. Now that I realized it, sorry, I guess this is the wrong channel to ask that. [13:18] <Mavrik> huh [13:18] <Mavrik> ffmpeg doesn't support RGB encoding? O.o [14:39] <Slony> Hey, what's wrong here? Using ffmpeg-0.9 [14:39] <Slony> ffmpeg -f x11grab -r 30 -s 1280x760 -i :0.0 -an -vcodec libx264 -preset lossless_ultrafast -threads 0 /tmp/foo.mp4 [14:40] <Slony> Output #0, mp4, to '/tmp/foo.mp4': [14:40] <Slony> Stream #0:0: Video: h264, yuv420p, 1280x760, q=-1--1, 90k tbn, 30 tbc [14:40] <Slony> Stream mapping: [14:40] <Slony> Stream #0:0 -> #0:0 (rawvideo -> libx264) [14:40] <Slony> Error while opening encoder for output stream #0:0 - maybe incorrect parameters such as bit_rate, rate, width or height [14:41] <Slony> all lossless* presets do that, the others work fine [15:03] <Slony> and if i cant use those lossless* presets due to this bug, which other preset would you recommend as an intermediate format for recording video game screencasts? [15:03] <Mavrik> Slony, if you're doing video games, I suggest using very low crf [15:04] <Mavrik> no need for lossless [15:05] <Slony> Mavrik ok, and how do I find out what other presets I have? [15:05] <JEEB> Slony, check if the preset is setting any rate control >_> [15:05] <JEEB> also, generally now you can just use x264's internal presets [15:05] <Slony> the error i reported above seems to be the same as was reported and fixed in 32 bit ubuntu half a year ago [15:06] <Slony> JEEB how do i list x264 presets? [15:06] <JEEB> and set qp to 0 for lossless (or crf to 0 for lossless with 8bit x264) [15:06] <JEEB> Slony, x264 --fullhelp or http://mewiki.project357.com/wiki/X264_Settings#preset [15:07] <Slony> i don't have an 'x264' binary [15:07] <JEEB> then just check the latter page :) [15:07] <Slony> media-libs/x264-0.0.20111220 (gentoo) is installed, 8 bit [15:07] <Slony> sure thing, thanks :] [15:08] <JEEB> basically you should be able to do -vcodec libx264 -preset veryfast -crf 0 with an 8bit libx264 [15:08] <JEEB> for lossless [15:08] <JEEB> for >8bit you'd need explicitly qp 0 for lossless [15:08] <Slony> why the higher bit depth? [15:09] <JEEB> ? [15:09] <Slony> why 10 instead of 8? [15:09] <Mavrik> JEEB, he's doing video game screencasts, >8bit is kinda pointless :) [15:09] <Slony> is that for better color fidelity during things like gamma conversions etc? [15:10] <JEEB> Mavrik, yes -- I just have to mention that crf 0 is not lossless with >8bit [15:10] <JEEB> I didn't mean to recommend them or anything [15:10] <Mavrik> ah, I see [15:11] <JEEB> Slony, 10 is a separate H.264 profile and all that stuff + the last bit depth where certain calculations can still be kept within 16 bits. Anyways, the benefits are mostly with the bigger precision leading to less problems with artifacts that would come from general errors in lossy compression [15:12] <JEEB> But... it's not really recommended to use >8bit H.264 when delivering video to most users [15:12] <JEEB> as it is both slower to decode, doesn't have (pretty much any) end-user level hardware decoding support and all those other things :3 [15:14] <Slony> aha, thank you [16:09] <cn1209> Hello. I'm trying to see how I can capture a frame based on it's time. The file type is flv [16:10] <cn1209> Currently using php [16:12] <Slony> JEEB Mavrik I just read about tune. Do you recommend using --tune animation when recording video games and software video tutorials? [16:13] <Mavrik> Slony, not really [16:13] <Mavrik> Animation is meant for videos that feature strong edges [16:13] <Mavrik> video games are exact opposite of that [16:14] <Slony> ok [16:15] <iive> at least most of them. 3d shooters and such :) [16:15] <Mavrik> not much cell-shaded games these days :) [16:18] <iive> you should still have some time to get humble budle 4. [16:18] <iive> few hours probably. [16:19] <Mavrik> meh, nothing fun there. [16:20] <iive> you've played them all before? [16:21] <Mavrik> I played one or two, rest of them don't look all that fun. [16:23] <Slony> Mavrik you suggested a very low crf for video games for the initial capture, how low did you have in mind? [16:24] <Mavrik> Slony, that's something you have to decide for yourself... for intermediate you can do like 18 and go up from that to get optimal quality/size ratio [16:25] <Slony> aight [16:25] <Slony> i did this a few months ago, if anyone's interested [16:25] <Slony> www.seeitmyway.org/articles/ffmpeg_compression_preset_benchmark.pdf [16:26] <Mavrik> that's a terrible comparison [16:26] <Slony> in what way terrible do you mean? [16:27] <Mavrik> there's no indicator of what settings were used, what the target was and it doesn't even attempt to address quality differene [16:27] <Mavrik> *difference [16:28] <Slony> default settings [16:28] <Slony> x264 [16:29] <Mavrik> there's no such thing as default settings [16:30] <Slony> i didnt specify anything other than -vpre * and crf=23 iirc [16:34] <Mavrik> well that's an important detail [16:34] <Mavrik> since without that all data listed have no meaning [16:34] <Mavrik> same without some kind of visual comparison [16:46] <Slony> ok, i'll add that if i ever reat the tests [16:47] <Slony> what should one use with ffmpeg-0.9 as acodec when the intended target is vimeo or youtube, libmp3lame or mp3lame or aac? [16:55] <JEEB> Slony, youtube re-encodes everything you push to it [16:55] <JEEB> so the optimum thing to push them is something lossless :P [16:56] <JEEB> be it PCM, flac, wavpack or whatever [17:01] <Slony> JEEB ok, what about non-youtube stuff that will get played back on my own machine? [17:04] <JEEB> you'll prolly want AAC or vorbis (qtaacenc/neroaacenc/libvo-aacenc and aotuv), depending on the container etc. you're aiming for [17:05] <Mavrik> hmm... how does libvo_aacenc in comparison to libfaac? [17:11] <supernull> Hey All, anyone using ffmpeg for iptv transcoding ? trying to figure out some of the short comings.. [17:15] <JEEB> Mavrik, they seemingly both base more or less on the standard's example encoder :D [17:15] <JEEB> so I'd say they're pretty similar [17:15] <Slony> thanks [17:15] <JEEB> and libvo-aacenc is still distributable because <things> [17:15] <Mavrik> JEEB, I've just noticed slow switch from libfaac to libvo_aacenc that's why I'm asking [17:16] <Mavrik> and most people agree quality of libfaac is abysmal :\ [17:16] <JEEB> the move away from libfaac is purely because of the fact that it has CLEAR licensing issues [17:16] <JEEB> and thus it's not distributable [17:16] <Mavrik> oh, I see [17:16] <JEEB> libvo-aacenc is quite similar, but still distributable [17:17] <Mavrik> I find it funny though, IIRC the libfaac licensing problems come from the fact it's based on a TI reference encoder... and now you say libvo si based on the same O.o [17:17] <JEEB> yes, but it seems like the company that did libvo-aacenc seemingly had the OK for it [17:18] <Mavrik> ok, good :) [17:20] <Slony> what container should i use for an mp4 video stream and aac audio? [17:20] <Mavrik> mp4. [17:21] <Mavrik> ;) [17:21] <JEEB> mp4 most probably >_> [17:21] <Slony> good, thx [17:21] <JEEB> although matroska is mostly just as fine as long as you don't need support from flash's and hardware players' side [17:21] <JEEB> and flv as well I guess [17:21] <JEEB> also [17:21] <Mavrik> I've never found any real plus sides to using matroska over mp4 [17:21] <JEEB> don't call anything "an mp4 video stream" [17:22] <JEEB> Mavrik, mostly when you need more tracks [17:22] <Mavrik> especially since mp4 is widely supported and mkv... isn't. [17:22] <JEEB> like subtitles etc. [17:22] <JEEB> and yes [17:22] <JEEB> if you need hw etc. support [17:22] <JEEB> of course mp4 is preferable [17:22] <JEEB> Slony, there are two MPEG-4 video standards, always say exactly which of them you're dealing with :P [17:23] <Shimmy> Anyone please help me: https://ffmpeg.org/trac/ffmpeg/ticket/747 [17:23] <JEEB> MPEG-4 Part 2 (xvid, divx etc.) and MPEG-4 Part 10 | ITU-T H.264/AVC (x264, mainconcept etc.) [17:23] <Shimmy> If you know C this is an easy job [17:23] <Shimmy> I'm willing to promote with a little [17:25] <Slony> JEEB ok [17:35] <JodaZ> so in the ffmpeg faq i see [17:35] <JodaZ> mkfifo temp1.a [17:35] <JodaZ> ffmpeg -i input1.flv -vn -f u16le -acodec pcm_s16le -ac 2 -ar 44100 - > temp1.a < /dev/null & [17:36] <JodaZ> why is "- > temp1.a" used [17:36] <JodaZ> and not just temp1.a ? [17:36] <Mavrik> probably to force ffmpeg not to mux output and just dump it raw [17:37] <JodaZ> strange way [18:36] <praedo> hello [18:37] <praedo> iive are you online? [18:37] <praedo> or mystica555_ [18:37] <praedo> or cbsrobot [18:37] <praedo> i need your help [18:40] <ePirat> hello [18:42] <ePirat> is there a option that ffmpeg ignores encoding errors? [18:57] <iive> praedo: i'm a little busy atm. but you should never ask to ask, just ask [19:15] <praedo> iive very good [19:15] <praedo> i explain the problem [19:15] <praedo> the other day you helped me with a video that was unable to be deinterlaced [19:15] <praedo> the reason was that it was captured from a bad tape and premiere failed to detect the interlace [19:16] <praedo> i cut the first corrupt seconds and the rest could be correctly exported and deinterlaced [19:16] <praedo> now the problem is very similar but there isn't any corrupt portion or anything [19:16] <praedo> i don't have the raw data because it's not me who did the video [19:16] <praedo> but i was asked to fix it [19:17] <praedo> it's a short video where you can see some people running and in those fast motion scenes everything looks interlaced [19:17] <praedo> maybe it isn't but i'm not sure [19:17] <praedo> could I post it for you and you have a look with ffplay please? [19:18] <praedo> i did -deinterlace and result is exactly the same: ugly lines at fast scenes [19:18] <praedo> i'd like to know how i could fix that result in post-productiong without having to tell my friend to recapture [19:18] <Slony> does smplayer work with ffmpeg0.9? [19:18] <Slony> *compile with [19:37] <Slony> nvmd [19:49] <iive> praedo: same problem. the interlaced frame have been resized as progressive and fields have bleed into each-other. no easy way to fix it without the original. [19:51] <relaxed> praedo: also, "-vf yadif" is the recommended way to deinterlace [20:33] <cbsrobot> praedo: share a little snipped of your video [20:47] <cn1209> hello [20:49] <cn1209> I'm getting this error: The image & cannot be displayed because it contains errors. [20:49] <cn1209> Here's what I have so far: http://pasteit.com/16438 [21:08] <JodaZ> is there any way to have ffmpeg output video info in a easily machine readable way ? [21:13] <sacarasc> JodaZ: Try using mediainfo instead. [21:14] <JodaZ> is that ffmpeg ? [21:18] <sacarasc> No, it's a different program that does what you want to do. [21:21] <ePirat> i have a problem when i try to convert a wmv (created with mimms stream-ripper) to mp4& http://pastebin.com/f8PYHs1x [21:23] <JodaZ> ePirat, the wmv works ? [21:23] <ePirat> vlc can play it, but time codes are broken [21:24] <JodaZ> ccc puts no good streams out ? [21:25] <JodaZ> those fn candy asses [21:25] <ePirat> i wanted to dump h264 streams which didn't worked, so i used mimms to dump the wmv streams [21:25] <JodaZ> how did you try dumping ? [21:26] <ePirat> the h264 streams? [21:26] <JodaZ> yes [21:26] <ePirat> i tried rtmpdump [21:26] <JodaZ> but they should have their stuff archived too so you can pull it later anyways [21:26] <JodaZ> no http stream ? [21:27] <JodaZ> yeah, no wonder seeing this proprietary dependence from those kids [21:28] <ePirat> anyway is there a chance to somehow convert my fucked up wmv's to mp4 or at least correct the broken timestamps? [21:29] <JodaZ> propably better off redownloading it or trying some other tool while noone answers here [21:35] <cbsrobot> ePirat: beeing at ccc and recording in wmv3 is rather offensive .... [21:37] <cbsrobot> ePirat: looks rather like a audio decoding problem [21:38] <cbsrobot> JodaZ: you could also try ffprobe [21:39] <cbsrobot> for ... video info in a easily machine readable way ... [21:43] <JodaZ> the "..." make me suspicious, cbsrobot [21:45] <cbsrobot> a few minutes ago: JodaZ: is there any way to have ffmpeg output video info in a easily machine readable way ? [21:46] <JodaZ> wouldn't " " be more appropriate there ? [21:46] <JodaZ> maybe with a [sic] if something it wrong [21:47] <cbsrobot> ah - may I should have used [...] or <quote></quote> but who cares [21:48] <darkstarbyte> Can anyone help me with some xvid options for converting? [22:01] <Big_Brother> is there a way to avoid dropped frames or is it simply inevitable? [22:04] <darkstarbyte> ? [22:05] <darkstarbyte> It depends on what you mean [22:05] <darkstarbyte> you can set it to the frame rate of the video [22:06] <darkstarbyte> I think that would be -r [22:08] <Big_Brother> I was thinking only half the frames would be dropped if i was encoding a video from 60fps to 30fps but i have MORE dropped frames than i do processed frames. [22:12] <Big_Brother> frame= 2140 fps= 3 q=28.0 Lsize= 44065kB time=00:01:29.08 bitrate=4052.2kbits/s dup=0 drop=3207 [22:35] <BusError> hi guys. I'm compiling/installing ffmpeg on a custom distro, but somehow libavcodec doesn't get installed. I have --enable-shared, and the lib is definitely compiled... but not installed; so VLC doesn't find it. Is there a magic makefile target to install it ? whats the trick ? [22:37] <Big_Brother> is there a straightforward way to take 20 videos and encode them into one video? (20x 1-1.5min videos into one video with no gaps between each clip) [22:45] <ePirat> why does ffmpeg makes these squared artefacts and can i prevent it somehow? https://img.skitch.com/20111227-d6dxf8fr1r26fmtd1mnwgrm7m1.png [22:59] <JodaZ> how does it matter again where i use -ss ? was before -i seek before decode and after decode before seek ? [23:07] <keyvan> JodaZ: i didnt know about this, is this true? [23:07] <JodaZ> keyvan, i think so >_> [23:07] <keyvan> thanks [23:07] <keyvan> i was scratching my head for weeks thinking why is -ss so damn slow [23:08] <JodaZ> in case you try now, tell me which one is the seek before decode one [23:08] <keyvan> ok i'll look [23:40] <keyvan> JodaZ: looks like doing -ss before -i is a lot faster than doing -i before -ss [23:40] <keyvan> so i guess -ss before -i seeks before decoding whereas -i before -ss decodes then seeks [23:42] <JodaZ> kk [00:00] --- Wed Dec 28 2011

1 0

ffmpeg-devel.log.20111226
by burek 27 Dec '11

27 Dec '11

[04:14] <CIA-49> ffmpeg: 03Anton Khirnov 07master * r2bbb472016 10ffmpeg/doc/faq.texi: FAQ: add an entry for common error when using -profile [04:14] <CIA-49> ffmpeg: 03Stefano Sabatini 07master * r3a2ddf7c2c 10ffmpeg/libavcodec/ (avcodec.h pthread.c utils.c): [04:14] <CIA-49> ffmpeg: lavc: add width and height fields to AVFrame [04:14] <CIA-49> ffmpeg: width and height are per-frame properties, setting these values in [04:14] <CIA-49> ffmpeg: AVFrame simplify the operation of extraction of that information, [04:14] <CIA-49> ffmpeg: since avoids the need to check the codec/stream context. [04:14] <CIA-49> ffmpeg: 03Stefano Sabatini 07master * rb58dbb5b03 10ffmpeg/ (6 files in 3 dirs): [04:14] <CIA-49> ffmpeg: lavc: add a sample_aspect_ratio field to AVFrame [04:14] <CIA-49> ffmpeg: The sample aspect ratio is a per-frame property, so it makes sense to [04:14] <CIA-49> ffmpeg: define it in AVFrame rather than in the codec/stream context. [04:14] <CIA-49> ffmpeg: Simplify application-level sample aspect ratio information extraction, [04:14] <CIA-49> ffmpeg: and allow further simplifications. [04:14] <CIA-49> ffmpeg: 03Gautam Gupta 07master * r671005558a 10ffmpeg/ (4 files in 2 dirs): [04:14] <CIA-49> ffmpeg: doxy: add website-alike style to the html output [04:14] <CIA-49> ffmpeg: Signed-off-by: Luca Barbato <lu_zero(a)gentoo.org> [04:14] <CIA-49> ffmpeg: 03Stefano Sabatini 07master * r8a4a5f6ff7 10ffmpeg/libavcodec/ (avcodec.h pthread.c utils.c): [04:14] <CIA-49> ffmpeg: lavc: add format field to AVFrame [04:14] <CIA-49> ffmpeg: The format is a per-frame property, having it in AVFrame simplify the [04:14] <CIA-49> ffmpeg: operation of extraction of that information, since avoids the need to [04:14] <CIA-49> ffmpeg: access the codec/stream context. [04:14] <CIA-49> ffmpeg: 03Anton Khirnov 07master * r64dca32cdf 10ffmpeg/avconv.c: [04:14] <CIA-49> ffmpeg: avconv: implement get_buffer()/release_buffer(). [04:14] <CIA-49> ffmpeg: This will allow memcpy-free passing frames to lavfi. [04:14] <CIA-49> ffmpeg: 03Anton Khirnov 07master * r04a14d4d25 10ffmpeg/avconv.c: [04:14] <CIA-49> ffmpeg: avconv: avoid memcpy in vsrc_buffer when possible. [04:14] <CIA-49> ffmpeg: Decoding Sintel.2010.1080p.mkv with 4 threads on an i7 2600K goes from [04:14] <CIA-49> ffmpeg: 260s to 244s. [04:14] <CIA-49> ffmpeg: 03Anton Khirnov 07master * re1d9dbf2d4 10ffmpeg/libavfilter/ (Makefile buffersrc.h vsrc_buffer.c): [04:14] <CIA-49> ffmpeg: lavfi: add a new function av_buffersrc_buffer(). [04:14] <CIA-49> ffmpeg: It can be used to directly pass a AVFilterBufferRef to lavfi, avoiding a [04:14] <CIA-49> ffmpeg: memcpy. [04:14] <CIA-49> ffmpeg: 03Konstantin Todorov 07master * rc65dfac466 10ffmpeg/libavcodec/mpegvideo.c: [04:14] <CIA-49> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje(a)gmail.com> [04:14] <CIA-49> ffmpeg: 03Anton Khirnov 07master * r06e8d38f2b 10ffmpeg/ (doc/APIchanges libavcodec/version.h libavfilter/avfilter.h): Add minor bumps and APIchanges entries for lavc/lavfi changes. [04:14] <CIA-49> ffmpeg: 03Stefano Sabatini 07master * r1c9e340d35 10ffmpeg/ (5 files in 2 dirs): [04:14] <CIA-49> ffmpeg: lavfi: add avfilter_copy_frame_props() [04:14] <CIA-49> (12 lines omitted) [06:05] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * ra45880dcdb 10ffmpeg/ffmpeg.c: [06:05] <CIA-49> ffmpeg: ffmpeg: fix non ANSI C for(int [06:05] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [06:05] <CIA-49> ffmpeg: 03Matthieu Bouron 07master * r357cb53306 10ffmpeg/ (Changelog libavformat/dv.c): [06:05] <CIA-49> ffmpeg: dv: add timecode to metadata [06:05] <CIA-49> ffmpeg: Reviewed-by: Reimar [06:05] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [11:17] <CIA-49> ffmpeg: 03Clément BSsch 07master * r831f6cd605 10ffmpeg/libavcodec/timecode.c: timecode: fix drop frame in avpriv_timecode_to_string(). [11:22] <CIA-49> ffmpeg: 03Clément BSsch 07master * r3701d547ac 10ffmpeg/libavformat/gxf.c: [11:22] <CIA-49> ffmpeg: gxf: remove gxf_ prefix to timecode metadata keys. [11:22] <CIA-49> ffmpeg: This will allow a shared "timecode" key metadata between formats to ease [11:22] <CIA-49> ffmpeg: third-party extraction. [15:30] <Compn> later [22:14] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r084bd109fc 10ffmpeg/libavformat/id3v2.c: [22:14] <CIA-49> ffmpeg: id3v2: fix variable type for uncompress() [22:14] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [22:14] <CIA-49> ffmpeg: 03Adrian Drzewiecki 07master * r7e09fe15d5 10ffmpeg/libavformat/id3v2.c: [22:14] <CIA-49> ffmpeg: Process compressed id3v2 tags. [22:14] <CIA-49> ffmpeg: ID3v2.4 allows for zlib compressed tags, but libavformat skips them. [22:14] <CIA-49> ffmpeg: Implement code to inflate compressed tags. [22:14] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [22:14] <CIA-49> ffmpeg: 03Marton Balint 07master * r7b2dba1c50 10ffmpeg/ (libavcodec/dvdata.c libavformat/dv.c): [22:14] <CIA-49> ffmpeg: dv: fix avpriv_dv_codec_profile with lowres > 0 [22:14] <CIA-49> ffmpeg: Fixes ticket #44. [22:14] <CIA-49> ffmpeg: Signed-off-by: Marton Balint <cus(a)passwd.hu> [22:14] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [22:37] <CIA-49> ffmpeg: 03Clément BSsch 07master * r2ffee0f450 10ffmpeg/tests/fate/ (14 files): fate: add generic rules for most of the tests/fate/*.mak files. [22:37] <CIA-49> ffmpeg: 03Clément BSsch 07master * r3d38a7260b 10ffmpeg/tests/fate-run.sh: fate: give stderr hint when a test fails. [22:37] <CIA-49> ffmpeg: 03Clément BSsch 07master * rf51469538f 10ffmpeg/tests/Makefile: fate: warn if SAMPLES is not specified. [00:00] --- Tue Dec 27 2011

1 0

ffmpeg.log.20111226
by burek 27 Dec '11

27 Dec '11

[00:48] <ffmpeguser> cbsrobot: thank you for helping out man, i'll try again with a new build once a new one is released, maybe in the one i have these functions are broken [00:48] <ffmpeguser> i'll disconnect now [05:23] <CarlFK> I have a dir full of jpegs named 100-capture.jpg - 3172-capture.jpg - would like to encode them into something not very lossy [05:25] <CarlFK> they are from a webcam, 6fps, most of it is no action, so not much change - hoping to delta that away [06:03] <zap0> CarlFK, ffmpeg docs have examples [06:04] <CarlFK> zap0: man or online? [06:04] <zap0> online [06:04] <zap0> ffmpeg.org [06:06] <CarlFK> http://ffmpeg.org/faq.html#toc-How-do-I-encode-single-pictures-into-movies_… thanks [06:15] <CarlFK> ffmpeg -f image2 -i %d-capture.jpg /tmp/a.mpg; %d-capture.jpg: No such file or directory log: http://dpaste.de/0xwWJ/ [06:19] <CarlFK> -i %03d-capture.jpg did something, but my guess is it stopped at 999? [07:04] <zap0> CarlFK, be careful, it might start at 0 or 1. [07:05] <zap0> CarlFK, yes.. %03 requires leading zero, and only 3 digits. [07:05] <CarlFK> zap0: my range is 100 to 3172 [07:05] <zap0> i do not think you can start at 100. [07:05] <CarlFK> it encoded 3172 frames. kinda weird [07:08] <CarlFK> Input #0, image2, from '%03d-capture.jpg': ... frame= 3172 http://dpaste.de/vgwFt/ [07:10] <CarlFK> I got what I needed... so all is well for now [07:10] <zap0> :) [08:09] <shifter1> Im getting a lot of "[mpeg2video @ 0x86e0100] mpeg_decode_postinit() failure" messages when trying to read from a dvb input source [08:09] <shifter1> mplayer shows the dvb source fine, but dumping the ts and having ffmpeg read from it produces a lot of errors [08:09] <shifter1> any ideas? [08:29] <shifter1> ping [10:00] <Mmike> Jel' mogu flv stavljati u mp4 container? [10:00] <Mmike> Erm, ups :) [10:01] <Mmike> Can a flv codec be inside a mp4 container? ffmpeg complains with: "[mp4 @ 0x25aac40] track 0: could not find tag, codec not currently supported in container ", not sure if I need to compile-in something, or I just have to use -vcoded mpeg4 for mp4 containters? [10:01] <Mmike> do I even make sense? :) [11:25] <Whitewizard> is there a wildcard key for input files? [11:57] <z_eos> hi all [11:57] <z_eos> may somebody help with "bind(port 8090): Address family not supported by protocol" issue, please? [15:08] <aiwass> hi, can anyone tell me if that ffmpeg's m4a decode process is lossless, so it can be used to convert m4a to another lossless format? [15:11] <mystica555_> m4a isn't lossless.. unless its ALAC. [15:11] <LexSfX> ffmpeg will properly and losslessly convert between lossless codecs, assuming the source and target codecs are both lossless [15:12] <LexSfX> and assuming the target codec supports the source sample rate [15:14] <aiwass> yes, if it's ALAC [15:15] <aiwass> thaks, so converting m4a to flac will produce same lossless output [15:15] <aiwass> *thanks [15:18] <LexSfX> assuming the source sample rate is 655350Hz or lower (FLAC's maximum sample rate), that is true. [15:20] <LexSfX> ah, ALAC has a max sample rate of 384000Hz anyway [15:21] <aiwass> so, it's no problem converting. thanks [15:21] <LexSfX> also, both codecs support up to 32 bits per sample, so there should never be a problem from ALAC to FLAC [15:22] <LexSfX> there may be a problem from FLAC to ALAC at high sample rates or 8-bit depth though [17:34] <ami_> can anybody tell me how we set the frame rate for video using avcodeccontext structure in ffmpeg, beacuse there is no structure member like frame rate in ffmpeg library [17:35] <ami_> I want to play h.264 video it look to play very fast so I want to control the frame rate [19:08] <setmeaway> i want to volume up input audio by 30dB, am i doing correctly? -i input -filter "volume=30dB" output [19:11] <setmeaway> it seems -filter "volume=30dB" doesn't work [20:43] <user82> Hi. i tried to read the documentation but i am not sure about the "-r" option. if i got 50fps input and add "-r 25" will it slow the video down or drop half the input frames? [20:47] <JEEB> I think that depends on if you set it before or after -i [20:47] <JEEB> I think before is "override input stuff" and after is "drop/duplicate to make output of this fps" [20:49] <user82> ah okay. thanks ill give it a try! [20:59] <fixxxermet> Getting an error during make while following the ubuntu compilation guide: http://pastebin.com/MWCTUKyX [22:54] <ThiefMaster> hi, i'm trying to convert ac3 audio to wav while mixing all 6 channels into a single one. ffmpeg -i webdl.ac3 -ac 1 webdl.wav seems to create an 1-channel wav but it sounds like some audio channels are lost instead of mixed [00:00] --- Tue Dec 27 2011

1 0

ffmpeg-devel.log.20111225
by burek 26 Dec '11

26 Dec '11

[00:23] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r2278ecc434 10ffmpeg/libavcodec/vmdav.c: [00:23] <CIA-49> ffmpeg: vmdav: check that theres enough space for a chunk remaining. [00:23] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [00:23] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r43abef9fde 10ffmpeg/libavformat/rpl.c: [00:23] <CIA-49> ffmpeg: rpl: Fix near infinite loop in index reading due to missing eof check. [00:23] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [00:23] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * rfc5c49ab32 10ffmpeg/libavcodec/mpeg4videodec.c: [00:23] <CIA-49> ffmpeg: mpeg4videodec: Fix division by zero in mpeg4_decode_sprite_trajectory() [00:23] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [00:23] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r66f71f3b5e 10ffmpeg/libavformat/rmdec.c: [00:23] <CIA-49> ffmpeg: rmdec: Avoid allocating huge packets [00:23] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [00:23] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r47c4713a23 10ffmpeg/libavformat/sierravmd.c: [00:23] <CIA-49> ffmpeg: sierravmd: limit packetsize to the amount that could be read. [00:23] <CIA-49> ffmpeg: Fixes huge allocations. [01:22] <CIA-49> ffmpeg: 03Baptiste Coudurier 07release/0.5 * r30ee6c1995 10ffmpeg/libavcodec/qdm2.c: [01:22] <CIA-49> ffmpeg: Fix qdm2 decoder packet handling to match the api [01:22] <CIA-49> ffmpeg: Originally committed as revision 25767 to svn://svn.ffmpeg.org/ffmpeg/trunk [01:22] <CIA-49> ffmpeg: (cherry picked from commit b26c1a8b7ed1a199b19f92bb5d62c61f1c149215) [01:22] <CIA-49> ffmpeg: Signed-off-by: Reinhard Tartler <siretart(a)tauware.de> [01:22] <CIA-49> ffmpeg: 03Justin Ruggles 07release/0.5 * r60eebf5c12 10ffmpeg/libavcodec/qdm2.c: (log message trimmed) [01:22] <CIA-49> ffmpeg: qdm2: check output buffer size before decoding [01:22] <CIA-49> ffmpeg: (cherry picked from commit 7d49f79f1cd47783a963a757a6563b9cac29db62) [01:22] <CIA-49> ffmpeg: Signed-off-by: Reinhard Tartler <siretart(a)tauware.de> [01:22] <CIA-49> ffmpeg: (cherry picked from commit 73472053516f82b7d273a3d42c583f894077a191) [01:22] <CIA-49> ffmpeg: Conflicts: [01:22] <CIA-49> ffmpeg: libavcodec/qdm2.c [01:22] <CIA-49> ffmpeg: 03Laurent Aimar 07release/0.5 * r8abf1d882e 10ffmpeg/libavcodec/qdm2.c: (log message trimmed) [01:22] <CIA-49> ffmpeg: Fix out of bound reads in the QDM2 decoder. [01:22] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [01:22] <CIA-49> ffmpeg: Signed-off-by: Justin Ruggles <justin.ruggles(a)gmail.com> [01:22] <CIA-49> ffmpeg: (cherry picked from commit 5a19acb17ceb71657b0eec51dac651953520e5c8) [01:22] <CIA-49> ffmpeg: Signed-off-by: Reinhard Tartler <siretart(a)tauware.de> [01:22] <CIA-49> ffmpeg: (cherry picked from commit 0d93d5c4614fafea74bdac681673f5b32eb49063) [01:22] <CIA-49> ffmpeg: 03Alex Converse 07release/0.5 * r7739947671 10ffmpeg/libavcodec/vp6.c: (log message trimmed) [01:22] <CIA-49> ffmpeg: vp6: Fix illegal read. [01:22] <CIA-49> ffmpeg: (cherry picked from commit 2a6eb06254df79e96b3d791b6b89b2534ced3119) [01:22] <CIA-49> ffmpeg: Signed-off-by: Reinhard Tartler <siretart(a)tauware.de> [01:22] <CIA-49> ffmpeg: (cherry picked from commit 67a7ed623b678a84c992dd7bf3e3d0329f83621b) [01:22] <CIA-49> ffmpeg: Signed-off-by: Reinhard Tartler <siretart(a)tauware.de> [01:22] <CIA-49> ffmpeg: (cherry picked from commit 8d68083298e2481669de4db0b7b86c915119df6d) [01:22] <CIA-49> ffmpeg: 03Thierry Foucu 07release/0.5 * r771ceb19f2 10ffmpeg/libavcodec/vp6.c: (log message trimmed) [01:23] <CIA-49> ffmpeg: vp6: Fix illegal read. [01:23] <CIA-49> ffmpeg: Found with Address Sanitizer [01:23] <CIA-49> ffmpeg: Signed-off-by: Alex Converse <alex.converse(a)gmail.com> [01:23] <CIA-49> ffmpeg: (cherry picked from commit e0966eb140b3569b3d6b5b5008961944ef229c06) [01:23] <CIA-49> (50 lines omitted) [01:35] <CIA-49> ffmpeg: 03Tomas Härdin 07release/0.6 * r2fc39af1de 10ffmpeg/libavformat/riff.c: [01:35] <CIA-49> ffmpeg: Added M701 codec_tag for mpeg2video [01:35] <CIA-49> ffmpeg: Originally committed as revision 23513 to svn://svn.ffmpeg.org/ffmpeg/trunk [01:35] <CIA-49> ffmpeg: (cherry picked from commit 77d3f1f792c2a0bd455c75b0aa8e356b3a470ea7) [01:35] <CIA-49> ffmpeg: Signed-off-by: Reinhard Tartler <siretart(a)tauware.de> [01:36] <CIA-49> ffmpeg: 03Alex Converse 07release/0.6 * r394c3e78d5 10ffmpeg/libavformat/riff.c: [01:36] <CIA-49> ffmpeg: riff: Add mpgv MPEG-2 fourcc [01:36] <CIA-49> ffmpeg: Supported by mplayer and seen in the wild. [01:36] <CIA-49> ffmpeg: (cherry picked from commit 505345ed5d180093a44da8d70ac541898c31c22f) [01:36] <CIA-49> ffmpeg: Signed-off-by: Reinhard Tartler <siretart(a)tauware.de> [01:36] <CIA-49> ffmpeg: 03Michael Niedermayer 07release/0.6 * rfeca20021c 10ffmpeg/libavcodec/mjpegdec.c: (log message trimmed) [01:36] <CIA-49> ffmpeg: jpegdec: actually search for and parse RSTn [01:36] <CIA-49> ffmpeg: Fixes decoding of MJPEG files produced by some UVC Logitec web cameras, [01:36] <CIA-49> ffmpeg: such as "Notebook Pro" and "HD C910". [01:36] <CIA-49> ffmpeg: References: [01:36] <CIA-49> ffmpeg: 03Piotr Kaczuba 07release/0.6 * r7487d53d01 10ffmpeg/libpostproc/postprocess.c: [01:36] <CIA-49> ffmpeg: postprocess.c: filter name needs to be double 0 terminated [01:36] <CIA-49> ffmpeg: Signed-off-by: Reinhard Tartler <siretart(a)tauware.de> [01:36] <CIA-49> ffmpeg: (cherry picked from commit f4f3300c09bb13eb7922e60888b55e3e0fb325e7) [01:36] <CIA-49> ffmpeg: (cherry picked from commit 20ca827019a72bfacb38e73d0b8590e651818272) [01:36] <CIA-49> ffmpeg: Signed-off-by: Reinhard Tartler <siretart(a)tauware.de> [01:36] <CIA-49> ffmpeg: 03Jason Garrett-Glaser 07release/0.6 * r572e94bc51 10ffmpeg/libavcodec/ (mpegvideo.c mpegvideo.h): [01:36] <CIA-49> ffmpeg: H.264: fix overreads of qscale_table [01:36] <CIA-49> ffmpeg: filter_mb_fast assumed that qscale_table was padded like many of the other tables. [01:36] <CIA-49> ffmpeg: (cherry picked from commit 5029a406334ad0eaf92130e23d596e405a8a5aa0) [01:36] <CIA-49> ffmpeg: Signed-off-by: Anton Khirnov <anton(a)khirnov.net> [01:36] <CIA-49> ffmpeg: Signed-off-by: Reinhard Tartler <siretart(a)tauware.de> [01:36] <CIA-49> ffmpeg: 03Clément BSsch 07release/0.6 * r86849f0976 10ffmpeg/libavformat/mxfenc.c: [01:36] <CIA-49> ffmpeg: mxfenc: fix ignored drop flag in binary timecode representation. [01:36] <CIA-49> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje(a)gmail.com> [01:36] <CIA-49> ffmpeg: (cherry picked from commit 4d5e7ab5c48451404038706ef3113c9925a83087) [01:36] <CIA-49> ffmpeg: Signed-off-by: Anton Khirnov <anton(a)khirnov.net> [01:36] <CIA-49> ffmpeg: Signed-off-by: Reinhard Tartler <siretart(a)tauware.de> [01:36] <CIA-49> ffmpeg: 03Justin Ruggles 07release/0.6 * r81d5ceff04 10ffmpeg/libavdevice/alsa-audio-common.c: [01:36] <CIA-49> ffmpeg: alsa: fallback to buffer_size/4 for period_size. [01:36] <CIA-49> ffmpeg: buffer_size/4 is the value used by aplay. This fixes output to null [01:37] <CIA-49> (180 lines omitted) [01:37] <CIA-49> ffmpeg: 03Laurent Aimar 07release/0.6 * r8856c4c5c9 10ffmpeg/libavcodec/qdm2.c: (log message trimmed) [01:37] <CIA-49> ffmpeg: Fix out of bound reads in the QDM2 decoder. [01:37] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [01:37] <CIA-49> ffmpeg: Signed-off-by: Justin Ruggles <justin.ruggles(a)gmail.com> [01:37] <CIA-49> ffmpeg: (cherry picked from commit 5a19acb17ceb71657b0eec51dac651953520e5c8) [01:37] <CIA-49> ffmpeg: Signed-off-by: Reinhard Tartler <siretart(a)tauware.de> [01:37] <CIA-49> ffmpeg: (cherry picked from commit 0d93d5c4614fafea74bdac681673f5b32eb49063) [01:37] <CIA-49> ffmpeg: 03Dustin Brody 07release/0.6 * r201fcfb894 10ffmpeg/libavcodec/vp6.c: [01:37] <CIA-49> ffmpeg: vp6: partially propagate huffman tree building errors during coeff model parsing and fix misspelling [01:37] <CIA-49> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje(a)gmail.com> [01:37] <CIA-49> ffmpeg: (cherry picked from commit f913eeea43078b3b9052efd8d8d29e7b29b39208) [01:37] <CIA-49> ffmpeg: Signed-off-by: Reinhard Tartler <siretart(a)tauware.de> [01:37] <CIA-49> ffmpeg: (cherry picked from commit 7367cbec1b8cf0cbb49707fb0fdfded8ec397b0d) [01:37] <CIA-49> ffmpeg: Signed-off-by: Reinhard Tartler <siretart(a)tauware.de> [05:02] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r27d323577c 10ffmpeg/libavformat/utils.c: [05:02] <CIA-49> ffmpeg: avio: Fix ffio_limit() when due to seeking past the end less than 0 bytes remain. [05:02] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [05:02] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r964506bb97 10ffmpeg/libavcodec/golomb.h: [05:02] <CIA-49> ffmpeg: golomb: Fix infinite loop in svq3_get_ue_golomb() [05:02] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [05:02] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r1795fed7bc 10ffmpeg/libavformat/segafilm.c: [05:02] <CIA-49> ffmpeg: segafilm: fail earlier in case theres not enough bytestream left for a [05:02] <CIA-49> ffmpeg: audio packet. [05:02] <CIA-49> ffmpeg: This prevents a potentially large memory allocation. [05:02] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [17:07] <Compn> is there anyway to get rid of non monotonically increasing dts message when vcodec copy to mp4 from mkv ? [17:07] <Compn> ffmpeg -i crap.mkv -acodec copy -vcodec copy out.mp4 [17:08] <Compn> a video filter that changes dts perhapse ? [17:09] <Compn> oh i'll just use mencoder, nevermind [17:20] <Compn> michaelni : everyone just wants to streamcopy from mkv to mp4 and mp4 to mkv . big feature request that is [17:20] <Compn> fixing it by default would be nice [19:07] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r18bcfc912e 10ffmpeg/libavcodec/shorten.c: [19:07] <CIA-49> ffmpeg: shorten: Fix invalid free() [19:07] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [19:07] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r5e9a56a035 10ffmpeg/libavcodec/shorten.c: [19:07] <CIA-49> ffmpeg: shorten: validate values in fmt chunk search [19:07] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [19:07] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r5a8e994287 10ffmpeg/libavutil/ (avutil.h mem.c mem.h): [19:07] <CIA-49> ffmpeg: mem: add av_max_alloc() to limit the maximum amount that may be allocated in one piece [19:07] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [19:07] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * rc48f67f06e 10ffmpeg/ (cmdutils.c cmdutils.h cmdutils_common_opts.h): [19:07] <CIA-49> ffmpeg: ff*: add -max_alloc command line option [19:07] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [19:28] <Compn> i will be afk for 3 weeks, bbl [19:39] <kierank> bye [19:56] <CIA-49> ffmpeg: 03Reinhard Tartler 07release/0.8 * rd61b38b9db 10ffmpeg/Changelog: Update Changelog for 0.7.3 release [19:57] <CIA-49> ffmpeg: 03Michael Niedermayer 07release/0.8 * r8413f12e1b 10ffmpeg/: [19:57] <CIA-49> ffmpeg: Merge remote-tracking branch 'qatar/release/0.7' into release/0.8 [19:57] <CIA-49> ffmpeg: * qatar/release/0.7: [19:57] <CIA-49> ffmpeg: Update Changelog for 0.7.3 release [19:57] <CIA-49> ffmpeg: Conflicts: [19:57] <CIA-49> ffmpeg: Changelog [19:57] <CIA-49> ffmpeg: Merged-by: Michael Niedermayer <michaelni(a)gmx.at> [20:09] <CIA-49> ffmpeg: 03Michael Niedermayer 07release/0.7 * re66860a66b 10ffmpeg/ (Doxyfile RELEASE VERSION): [20:09] <CIA-49> ffmpeg: Update for 0.8.7 [20:09] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [20:09] <CIA-49> ffmpeg: 03Sergiy Gur'yev 07release/0.7 * r4f58d8ebc1 10ffmpeg/libavcodec/libaacplus.c: [20:09] <CIA-49> ffmpeg: Fix adts format creation in aac+ encoder modified: libavcodec/libaacplus.c [20:09] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [20:09] <CIA-49> ffmpeg: (cherry picked from commit 32ed7da1350e551ec005b75e482da74f2e93fbb9) [20:09] <CIA-49> ffmpeg: 03Laurent Aimar 07release/0.7 * ra31ccacb1a 10ffmpeg/libavcodec/qdm2.c: [20:09] <CIA-49> ffmpeg: Check for out of bound writes in the QDM2 decoder. [20:09] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [20:09] <CIA-49> ffmpeg: Signed-off-by: Justin Ruggles <justin.ruggles(a)gmail.com> [20:09] <CIA-49> ffmpeg: (cherry picked from commit 291d74a46d32183653db07818c7b3407fd50a288) [20:09] <CIA-49> ffmpeg: Signed-off-by: Reinhard Tartler <siretart(a)tauware.de> [20:09] <CIA-49> ffmpeg: 03Justin Ruggles 07release/0.7 * r7347205351 10ffmpeg/libavcodec/qdm2.c: [20:09] <CIA-49> ffmpeg: qdm2: check output buffer size before decoding [20:09] <CIA-49> ffmpeg: (cherry picked from commit 7d49f79f1cd47783a963a757a6563b9cac29db62) [20:09] <CIA-49> ffmpeg: Signed-off-by: Reinhard Tartler <siretart(a)tauware.de> [20:09] <CIA-49> ffmpeg: 03Thierry Foucu 07release/0.7 * r28acce2861 10ffmpeg/libavutil/imgutils.c: [20:09] <CIA-49> ffmpeg: imgutils: Fix illegal read. [20:09] <CIA-49> ffmpeg: Found with address sanitizer. [20:09] <CIA-49> ffmpeg: Signed-off-by: Alex Converse <alex.converse(a)gmail.com> [20:09] <CIA-49> ffmpeg: (cherry picked from commit c693aa6f71b4f539cf9df67ba42f4b1932981687) [20:10] <CIA-49> ffmpeg: vp6: Reset the internal state when aborting key frames header parsing [20:10] <CIA-49> ffmpeg: It prevents leaving the state only half initialized. [20:10] <CIA-49> ffmpeg: Signed-off-by: Janne Grunau <janne-libav(a)jannau.net> [20:10] <CIA-49> ffmpeg: (cherry picked from commit a72cad0a6c05aa74940101e937cb3dc602d7d67b) [20:10] <CIA-49> ffmpeg: Signed-off-by: Reinhard Tartler <siretart(a)tauware.de> [20:10] <CIA-49> ffmpeg: 03Dustin Brody 07release/0.7 * r7367cbec1b 10ffmpeg/libavcodec/vp6.c: [20:10] <CIA-49> ffmpeg: vp6: partially propagate huffman tree building errors during coeff model parsing and fix misspelling [20:10] <CIA-49> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje(a)gmail.com> [20:10] <CIA-49> ffmpeg: (cherry picked from commit f913eeea43078b3b9052efd8d8d29e7b29b39208) [20:10] <CIA-49> ffmpeg: Signed-off-by: Reinhard Tartler <siretart(a)tauware.de> [20:10] <CIA-49> ffmpeg: 03Laurent Aimar 07release/0.7 * r30c08e2261 10ffmpeg/libavcodec/ (vp5.c vp56.c vp56.h vp6.c): [20:10] <CIA-49> ffmpeg: vp6: Check for huffman tree build errors [20:10] <CIA-49> ffmpeg: Signed-off-by: Janne Grunau <janne-libav(a)jannau.net> [20:10] <CIA-49> (106 lines omitted) [20:16] <CIA-49> ffmpeg: 03Reinhard Tartler 07release/0.6 * r6b156c4563 10ffmpeg/ (Changelog RELEASE): Release notes and changelog for 0.6.4 [20:16] <CIA-49> ffmpeg: 03Michael Niedermayer 07release/0.6 * rd32ea79ea2 10ffmpeg/: [20:16] <CIA-49> ffmpeg: Merge remote-tracking branch 'qatar/release/0.6' into release/0.6 [20:16] <CIA-49> ffmpeg: * qatar/release/0.6: [20:16] <CIA-49> ffmpeg: Release notes and changelog for 0.6.4 [20:16] <CIA-49> ffmpeg: Conflicts: [20:16] <CIA-49> ffmpeg: Changelog [20:16] <CIA-49> ffmpeg: Merged-by: Michael Niedermayer <michaelni(a)gmx.at> [20:30] <CIA-49> ffmpeg: 03Reinhard Tartler 07release/0.5 * rd0688fdd31 10ffmpeg/ (Changelog RELEASE): Release notes and changelog for 0.5.6 [20:30] <CIA-49> ffmpeg: 03Michael Niedermayer 07release/0.5 * r8755a7890e 10ffmpeg/: [20:30] <CIA-49> ffmpeg: Merge remote-tracking branch 'qatar/release/0.5' into release/0.5 [20:30] <CIA-49> ffmpeg: * qatar/release/0.5: [20:30] <CIA-49> ffmpeg: Release notes and changelog for 0.5.6 [20:30] <CIA-49> ffmpeg: Conflicts: [20:30] <CIA-49> ffmpeg: RELEASE [20:30] <CIA-49> ffmpeg: Merged-by: Michael Niedermayer <michaelni(a)gmx.at> [20:32] <Compn> thats an oooold release [20:33] <spaam> some ppl like old stuff [20:34] <spaam> just look at CrapOS also known as CentOS [21:37] <JEEB> don't remind me I have at least one CentOS machine under my administration >_> [22:15] <CIA-49> ffmpeg: 03Michael Niedermayer 07release/0.7 * rccdc68eb35 10ffmpeg/ (Doxyfile RELEASE VERSION): [22:15] <CIA-49> ffmpeg: Update for 0.7.9 [22:15] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [22:15] <CIA-49> ffmpeg: 03Michael Niedermayer 07release/0.5 * re75056bc54 10ffmpeg/VERSION: [22:15] <CIA-49> ffmpeg: Update for 0.5.7 [22:15] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [22:15] <CIA-49> ffmpeg: 03Michael Niedermayer 07release/0.8 * rd75909f247 10ffmpeg/ (Doxyfile RELEASE VERSION): [22:15] <CIA-49> ffmpeg: Update for 0.8.8 [22:15] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [23:05] <spaam> JEEB: :( what os do you have on the other ones? [23:11] <JEEB> debian/ubuntu [00:00] --- Mon Dec 26 2011

1 0

ffmpeg.log.20111225
by burek 26 Dec '11

26 Dec '11

[00:00] <mystica555_> this to me seems to potentially be something even more.. a framerate conversion in the middle [00:00] <praedo> maybe it can be cause because the tape was overwritten 4 times? [00:00] <iive> praedo: there is dvgrab program for linux, that should be able to grab the dv stream, in the way it appears on the camera [00:00] <mystica555_> no [00:00] <mystica555_> there is no way this is a tape issue [00:00] <iive> you MUST get that raw dv stream. [00:01] <praedo> then what can be the cause of it? [00:01] <mystica555_> this is 95% guaranteed premiere dealing with a 24>25 or 30>25 framerate conversion at the field level [00:01] <mystica555_> 4% premiere simply mucking with interlacing *very* badly [00:01] <iive> the exporting is probably ruining it. [00:01] <mystica555_> 1% anything else [00:01] <iive> mystica555_: I said above, this happens when you scale image as progressive, when it is actually interlaced. [00:02] <praedo> i always encode 24p footage using premiere and exported to pal dv at 25 [00:02] <mystica555_> iive: ah; i didnt know if that was the source of premiere's issues, but, i know premiere is an issue. [00:02] <praedo> then i deinterlace with ffmpeg and it's always good [00:02] <iive> the fields are bleeding into each other, in patterns that depend on scale algorithm and parameters. [00:02] <mystica555_> so what you're doing is an upconversion at the field level [00:02] <mystica555_> if you are going to do 24p, speed the film up and do a pitch-corrected new audio track [00:02] <mystica555_> dont try and have it convert by mangling it [00:03] <mystica555_> if anything, dont even pitch correct [00:03] <mystica555_> all the brits watching Star Trek don't realize picard sounds like hes on helium over there [00:04] <mystica555_> being 24p>25i, its trying to keep the motion 'fluid' and interpolating fields and weird crap on you right now. [00:04] <iive> praedo: once again. it seems that your capture program is doing some conversion. you should not allow it to do that. [00:04] <iive> you MUST get the raw dv stream, in the way it is stored on the camera. [00:05] <mystica555_> and even then, with said raw dv stream at 24p, the question is, is it literally 24p, or is it telecine'd 29.97i [00:05] <mystica555_> ive seen cams do both [00:05] <praedo> but it only happened for this footage [00:05] <cbsrobot> why not deinterlace first with ffmpeg and then import to premiere ? [00:05] <praedo> i did hundreds of captured before from 24p to 25i and it never happened before [00:05] <mystica555_> can you, somehow, export a 30-50 megabyte raw video so at least one of us can figure out exactly what you are starting with? [00:05] <iive> cbsrobot: because he needs the raw dv stream! [00:05] <mystica555_> dvgrab for nix etc [00:06] <mystica555_> praedo: then you must find out why this stream is unique [00:06] <mystica555_> it may be that interesting 29.97i conversion considering thats a truely interlaced downscale. [00:06] <iive> praedo: most probably, the previous streams have been properly marked as interlaced, so premiere used proper scaling. [00:06] <mystica555_> 59.94 fields per second meshing into 50i fields [00:07] <praedo> how can i cut a piece of the raw video without having to export it with premiere? [00:07] <iive> praedo: try to find the simplest and dumbest program possible that would capture the dv stream. [00:08] <praedo> no linux on the comptuter where i have the 10gb file [00:08] <praedo> any other fast way to do it? [00:09] <praedo> instead of premiere [00:09] <mystica555_> what made the 10gb file? [00:09] <iive> praedo: the sample you gave me, did you cut it from that file? [00:09] <mystica555_> or is it on a tape at that point? [00:09] <cbsrobot> praedo: premiere captures the file into a project or temp folder [00:09] <cbsrobot> you can use ffmpeg on it to cut the file [00:09] <praedo> that's it [00:09] <iive> praedo: did you cut it with premiere? [00:10] <praedo> the sample is cut from the captured file using premiere [00:10] <praedo> and it was capture also using premiere [00:10] <iive> the cutting may be what created these artifacts. [00:10] <praedo> the whole video also has the same bad thing [00:10] <praedo> after exported [00:10] <praedo> the captured video just plays fine (normally interlaced as usual) [00:11] <iive> ok, don't export it. [00:11] <mystica555_> save the captured dv raw somehow if possible? [00:11] <iive> try to cut a sample using ffmpeg and -vcodec copy -t 11 [00:11] <praedo> instead of grabbing from tape again, i think the best thing i can try is to -deinterlace the captured file before being edited [00:11] <cbsrobot> praedo: I still suggest you deinterlace the captured file with ffmpeg and cut it in premiere afterwards [00:12] <praedo> should i? [00:12] <iive> using the capture, not export. [00:12] <mystica555_> praedo: no [00:12] <mystica555_> figure the reason its being flaky [00:12] <mystica555_> ive got 4 potential ideas in my head right now [00:12] <iive> praedo: yes, use the captured file. [00:12] <mystica555_> and if you deinterlace, it'll a: never properly diagnose the reason for a proper fix and b: may make it worse [00:12] <mystica555_> thus wasting more time [00:13] <praedo> okay, so i transfer the captured file (29i) to the linux box and i cut a sample with ffmpeg, yes? [00:13] <iive> praedo: you don't have ffmpeg for windows? [00:13] <mystica555_> there are ffmpeg binaries for both platforms that have premiere.. [00:13] <cbsrobot> 29i ? or PAL ? [00:13] <praedo> it's recorded as 24p but always captured as 29i and then exported to 25 pal dv [00:13] <praedo> this is what i always do and results are fine [00:13] <iive> praedo: DO NOT EXPORT. [00:14] <iive> work with the raw capture. [00:14] <praedo> no, now i will just cut the captured file [00:14] <praedo> raw = the file captured with premiere? [00:14] <iive> i guess. the file as it is grabbed from the dv camera [00:15] <praedo> exactly [00:15] <praedo> so that i don't have to transfer 10gb from pc to pc, how can i do ffmpeg and -vcodec copy -t 11 using windows? [00:15] <praedo> i never found working binary [00:16] <iive> the ffmpeg site points to this site http://ffmpeg.zeranoe.com/builds/ [00:16] <mystica555_> random comment: IMHO the best way to do a framerate conversion (which i dont know if anything even does, but to me, this seems reasonable) is: motion estimate and upconvert like 240hz LCDs to the greatest common factor of 24, 25, 50 and 60... 600fps. then select one frame at a proper fraction of those 600fps... [00:16] <praedo> avidemux could do it too? [00:17] <iive> get the static builds. [00:18] <praedo> okay... please wait some minutes... [00:19] <praedo> fine, i got the binary [00:19] <cbsrobot> wiki says The camera can record 24p video in "telecined" fashion (2-3 pulldown for 24P and 2-3-3-2 pulldown for 24PA) [00:19] <cbsrobot> what is 24PA ? [00:19] <iive> btw, one unrelated question. where are you going to use the end result, is it going to be burnt on dvd or just uploaded online? [00:19] <praedo> what's the command to cut from second 0 to second 15? [00:20] <mystica555_> it may have done that 2332 pattern..... [00:20] <mystica555_> thats annoying [00:20] <cbsrobot> -t 15 [00:20] <praedo> the end result will be both delivered on dvd (as mp4 file) and posted to be viewed with a flash player [00:21] <praedo> i got paid 40 eur for it! :) [00:21] <mystica555_> so its not even needed to be anything but 24p [00:21] <mystica555_> its not playing back in a PAL dvd player [00:21] <praedo> yes, it was recored using 24PA mode in the dvx100 camera [00:21] <mystica555_> 24pa mode would definitely figure into the weird shuddering artifacts every few seconds [00:22] <mystica555_> if its not expecting the reversal of field cadance [00:22] <mystica555_> so ideally, find something that plays this 24pa format properly and make a 24p output at 720x576 framesize [00:23] <mystica555_> or scale for whatever media you need; its 24p [00:23] <mystica555_> 25fps PAL output is going to convolute the process any which way if your output is 100% not a PAL DVD player circa 1999. [00:24] <iive> praedo: maybe it would be better to record in 24p in future. [00:24] <iive> telecine and interlace work properly only on CRT TV [00:25] <iive> avoiding them whenever possible is good thing (tm) [00:26] <iive> and yes, adobe seems a little too fixated on proper support for PAL/NTSC, that are also things that matter on analog CRT TV. [00:27] <praedo> so no need to record in 24pa [00:27] <praedo> 24p is fine [00:29] <mystica555_> technically from what im reading, if the algorithm detects 24pa properly, it'll work just fine and produce a bit better result even than 24p. just need that detection and proper pulldown step. [00:29] <praedo> okay, i have cut 20 seconds from the raw captured file [00:29] <praedo> uploading now... will take a while [00:29] <praedo> 50mb [00:30] <cbsrobot> praedo: use another filehoster [00:30] <praedo> no... 70mb [00:30] <praedo> okay, which one? [00:30] <cbsrobot> http://www.1fichier.com/ [00:30] <iive> ffmpeg doesn't have any ivtc filters on its own, i hope some of the mp= imports would work correctly. [00:31] <praedo> i played the file with windows media and no artifacts there [00:31] <praedo> problem starts after export [00:31] <praedo> but as you will see in this raw file, the beginning of the tape is corrupt and then video plays fine [00:31] <iive> praedo: it would be quite annoying, if premiere is the one creating the interlace/telecine. [00:32] <praedo> uploading to 1fichier, ok [00:32] <iive> praedo: do you know what telecine is? [00:32] <praedo> why not megaul? [00:32] <praedo> more or less, yes [00:32] <mystica555_> does premiere understand 24pa i wonder.. [00:33] <iive> mystica555_: mpeg1/2 have flags for soft telecine, i won't be surprised if dv have something similar. and premiere may turn it from soft into hard. [00:33] <cbsrobot> megaupload is dogslow [00:34] <iive> donno... for some reason I got it before I hit the save button. :E [00:34] <praedo> yes, premiere does 2:3:3:2 pulldown automatically [00:39] <iive> hum... maybe premiere test just the first minute, so the detection fails, that leads to the rest of your problems. [00:40] <cbsrobot> iive: nice idea [00:41] <iive> at least it sounds plausible. [00:45] <praedo> iive, that could be very possible, yes [00:45] <praedo> 90% of the uploaded file... [00:46] <mystica555_> iive: its not soft.. there is nothing in the dv format to repeat a field. its literally encoding 29.97/59.94 fields and how it combines them is the question. the pulldown method is key [00:47] <iive> :/ [00:47] <praedo> mystica555_, but i always do this 29i capture from 24pa and this never happened before [00:47] <mystica555_> praedo: try exporting as 24p without going to 25i [00:48] <mystica555_> if premiere understands what the 2:3:3:2 pulldown format is, then getting it to 24p as soon as possible in the chain is beneficial to you. [00:48] <iive> i'd like to see the sample before giving advice. [00:48] <praedo> pal dv is 25 [00:48] <mystica555_> why are you even putting pal into it? [00:48] <mystica555_> youve already stated your output format is MP4 in ipod and data-dvd presentation formats [00:48] <praedo> no exporting option i premiere produces a suitable 24 frames file [00:49] <mystica555_> ^^^ this is why i loathe premiere with such fervor. it can't do what i consider a *basic* operation. [00:49] <mystica555_> "plz don't screw up the file kthx" [00:50] <cbsrobot> praedo: why not record in 29i ? [00:50] <praedo> http://www.1fichier.com/remove/rnn9ja/MmMWK [00:50] <cbsrobot> nooooo [00:50] <cbsrobot> not the remve link [00:50] <cbsrobot> the other one [00:50] <praedo> no, not this link [00:50] <praedo> okay [00:50] <praedo> http://rnn9ja.1fichier.com [00:50] <cbsrobot> I'm sure somebody clicked [00:50] <praedo> not deleted yet? [00:50] <cbsrobot> no [00:50] <cbsrobot> password ? [00:51] <praedo> notcorrupt [00:51] <praedo> hehehe [00:51] <praedo> but in fact it is, at least the tape [00:51] <mystica555_> strange; i got 12 megabit from megaupload... 1.5 from 1fichier [00:51] <mystica555_> maybe closer to 1.2 if i dont do the /10 rounding [00:52] <cbsrobot> I hate the 1 min waiting time at megaupload [00:52] <mystica555_> last one he uploaded was 8 seconds wait for me [00:52] <iive> cbsrobot: i got 4 seconds, you must be regular there :P [00:52] <mystica555_> then 1.3 MB/s download [00:52] <praedo> iive, downloading [00:52] <praedo> ? [00:53] <mystica555_> id rather wait upto 1 minute if my download is 10x faster [00:53] <iive> waiting. It got downloaded while i was navigating directories. [00:53] Action: mystica555_ stares at cbsrobot for an apparent bad suggestion in uploader [00:54] Action: cbsrobot is sorry for all the kittens killed while downloading [00:54] <praedo> video starts at second 12 [00:54] <praedo> that's the beginning of the raw capture [00:56] <mystica555_> on the up-side, that uploader service doesn't mind 10 down-them-all threads......... [00:57] Action: cbsrobot has finally time to watch the megaupload song [00:59] <iive> 4p frames 1i [00:59] <mystica555_> right, and those 2 interlaced ones get discarded according to what wiki says. [00:59] <mystica555_> or should rather [01:04] <praedo> have you all watched the sample? [01:04] <praedo> from 1fichier [01:05] <iive> yes, the interlaced frame have one field from the previous and one from the next frame. [01:05] <praedo> but from second 12 to 20 it looks fine as the rest of my captured 29i videos [01:06] <praedo> so then problem is when premiere tries to detect the interlace in the first seconds? [01:07] <praedo> iive, what did you use to analize that? [01:08] <cbsrobot> well try to cut the first few seconds and see in a quick test if premiere can export it the way you want [01:08] <iive> praedo: i told you, mplayer -vf il=d:d and my eyes :) [01:08] <mystica555_> iive: i can't even get mplayer to play anything correctly.. heh. i just see the repeating bad test pattern image [01:08] <praedo> okay, i'll do that... what's the command with ffmpeg to cut from second 11 to the end? [01:08] <praedo> -t ...? [01:09] <iive> mystica555_: i hope you are not using any forks :P [01:09] <cbsrobot> ffmpeg -i model.avi -ss 20 -t 60 -vcodec copy -acodec copy output.avi [01:09] <praedo> mystica555_, no video in second 12 to 20? [01:09] <mystica555_> iive: no, using about a 2 or 3 week old compile though [01:09] <cbsrobot> ^ just a test file - not the whole ... [01:10] <praedo> okay, i'll do that test now [01:11] <praedo> [dvvideo @ 003F8CC0] AC EOB marker is absent pos=64 [01:11] <praedo> [dvvideo @ 003F8CC0] AC EOB marker is absent pos=66 [01:11] <praedo> [dvvideo @ 003F8CC0] AC EOB marker is absent pos=64 [01:11] <praedo> [dvvideo @ 003F8CC0] AC EOB marker is absent pos=68 [01:11] <praedo> while cutting [01:12] <cbsrobot> thats your damaged file ... [01:15] <mystica555_> its fun; i just got a brand new pull of git ffmpeg / svn mplayer; need to force -demuxer lavf to have it pass 1.2 seconds. then it gets to 19.7s and quits, still showing the same crap from the beginning here. [01:15] <praedo> so you can confirm that the imposibility to correctly deinterlace is because of these first 10 seconds of noise? [01:15] <Howie69> OK after 2 hours of looking and such, I guess I will follow the docs and ask here :) [01:15] <Howie69> It says to come here to get help for compiling/installing handbrake [01:15] <Howie69> is that correct? [01:15] <mystica555_> praedo: if you can get proper video sync later in the file, perhaps start editing at that point. [01:16] <mystica555_> it may be confusing the way premiere handles the file [01:16] <iive> btw, it is not 4p1i, it seems to be 3:2 pulldown. [01:16] <mystica555_> i don't know; but regardless, maybe try and make a new 24p profile yourself.. [01:16] <mystica555_> it shouldnt be 25fps pal [01:16] <iive> Howie69: if it is ffmpeg question then yes. handbrake... no idea... [01:17] <iive> there seems to be #handbrake channel too :) [01:17] <Howie69> iive: Yeah, I found it odd for one project to refer to another project's IRC... [01:18] <iive> Howie69: well, handbrake uses x264 encoder, so it may also use ffmpeg as engine core, while HB provide only the GUI. [01:19] <Howie69> iive: Heh, in one place on their webpage it says to come here for help, in another it says use #handbrake. Oh well [01:19] <Howie69> Thanks for the help [01:19] <iive> i guess, it depends what is the problem :) [01:22] <praedo> mystica555_, how would you export if source is 24p and it's to be viewed in web? [01:22] <mystica555_> as 24p [01:22] <praedo> not 25 [01:22] <mystica555_> the web doesnt care about PAL vs FILM vs NTSC framerates [01:23] <mystica555_> it cares somewhat about frame sizes, so 720x756 would still work, but 640x480 may be better for some devices. [01:23] <sacarasc> 27.4fps or death! [01:23] <praedo> 27.4? [01:23] <praedo> never heard that [01:23] <mystica555_> hes being sarcastic ;) [01:23] Action: iive gives death to sacarasc [01:24] <praedo> iive, 3:2 or 2:3:3:2? [01:24] <praedo> if it's 24p it would be the 2nd [01:24] <praedo> btw, how do you analize it? [01:24] <iive> praedo: told ya... watching it. [01:24] <iive> and yes, it could be 2:3:3:2 [01:25] <praedo> i see... mplayer -vf il=d:d [01:25] <cbsrobot> it is 2:3:3:2 [01:25] <praedo> so premiere handles correctly the pulldown? [01:25] <praedo> it detected it as 24pa [01:27] <cbsrobot> praedo: just know - you do some really nasty video porn ! [01:27] <iive> mplayer -vf pullup seems to work nicely with it. [01:27] <praedo> i imported the cut file into premiere and when i try to export, it gives error with a yellow warning triangle [01:27] <cbsrobot> recording with a 29i to a 24pa and exporting to 25p ... [01:27] <iive> but for some reason ffplayer -vf mp=pullup doesn't do the same. [01:27] <iive> ffplay [01:28] <praedo> i did ffmpeg -i model.avi -ss 20 -t 60 -vcodec copy -acodec copy output.avi [01:28] <praedo> it doesn't even start encoding [01:30] <praedo> cbsrobot: source is 24pa, import is done to 29i (unwanted) and export should be 24p again but no option in premiere produces a correct 24p file [01:30] <cbsrobot> praedo: just saying ... [01:30] <praedo> strange i can't encode that cut sample... what should i do next? [01:30] <iive> praedo: there probably is a way to do it, it is just well hidden. [01:31] <cbsrobot> this works for me: ffmpeg -i model.avi -ss 18 -vcodec copy -acodec copy output.avi [01:31] <praedo> yes, the output.avi is fine [01:32] <cbsrobot> ah you mean premiere ... [01:32] <praedo> but once imported and moved to the time line, the export to PAL DV fails before it even starts [01:32] <praedo> i tried quicktime and fails too [01:33] <praedo> in fact the output.avi can't be played in premiere [01:33] <praedo> but plays in windows media fine [01:34] <iive> let's try to keep it simple [01:34] <iive> can you capture again, but not from the start, but after the broken portion? [01:35] <praedo> yes, that will be the easier thing [01:35] <praedo> i'll capture 15min. or so [01:35] <iive> and don't delete that 20gb file... just in case [01:35] <praedo> yes [01:35] <praedo> i'll let you know something [01:36] <praedo> btw, why that broken portion happened if the tape was new? [01:36] <praedo> bad media? [01:37] <praedo> only 10 seconds bad in 60min. of good tape [01:37] <mystica555_> bad quality control perhaps [01:37] <mystica555_> *shrug* [01:37] <praedo> it's panasonic [01:37] <mystica555_> the proverbial 0.01% defect [01:38] <cbsrobot> maybe some cosmic beams - who knows [01:42] <praedo> good news [01:42] <praedo> no need to recapture those 15min. [01:42] <praedo> the ffmpeg cut was encoded successfully and no artifacts at all this time in the PAL DV [01:42] <praedo> it failed to encode for some reason and it worked after closing and starting premiere again [01:43] <praedo> now i will just need to cut from second 11 to the end [01:43] <praedo> how can i specify that in -t ? [01:44] <praedo> for the whole file [01:44] <praedo> instead of ffmpeg -i model.avi -ss 20 -t 60 -vcodec copy -acodec copy output.avi [01:46] <mystica555_> if youre skipping 20 seconds (-ss 20) the -t 60 is only a limit; removing it entirely should give you the whole thing [01:47] <mystica555_> thats saying 'skip 20 seconds, output 60 then stop' [01:47] <praedo> very good [01:47] <praedo> i'm going for it now... let's see [01:48] <praedo> thanks a lot everyone [03:33] <Howie69> xes [04:16] <bcardarella> I just encoded a MOV file and the resulting file was flipped 180 degrees... any reason for this? [04:20] <bcardarella> it seems to be happening when encoding from MOV to MPEG [04:20] <bcardarella> I don't think I'm doing anything crazy, here is the command with options: [04:20] <bcardarella> ffmpeg -i input/IMG_0194.MOV -qscale 5 -t 10 -threads 14 -y -loglevel quiet output/IMG_0194_temp_2.mpeg [05:20] <relaxed> bcardarella: look at the transpose filter in the man page [11:44] <RoyK> can ffmpeg add subtitles to a video steam? [13:29] <burek> RoyK, yes it can [15:13] <Ginks> question, and maybe this is better suited for a x264 channel. Does ffmpeg default to slow-firstpass now? [15:13] <Ginks> because when I invoke it with --slow-firstpass it says the option isn't valid [15:13] <Ginks> running from the latest git [15:16] <khali> hi there [15:16] <khali> is it possible to pass several -fflags options together? I can't find the right syntax [15:24] <khali> ah, apparently they should be prefixed with + [15:24] <khali> didn't help though... I'm still stuck with: "pts < dts in stream 0" [15:24] <khali> no idea how do go past this [15:25] <iive> looks like there is bug in there. [15:25] <khali> I'm able to play the input video file so I know it's not all borked [16:53] <gdp16> hello [16:54] <gdp16> I have question. [16:54] <gdp16> I want to enable yasm in cygwin build [16:54] <gdp16> do anyone have sugestion to compile ffmpeg with yasm in cygwin environemnt? [17:08] <JEEBcz> gdp16, build yasm and use it as usual? I didn't know cygwin was SuperSpecial in any way that building with yasm would be impossible or whatever [17:12] <gdp16> JEEBcz, I just realize I got woring configuration in my cygwin [18:03] <Yoda----> hello, I'm using libav* and I often have error messages that breaks the decoding about missing headers [18:03] <Yoda----> does it ring a bell to anyone and is there a way to avoid this ? [18:07] <caytchen> it probably means your files are corrupt [18:07] <Yoda----> yes they are, reending it fix the problem [18:08] <Yoda----> bu I wonder how to bypass this because I don't think any other program using ffmpeg has the same problem [18:08] <Yoda----> seeking in the midle of the file is ok AFAIR (using ffmpeg seek) [18:20] <burek> khali, use something like this: -flags +ilme+ildct [18:54] <khali> burek: thanks, I figured it out meanwhile, but the flags did not fix my issue anyway [22:24] <ffmpeguser> hi all, is it possible to remux a video stream with -vcodec copy, changing only the aspect ratio of it? [22:28] CTCP VERSION: from ffmpeguser (ffmpeguser!5e44f028(a)gateway/web/freenode/ip.94.68.240.40) to #ffmpeg [22:31] <reflexrg> ? [22:32] <ffmpeguser> do you know reflexrg? [22:33] <cbsrobot> ffmpeguser: Idont think so - but I'm not sure - try it out [22:33] <reflexrg> (03:28:13 PM) ffmpeguser: Received CTCP 'VERSION' (to #ffmpeg) from ffmpeguser [22:34] <ffmpeguser> reflexrg: don't worry i won't steal your files [22:35] <reflexrg> lol [22:35] <ffmpeguser> :P [22:35] <reflexrg> I am on secure system [22:35] <reflexrg> not winblows [22:35] <ffmpeguser> there is always a weak link in every security chain, the human that is part of it [22:37] <kode54> channel version query, classy [22:38] <caytchen> theres always one moron in a channel that will exploit the missing +C mode and go on to annoy us all. [22:41] <ffmpeguser> cbsrobot: i'll try it in a few minutes, i'm still doing other things| kode54 always classy | caytchen i'm not a moron, but you might be one [22:49] <maujhsn> ePirat What version of ffmpeg are you using? [22:56] <ePirat> maujhsn, i don't know o_o why are you asking? [22:56] <ePirat> I'll have to look it up [22:56] <ePirat> on my mac: ffmpeg version 0.7.6 [22:57] <ePirat> and on my ubuntu ffmpeg version git-2011-10-31-c0dbab9 [22:58] <ffmpeguser> cbsrobot: i've been looking into it, and a few values about the video stream change even with -codec copy [22:58] <ffmpeguser> of the video* [22:58] <ffmpeguser> -vcodec* [22:58] <cbsrobot> f.ex. ? [22:59] <ffmpeguser> i'll copy-paste [22:59] <ffmpeguser> Stream #0:0(eng): Video: h264 (Main) (avc1 / 0x31637661), yuv420p, 720x390, 1564 kb/s, 25 fps, 25 tbr, 2500 tbn, 5k tbc | original [23:00] <ffmpeguser> Stream #0:0: Video: h264 (Main), yuv420p, 720x390, 1564 kb/s, 25 tbr, 1k tbn , 2k tbc | copied stream [23:01] <ffmpeguser> the tbn and the tbc values [23:02] <cbsrobot> 2500 tbn, 5k tbc are way off [23:03] <cbsrobot> I think they're limited by some sane values [23:03] <ffmpeguser> how does that affect the stream in the output file? [23:03] <cbsrobot> they do not affect the stream in any way [23:05] <ffmpeguser> you mean that the values of the stream in the original file are incorrect and/or abnormal? [23:07] <cbsrobot> tbn is the timebase from the container - and as I assume you coy the stream from one container into the other, the timebase is still off limit - but not insanely off limit [23:07] <ffmpeguser> flv to flv [23:07] <cbsrobot> tbc is the time base for the codec and it's even worse [23:08] <cbsrobot> tbr is guessed from the video stream and is normally the value you need [23:08] <cbsrobot> who made the first flv ? [23:08] <ffmpeguser> also the seeking is very slow at first in the output file [23:09] <cbsrobot> errr - what tool made the flv [23:09] <ffmpeguser> i only know that the file was created by someone in new zealand :P, i grabbed it with rtmpdump [23:10] <cbsrobot> ok [23:11] <cbsrobot> anyway these values do not affect the quality of the content [23:12] <cbsrobot> and try the setdar filter while copying the stream, maybe it works [23:12] <ffmpeguser> good. let me check the documentation [23:13] <cbsrobot> what aspect ratio you need ? [23:14] <ffmpeguser> give a few seconds, i'll find the right one again [23:16] <ffmpeguser> it's 2.something:1 [23:16] <ffmpeguser> give me* [23:16] <cbsrobot> lol [23:17] <ffmpeguser> 2:21:1 :P [23:17] <ffmpeguser> 2.21:1* [23:18] <cbsrobot> hmm 221:100 ? [23:18] <cbsrobot> I never tried that but here you go [23:18] <cbsrobot> ffmpeg -i dump.flv -vf setdar=221:100 -vcodec copy -acodec copy output.flv [23:18] <ffmpeguser> i've never used that parameter, i'll try it right now [23:22] <ffmpeguser> it still has the original aspect ratio [23:24] <cbsrobot> well .... [23:29] <ffmpeguser> no matter what i try, it doesn't change it [23:32] <ffmpeguser> even when the video stream is re-encoded [23:47] <cbsrobot> ffmpeguser: try setsar [23:48] <ffmpeguser> ok, i'll look into it [23:52] <ffmpeguser> no change [23:55] <ffmpeguser> even with both parameters :P [00:00] --- Mon Dec 26 2011

1 0

ffmpeg-devel.log.20111224
by burek 25 Dec '11

25 Dec '11

[00:07] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r320b8d6335 10ffmpeg/libswresample/ (swresample.c swresample.h): [00:07] <CIA-49> ffmpeg: swr: fix swresample_version() [00:07] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [00:07] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * rad9e0ed170 10ffmpeg/libavformat/avio.c: [00:07] <CIA-49> ffmpeg: avio: fix handling of , in urls [00:07] <CIA-49> ffmpeg: Fixes Ticket805 [00:07] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [01:00] <CIA-49> ffmpeg: 03Sergey Radionov 07master * r01eb9805f3 10ffmpeg/libavcodec/w32pthreads.h: [01:00] <CIA-49> ffmpeg: w32thread: call ResetEvent() in pthread_cond_broadcast(). [01:00] <CIA-49> ffmpeg: Also add "volatile" to broadcast flag (since it is used from [01:00] <CIA-49> ffmpeg: multiple threads). [01:00] <CIA-49> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje(a)gmail.com> [01:00] <CIA-49> ffmpeg: 03Anton Khirnov 07master * r82d05e78a6 10ffmpeg/libavfilter/allfilters.c: allfilters: fix type of avfilter_vsrc_buffer. [01:00] <CIA-49> ffmpeg: 03Anton Khirnov 07master * rd09298f0d6 10ffmpeg/libavcodec/4xm.c: 4xm: remove unused variables. [01:00] <CIA-49> ffmpeg: 03Alexander Strange 07master * rbc6a3bd4a5 10ffmpeg/libavcodec/h264.c: [01:00] <CIA-49> ffmpeg: h264: Fix a possible overread in decode_nal_units() [01:00] <CIA-49> ffmpeg: Signed-off-by: Anton Khirnov <anton(a)khirnov.net> [01:00] <CIA-49> ffmpeg: 03Janne Grunau 07master * r846dca1aa3 10ffmpeg/libavcodec/pthread.c: [01:00] <CIA-49> ffmpeg: pthread: include sys/types.h before sys/sysctl.h [01:00] <CIA-49> ffmpeg: Fixes compilation on FreeBSD with clang 3. [01:00] <CIA-49> ffmpeg: 03Reinhard Tartler 07master * rbe00d2e174 10ffmpeg/ (Changelog RELEASE doc/RELEASE_NOTES): Prepare for 0.8_beta1 snapshot release [01:00] <CIA-49> ffmpeg: 03Reinhard Tartler 07master * r8b94df0f20 10ffmpeg/libavcodec/vp3.c: [01:00] <CIA-49> ffmpeg: vp3dec: Check coefficient index in vp3_dequant() [01:00] <CIA-49> ffmpeg: Based on a patch by Michael Niedermayer <michaelni(a)gmx.at> [01:00] <CIA-49> ffmpeg: Fixes NGS00145, CVE-2011-4352 [01:00] <CIA-49> ffmpeg: Found-by: Phillip Langlois [01:00] <CIA-49> ffmpeg: Signed-off-by: Reinhard Tartler <siretart(a)tauware.de> [01:00] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r6e24b9488e 10ffmpeg/libavcodec/svq1dec.c: [01:00] <CIA-49> ffmpeg: svq1dec: call avcodec_set_dimensions() after dimensions changed. [01:01] <CIA-49> ffmpeg: Fixes NGS00148, CVE-2011-4579 [01:01] <CIA-49> ffmpeg: Found-by: Phillip Langlois [01:01] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [01:01] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * rbba6f5b77f 10ffmpeg/: (log message trimmed) [01:01] <CIA-49> ffmpeg: Merge remote-tracking branch 'qatar/master' [01:01] <CIA-49> ffmpeg: * qatar/master: [01:01] <CIA-49> ffmpeg: vp3dec: Check coefficient index in vp3_dequant() [01:01] <CIA-49> ffmpeg: svq1dec: call avcodec_set_dimensions() after dimensions changed. [01:01] <CIA-49> ffmpeg: Prepare for 0.8_beta1 snapshot release [01:01] <CIA-49> ffmpeg: threads: check defines before using them in automatic thread detection [01:01] <CIA-49> ffmpeg: 03Janne Grunau 07master * r7052618c7e 10ffmpeg/libavcodec/pthread.c: threads: check defines before using them in automatic thread detection [01:43] <CIA-49> ffmpeg: 03Martin Storsjö 07release/0.8 * r38a511e84c 10ffmpeg/libswscale/utils.c: (log message trimmed) [01:43] <CIA-49> ffmpeg: swscale: Readd #define _SVID_SOURCE [01:43] <CIA-49> ffmpeg: This was removed erroneously in [01:43] <CIA-49> ffmpeg: 046f081b46c8479820409cf8f530b988221bd15b. This define still is [01:43] <CIA-49> ffmpeg: necessary for getting MAP_ANONYMOUS defined on linux/glibc, [01:43] <CIA-49> ffmpeg: despite the define reshuffling done in that commit. [01:43] <CIA-49> ffmpeg: Without MAP_ANONYMOUS defined, the mprotect calls for setting the [01:44] <CIA-49> ffmpeg: 03Michael Niedermayer 07release/0.8 * r0eca0da06e 10ffmpeg/libavcodec/svq1dec.c: [01:44] <CIA-49> ffmpeg: svq1dec: call avcodec_set_dimensions() after dimensions changed. [01:44] <CIA-49> ffmpeg: Fixes NGS00148, CVE-2011-4579 [01:44] <CIA-49> ffmpeg: Found-by: Phillip Langlois [01:44] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [01:44] <CIA-49> ffmpeg: (cherry picked from commit 6e24b9488e67849a28e64a8056e05f83cf439229) [01:44] <CIA-49> ffmpeg: Signed-off-by: Reinhard Tartler <siretart(a)tauware.de> [01:44] <CIA-49> ffmpeg: 03Reinhard Tartler 07release/0.8 * r851098c9e0 10ffmpeg/libswscale/utils.c: (log message trimmed) [01:44] <CIA-49> ffmpeg: swscale: #include "libavutil/mathematics.h" [01:44] <CIA-49> ffmpeg: this file uses the M_PI macro since [01:44] <CIA-49> ffmpeg: 4e74187db2f5db52f88729efc662df9d6bc763e1, so include the correct header [01:44] <CIA-49> ffmpeg: directly. [01:44] <CIA-49> ffmpeg: Signed-off-by: Reinhard Tartler <siretart(a)tauware.de> [01:44] <CIA-49> ffmpeg: (cherry picked from commit 5089ce1b5abe2ecbbfd7235aeb0ad47ba38305c1) [01:44] <CIA-49> ffmpeg: 03Reinhard Tartler 07release/0.8 * r7ce728050b 10ffmpeg/RELEASE: Update RELEASE file for 0.7.3 [01:44] <CIA-49> ffmpeg: 03Shitiz Garg 07release/0.8 * rd912a30c7d 10ffmpeg/libavcodec/4xm.c: [01:53] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * rd7e11d3f74 10ffmpeg/doc/RELEASE_NOTES: [01:53] <CIA-49> ffmpeg: RELEASE_NOTES: remove possibly ofensive sounding wording [01:53] <CIA-49> ffmpeg: from after the merge that i missed and do other general cleanup [01:53] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [01:53] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r0d30c95394 10ffmpeg/doc/RELEASE_NOTES: [01:53] <CIA-49> ffmpeg: release_notes: merge back changes from our 0.9 to master [01:53] <CIA-49> ffmpeg: seems i have forgotten this [01:53] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [02:03] <CIA-49> ffmpeg: 03Mashiat Sarker Shakkhar 07master * r153a853718 10ffmpeg/libavcodec/wmalosslessdec.c: Use int for channel_coeffs instead of int16_t [02:03] <CIA-49> ffmpeg: 03Mashiat Sarker Shakkhar 07master * r4fa3f1c507 10ffmpeg/libavcodec/wmalosslessdec.c: Do not update buffers in case no speed change is necessary [02:03] <CIA-49> ffmpeg: 03Mashiat Sarker Shakkhar 07master * r10406b95ad 10ffmpeg/libavcodec/wmalosslessdec.c: Enable inverse-MCLMS filter [02:03] <CIA-49> ffmpeg: 03Mashiat Sarker Shakkhar 07master * r7409210647 10ffmpeg/libavcodec/wmalosslessdec.c: Implement revert_inter_ch_decorr() and revert_acfilter() [02:03] <CIA-49> ffmpeg: 03Mashiat Sarker Shakkhar 07master * rcf6284b273 10ffmpeg/libavcodec/wmalosslessdec.c: Perform inverse inter-channel decorrelation and ac-filter [02:03] <CIA-49> ffmpeg: 03Mashiat Sarker Shakkhar 07master * r70dd5a603c 10ffmpeg/libavcodec/wmalosslessdec.c: Perform dequantization of channel coefficients [02:03] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r28dff06176 10ffmpeg/: (log message trimmed) [02:03] <CIA-49> ffmpeg: Merge remote-tracking branch 'shariman/wmall' [02:03] <CIA-49> ffmpeg: * shariman/wmall: [02:03] <CIA-49> ffmpeg: Perform dequantization of channel coefficients [02:03] <CIA-49> ffmpeg: Perform inverse inter-channel decorrelation and ac-filter [02:03] <CIA-49> ffmpeg: Implement revert_inter_ch_decorr() and revert_acfilter() [02:03] <CIA-49> ffmpeg: Enable inverse-MCLMS filter [02:03] <CIA-49> ffmpeg: 03Mashiat Sarker Shakkhar 07master * r302d5df834 10ffmpeg/libavcodec/wmalosslessdec.c: Fix inverse-MCLMS filtering routines [02:42] <CIA-49> ffmpeg: 03Jean First 07master * r0d4a77472a 10ffmpeg/libavcodec/libopenjpegdec.c: [02:42] <CIA-49> ffmpeg: libopenjpeg: support decoding with bits per pixel greater than 8 [02:42] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [02:42] <CIA-49> ffmpeg: Approved-by: Alex Zhukov [02:49] <GordonFreeman> yes, v is double [02:52] <GordonFreeman> pengvado<< thanks lrintf() looks superior to round() for that purpose [03:05] <CIA-49> ffmpeg: 03Alexander Strasser 07master * rbd9d4a6bb9 10ffmpeg/Makefile: [03:05] <CIA-49> ffmpeg: build: Try to find fate_config.sh.template in source path [03:05] <CIA-49> ffmpeg: The FATE documentation depends on the mentioned file. But that [03:05] <CIA-49> ffmpeg: did break out of tree builds because the file was not found. [03:05] <CIA-49> ffmpeg: Signed-off-by: Alexander Strasser <eclipse7(a)gmx.net> [03:05] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [03:05] <CIA-49> ffmpeg: 03Alexander Strasser 07master * rc7062802d9 10ffmpeg/doc/Makefile: [03:05] <CIA-49> ffmpeg: doc: build: TEXIDEP: Recognize verbatim includes [03:05] <CIA-49> ffmpeg: When generating the .dep files for .texi sources, verbatim includes [03:05] <CIA-49> ffmpeg: (@verbatiminclude) should also be taken into account. [03:05] <CIA-49> ffmpeg: Signed-off-by: Alexander Strasser <eclipse7(a)gmx.net> [03:05] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [03:43] <CIA-49> ffmpeg: 03Carl Eugen Hoyos 07master * r64cafe340b 10ffmpeg/libavformat/avidec.c: Make the avi palette opaque. [03:55] <Daemon404> neat... ffmpeg with llvm LTO works [06:21] Action: relaxed likes the logo :) [12:55] <CIA-49> ffmpeg: 03Carl Eugen Hoyos 07master * rf2ed115899 10ffmpeg/libavcodec/xan.c: Make the palette in Wing Commander III mve files opaque. [12:55] <CIA-49> ffmpeg: 03Carl Eugen Hoyos 07master * r0d59ae32c2 10ffmpeg/libavformat/mov.c: Make mov palette opaque or load alpha if present. [12:55] <CIA-49> ffmpeg: 03Carl Eugen Hoyos 07master * r32c49389bc 10ffmpeg/libavformat/idcin.c: Make the palette in Quake II CIN files opaque and increase its dynamics. [16:27] <CIA-49> ffmpeg: 03Stefano Sabatini 07master * r1858a5c25e 10ffmpeg/libavfilter/vf_drawtext.c: [16:27] <CIA-49> ffmpeg: vf_drawtext: select YUV color for drawbox() in case YUV colorspace is used [16:27] <CIA-49> ffmpeg: Fix box alpha rendering when a YUV colorspace is selected, in particular [16:27] <CIA-49> ffmpeg: fix trac ticket #763. [18:53] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r58c41799ab 10ffmpeg/libavutil/lzo.c: [18:53] <CIA-49> ffmpeg: lzo: fix memcpy_backptr() with 0 offset [18:53] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [18:53] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r282bb02839 10ffmpeg/libavcodec/j2kdec.c: [18:53] <CIA-49> ffmpeg: j2kdec: Fix crash in get_qcx [18:53] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [18:53] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r628c9dcca3 10ffmpeg/libavcodec/j2kdec.c: [18:53] <CIA-49> ffmpeg: j2kdec: fix division by zero, check tile dimensions for validity [18:53] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [18:53] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r3eedf9f716 10ffmpeg/libavcodec/j2kdec.c: [18:53] <CIA-49> ffmpeg: j2kdec: Check curtileno for validity [18:53] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [18:53] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r3132999fdb 10ffmpeg/libavcodec/j2kdec.c: [18:53] <CIA-49> ffmpeg: j2kdec: Check for interger overflow in tile array allocation [18:53] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [00:00] --- Sun Dec 25 2011

1 0

ffmpeg.log.20111224
by burek 25 Dec '11

25 Dec '11

[00:03] <u0m3> caytchen: the lib... I mentioned the ffmpeg command line tool because I tryed to check out how it handles resampling when passed the argument -ar <val> within it's code. But as I said, the argument handling is well hidden [00:07] <caytchen> okay, then my hints still apply. the process of resampling (with ReSampleContext or swresample) is pretty straightforward and simple though. swresample is a bit more convoluted since its designed for multiple audio channels in multiple buffers, not the packed audio that I think every ffmpeg audio codec outputs. [00:09] <caytchen> theres one function that allocates the context and initializes it with your specifications and then one convert function that you use to move data in and get freshly resampled out. [00:10] <u0m3> caytchen: actually I need to check for multiple audio channels... I need to get sound data into a buffer at 48kHz / stereo / signed 16bit / pcm or wave ... so I guess I'll have to use swresample... [00:16] <caytchen> well yes, you have multiple channels, but if you're using ffmpeg to get the raw audio data, chances are it will be in one big buffer despite having multiple channels [00:16] <u0m3> caytchen: by big buffer you mean the whole file in a buffer at once? [00:21] <caytchen> say I have a stereo MP3 file. I have ffmpeg decode a frame into an AVFrame. AVFrame actually has seperate data buffers for seperate channels, so you would expect the two stereo channels to be decoded into their own buffers. [00:21] <caytchen> but what actually happens is that all the data ends up in the first buffer, samples for left and right are packed together. [00:23] <u0m3> caytchen: better that way... I need to send the buffer to another lib and it expects all the channels in a single buffer. [00:24] <caytchen> its kind of a moot point anyway, all I wanted to say is that swresample has facilities for audio data being split by channels, but you can still just give it the packed data, too. [00:25] <u0m3> ah, got it. [03:02] <shevy> [aac @ 0x893b7e0]Reserved SBR extensions are not implemented. Update your FFmpeg version to the newest one from Git. If the problem still occurs, it means that your file has a feature which has not been implemented. [03:02] <shevy> [aac @ 0x893b7e0]If you want to help, upload a sample of this file to ftp://upload.ffmpeg.org/MPlayer/incoming/ and contact the ffmpeg-devel mailing list. [03:02] <shevy> hmm [03:03] <shevy> is this bad? the file in question seems to play just fine, audio is also fine ... other than those warning messages I'd say everything is perfect [04:11] <aixenv> hey i just noticed that my faac doesn't work anymore, what is the new audio standard to use? [04:11] <aixenv> my encoding log shows 'usr/share/ffmpeg/libx264-slow.ffpreset: Invalid option or argument: 'rc_lookahead=50', parsed as 'rc_lookahead' = '50' [04:12] <aixenv> the params i was using were: ffmpeg -y -i "$file" -vcodec libx264 -r 30000/1001 -deinterlace -s 1024x576 -crf 21 -maxrate 2M -bufsize 6M -vpre slow -threads 0 -acodec libfaac -ar 48000 -ab 128k ${file%.*}-1.mp4 >> "${file%}"-encoding.out 2>&1 [04:15] <treund> aixenv: what does ffmpeg -codecs | grep faac show [04:17] <aixenv> that option doesnt work [04:17] <aixenv> the "-codecs" [04:17] <treund> ? i get as a last line after the configure info 128: EA libfaac libfaac AAC (Advanced Audio Codec) [04:18] <aixenv> ffmpeg: missing argument for option '-codec' [04:18] <caytchen> -codec_s_ [04:20] <aixenv> well -formats works [04:20] <aixenv> D aac raw ADTS AAC [04:20] <aixenv> D A aac Advanced Audio Coding [05:29] <aixenv> any ideas? [05:39] <relaxed> aixenv: about what? [05:43] <aixenv> my encoding log shows 'usr/share/ffmpeg/libx264-slow.ffpreset: Invalid option or argument: 'rc_lookahead=50', parsed as 'rc_lookahead' = '50' [05:43] <aixenv> the params i was using were: ffmpeg -y -i "$file" -vcodec libx264 -r 30000/1001 -deinterlace -s 1024x576 -crf 21 -maxrate 2M -bufsize 6M -vpre slow -threads 0 -acodec libfaac -ar 48000 -ab 128k ${file%.*}-1.mp4 >> "${file%}"-encoding.out 2>&1 [05:54] <relaxed> aixenv: replace '-vpre slow' with '-preset slow' [05:54] <aixenv> testing [05:57] <relaxed> and if possible, replace '-deinterlace -s 1024x576' with '-vf "yadif,scale=1024:576"' [05:57] <aixenv> ok i can do whatever you recommend, whats the reason on that one? [05:58] <relaxed> yadif is better at deinterlacing [05:59] <aixenv> ty - testing [06:01] <aixenv> i get unrecognized -vf [06:02] <aixenv> is my ffmpeg too old? i have an svn -r0.5.5-1 version [06:02] <relaxed> yes [06:02] <aixenv> ok let me update my svn [06:02] <relaxed> uh, no [06:02] <aixenv> is it ok to still use libfaac or should i be using something else now? [06:02] <relaxed> ffmpeg now uses git [06:02] <aixenv> ok [06:03] <aixenv> git://source.ffmpeg.org/ffmpeg.git <== right? [06:03] <relaxed> correct [06:03] <aixenv> ok hang tight let me grab the source [06:03] <aixenv> my config options are [06:03] <relaxed> libfaac is fine but you'll need to pass --enable-nonfree to configure [06:04] <aixenv> --prefix=/usr --enable-avfilter --enable-avfilter-lavf --enable-vdpau --enable-bzlib --enable-libdirac --enable-libgsm --enable-libopenjpeg --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-libfaad --enable-libdc1394 --enable-shared --disa [06:04] <aixenv> ble-static [06:04] <aixenv> any issues with that nowadays? [06:05] <relaxed> yes, one second [06:05] <aixenv> ok thank you for the help, much appreciated [06:08] <relaxed> --prefix=/usr --enable-vdpau --enable-libdirac --enable-libgsm --enable-libx264 --enable-libopenjpeg --enable-libschroedinger --enable-libspeex --enable-libtheora --enable-libvorbis --disable-stripping --enable-x11grab --enable-libfaac --enable-nonfree --enable-libdc1394 --enable-shared --disable-static [06:08] <relaxed> Do you need all of that? [06:10] <aixenv> yes i believe so [06:10] <aixenv> possibly now libspeez [06:10] <aixenv> *speex [06:10] <aixenv> sorry *not [06:11] <relaxed> you may need it for decoding [06:12] <aixenv> ok, so that line you gave me should suit my initial command right? [06:12] <relaxed> yes, if you're buuilding a package for use on other machines add --enable-runtime-cpudetect [06:12] <aixenv> ok [06:13] <aixenv> needed -enable-gp too [06:13] <aixenv> --enable-gpl [06:15] <aixenv> ERROR: libdc1394-2 not found ; finding and making [06:16] <relaxed> I doubt you need it. [06:16] <relaxed> Do you use ffmpeg to read DV over firewire? [06:17] <aixenv> no -ill omit [06:19] <aixenv> probably dont need libdirac either, never use that [06:26] <aixenv> what about openjpeg that needed for basic use? [06:37] <relaxed> no [06:55] <aixenv> i ended up using "--prefix=/usr --enable-vdpau --enable-libx264 --enable-libtheora --enable-libvorbis --disable-stripping --enable-x11grab --enable-libfaac --enable-nonfree --enable-shared --disable-static --enable-runtime-cpudetect --enable-gpl --enable-libmp3lame --enable-libspeex --enable-xvid --enable-pthreads --enable-libopencore-amrnb --enable-libopencore-amrwb --enable-postproc --enable-libvpx --enable-version3" [07:12] <aixenv> there we go - ffmpeg version N-36147-g64cafe3, Copyright (c) 2000-2011 the FFmpeg developers [07:12] <aixenv> now to test [07:35] <aixenv> WOOT fixed [08:24] <ablegreen> hi [08:24] <ablegreen> i have a question about live webcam streaming [08:26] <ablegreen> i guess no one is here [08:30] <ablegreen> styles, you there? [08:30] <ablegreen> anyoneeeee here? [17:45] <h_senior_paris> good evening , does anybody already succeeded on encoding a webcam to H264 format and flv container ? [19:31] <praedo> hello [19:31] <praedo> merry xmas first of all [19:31] <praedo> i have a problem deinterlacing a video [19:31] <praedo> the DV tape was corrupt at some point and the capture must be missing some bit [19:32] <praedo> the image is good but deinterlacing presents a weird blinking when the camera is moved [19:32] <praedo> still images are viewed fine [19:32] <praedo> i always did it the same way and deinterlacing worked [19:32] <praedo> now the difference is the corrupt tape [19:33] <praedo> i export with adobe premiere as i always do, using PAL DV [19:33] <praedo> the original DV file before being edited doens't present this blinking problem [19:33] <praedo> only after edited and exported [19:34] <praedo> i tried to encode into mp4 without using deinterlace and the image is interlaced but without the blinking problem [19:34] <praedo> so the problem is at the deinterlace point [19:34] <praedo> with this description of the problem, what kind of deinterlace technique do you suggest to fix it? [19:35] <praedo> thanks in advance for any advice [19:47] <praedo> maybe using yadif? [19:56] <iive> praedo: afaik dv is key-frame only, so there shouldn't be long lasting effects. my guess is that field dominance is guessed wrong. [19:56] <iive> aka whatever the upper or bottom field is displayed first. so in effect, it should appear like every 2 frames are swapped. [19:57] <iive> or in other words, the motion would be doing 2 steps forward and one backward. [21:57] <praedo> iive, yes, that seems to be the problem [21:57] <praedo> any solution for it? [22:00] <iive> praedo: not sure. I see there is -top option, but not sure how it works. I think yadif should have field dominance parameter too. [22:00] <praedo> what field dominance it should be? [22:01] <iive> donno... depends on the source [22:38] <praedo> i tried many deinterlacing options and none of them succees [22:39] <praedo> succeeds [22:39] <praedo> could you have a look with a 30sec. sample? [22:42] <praedo> iive, can you try to encode it for me? [22:42] <praedo> just the sample to check [22:48] <iive> ok [22:55] <praedo> thanks iive, where can i send it to you? [22:55] <iive> praedo: donno, try dcc [22:55] <praedo> i have just tried mpeg streamclip setting both upper field and then lower field for dominance and result is worse than seeing it interlaced [22:56] <praedo> it must be something else [22:56] <praedo> please have a look to it [22:56] <praedo> i upload to megaupload, okay? [22:57] <iive> ok. [22:57] <praedo> thanks [22:57] <iive> it does (still) allow download without buying premiums [23:02] <praedo> uploading... 39mb for 10 seconds [23:02] <praedo> recently exported into PAL DV from the edited piece [23:03] <praedo> as i always do it [23:12] <praedo> finished! [23:12] <praedo> http://www.megaupload.com/?d=F7NPMCII [23:12] <praedo> you can download it without need to have premium [23:15] <praedo> iive, let me know if it downloads fine [23:35] <praedo> any luck encoding, iive? [23:45] <iive> oops, got carried away. [23:46] <iive> downloaded. [23:48] <cbsrobot> praedo: my guess: broken ccd [23:49] <cbsrobot> every 8 to 9th frame the interlacing is broken [23:49] <iive> praedo: is that the original footage? [23:49] <iive> it looks like it have been scaled like it is progressive, thus bleeding interlace into the opposite fields [23:53] <praedo> as i explained before, the tape is corrupt at the beginning [23:53] <iive> no [23:54] <iive> this have nothing to do with corruption [23:54] <praedo> when i play the tape on the camera itself, it shows scrambled frames for some seconds, then the recording starts playing fine [23:54] <iive> as dv is keyframe only format, it should not be affected if portion of it is missing. [23:54] <praedo> the captured file also plays fine without this problem (being interlaced) [23:55] <praedo> the problem occurs after exporting to PAL DV as usual [23:55] <praedo> i always deinterlace that PAL DV material and plays fine [23:55] <praedo> but in this case it refused to be deinterlaced by all methods i tried [23:55] <praedo> maybe i'm just missing some option that i don't know [23:56] <iive> no [23:57] <praedo> then maybe i should try to swap opposite fields? [23:57] <iive> the sample you gave me is permanently damaged. [23:57] <praedo> all the 1-hour conference looks like this [23:57] <praedo> and i need to deliver it to the customer... [23:58] <iive> repairing it is enormous mathematical challenge [23:58] <praedo> still images aren't affected [23:58] <praedo> how come? [23:58] <mystica555_> praedo: is that sample you posted to megaupload, the raw footage? [23:58] <iive> i tried to play it with mplayer using -vf il=d:d , this separates both fields and makes them as 2 whole picture one above the other. [23:59] <praedo> no, that sample is the exported PAL DV after being edited with premiere [23:59] <mystica555_> whats the input source? [23:59] <praedo> the raw is 10gb [23:59] <iive> some of the fields are looking like sine waves. [23:59] <mystica555_> is it also pal-dv? [23:59] <mystica555_> premiere deals with interlacing......badly [23:59] <praedo> the input source is dv tape from dvx100 camera [00:00] --- Sun Dec 25 2011

1 0

ffmpeg-devel.log.20111223
by burek 24 Dec '11

24 Dec '11

[02:03] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r9eb0d8bab1 10ffmpeg/libavutil/fifo.c: [02:03] <CIA-49> ffmpeg: fifo: Make writes atomic. [02:03] <CIA-49> ffmpeg: Prior to this a X bytes write could be seen as less than X bytes being [02:03] <CIA-49> ffmpeg: available if the check was done at an unfortunate moment. [02:03] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [02:03] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r3dcbafc777 10ffmpeg/libavformat/udp.c: [02:03] <CIA-49> ffmpeg: udp: return circular buffer error if such error happened. [02:03] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [02:03] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * rbc900501e0 10ffmpeg/libavformat/udp.c: (log message trimmed) [02:03] <CIA-49> ffmpeg: udp: Replace double select() by select+mutex+cond. [02:03] <CIA-49> ffmpeg: When no data was available both the buffer thread as well as [02:03] <CIA-49> ffmpeg: the main thread would block in select(), when data becomes [02:03] <CIA-49> ffmpeg: available both should move forward and as data is read in the [02:03] <CIA-49> ffmpeg: buffer thread the main thread would block in select() later [02:03] <Daemon404> hmmm [02:03] Action: Daemon404 sees an error with fate.texi or doc/Makefile but cant decide which [02:14] <Daemon404> is alexander strasser ever on irc? [02:21] <iive> i think his nick is beastd [02:21] <Daemon404> oic [02:28] <Daemon404> iive, i just use the mailing list anyway. [03:38] <CIA-49> ffmpeg: 03Anton Khirnov 07master * r131609dc2a 10ffmpeg/libswscale/swscale_internal.h: [03:38] <CIA-49> ffmpeg: sws: readd PAL8 to isPacked() [03:38] <CIA-49> ffmpeg: Fixes PAL8 to YUV conversion. [03:38] <CIA-49> ffmpeg: 03Anton Khirnov 07master * r196c9e5c8c 10ffmpeg/ (configure libavfilter/Makefile libavfilter/allfilters.c): [03:38] <CIA-49> ffmpeg: lavfi: always build vsrc_buffer. [03:38] <CIA-49> ffmpeg: It's a part of public API. [03:38] <CIA-49> ffmpeg: 03Diego Biurrun 07master * raaf47bcde7 10ffmpeg/libavcodec/ (27 files): [03:38] <CIA-49> ffmpeg: Drop ALT_ prefix from BITSTREAM_READER_LE name. [03:38] <CIA-49> ffmpeg: The prefix is a historic remnant that probably meant "alternative". [03:38] <CIA-49> ffmpeg: Now that the A32 bitstream reader has been dropped it makes no sense anymore. [03:38] <CIA-49> ffmpeg: 03Justin Ruggles 07master * rf264d336fe 10ffmpeg/libavcodec/truespeech.c: [03:38] <CIA-49> ffmpeg: truespeech: fix invalid reads in truespeech_apply_twopoint_filter() [03:38] <CIA-49> ffmpeg: fixes Bug 171 [03:38] <CIA-49> ffmpeg: 03Alex Converse 07master * r4bf3c8f226 10ffmpeg/ (5 files in 3 dirs): [03:38] <CIA-49> ffmpeg: mov: Don't stick the QuickTime field ordering atom in extradata. [03:38] <CIA-49> ffmpeg: The 'fiel' atoms can be found in H.264 tracks clobbering the extradata. [03:38] <CIA-49> ffmpeg: MJPEG supports non field based extradata, and this data should be [03:38] <CIA-49> ffmpeg: preserved when copying. [03:38] <CIA-49> ffmpeg: 03Janne Grunau 07master * rc32e4029d9 10ffmpeg/libavcodec/options.c: [03:38] <CIA-49> ffmpeg: options: set minimum for "threads" to zero [03:38] <CIA-49> ffmpeg: A negative number of threads does not make sense and 0 is used for [03:38] <CIA-49> ffmpeg: autodetection. Adds a symbolic name for autodetection. [03:38] <CIA-49> ffmpeg: 03Janne Grunau 07master * rea4d5f4837 10ffmpeg/ (configure libavcodec/pthread.c): [03:38] <CIA-49> ffmpeg: linux: use number of CPUs as automatic thread count [03:38] <CIA-49> ffmpeg: Use sched_getaffinity to determine the number of logical CPUs. [03:38] <CIA-49> ffmpeg: Limits the number of threads to 16 since slice threading of H.264 [03:38] <CIA-49> ffmpeg: seems to be buggy with more than 16 threads. [03:39] <CIA-49> ffmpeg: 03Shitiz Garg 07master * r355d917c0b 10ffmpeg/libavcodec/4xm.c: [03:39] <CIA-49> ffmpeg: 4xm: Add a check in decode_i_frame to prevent buffer overreads [03:39] <CIA-49> ffmpeg: Fixes bugzilla #135 [03:39] <CIA-49> ffmpeg: Signed-off-by: Janne Grunau <janne-libav(a)jannau.net> [03:39] <CIA-49> ffmpeg: 03Gaurav Narula 07master * rad3161ec1d 10ffmpeg/libavcodec/kmvc.c: [03:39] <CIA-49> ffmpeg: kvmc: fix invalid reads [03:39] <CIA-49> ffmpeg: Signed-off-by: Janne Grunau <janne-libav(a)jannau.net> [03:39] <CIA-49> ffmpeg: 03Paul B Mahol 07master * r01a01bf8bd 10ffmpeg/libavcodec/adpcm.c: [03:39] <CIA-49> ffmpeg: adpcm: fix IMA SMJPEG decoding [03:39] <CIA-49> ffmpeg: Signed-off-by: Janne Grunau <janne-libav(a)jannau.net> [03:39] <CIA-49> (30 lines omitted) [03:46] <pasteeater> michaelni: the "clone git://source..." is fine with me. my download page suggestion wasn't that strong of a preference of mine. [03:48] <pasteeater> also, i'm out of town for a week, so the FFmpeg living spam filter might be a little slower than normal [03:49] <michaelni> np [03:49] <michaelni> ill think about how to best word it ... [03:50] <michaelni> iam not sure myself ATM what is best [03:50] <pasteeater> the original was good enough [06:48] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * rae7a6441df 10ffmpeg/libavcodec/h264_ps.c: [06:48] <CIA-49> ffmpeg: h264: Decode short VUI without error [06:48] <CIA-49> ffmpeg: Fixes 47j9R7PXBep.mov [06:48] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [06:48] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * rf810ab45e3 10ffmpeg/libavfilter/ (avfilter.c avfilter.h): [06:48] <CIA-49> ffmpeg: libavfilter: increase LIBAVFILTER_VERSION_MICRO to 100 like the other libs. [06:48] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [11:57] <CIA-49> ffmpeg: 03Carl Eugen Hoyos 07master * re5de928923 10ffmpeg/libavcodec/aacdec.c: [11:57] <CIA-49> ffmpeg: Fix a possible endless loop when decoding aac. [11:57] <CIA-49> ffmpeg: Fixes ticket #789. [11:58] <CIA-49> ffmpeg: 03Carl Eugen Hoyos 07master * r8e88145d0b 10ffmpeg/libavcodec/svq1enc.c: [11:58] <CIA-49> ffmpeg: Release buffers after encoding svq1. [11:58] <CIA-49> ffmpeg: Fixes ticket #820. [11:58] <CIA-49> ffmpeg: 03Carl Eugen Hoyos 07master * rc25c7ba87c 10ffmpeg/libavutil/imgutils.c: Make systematic palette opaque. [16:10] <CIA-49> ffmpeg: 03Janne Grunau 07master * r8d4668cc5f 10ffmpeg/libavcodec/pthread.c: [16:10] <CIA-49> ffmpeg: pthread: include sys/types.h before sys/sysctl.h [16:10] <CIA-49> ffmpeg: Fixes compilation on FreeBSD with clang 3. [16:10] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [16:10] <CIA-49> ffmpeg: 03Janne Grunau 07master * r5c33a9a5c1 10ffmpeg/libavcodec/pthread.c: [16:10] <CIA-49> ffmpeg: threads: check defines before using them in automatic thread detection [16:10] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [16:37] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r7de733b752 10ffmpeg/libavcodec/pthread.c: [16:37] <CIA-49> ffmpeg: pthread: sysctl.h needs param.h on openbsd [16:37] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [18:18] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * rc2c284b3eb 10ffmpeg/doc/Makefile: [18:18] <CIA-49> ffmpeg: Revert "doc: build: TEXIDEP: Recognize verbatim includes" [18:18] <CIA-49> ffmpeg: This reverts commit e5cbf24e45ed21178674fff8f485d919128b793c. [18:18] <CIA-49> ffmpeg: This commits breaks out of tree builds [22:24] <GordonFreeman> hi [22:24] <GordonFreeman> round(y+v*8) or y+round(v*8) for performance? (y is int) [22:27] <iive> integer addition is considered faster, but don't forget that round still returns double [22:28] <cbsrobot> I guess you could use some shift operations [22:28] <iive> if v is integer, then i don't see point in using round. [22:29] <cbsrobot> v<<3 = v*8 [22:29] <caytchen> the compiler will already do that. [22:29] <cbsrobot> but I think the compiler will optimize it [22:30] <iive> cbsrobot: you are missing the point, you can't shift floats [22:30] <cbsrobot> is v a float ? [22:30] <iive> "if v is integer, then i don't see point in using round." [22:31] <cbsrobot> true [22:31] <Daemon404> [16:30] <@iive> cbsrobot: you are missing the point, you can't shift floats <-- on a bit-level, you could if you really wanted to [22:31] <Daemon404> it'd just break shit [22:34] <iive> Daemon404: the compiler won't accept it. you need to cast or alias it first. [22:34] <Daemon404> iive, merely meant in theory you can [22:34] <Daemon404> as a float is still a bunch of bits [22:35] <iive> if it is in memory, there are special registers for floats. aka the same used by mmx. [22:35] <Daemon404> not talking about in practice [22:38] <iive> ~~ [23:08] <CIA-49> ffmpeg: 03Michael Niedermayer 07master * r8ff4fff774 10ffmpeg/libavformat/flvdec.c: [23:08] <CIA-49> ffmpeg: flvdec: disable new midstream param change code not only for h264 but all cases. [23:08] <CIA-49> ffmpeg: It breaks some samples and iam not aware of one that it fixes. [23:08] <CIA-49> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [23:37] <pengvado> GordonFreeman: y+lrintf(v*8) [00:00] --- Sat Dec 24 2011

1 0