Ffmpeg-devel-irc

Download

ffmpeg-devel-irc@ffmpeg.org

April 2012

1 participants
60 discussions

ffmpeg-devel.log.20120429
by burek 30 Apr '12

30 Apr '12

[10:55] <Shimmy> https://ffmpeg.org/trac/ffmpeg/ticket/747 [11:42] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r5932938c9a 10ffmpeg/libswresample/x86/ (audio_convert.asm swresample_x86.c): [11:42] <CIA-17> ffmpeg: swr: float_to_int32_sse2() [11:42] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [12:33] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r73edb58c3c 10ffmpeg/libswresample/x86/ (audio_convert.asm swresample_x86.c): [12:33] <CIA-17> ffmpeg: swr: float_to_int16_sse2() [12:33] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [13:03] <CIA-17> ffmpeg: 03Reimar Döffinger 07master * r1d128e5814 10ffmpeg/libavformat/nutdec.c: [13:03] <CIA-17> ffmpeg: nutdec: replace assert with av_assert0. [13:03] <CIA-17> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger(a)gmx.de> [13:03] <CIA-17> ffmpeg: 03Reimar Döffinger 07master * r370211f470 10ffmpeg/libavformat/nutdec.c: [13:03] <CIA-17> ffmpeg: nutdec: minor simplification. [13:03] <CIA-17> ffmpeg: Also fixes an (incorrect) "control reaches end of non-void function" [13:03] <CIA-17> ffmpeg: warning with some compilers. [13:03] <CIA-17> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger(a)gmx.de> [13:03] <CIA-17> ffmpeg: 03Reimar Döffinger 07master * r755667eebc 10ffmpeg/libavformat/rmenc.c: (log message trimmed) [13:03] <CIA-17> ffmpeg: rmenc: more than 2 streams are not currently possible. [13:03] <CIA-17> ffmpeg: It is possible that just extending the RMMuxContext.streams [13:03] <CIA-17> ffmpeg: array would avoid it. [13:03] <CIA-17> ffmpeg: It is also possible that two audio streams will fail to mux [13:03] <CIA-17> ffmpeg: correctly as well, though at least it should not crash for [13:03] <CIA-17> ffmpeg: this reason. [14:26] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r65722e7fc5 10ffmpeg/libswresample/x86/ (audio_convert.asm swresample_x86.c): [14:26] <CIA-17> ffmpeg: swr: int32_to_int16_mmx/sse [14:26] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [14:28] <Shimmy> https://ffmpeg.org/trac/ffmpeg/ticket/747 [14:44] <burek> Shimmy, http://ffmpeg.org/consulting.html [15:45] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r2d6c29f566 10ffmpeg/libswresample/ (swresample.c swresample_internal.h): [15:45] <CIA-17> ffmpeg: swr: add set_audiodata_fmt() and use it to simplify code [15:45] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [15:45] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r106789dfa0 10ffmpeg/libswresample/ (swresample.c swresample_internal.h): [15:45] <CIA-17> ffmpeg: swr: add sample format to AudioData [15:45] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [15:45] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r9f98a8e9bc 10ffmpeg/libswresample/audioconvert.c: [15:45] <CIA-17> ffmpeg: swr: use memcpy when input and output match [15:45] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [17:17] <burek> one noob question.. if I want to use 1 ffmpeg git cloned on my machine to produce several different builds (one gpl, another non-free), how can I specify in the configure script to put all object/binary/make files into separate directory for each build? [17:23] <Compn> set the bin dir ? [17:23] <Compn> ./configure --help | grep -i bin [17:24] <burek> --bindir=DIR install binaries in DIR [PREFIX/bin] <--- that? [17:24] <burek> well, that installs binaries into some dir [17:24] <burek> I need all *.o files to put into /obj/ [17:27] <Compn> oh [17:27] <Compn> no clue [17:27] <Compn> yeah that [17:29] <burek> would it be enough, for example, to just cd into the obj/ directory and run ../configure ...? [17:33] <Compn> sounds like a user question :P [17:33] Action: Compn is helpful! [17:33] <burek> :beer: :) [17:33] <Compn> sounds like you want 'to build ffmpeg out of tree' [17:33] <Compn> try google that [17:36] <burek> that's it :) thanks a lot! :) [18:39] <CIA-17> ffmpeg: 03Carl Eugen Hoyos 07master * r3dafde0b96 10ffmpeg/libavcodec/ (cscd.c version.h): [18:39] <CIA-17> ffmpeg: Simplify camstudio decoding and fix odd 16bit decoding. [18:39] <CIA-17> ffmpeg: Fixes ticket #1220. [18:47] <CIA-17> ffmpeg: 03Stefano Sabatini 07master * r46eba43e0e 10ffmpeg/libavfilter/buffersink.h: lavfi/buffersink: fix header inclusion guard name [19:30] <burek> Compn, do you know maybe how can I compile libx264 in some tmp directory without make install and then tell ffmpeg to find what it needs for libx264 in that tmp directory? [19:30] <burek> something like ./configure --enable-libx264=/my/tmp/dir/x264 :) [20:40] <burek> also, is it more appropriate (if I'd like to compile 32bit ffmpeg on 64bit machine) to use "--arch=i386" or "--cc='gcc -m32'" [20:41] <Daemon404> --extra-cflags and --extra-ldflags [20:42] <burek> i see, it's better to use those 2 instead of -cc, but does --arch set up all that automatically? [20:43] <Daemon404> that was in response to the part about a local libx264 [20:43] <burek> oh, ok :) thanks :) [21:31] <burek> hm, I hit a wall now.. I've built libx264 as 32bit lib/bin and it works on 32bit machine [21:32] <burek> I've also built bare ffmpeg (just ./configure) as 32bit bin and it also works on 32bit machine [21:32] <burek> but trying now to build ffmpeg+libx264 as 32bit bin gives me this: [21:33] <burek> http://pastebin.com/kKpuGtwn [21:33] <burek> shortly, undefined reference to `x264_encoder_open_124' [21:33] <burek> :S [21:49] <michaelni> burek, you compile ffmpeg against the headers for the x264 you link it too ? aka they are matching ? [21:49] <michaelni> both version 124 ? [21:49] <burek> I figured I've put --arch=i386 which I guess is invalid.. I put --arch=x86 and doing make again [21:50] <burek> both x264 and ffmpeg are from latest git if that's what you are asking :) [21:59] <burek> woo hoo it works \o/ :))))))))) [22:01] <burek> or not :/ [22:01] <burek> ./ffmpeg: error while loading shared libraries: libx264.so.124: cannot open shared object file: No such file or directory [22:01] <burek> back to the lab.. [22:09] <Daemon404> protip: have the shared lib somewhere ffmpeg can see [22:10] <burek> I figured it's looking in usr/local/lib and /usr/lib but I've used --prefix=/custom/dir to install compiled x264 [22:10] <burek> I'm trying static build now and if it works, that will be ok too :) [22:11] <burek> it would be cool if I could just speed up make like not to build everything, just the very basic set fo formats, codecs, etc [22:13] <Daemon404> burek, you can enable/disable anything you want [22:13] <Daemon404> configure --help [22:13] <Daemon404> also, this may be mroe suited to #ffmpefg [22:13] <Daemon404> er [22:13] <Daemon404> #ffmpeg [22:13] <burek> ok, sorry [22:13] <burek> :X [22:52] <CIA-17> ffmpeg: 03Christophe Gisquet 07master * rfe5ed69dc7 10ffmpeg/libavcodec/x86/ (rv34dsp.asm rv34dsp_init.c): [22:52] <CIA-17> ffmpeg: rv34dsp x86: implement MMX2 inverse transform [22:52] <CIA-17> ffmpeg: 141 cycles down to 51. [22:52] <CIA-17> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje(a)gmail.com> [22:52] <CIA-17> ffmpeg: 03Christophe GISQUET 07master * re75d1d4f73 10ffmpeg/libavcodec/x86/dsputil_mmx.c: [22:52] <CIA-17> ffmpeg: dsputil x86: revert a test back to its previous value [22:52] <CIA-17> ffmpeg: Commit 356ee8d caused the initial inversion. [22:52] <CIA-17> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje(a)gmail.com> [22:52] <CIA-17> ffmpeg: 03Christophe Gisquet 07master * r8802799991 10ffmpeg/libavcodec/rv34.c: [22:52] <CIA-17> ffmpeg: rv34: remove constant parameter [22:52] <CIA-17> ffmpeg: is_block2 was always 0, so just remove it, and change accordingly the code. [22:52] <CIA-17> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje(a)gmail.com> [22:52] <CIA-17> ffmpeg: 03Ronald S. Bultje 07master * r774b681b80 10ffmpeg/libavcodec/rv34.c: [22:52] <CIA-17> ffmpeg: rv34: remove inline keyword from rv34_decode_block(). [22:52] <CIA-17> ffmpeg: Force-inlining increases object size by 8kB and causes a 2% slowdown. [22:52] <CIA-17> ffmpeg: 03Christophe Gisquet 07master * rd834ab0f2a 10ffmpeg/libavcodec/rv40.c: [22:52] <CIA-17> ffmpeg: rv40: perform bitwise checks in loop filter [22:52] <CIA-17> ffmpeg: Down from 95 kcycles to 93 (including all called functions). [22:53] <CIA-17> ffmpeg: Signed-off-by: Ronald S. Bultje <rsbultje(a)gmail.com> [00:00] --- Mon Apr 30 2012

1 0

ffmpeg.log.20120429
by burek 30 Apr '12

30 Apr '12

[00:11] <robertzaccour> I'm recording with ffmpeg and wanna record the audio from my webcam instead of internal mic. How do I know what to input for my webcam? [00:32] <burek> robertzaccour, type: arecord -L [00:32] <burek> and use it like: ffmpeg -f alsa -i <here> [00:32] <burek> like -i hw:0 [00:32] <burek> or -i default:front [00:33] <burek> etc [00:36] <robertzaccour> burek, If I paste the command I know to use would you be able to help me adjust it accordingly? [00:37] <robertzaccour> burek, ffmpeg -f alsa -ac 2 -i pulse -r 30 -s 1366x768 -f x11grab -i :0.0 -vcodec msmpeg4v2 -qscale 2 screencast.avi how do I adjust the input from here? [00:41] <robertzaccour> burek, I did ffmpeg -f arecord -L -i pulse -r 30 -s 1366x768 -f x11grab -i :0.0 -vcodec msmpeg4v2 -qscale 2 screencast.avi and it said this version has nonfree parts and not legally distributable but didn't record [00:43] <ZodiacMentor> http://www.rizeofthefenix.com/ [00:45] <burek> robertzaccour, if you use -i pulse, then you need to configure pulse [00:46] <burek> ZodiacMentor, what's the point of your spam? [00:46] <robertzaccour> burek, I don't know a lot about ffmpeg I just need to switch the audio recording from the internal mic to the webcam mic. would you be able to tell me the the input needed to capture from the webcam mic instead of the internal? [00:46] <burek> robertzaccour, first of all type: arecord -l [00:46] <burek> (small L) [00:47] <burek> and paste the output to pastebin.com [00:47] <ZodiacMentor> burek, no point. no more spam if that's the rule/general idea of the channel. sorry. [00:47] <burek> ZodiacMentor, do you need any help with ffmpeg or what? [00:47] <robertzaccour> burek, http://pastebin.com/j10UssnH [00:48] <ZodiacMentor> had once, the problems have been solved since then... decided to stay, just in case I need help and/or I could help [00:48] <burek> robertzaccour, try -i hw:1 [00:49] <burek> ZodiacMentor, ok :) [00:49] <robertzaccour> burek, command not found [00:50] <burek> :) [00:50] <burek> ffmpeg -f alsa -ac 2 -i hw:1 -r 30 -s 1366x768 -f x11grab -i :0.0 -vcodec msmpeg4v2 -qscale 2 screencast.avi [00:51] <burek> also, if your mic is mono, you might change -ac 2 to -ac 1 [00:51] <robertzaccour> http://pastebin.com/0dCGji1R [00:52] <robertzaccour> trying it brb to let you know thanks [00:52] <burek> "cannot set channel count to 2" [00:52] <burek> as I said, try -ac 1 [00:55] <robertzaccour> burek, the sound is clearer but now the audio is out of sync with the video [00:56] <robertzaccour> burek, perhaps there's a way to sync it or record both audio and video directly from the webcam? [00:56] <burek> robertzaccour, if you are recording both from webcam (audio/video) there should be no sync issues [00:57] <burek> but since you are using x11grab (grabbing your desktop) some sync issues can exist [00:57] <burek> which you can resolve with -async or -vsync or -isync [00:57] <burek> http://ffmpeg.org/ffmpeg.html [00:57] <robertzaccour> burek, where would I input -async or -vsync? [00:58] <burek> I'm not quite sure, I didn't use it a lot [00:58] <burek> try and see :) [01:03] <robertzaccour> :( no dice [01:03] <robertzaccour> the tutorials are confusing to me [01:06] <robertzaccour> burek, do you know how to capture both audio and video directly from the webcam? instead of x11grab what should I do? [01:06] <burek> ffmpeg -f alsa -ac 2 -i pulse -r 30 -s 1366x768 -f video4linux2 -i /dev/video0 -vcodec msmpeg4v2 -qscale 2 screencast.avi [01:07] <burek> or /dev/video1 if you have multiple devices [01:09] <burek> need to go, I hope you'll make it :) [01:10] <robertzaccour> oh didn't yet but thanks [01:34] <praedo> hello [01:35] <praedo> i try to catch a png photo from a mp4 video and being ntsc it shows with blue and red lines [01:35] <praedo> i use this: [01:35] <praedo> ffmpeg -i file.mp4 -vframes 1 -ss 00:01:00 -an -vcodec png -f rawvideo -s 93x70 file.png [01:36] <praedo> do i need to specify something for ntsc input? [01:40] <pasteeater> praedo: what are you useing to view the output? [01:40] <pasteeater> also "-an -vcodec png -f rawvideo" are unnecessary [02:00] <praedo> why? [02:00] <praedo> i just double click to open the png [02:01] <praedo> http://www.ixac.tv/tramuntana_live/reportatges_sports/load/images/uef_loren… [02:01] <praedo> this is ranamed to jpg but it's png [02:01] <praedo> no problem there [02:01] <praedo> but you can see how it looks [02:31] <elkng> using 2 pass ecoding ensures exact bitrait but is there different in CPU load if I create video using regular one pass way or using 2 pass way ? will 2 pass encoding generate video that will load CPU less while playing it ? [05:52] <Shimmy> I have a video that is working on my device. Here is its details: http://pastebin.com/10AigrQ1 [05:53] <Shimmy> I want to convert videos to match its same details so the converted ones should also work on my device. I'm an ffmpeg noob, please help me with what params I should pass. [07:34] <elkng> I encoded video using libvorbis for sound using "ffmpeg version 0.10.2" and when I played it on another PC with "MPlayer 1.0rc1" there [07:34] <elkng> is a problem with a sound its appeared every few seconds for half a second is that problem with incorrectly encoded libvorbis or with version of mplayer ? I've no problems using "libmp3lame" for sound [08:18] <Shimmy> Anyone around? [08:19] <Shimmy> Is there anything in ffmpeg that can mimic params from a different file? [08:19] <Shimmy> I need to convert a file so it should have the same configurations like a third file, can i extract a preset/params list from it? [08:22] <elkng> ffprobe ? [08:28] <Shimmy> what's ffprobe? [08:29] <Shimmy> second, im checking [08:35] <Shimmy> elkng, it doesn't do what I want, I want to get a ready command line configuration/preset for a specific file, i.e. if the video size is WxH, it should add a parameter -s WxH and so on, is there anything that does it? [08:36] <Shimmy> This file: http://pastebin.com/10AigrQ1 works on my device, I'm trying to convert another video to have the same configuration but it doesn't work on my device [08:37] <elkng> I want the same [08:42] <elkng> "ffwavesynth Internal wave synthetizer. This decoder generates wave patterns according to predefined sequences. Its use is purely internal and the format of the data it accepts is not publicly documented.", what does "not publicly documented" means ? [08:57] <Plorkyeran> it means it's not documented and don't use it [09:31] <Shimmy> How to convert a video it should be like his one: http://pastebin.com/10AigrQ1 [10:33] <elkng> -vcodec mpeg4 -vb 400k -s 160x128 -ab 128k -ar 44100 -ac 2 -acodec mp2 [10:43] <Shimmy> elkng, it doesn't work on my device. [10:45] <Shimmy> Here is the converted file. [10:46] <Shimmy> The converted file doesn't work: http://pastebin.com/r5825AVF [10:46] <Shimmy> This file however, works on the device: http://pastebin.com/10AigrQ1 [10:46] <Shimmy> What did I miss out? [10:54] <Shimmy> Anyone? [11:05] <elkng> try to add "-vtag xvid" option [12:38] <fantastic001> hi, i am trying to compile opencv but i get error, there is not any function/variable needed for opencv in ffmpeg, i need to compile ffmpeg with something but what ? i asked on #opencv but no answer, this is error /tmp/fantbuild-work/opencv-2.3.1-1/OpenCV-2.3.1/modules/highgui/src/cap_ffmpeg_impl.hpp:1485:25: error: av_write_header was not declared in this scope [12:39] <fantastic001> i compiled ffmpeg without libavdevice [12:40] <burek> it could be that you need to use exact version of opencv and ffmpeg [12:40] <burek> in order to bind those together [12:41] <fantastic001> i use ffmpeg from git [12:41] <fantastic001> and opencv 2.3.1a [12:44] <burek> you need to read documentation of opencv [12:44] <burek> to see what exact version of ffmpeg you need [12:46] <fantastic001> i'll try opencv from svn [12:48] <fantastic001> and when i was compiling ffmpeg i set --libdir=/usr/lib64 but make install installed .so libraries to /usr/lib, why ? [12:49] <burek> you need to use --prefix [12:49] <fantastic001> i used [12:49] <burek> type ./configure --help [12:49] <burek> and read it [12:51] <fantastic001> this is my ./configure command http://pastebin.com/dWLxpc8b [12:51] <fantastic001> $LIBSUFFIX is set to 64 [12:53] <burek> Shimmy, try using MEncoder [12:53] <burek> Metadata: encoder : MEncoder dev-SVN-r27958-4.2.2 [12:54] <burek> fantastic001, read ./configure --help [12:54] <burek> nobody is gonna solve your problem for you [12:54] <burek> if you don't want to read, well be it [12:56] <fantastic001> agh, there is --shlibdir :D [13:08] <ashka> hello [13:09] <ashka> as you may know, ffmpeg introduced a support to stream with rtmp [13:09] <ashka> is it possible to get a avi/mkv file from a live rtmp stream ? [13:12] <ashka> nvm [17:56] <Adys> Trying to record from webcam and keep audio in sync, this has audio running behind by 3-4 seconds [17:56] <Adys> ffmpeg -f alsa -i pulse -f video4linux2 -s 800x600 -i /dev/video0 -sameq -acodec libmp3lame -ab 96k out.mkv [17:58] <burek> can you please use pastebin.com, to show your command line and its output? [18:01] <Adys> burek: http://paste2.org/p/2000408 - first part recording it with ffmpeg, 2nd part playing it with mplayer [18:02] <burek> is this an usb webcam [18:02] <Adys> yes [18:02] <burek> i suppose usb 2.0 [18:02] <Adys> ya [18:03] <burek> 153600 kb/s [18:03] <Adys> input comes from a separate usb mic if that makes any difference [18:03] <burek> im not sure if 2.0 can support such speed [18:03] <burek> well audio is not a problem [18:03] <burek> but video is [18:03] <burek> although they most probably share the same usb host controller [18:03] <burek> try lowering your video resolution to 320x240 [18:04] <burek> just to see if it works [18:04] <Adys> hmm [18:04] <Adys> i get some weird speedup [18:04] <Adys> and lots of PTS 4254653, next:4716287 invalid dropping st:0 [18:05] <burek> try 400x300 [18:06] <Adys> same [18:06] <burek> try using v4l2-ctl [18:06] <burek> to see webcam capabilities [18:06] <burek> which resolutions it supports [18:07] <Adys> gives me usage, what do I use? [18:08] <burek> v4l2-ctl --list-formats-ext [18:08] <Adys> burek: http://paste2.org/p/2000416 [18:10] <burek> ffmpeg -f video4linux2 -s 320x240 -r 25 -i /dev/video0 -f alsa -i pulse -sameq -acodec libmp3lame -ab 96k out.mkv [18:10] <Adys> no errors but sound still lagging about 1 second behind [18:11] <burek> well, it's a bug then :/ and you should report it if you can :) [18:11] <burek> http://ffmpeg.org/bugreports.html [18:12] <burek> several people experienced this sync problem between audio/video, myself including, but it wasn't fixed so far :s [18:12] <burek> I'm not sure if it's an ffmpeg bug at all, maybe even alsa or video4linux2.. [18:13] <Adys> using mp4 solves it [18:13] <Adys> hm [18:14] <Adys> even works in 1280x720 [18:14] <burek> cool :) [18:14] <burek> still you could report it for mkv to fix it if there is a problem [18:15] <Adys> will do [18:15] <burek> thanks :) [18:16] <drewfulton> hello, i'm looking for some help compiling from the git, i keep getting an error that kills it during the make ->http://pastebin.com/r701A4SS [18:16] <Adys> burek: though wouldnt mind if you could file it for me, I have a strong distaste for creating new accounts nowadays [18:17] <burek> ok [18:23] <burek> drewfulton, just a sec [18:24] <drewfulton> no hurry [18:25] <burek> drewfulton, are you cross-compiling [18:26] <burek> Adys, this is for your reference: https://ffmpeg.org/trac/ffmpeg/ticket/1259 [18:26] <drewfulton> to be honest, i'm new to git and this process as I usually use MacPorts but needed some external AAC libraries [18:26] <Adys> burek: keeping in mind i use pulse [18:27] <drewfulton> this is my configure line -> http://pastebin.com/EtW2jBHc [18:28] <drewfulton> and then i do 'make' and it runs through a ton of code and ends with that error from the first paste above [18:29] <burek> Adys, yes but -f alsa :) [18:29] <Adys> mmh yeah [18:29] <burek> drewfulton, try running: ldconfig [18:29] <burek> and then repeat make, to see if it will succeed [18:30] <drewfulton> burek, just type ldconfig in the command line? [18:30] <burek> yes [18:31] <drewfulton> command not found [18:32] <burek> as root? [18:34] <drewfulton> yup, BTW, i'm on a Mac running Lion... [18:35] <burek> I really wouldn't know then.. Never used it.. [18:35] <kami8785> how can i get ffmpeg to give me a 5 second clip of a flv? ``ffmpeg -i ogs.flv -ss 1000 -t 5 -acodec copy -vcodec copy -scodec copy s.flv`` gives me a broken 5 sec clip (probably because there's no keyframes) and putting the -ss before the -i gives me a working clip that is 12 seconds long.. [18:36] <drewfulton> looks like dyld is the mac's version of it, one sec [18:36] <burek> kami8785, maybe the original file has got key frames which are not in those 5 seconds [18:36] <burek> so you naturally get defective output [18:37] <burek> for 5 seconds it is wise to reencode anyway [18:38] <kami8785> hi again burek :) [18:38] <kami8785> yeah i want to crop it and then resize that crop /2 [18:39] <kami8785> so i cant get around reencoding? [18:39] <drewfulton> burek, thanks for your help, i'll keep trying things and see what I can figure out. i might be back later [18:42] <drewfulton> burek, separate question taht might end up avoiding all of this to start with, what is the best AAC encoder to use for x264 video? [18:42] <burek> NeroAACEnc [18:42] <burek> :) [18:43] <burek> "to use for x264 video" ? [18:43] <burek> I don't understand [18:43] <burek> kami8785, well, you'll need to reencode anyway :) [18:44] <drewfulton> when i'm encoding something for web display, and i want to set my vcodec to x264, don't i also need to set a separate acodec? [18:46] <burek> well, there is no "best match for x264" if that's what you are asking me [18:46] <burek> there are several aac encoders ffmpeg supports [18:46] <burek> internal aac, libvo-aaenc [18:47] <burek> faac (I'm not sure if this has been deprecated) [18:47] <burek> and libaacplus (non-redistributable) [18:47] <burek> with libaacplus you can create 32kbps audio, which most radio stations online use for broadcasting, because of its superb qualityu [18:48] <drewfulton> ok, let me step back a minute... if I want to create HTML5 compatible video, what would you recommend for the mp4 side of things? [18:49] <drewfulton> sorry for my ignorance, i've been doing a lot of research and feel like i know just enough to get myself in trouble but not have a really full understanding [18:49] <burek> :) [18:50] <burek> my understanding of html5 is that you just create your video/audio and let the client's machine decide how to play it [18:50] <burek> so, it really depends on what client is accessing your media [18:50] <burek> flash player was so far the best choice [18:51] <burek> and flv container with libx264 video + aac audio in it [18:51] <drewfulton> right, one of those optoins needs to be a H.264 compatable file with a H.264 video stream and an AAC audio stream [18:51] <drewfulton> exactly, so what is my best option using ffmpeg to create that file [18:52] <burek> well you can try ffmpeg -i input.wav -f aac -strict experimental -ab 96k -ac 2 -ar 44100 out.aac [18:52] <burek> or use NeroAacEnc and create your audio.aac file [18:53] <burek> and later merge it into media with: ffmpeg -i video.flv -i audio.aac -vcodec copy -acodec copy out.flv [18:53] <drewfulton> alright, i'll give it shot [18:53] <drewfulton> thanks for your help [18:56] <burek> :beer: :) [19:18] <echelon> how do i record a stream that i'm using ffplay to play [19:19] <burek> ? [19:21] <echelon> i use "ffplay -rtsp_transport tcp -probesize 20480 rtsp://208.86.19.32:8000/crtn" to watch a stream [19:21] <echelon> how do i save it [19:22] <burek> ffmpeg rtsp://208.86.19.32:8000/crtn -acodec copy -vcodec copy -scodec copy save.flv [19:23] <echelon> so i don't use -i? [19:23] <burek> pardon, use -i [19:25] <echelon> Output file is empty, nothing was encoded (check -ss / -t / -frames parameters if used) [19:26] <burek> try: ffmpeg -rtsp_transport tcp -probesize 20480 -i rtsp://208.86.19.32:8000/crtn out.mpg [19:26] <burek> or mkv, mp4, avi [19:26] <burek> can you please use pastebin.com, to show your command line and its output? [19:29] <echelon> ah, it worked [19:29] <echelon> thanks :) [19:29] <burek> :beer: :) [19:29] <echelon> ^_^ [19:30] <kami8785> ok if i want to reencode a flv file like this http://pastebin.com/tJNiUEQ8, what would be some good settings so it doesnt look like crap? [19:32] <burek> kami8785, do you see the line (for video) that starts with: Encoding settings: [19:33] <kami8785> yes, but honestly that looks like chinese to me :| can i plug that stuff into ffmpeg? [19:35] <burek> well, I'm not aware of any on-the-fly way to re-encode a file with exact same options [19:36] <burek> but you can try ffmpeg -i ogs.flv -vcodec libx264 -crf 20 -b 888k out.flv [19:36] <kami8785> k cool i will try :) [19:37] <burek> and -acodec copy [19:37] <kami8785> it has to play in flash player. is x264 a good choice for that? [19:37] <burek> ffmpeg -i ogs.flv -acodec copy -vcodec libx264 -crf 20 -b 888k -t 60 out.flv [19:37] <kami8785> something like this would be good http://pastebin.com/2E0CSUXB [19:37] <burek> that would give you 60 seconds of output, just for test [19:37] <kami8785> but ill try yours first [19:37] <burek> which flash player exactly [19:39] <kami8785> the one these guys are using for mouseover previews http://www.videocopilot.net/tutorials/ : http://www.videocopilot.net/assets/public/flash/player.swf [19:45] <burek> web flv player I guess [19:45] <burek> well flv with h264+aac will work generally [19:47] <kami8785> k im recompiling ffmpeg with x264 now [19:48] <burek> good idea :) [21:20] <kami8785> thanks burek, that worked very well, if i want to crop a video and then resize it, what would be the best approach? do it in 2 commands or do one before the other? [21:21] <kami8785> this ` ffmpeg -i s.flv -an -vcodec libx264 -crf 20 -b 888k -vf crop=760:420:260:100 -s 380x210 s2before.flv` doesnt work [21:21] <burek> what exactly doesn't work [21:22] <kami8785> http://pastebin.com/RP35bGyb [21:22] <burek> oh [21:23] <burek> well, ffmpeg first resizes the input [21:23] <burek> and then proceeds to your -vf [21:23] <kami8785> ah ok [21:23] <kami8785> so rework those crop params for the resized version [21:23] <burek> or remove -s [21:23] <burek> and add 'scale' into -vf [21:23] <burek> http://ffmpeg.org/ffmpeg.html#Video-Filters [21:24] <kami8785> oh cool :) [21:26] <kami8785> mhh that doesnt work [21:26] <kami8785> ffmpeg -i s.flv -an -vcodec libx264 -crf 20 -b 888k -vf crop=760:420:260:100 -vf scale=380:210 s2before.flv [21:26] <kami8785> video comes out uncropped [21:28] <burek> no [21:28] <burek> one -vf [21:28] <burek> http://ffmpeg.org/libavfilter.html [21:28] <burek> ffmpeg -i s.flv -an -vcodec libx264 -crf 20 -b 888k -vf 'crop=760:420:260:100,scale=380:210' s2before.flv [21:28] <kami8785> i tried that first but it gave me Unable to find a suitable output format for 'scale=380:210' [21:28] <kami8785> oh [21:28] <kami8785> ok :D [21:29] <kami8785> anyway i got the other way to work already [21:29] <kami8785> thanks so much burek [21:29] <burek> ok [21:29] <burek> :beer: :) [21:29] <kami8785> ill give you a shoutout on this project's website when it's done :) [21:29] <burek> does it include a free beer? O.O :) [21:29] <kami8785> seriously i wouldnt be able to do it without your help [21:29] <kami8785> hmm im afraid my website doesnt service beer t_t [21:30] <burek> :) we're glad we can help :) [23:29] <burek> what is a proper way to compile ffmpeg as 32bit lib/bin using 64bit linux (debian)? [23:41] <cbreak> doesn't debian's gcc support -m32? [23:41] <MooingLemur> burek: well, you need 32 bit versions of the dependency libs as well. Depending on your objectives, you might want to create a full-on 32 bit chroot on your system (see the debootstrap and setarch utils). [23:42] <cbreak> afaik, linux doesn't support multi-arch libs like mac os [23:42] <cbreak> so you probably need to put all your 32bit libs into a different prefix [23:42] <cbreak> (including those required by your compiler) [23:42] <burek> cbreak, that's what I did [23:42] <burek> in libx264 only configure --host=.. was needed to set the output to 32bit [23:43] <burek> but in ffmpeg there are obviously some more things I need to set, to make it 32bit [23:43] <burek> I've set --arch=x86_32 --cc='gcc -m32' [23:44] <burek> is that all or there is something else that is needed to be set? [23:44] <cbreak> you forgot the linker [23:44] <burek> is there any doc/tutorial or something? [23:44] <cbreak> I'd put -m32 into LDFLAGS and CFLAGS [23:44] <cbreak> but I have no idea if ffmpeg uses those [23:44] <burek> well, I think gcc is used for linking too [23:44] <MooingLemur> not sure, but some things look at the machine type (uname -m) [23:45] <cbreak> burek: yes, but not cc [23:45] <cbreak> cc stands for c compiler [23:45] <MooingLemur> setarch (or linux32) might help there [23:45] <burek> I've looked into config.log and it puts -m32 in both compile and link lines [23:45] <burek> MooingLemur, yes, I'll take a look at that too :) [00:00] --- Mon Apr 30 2012

1 0

ffmpeg-devel.log.20120428
by burek 29 Apr '12

29 Apr '12

[10:07] <CIA-17> ffmpeg: 03Nicolas George 07master * r7ad0d9864e 10ffmpeg/libavfilter/ (Makefile asrc_abuffer.c vsrc_buffer.c): lavfi: merge asrc_abuffer into vsrc_buffer. [10:07] <CIA-17> ffmpeg: 03Nicolas George 07master * r2ae5af8094 10ffmpeg/libavfilter/ (Makefile vsrc_buffer.c): lavfi: rename vsrc_buffer.c into src_buffer.c. [10:47] <CIA-17> ffmpeg: 03Nicolas George 07master * r8f19483d06 10ffmpeg/doc/examples/filtering_audio.c: examples/filtering_audio: fix a memory leak. [10:48] <CIA-17> ffmpeg: 03Nicolas George 07master * ra0a0199da3 10ffmpeg/doc/examples/filtering_audio.c: [10:48] <CIA-17> ffmpeg: examples/filtering_audio: do not stop on decode error. [10:48] <CIA-17> ffmpeg: Decode errors can happen with concatenated MP3s [10:48] <CIA-17> ffmpeg: with different formats. [12:47] <aquarat> hi, does anyone know if ffmpeg supports MPEG multichannel audio ? [12:47] <aquarat> http://en.wikipedia.org/wiki/MPEG_Multichannel [12:47] <aquarat> it's a bit of a weird format [12:48] <j-b> did you tried it? [12:48] <Compn> did you have sample ? [12:48] <aquarat> if I hand ffmpeg an mpeg multichannel file it reads it as having 2 chans [12:48] <aquarat> yes I do [12:48] <aquarat> they're fairly large though [12:48] <aquarat> erm [12:48] <aquarat> I have a sample of code [12:48] <aquarat> that can decode it [12:48] <aquarat> a project called mctoolamed [12:49] <aquarat> can decode it with some mods [12:49] <Compn> ok [12:49] <aquarat> btw [12:49] <Compn> i dont remember if ffmpeg has support for it [12:49] <Compn> sounds rare :) [12:49] <aquarat> thatnk you so much for prores support [12:49] <Compn> have to check mpegaudiodec.c or whatnot [12:49] <aquarat> it is rare [12:49] <aquarat> it was originally supposed to be part of the DVD spec [12:50] <aquarat> but was dropped in 1997 [12:50] <Compn> they went with ac3 multichannel :P [12:50] <aquarat> Canon XLH1 cameras generate those files in 4 channel mode [12:50] <aquarat> it's part of the HDV spec [12:50] <aquarat> but very few cams support it [12:50] <aquarat> lol [12:50] <aquarat> yeah :) [12:50] <aquarat> mctoolame.sourceforge.net [12:50] <Compn> but uh, if you upload a sample , and a link to the code that works, someone may add it :) [12:50] <aquarat> kew [12:51] <j-b> does libmad decode it? [12:51] <aquarat> my C++ skills are useless :/ [12:51] <aquarat> haven't tried [12:51] <Compn> j-b works on videolan , so he is also interested in a sample and such :) [12:51] <aquarat> can I cut up the sample using dd? [12:51] <j-b> http://mctoolame.sourceforge.net/test.aiff-mpeg1-192.mp2 ? [12:51] <Compn> yep [12:51] <aquarat> each sample is about an hour long [12:51] <aquarat> kew [12:51] <aquarat> I used ffmpeg to demux the mp2 from the hdv files [12:52] <Compn> ftp://upload.ffmpeg.org/incoming [12:52] <aquarat> that'll probably work @ j-b [12:52] <j-b> so, this is easy [12:52] <j-b> a sample is alredy there [12:52] <j-b> and ffplay, VLC and Mplayer says it is stereo [12:52] <aquarat> from 1996 [12:52] <j-b> roflillol [12:52] <aquarat> yes [12:52] <aquarat> but it's not :P [12:52] <aquarat> hdv cams generate 4 channels [12:52] <aquarat> mctoolamed dies if you hand it a four channel file [12:52] <Compn> oh so one of you tested if that sample is multichannel ? :) [12:53] <Compn> making sure its not just a 2ch sample [12:53] <j-b> Compn: lol [12:53] <aquarat> I haven't recently tested [12:55] Action: aquarat compiles mctoolamed [12:57] <j-b> mediainfo says stereo [12:57] <j-b> but it could be worong too [12:57] Action: aquarat changes arch to "core2" [12:59] <Compn> how can you guys just pick a sample and think its correct? :P [13:00] <aquarat> that mp2 file has 5 channels in it [13:00] <Compn> yay [13:00] <aquarat> I just tested [13:00] <Compn> ok [13:00] <Compn> if you want to make a report, you can file it here > http://ffmpeg.org/trac/ffmpeg/ [13:00] <aquarat> the code only make provision for 2,6 or 8 channels [13:00] <Compn> or i can do it if you dont feel like registering [13:00] <aquarat> I had to mod it for 4 chan [13:01] <Compn> thanks for testing sample [13:01] <aquarat> you're probably better equipped to submit a report? [13:01] <aquarat> I can try [13:02] <CIA-17> ffmpeg: 03Reimar Döffinger 07master * r9d4a1aca88 10ffmpeg/libavresample/x86/audio_mix.asm: [13:02] <CIA-17> ffmpeg: Fix compilation with YASM/NASM without AVX support. [13:02] <CIA-17> ffmpeg: Signed-off-by: Reimar Döffinger <Reimar.Doeffinger(a)gmx.de> [13:03] Action: aquarat creates new ticket [13:04] <Compn> dont forget to link or attach the code you hacked for 4 channel. [13:10] <aquarat> wow [13:10] <aquarat> it's really really nasty [13:11] <aquarat> but okay [13:11] <aquarat> mctoolame doesn't handle sample rates properly either [13:11] <aquarat> it assume 44.1kHz [13:11] <aquarat> HDV cameras shoot 48kHz [13:12] <aquarat> https://ffmpeg.org/trac/ffmpeg/ticket/1258 [13:12] <aquarat> ^^^ I hope I did that correctly [13:15] <Compn> yeha, what about your hacks for the 4channel ? [13:15] <aquarat> I'm looking for it ;) [13:15] <aquarat> I last used this code 2 years ago [13:16] <Compn> oh :) [13:17] <aquarat> I thought I'd be clever... someone gave me an HDV camera to shoot part of a documentary on [13:17] <aquarat> and I switched it to 4 channel mode [13:18] <j-b> but who used those mpeg multichannel codec? [13:18] <aquarat> it's part of the HDV specification [13:18] <aquarat> in four channel mode HDv cameras switch the audio stream from MPEG2 layer 1 to layer 2 [13:18] <Compn> j-b : <aquarat> Canon XLH1 cameras generate those files in 4 channel mode [13:18] <aquarat> and encode the audio using this method [13:18] <Compn> canon cameras i guess :P [13:18] <aquarat> yes [13:18] <j-b> morons [13:18] <aquarat> the xl h1 in particular [13:19] <aquarat> welll [13:19] <aquarat> it's quite cool [13:19] <j-b> riiight... [13:19] <aquarat> if the decoding ability is available :P [13:19] <aquarat> it is [13:19] <Compn> is your camera from 1996 too ? [13:19] <aquarat> you basically have a surround set of channels [13:19] <j-b> there are not enough 5.1 codecs? [13:19] <aquarat> er [13:19] <Compn> and what documentary are you working on ? :) [13:19] <aquarat> the xl h1 is a 2005 camera [13:19] <aquarat> still being produced [13:19] <Compn> ah [13:19] <aquarat> newer cameras like the EX and EX3 use XDCAM [13:19] <aquarat> which stores channels are pcm [13:20] <aquarat> in seperate tracks [13:20] <aquarat> which is, of course, lovely :) [13:20] <aquarat> lol [13:20] <aquarat> *as [13:20] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * redbde52226 10ffmpeg/libswresample/ (6 files): [13:20] <CIA-17> ffmpeg: swr: change sample format representation so as to maintain the planer/packed distinction. [13:20] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [13:20] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r06b62cb8f2 10ffmpeg/libswresample/ (audioconvert.c audioconvert.h): [13:20] <CIA-17> ffmpeg: swr: move AudioConvert to header [13:20] <CIA-17> ffmpeg: the next commits will need it [13:20] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [13:20] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rbcc66ff0e4 10ffmpeg/libswresample/ (6 files in 2 dirs): [13:20] <CIA-17> ffmpeg: swr: add int16_to_int32_mmx/sse [13:20] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [13:20] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rfa5daaca0d 10ffmpeg/libswresample/x86/ (audio_convert.asm swresample_x86.c): [13:20] <CIA-17> ffmpeg: swr: seperate functions for aligned & unaligned [13:20] <CIA-17> ffmpeg: If someone has an idea on how to do this cleaner, its welcome [13:20] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [13:21] <aquarat> it's so weird that HDV wasn't designed with proper multichannel in mind [13:21] <aquarat> they had the bitrate for it [13:22] <Compn> thanks for reporting bug :) [13:22] Action: Compn goes afk [13:22] <aquarat> lol [13:22] <aquarat> thanks for looking at it [13:22] <aquarat> :) [16:24] <Compn> aquarat : carl requested 4channel sample if you can make one :) [16:46] <aquarat> I'll get it on it [16:55] <CIA-17> ffmpeg: 03Nicolas George 07master * r12eeced8b4 10ffmpeg/libavformat/avio.c: [16:55] <CIA-17> ffmpeg: avio: change ffurl_alloc return code. [16:55] <CIA-17> ffmpeg: If the designated protocol is not found, return [16:55] <CIA-17> ffmpeg: AVERROR_PROTOCOL_NOT_FOUND instead of AVERROR(ENOENT). [16:58] <burek> if anyone has a spare minute/two please take a look at this compile error: http://pastebin.com/MkraQvRL [16:58] <burek> In file included from libavfilter/vf_deshake.c:56:0: ./libavcodec/dsputil.h:448:47: error: expected ;, , or ) before v1 [16:59] <burek> the line in question is this (dsputil.h): void (*butterflies_float)(float *restrict v1, float *restrict v2, int len); [16:59] <burek> is it some kind of typo, or some missing #include or what? [16:59] <burek> it's the latest ffmpeg git in question [17:03] <aquarat> done [17:19] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r113738d6c2 10ffmpeg/libswresample/x86/audio_convert.asm: [17:19] <CIA-17> ffmpeg: swr: more correct cglobal parameters to int16->int32 [17:19] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [17:19] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r832c3b10d2 10ffmpeg/libswresample/x86/ (audio_convert.asm swresample_x86.c): [17:19] <CIA-17> ffmpeg: swr: add int32_to_float_sse2 [17:19] <CIA-17> ffmpeg: could be done for sse/3dnow too if someone wants [17:19] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [17:19] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r95057b1972 10ffmpeg/libswresample/x86/audio_convert.asm: [17:19] <CIA-17> ffmpeg: swr: int16->int32: use the old index negate trick to avoid 2 adds [17:19] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [18:58] <CIA-17> ffmpeg: 03Nicolas George 07master * r75d5624cb2 10ffmpeg/libavfilter/avfilter.c: [18:58] <CIA-17> ffmpeg: avfilter: filter_samples: read pts before filtering. [18:58] <CIA-17> ffmpeg: The call to the next filter_sample will likely unref the [18:58] <CIA-17> ffmpeg: current buffer, so it is not possible to read it afterwards. [19:12] <CIA-17> ffmpeg: 03Jean First 07master * r64943b47e7 10ffmpeg/libavformat/utils.c: [19:12] <CIA-17> ffmpeg: lavf remove duplicated check in has_duration [19:12] <CIA-17> ffmpeg: this hunk was merged in 8b97ae64 and cbf767a8 although the check was there a [19:12] <CIA-17> ffmpeg: few lines above since cdced09e. I removed the first check to reduce the differences [19:12] <CIA-17> ffmpeg: to libav. [19:12] <CIA-17> ffmpeg: Signed-off-by: Jean First <jeanfirst(a)gmail.com> [19:12] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [19:12] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rb72a0f9c23 10ffmpeg/libswresample/x86/ (audio_convert.asm swresample_x86.c): [19:12] <CIA-17> ffmpeg: swr: add int16_to_float_sse2() [19:12] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [19:47] <CIA-17> ffmpeg: 03Nicolas George 07master * rf1b82099cd 10ffmpeg/libavfilter/src_buffer.c: src_buffer: add forgotten avfilter_unref_buffer. [20:07] <CIA-17> ffmpeg: 03Roland Scheidegger 07master * r5520df6a8f 10ffmpeg/libavcodec/x86/cabac.h: [20:07] <CIA-17> ffmpeg: h264: (trivial) remove unneeded macro argument in x86/cabac.h [20:07] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [20:07] <CIA-17> ffmpeg: 03Roland Scheidegger 07master * re52b96187b 10ffmpeg/libavcodec/cabac.c: [20:07] <CIA-17> ffmpeg: h264: (trivial) make ff_h264_lps_state static [20:07] <CIA-17> ffmpeg: not used outside the cabac test functions (which probably means it's [20:07] <CIA-17> ffmpeg: a bad test if it doesn't use the same tables as the real functions?) [20:07] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [20:07] <CIA-17> ffmpeg: 03Roland Scheidegger 07master * r7f668cd2b5 10ffmpeg/libavcodec/ (6 files in 2 dirs): (log message trimmed) [20:07] <CIA-17> ffmpeg: h264: use one table instead of several for cabac functions [20:07] <CIA-17> ffmpeg: The reason is this is easier for PIC code (in particular on darwin...). [20:07] <CIA-17> ffmpeg: Keep the old names as pointers (static in cabac_functions.h so gcc [20:07] <CIA-17> ffmpeg: knows these are just immediate offsets) so the c code can nicely stay the same [20:07] <CIA-17> ffmpeg: (alternatively could use offsets directly in the functions needing the [20:30] <CIA-17> ffmpeg: 03Nicolas George 07master * r4b0521eca9 10ffmpeg/libavcodec/ (avcodec.h options.c utils.c): [20:30] <CIA-17> ffmpeg: lavc: add a channel_layout field to AVFrame. [20:30] <CIA-17> ffmpeg: The field is filled with the codec context information. [20:30] <CIA-17> ffmpeg: FIXME need a minor version bump. [20:30] <CIA-17> ffmpeg: 03Nicolas George 07master * r128dda7016 10ffmpeg/libavcodec/ (avcodec.h options.c utils.c): [20:30] <CIA-17> ffmpeg: lavc: add a sample_rate field to AVFrame. [20:30] <CIA-17> ffmpeg: The field is filled with the codec context information. [20:43] <CIA-17> ffmpeg: 03Nicolas George 07master * rc31be45e14 10ffmpeg/libavcodec/version.h: lavc: minor bump for the new fields in AVFrame. [21:10] <CIA-17> ffmpeg: 03Yusuke Nakamura 07master * r546adc1fee 10ffmpeg/libavformat/mov.c: [21:10] <CIA-17> ffmpeg: mov: Parse EC3SpecificBox (dec3 atom). [21:10] <CIA-17> ffmpeg: Skip to parse fields for additional independent substreams and its [21:10] <CIA-17> ffmpeg: associated dependent substreams since libavcodec's E-AC-3 decoder does not [21:10] <CIA-17> ffmpeg: support them yet. [21:10] <CIA-17> ffmpeg: Signed-off-by: Justin Ruggles <justin.ruggles(a)gmail.com> [21:10] <CIA-17> ffmpeg: 03Luca Barbato 07master * re1e146a2d1 10ffmpeg/libavformat/aviobuf.c: [21:10] <CIA-17> ffmpeg: avio: make avio_close(NULL) a no-op [21:10] <CIA-17> ffmpeg: Its behaviour in line with ffurl_close(NULL). [21:10] <CIA-17> ffmpeg: 03Roland Scheidegger 07master * r14e9ffc1e4 10ffmpeg/libavcodec/ (6 files in 2 dirs): (log message trimmed) [21:10] <CIA-17> ffmpeg: h264: use one table instead of several for cabac functions [21:10] <CIA-17> ffmpeg: The reason is this is easier for PIC code (in particular on darwin...). [21:10] <CIA-17> ffmpeg: Keep the old names as pointers (static in cabac_functions.h so gcc [21:10] <CIA-17> ffmpeg: knows these are just immediate offsets) so the c code can nicely stay the same [21:10] <CIA-17> ffmpeg: (alternatively could use offsets directly in the functions needing the [21:11] <CIA-17> ffmpeg: h264: new assembly version of get_cabac for x86_64 with PIC [21:11] <CIA-17> ffmpeg: This adds a hand-optimized assembly version for get_cabac much like the [21:11] <CIA-17> ffmpeg: existing one, but it works if the table offsets are RIP-relative. [21:11] <CIA-17> ffmpeg: Compared to the non-RIP-relative version this adds 2 lea instructions [21:11] <CIA-17> ffmpeg: and it needs one extra register. get_cabac() gets about 40% faster, for [21:11] <CIA-17> ffmpeg: an overall speedup of about 5%. [21:11] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r9d7c1b4cf3 10ffmpeg/: (log message trimmed) [21:11] <CIA-17> ffmpeg: Merge remote-tracking branch 'qatar/master' [21:11] <CIA-17> ffmpeg: * qatar/master: [21:11] <CIA-17> ffmpeg: h264: new assembly version of get_cabac for x86_64 with PIC [21:11] <CIA-17> ffmpeg: h264: use one table instead of several for cabac functions [21:12] <CIA-17> ffmpeg: h264: (trivial) remove unneeded macro argument in x86/cabac.h [21:12] <CIA-17> (8 lines omitted) [00:00] --- Sun Apr 29 2012

1 0

ffmpeg.log.20120428
by burek 29 Apr '12

29 Apr '12

[01:01] <Diogo> hi again..one question this is possible to stream a videofile do windows media server? [01:02] <Diogo> http://forum.videohelp.com/threads/326575-ffmpeg-to-windows-media-server-wh… --- i saw this post talking about that...thanks [10:03] <elkng> is mp3 free to use ? can I create audio files in *.mp3 formats ? [10:34] <burek> I think not [11:48] <elkng> is "-vcodec mpeg4" the same as "-c:v mpeg4" ? [11:52] <sacarasc> I think so. [11:55] <elkng> I have video "h264 1280x720" what is the fastest way to rescale it to "640x480" ? entire recoding ? [11:57] <sacarasc> Yes, you'd have to do that. [12:17] <elkng> when I use option "-pass 1" and it generates file 1.avi can I use same output file 1.avi with option "-pass 2" ? the important file after making "-pass 1" is ffmpeg2pass-0.log ? [12:18] <sacarasc> If you want to use the same output file, you'll need to use -y on the command line too. [12:18] <sacarasc> Or, you could push the first pass output to /dev/null. [12:20] <Vigud> Hi, I want to convert a video so that if in the source video something is barely moving, then in the output video it's not moving at all. [12:21] <Vigud> Can anyone here help me? I'm new to video encoding/decoding. [12:22] <Vigud> I have tried noise reduction but it didn't give me expected results. [12:38] <aquarat> Hi... does FFMpeg support multi-channel mpeg2 audio files ? ie. a single mp2 audio file that contains two or more channels of discrete audio? [12:53] <Doxin> I've got a program that puts out raw RGB24 frames, and i'd like to pipe those to ffmpeg to encode it. any pointers? [13:15] <k-man> what preset should i use to encode for iphone 4s/ [13:15] <k-man> ? [15:43] <Adys> while recording from webcam, I get errors such as "The v4l2 frame is 614396 bytes, but 614400 bytes are expected" [15:43] <Adys> after a few seconds and it stops recording.any way to avoid that? [15:47] <burek> can you please use pastebin.com, to show your command line and its output? [16:15] <howdy> hi. can anyone help me with this error? http://pastebin.com/iA2jneY3 [16:19] <howdy> anyone here use ffmpeg? [16:20] <Vigud> Most people here don't answer questions, they ask them. [16:20] <howdy> Vigud: how about you? do you answer? [16:22] <howdy> cbreak, my main man from amsterdam. you around? [16:22] <cbreak> I am not from that place. [16:22] <Vigud> No, I am one of many askers here. [16:23] <howdy> cbreak: oh, thought you were. you usually help me out. can you help me with this error? [16:23] <cbreak> no idea. FFMPEG seems to think that your input stream is broken [16:24] <cbreak> "Application provided invalid, non monotonically increasing dts to muxer in stream 0: 67 >= 33" [16:24] <howdy> so there's no way around this? [16:25] <howdy> like a way to force it or something? [16:25] <cbreak> no idea. I would try to put the -vcodec copy after the -acodec so that you don't reorder the streams, but I doubt that that's the problem's cause [16:27] <howdy> cbreak: ya, i still get the same error [16:32] <howdy> cbreak: do you possibly know of something aside from ffmpeg that could get the job done? [16:32] <cbreak> maybe using MP4Box directly to mux, after using ffmpeg or what ever to transcode the audio [16:33] <cbreak> but I haven't compiled MP4Box in ages and it was a major pain to do it on OS X [16:33] <howdy> cbreak: i'm on os x also. i've tried that and it works but the audio is way out of sync [16:36] <burek> howdy, the error is produced in mp4 muxer [16:36] <burek> so try another format [16:36] <burek> and see if the error persists [16:37] <howdy> burek: i tried mkv and i get the same error. is there another container you suggest? [16:37] <burek> then your input format (mkv) might be damaged/invalid [16:38] <burek> did you try extracting raw audio/video to dev/null [16:38] <burek> just to see if everything is fine with the input [16:38] <howdy> burek: how do i do that? [16:38] <burek> -vn -an [16:39] <howdy> so just like ffmpeg -i movie.mkv -vn -an ?? [16:39] <burek> ffmpeg -i input -vn -f wav /dev/null [16:39] <burek> or something similar [16:40] <howdy> just to be sure, i get "File '/dev/null' already exists. Overwrite ? [y/N]". select yes? [16:40] <burek> also try updating your ffmpeg, it might help [16:40] <burek> yes, or use -y [16:41] <fantastic001> hi, when i try to compile ffmpeg i get this error ./libavcodec/dsputil.h:448:47: error: expected ;, , or ) before v1 [16:41] <howdy> burek: it seems to have worked. this is what i got http://pastebin.com/FrxrGCXJ [16:42] <burek> now try with video [16:42] <burek> maybe -f rawvideo or something [16:42] <burek> not sure.. type ffmpeg -codecs [16:43] <burek> no ffmpeg -formats [16:43] <burek> fantastic001, did you change the source code in any way? [16:43] <howdy> burek: so just add -f rawvideo to the syntax? [16:43] <cbsrobot> howdy: you can also try to use the following command to see where problem occures: ffmpeg -i input.mkv -vf showinfo -f null /dev/null [16:43] <howdy> or remove -vn also? [16:43] <burek> howdy, ffmpeg -i input -an -f rawvideo /dev/null [16:43] <fantastic001> burek: no [16:44] <burek> fantastic001, where did you get the sourcecode from? [16:44] <fantastic001> from git repo [16:44] <burek> let me take a look, just a sec [16:44] <howdy> burek: it still has a while to go. but it is working without error [16:45] <cbsrobot> fantastic001: do a "git pull" again [16:45] <howdy> cbsrobot: thanks. i'll try that as soon as this finishes [16:45] <cbsrobot> make clean [16:45] <cbsrobot> -configure again [16:45] <cbsrobot> and if it breaks try "git diff" [16:45] <fantastic001> cbsrobot: i did git clone today [16:46] <cbsrobot> what compiler ? [16:46] <burek> void (*butterflies_float)(float *restrict v1, float *restrict v2, int len); [16:47] <fantastic001> gcc [16:47] <fantastic001> 4.7 [16:47] <cbreak> interesting... --enable-nonfree [16:47] <burek> howdy, then your input is ok [16:47] <fantastic001> i enabled nonfree [16:48] <burek> it's probably mp4 muxer having trouble with your a/v streams [16:48] <burek> try using mkv for the output too [16:48] <howdy> burek: here's the final output http://pastebin.com/Cztm0D9M [16:49] <cbreak> fantastic001: that's not for you :) [16:49] <howdy> burek: i get the same error when i try mkv [16:49] <cbreak> I was just surprised that I had to enable nonfree for openssl [16:49] <howdy> burek: i also get an error even if i try to copy both the -acodec and -vcodec [16:50] <burek> cbreak, yeah, that's really weird :) [16:50] <cbreak> maybe appache license isn't free enough :) [16:50] <fantastic001> does anyone knows why i get this error ? :D [16:51] <burek> howdy, it's most probably a bug in ffmpeg, but in order to be sure, you need to update your ffmpeg to the latest git [16:51] <burek> fantastic001, can you paste 10-20 lines above that one on pastebin.com [16:53] <howdy> burek: is there an easy way to update? or do i just delete everything and start over? [16:53] <burek> in the folder where you extracted your ffmpeg source code [16:53] <burek> say /tmp/bla/ffmpeg [16:54] <burek> just cd into it [16:54] <burek> and do make uninstall; make distclean; git pull [16:54] <burek> and then do ./configure again [16:54] <burek> make && make install [16:55] <fantastic001> burek: yes, wait please [16:55] <fantastic001> http://pastebin.com/MkraQvRL [17:00] <burek> fantastic001, just a sec [17:00] <fantastic001> ok.... [17:05] <burek> fantastic001, is there any way of downgrading your gcc [17:07] <fantastic001> burek: hmm [17:07] <fantastic001> why ? [17:07] <burek> well, I'm looking at google's results regarding this message [17:07] <burek> I hope he will find his terminal again :) [17:08] <burek> fantastic001 :) [17:08] <burek> I've tried looking at google: http://www.google.com/search?q=libavcodec%2Fdsputil.h+error+expected+before… [17:08] <fantastic001> :) [17:09] <fantastic001> sorry [17:09] <burek> and most of those results point to the broken versions of gcc [17:09] <burek> so my wild guess is that it's the culprit [17:09] <burek> so, if you can, try installing some other version of gcc, either newer or older [17:09] <fantastic001> newer version does not exists [17:10] <fantastic001> this is case only for gcc <4.3 [17:10] <fantastic001> in google results [17:10] <fantastic001> i googled before i asked question :) [17:10] <burek> yes, but still.. [17:10] <cbreak> 4.3 is quite old [17:10] <burek> there are no other references to the solution of that kind of problem [17:10] <fantastic001> cbreak: i know, i am using 4.7 [17:10] <burek> so I guess they've made a regression (again) [17:10] <cbreak> isn't 4.7 a beta? [17:11] <fantastic001> no [17:11] <fantastic001> gcc (GCC) 4.7.0 [17:11] <burek> there is 4.8 [17:13] <fantastic001> hmm [17:13] <fantastic001> i'll try [17:13] <fantastic001> but before i set --enable-shared option in ./configure i compiled ffmpeg succesfully [17:13] <fantastic001> but now with --enable-shared i get error [17:13] <burek> you might check the channel #gcc here [17:14] <burek> and ask there? [17:15] <fantastic001> ok [17:17] <howdy> burek: now it works! thanks bro [17:19] <burek> :beer: :) [17:53] <malinens> hi guys! yadif sucks. with older ffmpeg version I could use -deinterlace option and it worked well but now I can try any combination of yadif option and I still bad results :/ Is it possible to deinterlace video in the same manner as with older ffmpeg? [17:53] <JEEB> malinens, I really don't see how you could've possibly gotten better results with -deinterlace than with yadif [17:54] <malinens> old ffmpeg: http://bwap.org/a.mp4 [17:54] <JEEB> what kind of a source were you dealing with? [17:54] <JEEB> can you post a sample? [17:54] <malinens> new ffmpeg: http://video.tiesraides.lv/flv/20120428_1804_blogs_3_a.mp4 [17:55] <malinens> http://pastebin.com/TwAk1bZi [17:58] <JEEB> anyways, yadif has been so good even in cases where it shouldn't be used that I'd be surprised to see something as hated as -deinterlace to be able to beat it [17:59] <JEEB> in any case, a sample would be nice of the source :P [18:01] <malinens> sample bwap.org/d.mp4 [18:01] <malinens> http://bwap.org/d.mp4 [18:02] <malinens> sorry [18:02] <malinens> http://bwap.org/d.mpg [18:05] <JEEB> ooh, looking at the latter sample I guess you have just selected the wrong field order [18:06] <JEEB> or something along the lines of that :) [18:20] <malinens> yup yadif=0:0 seems ok when checking manually. maybe something with my software [18:33] <malinens> nope :/ http://video.tiesraides.lv/flv/20120428_1904_d.mp4 I use 2-pas encoding with yadif=0:0. command I used: http://pastebin.com/TitcmNRB [18:34] <malinens> ignore -deinterlace. not used in my app... [21:03] <dalkor> http://pastebin.com/EyqwsEfU Any ideas? I have both codecs [21:04] <sacarasc> Sounds like a broken file. [21:06] <dalkor> I figured as much >.> Anyone have experience with rtmpdump? [21:06] <dalkor> Or know where I could get support. [22:33] <odie5533> What is the difference between having vcodec before or after -i or the -ss before or after -i? [22:40] <sacarasc> odie5533: If you have -vcodec before -i, you're telling ffmeg that the input file should use that decoder. [22:40] <sacarasc> I forget about -ss... [22:40] <odie5533> I found this on it: http://ffmpeg.org/pipermail/ffmpeg-user/2011-November/003221.html [22:41] <odie5533> which explains it well enough. Thanks for explaining the vcodec though. I think that might be why I was getting an error [00:00] --- Sun Apr 29 2012

1 0

ffmpeg-devel.log.20120427
by burek 28 Apr '12

28 Apr '12

[00:40] <CIA-17> ffmpeg: 03Luca Barbato 07master * r8b97ae6484 10ffmpeg/libavformat/utils.c: [00:40] <CIA-17> ffmpeg: avf: fix faulty check in has_duration [00:40] <CIA-17> ffmpeg: An invalid duration is AV_NOPTS_VALUE not 0. [00:40] <CIA-17> ffmpeg: 03Hendrik Leppkes 07master * r949d942eef 10ffmpeg/libavformat/isom.c: [00:40] <CIA-17> ffmpeg: mov: support eac3 audio [00:40] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [00:40] <CIA-17> ffmpeg: Signed-off-by: Derek Buitenhuis <derek.buitenhuis(a)gmail.com> [00:40] <CIA-17> ffmpeg: 03Mans Rullgard 07master * rdac78fd1d7 10ffmpeg/libavcodec/arm/asm.S: [00:40] <CIA-17> ffmpeg: ARM: add some compatibility macros [00:40] <CIA-17> ffmpeg: This adds some macros simplifying Thumb and pre-v6T2 compatibility. [00:40] <CIA-17> ffmpeg: Signed-off-by: Mans Rullgard <mans(a)mansr.com> [00:40] <CIA-17> ffmpeg: 03Mans Rullgard 07master * rb692d246ea 10ffmpeg/libavcodec/arm/ (7 files): [00:40] <CIA-17> ffmpeg: vp8: arm: separate ARMv6 functions from NEON [00:40] <CIA-17> ffmpeg: This is a preparation for complete ARMv6 optimisations. [00:40] <CIA-17> ffmpeg: Signed-off-by: Mans Rullgard <mans(a)mansr.com> [00:40] <CIA-17> ffmpeg: 03Mans Rullgard 07master * re4ac031233 10ffmpeg/libavcodec/arm/ (vp8dsp_armv6.S vp8dsp_init_armv6.c): [00:40] <CIA-17> ffmpeg: vp8: armv6 optimisations [00:40] <CIA-17> ffmpeg: Based on patch by Ronald S. Bultje <rsbultje(a)gmail.com>, [00:40] <CIA-17> ffmpeg: partially ported from libvpx. [00:40] <CIA-17> ffmpeg: Signed-off-by: Mans Rullgard <mans(a)mansr.com> [00:40] <CIA-17> ffmpeg: 03Mans Rullgard 07master * r3d11c2d76d 10ffmpeg/libavcodec/arm/vp8dsp_armv6.S: [00:40] <CIA-17> ffmpeg: vp8: armv6: fix non-armv6t2 build [00:40] <CIA-17> ffmpeg: The assembler may fail to place literal pools close enough to [00:41] <CIA-17> ffmpeg: instructions referencing them. An explicit .ltorg directive [00:41] <CIA-17> ffmpeg: fixes this. [00:41] <CIA-17> ffmpeg: Signed-off-by: Mans Rullgard <mans(a)mansr.com> [00:41] <CIA-17> ffmpeg: 03Mans Rullgard 07master * r3527a73933 10ffmpeg/ (5 files in 3 dirs): [00:41] <CIA-17> ffmpeg: avutil: add av_parse_cpu_flags() function [00:41] <CIA-17> (32 lines omitted) [00:42] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r76d376d99b 10ffmpeg/libswresample/swresample_test.c: [00:42] <CIA-17> ffmpeg: swr-test: fix declaration statement ordering. [00:42] <CIA-17> ffmpeg: swr-test is now warning free. [00:42] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [00:42] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r7b2ce50701 10ffmpeg/libswresample/swresample_test.c: [00:42] <CIA-17> ffmpeg: swr-test: fix unused var warnings [00:42] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [00:42] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * ra0dd81dd95 10ffmpeg/libswresample/swresample_test.c: [00:42] <CIA-17> ffmpeg: swr-test:indent [00:42] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [00:42] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rf75169b251 10ffmpeg/libswresample/swresample_test.c: [00:42] <CIA-17> ffmpeg: swr-test: fix pointer type warnings [00:42] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [04:58] <Daemon404> lovely [04:58] <Daemon404> ffmpeg fails to build with mingw/msys due to library.mak [05:00] <Daemon404> olol, it's a problem with line endings [05:00] <Daemon404> stupid msysgit [05:06] <Compn> hehehe [05:06] <Compn> ran into that once or twice :P [05:11] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rbbd3dc49ff 10ffmpeg/libswresample/audioconvert.c: [05:11] <CIA-17> ffmpeg: swr: unroll audioconvert core C function [05:11] <CIA-17> ffmpeg: 36k->32k decicycles [05:11] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [05:11] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r4119fa7023 10ffmpeg/libswresample/audioconvert.c: [05:11] <CIA-17> ffmpeg: swr: audioconvert, mark some constants as floats. [05:11] <CIA-17> ffmpeg: idea from avr [05:11] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [06:06] <Compn> dang [06:06] <Compn> http://stream-recorder.com/forum/adobe-hds-fragment-joiner-t12074.html [06:07] <Compn> ksv coded up a adobe f4f demuxer :) [06:07] <Daemon404> f4f? [06:08] <Compn> stupid adobe http dynamic streaming fragmented file [06:08] <Daemon404> oic [06:08] <Daemon404> also lol [06:08] <Daemon404> its written in php [06:09] <Compn> cross-platform :P [06:09] <Compn> ehe [06:09] <Daemon404> it's using a regex to match hex chars [06:09] <Daemon404> o god [06:13] <Compn> Daemon404 : looks like you want to make an f4f demuxer for ffmpeg :) [06:14] <Compn> i think i have samples if you want [06:14] <Daemon404> it looks simple [06:14] <Daemon404> i certainly dont have time right now though [06:14] <Daemon404> moving to NYC [06:14] <Compn> sounds like funm [06:15] Action: Compn afk [06:15] <Daemon404> lul [11:20] <CIA-17> ffmpeg: 03Nicolas George 07master * re5474f5c48 10ffmpeg/doc/ (ffmpeg.texi filters.texi): [11:20] <CIA-17> ffmpeg: doc: fix syntax for durations and timestamps. [11:20] <CIA-17> ffmpeg: Omitting the seconds has not worked for a long time, if ever. [11:20] <CIA-17> ffmpeg: Omitting the minutes too is just nonsensical for a duration [11:20] <CIA-17> ffmpeg: (it is indistinguishable from just seconds). [12:29] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rdcfc12e86a 10ffmpeg/libswresample/swresample_test.c: [12:29] <CIA-17> ffmpeg: swr-test: simplify prng code [12:29] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [13:45] <Shimmy> Any please help me urgently: https://ffmpeg.org/trac/ffmpeg/ticket/747 [13:54] <Compn> why not use the amv-codec-tools to convert your files ? [13:55] <michaelni> Shimmy, id love to but i have no hardware to test the validity of such encoded streams [13:55] <michaelni> and i dont want to just write it blindly and hope it wont be totally buggy [13:56] <Compn> michaelni : just mark it experimental and wait for bug reports ;) [13:56] <Shimmy> What do you mean you don't have hardware [13:56] <Compn> Shimmy : he doesnt have amv player [13:56] <Compn> do you have amv player software he can test his encoder with ? [13:56] <ubitux> send him a s1mp3 player :) [13:56] <Shimmy> I will give you everything, yes, I have. [13:56] <Compn> michaelni : do you mind testing an encoder against a software amv player ? [13:57] <JEEB> also, I think that's just one of the Chinese "amv" formats :D (IIRC there was some other format, too) [13:57] <Shimmy> Also take into account that it's already done, as I said in that post there is already an open source amv converter, the problem is it doens't work with the new versions of ffmpeg, I need someone to merge the encoder from that open source project (it's ffmpeg-based) into ffmpeg itself. [13:58] <JEEB> Shimmy, as far as I read the old stuff, it wasn't exactly nicely written :P [13:58] <JEEB> so a partial/full rewrite would be needed [13:59] <Shimmy> I have a closed-source windows software that plays amv files, if that player plays the file, I can assum it's gonna play on my hardware (i will convert using the required parameters necessary for my device) amv support a limited range anyway, so the testing job is minor. [14:00] <Shimmy> @JEEB what wasn't nicely written, that open source project? [14:00] <JEEB> the amv-codec-tools thingy [14:00] <Shimmy> aha. [14:01] <JEEB> also, never ever underestimate testing [14:02] <JEEB> the format has no open spec, and it's just a weirdly limited version of MJPEG + ADPCM [14:02] <Shimmy> if i would only know C i would get into it and do it myself. I learned some C, but from learning a bit C to developing an encoder is huge distance :) [14:02] <Shimmy> BTW, the issue, is only with the audio, it's "adpcm_ima_amv" [14:03] <michaelni> Shimmy, ill try to look into the stuff when i have time and post a version for testing [14:03] <Shimmy> and it has decoding support in ffmpeg, but no encoding [14:03] <michaelni> but dunno when ive time [14:03] <Shimmy> michalni, can i contact u privately [14:03] <Shimmy> here in irc [16:56] <ubitux> i'm looking at porting the -async option to an audio filter [16:56] <ubitux> from what i can tell, it just needs the input pts, which lavfi is aware of [16:56] <ubitux> the filter would just "scale" the audio according to this [16:57] <ubitux> (drop samples or insert silence) [16:57] <ubitux> am i missing something? [16:58] <av500> insert silence? [16:58] <av500> that would sound awesome [16:58] <av500> or awful? [17:00] <ubitux> i just want to move the -async option to lavfi [17:00] <ubitux> because it's blocking for the audio filtering in ffmpeg [17:00] <ubitux> there is also the filter reconfiguration to do afaik [17:01] <ubitux> but async is one of the blocking thing [17:01] <ubitux> and i want to move on on this [17:11] <ubitux> also the 3nd thing to do (but not blocking for -af) would be to have the asetframesize filter proposed a while ago, which would add silence padding for the last frame [17:12] <ubitux> it would simplifies the flushing in ffmpeg.c, as well as the frame writing in do_audio_out [17:14] <saste> ubitux: what was the issue with asetframesize? [17:14] <ubitux> dunno, i think there was a tricky poll_frame() [17:14] <ubitux> also, it didn't add silence iirc [17:16] <ubitux> anyway, if all of this is ported to lavfi, we would just need to have an audio configuration like abuffer=...,aresample=...,async=...,volume=...,pan=...,asetframesize=... and all the audio stuff in ffmpeg.c could be drop [17:16] <ubitux> and we will be able to add -af, audio filtering in complex filters etc [17:19] <ubitux> right now we have most of the features duplicated (resample, volume, channel mapping, frame sample buffering, etc) [18:16] <CIA-17> ffmpeg: 03Nicolas George 07master * r3e1d3ebf5e 10ffmpeg/libavutil/parseutils.c: [18:16] <CIA-17> ffmpeg: parseutils: replace setenv() by putenv(). [18:16] <CIA-17> ffmpeg: putenv() seems to be more portable. [19:02] <ubitux> internet is back, i can live again [19:04] <burek> :) [19:30] <burek> what tool can generally be used to copy metadata from one media file to another, like copy/paste [19:40] <saste> burek: http://ffmpeg.org/ffmpeg.html#Metadata [19:40] <saste> there is a metadata muxer/demuxer, you need some layer of scripting though [19:41] <ubitux> when ffmpeg will support json as input metadata file, you should be able to do nasty piping with ffprobe [19:41] <ubitux> shouldn't happen soon though [19:42] <burek> saste, thanks :) this is really cool :) [19:43] <burek> err.. how is this thing used? ffmpeg -i ... -f metadata out.txt ? [19:44] <ubitux> i'd say you extract the metadata with ffprobe [19:44] <ubitux> and then convert it to a ffmpeg cmd line with -metadata title=... or stuff like that [19:44] <burek> oh.. :) [19:45] <saste> burek: indeed the metadata demuxer/muxer is not documented, only the supported format [19:45] <saste> i suppose you need to use the demuxer for extracting metadata, then use the muxer for writing to a file, using ffmpeg -codec copy -acodec copy [19:45] <saste> never done that so i can only guess [19:46] <burek> -f metadata is invalid, or I'm using an old ffmpeg? [19:46] <saste> what about -f ffmetadata? [19:46] <burek> :beer: :) [19:47] <ubitux> ah nice thing [19:47] <burek> in the docs, there is "metadata", so I tried that first :) [19:48] <saste> ubitux: for *reading* from JSON you need a parser, writing that in ffmpeg would be a bit annoying (same for xml) [19:48] <saste> unless you rely on external libs, and bloat dependencies [19:48] <ubitux> it was just an insane suggestion [19:48] <ubitux> don't take it too seriously ;) [19:49] <burek> hm, I tried ffmpeg -i ... -f ffmetadata out.txt [19:49] <burek> but it just outputs ;FFMETADATA1 line [19:49] <burek> is this a placeholder for future mux/demux or it is a working one? [19:51] <CIA-17> ffmpeg: 03Clément BSsch 07master * re40981b8c4 10ffmpeg/libavformat/microdvddec.c: lavf/microdvd: set packet duration. [19:51] <CIA-17> ffmpeg: 03Clément BSsch 07master * r423b827668 10ffmpeg/ (6 files in 3 dirs): [19:51] <CIA-17> ffmpeg: lavc: add MicroDVD decoder. [19:51] <CIA-17> ffmpeg: Based on my MicroDVD->ASS conversion code from MPlayer [19:51] <CIA-17> ffmpeg: (sub/subassconvert.c). [19:52] <saste> burek: works here, are you sure you *have* metadata in the file? [19:52] <saste> also from what i see the demuxer only extracts metadata in the format, not in the streams [19:52] <burek> oh I see, this one is in the stream, you're right [19:53] <burek> it's ogg audio file, and the vorbis audio contains meta [19:53] <saste> ah... so it's more complicated [19:54] <burek> well, this is also something useful :) [19:54] <burek> at least for some meta I don't have to manually do copy/paste :) [19:56] <saste> how do you set metadata to a specific stream? [19:56] <saste> i suppose ffmpeg can do that, but i don't know how [19:57] <saste> then you can always parse the ffprobe output and insert metadata to the right place, via scripts [19:57] <burek> I tried -map_metadata [19:58] <saste> yep, that's even better [20:00] <burek> oh now I get it, first the output is specified and then input :) [21:40] <CIA-17> ffmpeg: 03Stefano Sabatini 07master * r6f3b1d7f86 10ffmpeg/libavfilter/vf_bbox.c: lavfi/bbox: add missing space between pts and x1 values in the log [21:40] <CIA-17> ffmpeg: 03Stefano Sabatini 07master * r6b55aab01a 10ffmpeg/libavfilter/ (5 files): [21:40] <CIA-17> ffmpeg: lavfi: create buffer reference in filters which need to access the ref later [21:40] <CIA-17> ffmpeg: Also add internal function ff_null_start_frame_keep_ref(). [21:40] <CIA-17> ffmpeg: Fix crash when a following filter (e.g. settb) will unref the reference [21:40] <CIA-17> ffmpeg: passed by start_frame(), and then the reference is accessed in [21:40] <CIA-17> ffmpeg: end_frame() through inlink->cur_buf. [21:58] <Compn> anyone remember the current status of the 3ivx decoder ? [21:58] <Compn> er 3iv1 [21:58] <Compn> last i remember it was disabled because it slowed down decoding of mpeg4 ? [21:59] <Compn> i'm wondering if i should make it into a small ffmpeg task [22:00] <Compn> michaelni : what should be done? split it out as a seperate file ? [22:00] <Compn> mpeg4video.h:#if 0 //3IV1 is quite rare and it slows things down a tiny bit [22:06] <michaelni> if the code still works when enabled then it could somehow be optimized by using always_inline [22:31] <CIA-17> ffmpeg: 03Matthieu Bouron 07master * r73c1e3baad 10ffmpeg/libavfilter/asrc_abuffer.c: [22:31] <CIA-17> ffmpeg: asrc_abuffer: return EAGAIN if no frame is available [22:31] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [22:47] <CIA-17> ffmpeg: 03Stefano Sabatini 07master * ra853e88af2 10ffmpeg/libavfilter/vf_tinterlace.c: lavfi/tinterlace: fix doxy in copy_picture_field() [22:47] <CIA-17> ffmpeg: 03Stefano Sabatini 07master * rcf9f7e40cc 10ffmpeg/libavfilter/vf_tinterlace.c: lavfi/tinterlace: use avfilter_unref_bufferp() [22:47] <CIA-17> ffmpeg: 03Stefano Sabatini 07master * r18c6bd098b 10ffmpeg/libavfilter/vf_tinterlace.c: lavfi/tinterlace: remove unnecessary NULL check in start_frame() [22:57] <ubitux> awesome downsampler: wget http://www.ioccc.org/2011/2011.tar.bz2 && tar xf 2011.tar.bz2 && cd 2011/akari && make [22:57] <ubitux> then ./akari example.ppm > out.ppm [22:58] <ubitux> or even ./akari akari.c > akari2.c [23:00] <Compn> michaelni : well i added 3iv1 decoder as a task to http://wiki.multimedia.cx/index.php?title=Small_FFmpeg_Tasks [23:00] <Compn> maybe someone will pick it up [23:00] <ubitux> holy shit we can build akari2.c [23:01] Action: Compn wonders why libav hasnt removed it so far :P [23:02] <CIA-17> ffmpeg: 03Diego Biurrun 07master * r1b4f085772 10ffmpeg/.gitignore: [23:02] <CIA-17> ffmpeg: gitignore: replace library catch-all pattern by more specific patterns [23:02] <CIA-17> ffmpeg: Ignoring all files that start with the name of a library matches some [23:02] <CIA-17> ffmpeg: files that are not generated. So replace libfoo/libfoo* with patterns [23:02] <CIA-17> ffmpeg: for static and shared libraries, pkg-config and version files. [23:02] <CIA-17> ffmpeg: 03Martin Storsjö 07master * rdf8aa4598c 10ffmpeg/libavformat/mpegts.c: [23:02] <CIA-17> ffmpeg: mpegts: Make sure we don't return uninitialized packets [23:02] <CIA-17> ffmpeg: This fixes crashes, where the demuxer could return 0 even [23:02] <CIA-17> ffmpeg: if the returned AVPacket isn't initialized at all. This [23:02] <CIA-17> ffmpeg: could happen if running into EOF or running out of probesize [23:02] <CIA-17> ffmpeg: with non-seekable sources. [23:02] <CIA-17> ffmpeg: Signed-off-by: Martin Storsjö <martin(a)martin.st> [23:03] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r28c5775183 10ffmpeg/: (log message trimmed) [23:03] <CIA-17> ffmpeg: Merge remote-tracking branch 'qatar/master' [23:03] <CIA-17> ffmpeg: * qatar/master: [23:03] <CIA-17> ffmpeg: mpegts: Make sure we don't return uninitialized packets [23:03] <CIA-17> ffmpeg: gitignore: replace library catch-all pattern by more specific patterns [23:03] <CIA-17> ffmpeg: Conflicts: [23:03] <CIA-17> ffmpeg: .gitignore [00:00] --- Sat Apr 28 2012

1 0

ffmpeg.log.20120427
by burek 28 Apr '12

28 Apr '12

[00:38] <Yuhjn> is anyone awake that can help me with an audio problem? [00:48] <Yuhjn> :( [02:40] <Yuhjn> having trouble building from source [02:40] <Yuhjn> anyone know of a mac build with libvo_aacenc? [02:42] <BoomerBile> what happened to the -vpre lossless_ultrafast preset? [02:42] <mjevans> I think that might be profile or something now [02:43] <sacarasc> It's not needed any more. -preset ultrafast -crf 0 does the same. [02:43] <BoomerBile> ok thanks much [02:53] <mjevans> Ah that was it preset [02:54] <luckz> http://pastebin.com/8tMKe1X6 , http://pastebin.com/DcW2nNrq -- my recorded .ts h264/ac3 file is identified properly but there's a tower of errors? [03:05] <Yuhjn> I wish someone could do a mac build for me, someone who's set up... there are a milion libs you need [03:06] <luckz> Yuhjn: be hilarious, run the windows build in wine [03:14] <aphid> Yuhjn - you tried brew? [03:15] <Yuhjn> brew? [03:19] <aphid> http://mxcl.github.com/homebrew/ [03:20] <aphid> but if you go to #homebrew, expect people to be talking about beer :D [03:21] <Yuhjn> lol [03:22] <Yuhjn> it's actually #machomebrew apparently [03:22] <aphid> yeah [03:22] <Yuhjn> yes that looks very nice... ugh [03:22] <aphid> i brewed ffmpeg, it came pretty loaded without any messing around. even librtmp [03:23] <Yuhjn> ok let me learn it :) [03:23] <aphid> install brew, "brew install ffmpeg", done [03:25] <Yuhjn> nice! [03:25] <Yuhjn> error [03:26] <aphid> :( [03:26] <Yuhjn> The linking step did not complete successfully The forumla built, but is not symlinked into /usr/local. try again using brew link yasm [03:31] <Yuhjn> ok back on track i think now [03:31] <Yuhjn> back to building ffmpeg [03:31] <Yuhjn> grrr [04:21] <buhman> I've got some dvd-video mpeg-2's that I encoded with ffmpeg, and when I try to dvduathor them, I get all sorts of http://sprunge.us/XWOQ [04:24] <buhman> I encoded the videos with "ffmpeg -i foo.avi -target dvd -aspect 16:9 -sameq foo.mpg" [04:25] <buhman> and that's the output of do dvdauthor -o ${i%.mpg} -t $i; dvdauthor -o ${i%.mpg} -T; [04:26] <buhman> interestingly, however, that hasn't happened in all of the videos; only about 1/8->1/6 of the time [04:29] <buhman> I'm not exactly sure whether dvdauthor is being stupid, or ffmpeg legitimately did something terribly wrong [04:30] <buhman> encoded with: http://sprunge.us/HZgf [04:39] <Yuhjn> damnit... brew ffmpeg doesnt have --enable libvo_aacenc [04:45] <aphid> looks like it uses libfaac [04:45] <aphid> you could probably 'brew edit ffmpeg' to change the args [04:51] <buhman> "brew"? [04:59] <Yuhjn> brew is a package manager for mac [04:59] <Yuhjn> homebrew [04:59] <Yuhjn> i got it working with libmp3lame [05:00] <Yuhjn> so dont need vo_aacenc [06:53] <Yuhjn> man quicktime really sucks :( [10:33] <larsemil> hi! i am new sysadmin on a server using ffmpeg for some videoconverting. its installscript says -enable-libfaad but ffmpeg configure script says it does not know of that option [10:34] <larsemil> is faad and faac not needed anymore? [10:35] <larsemil> this is what is in the guide from last sysadmin: [10:35] <larsemil> http://pastebin.com/yzP75rgq [10:36] <larsemil> but it seems to fail on so many levels [10:37] <JEEB> libfaad was removed as the libavcodec aac decoder gained a similar feature set (and in many ways surpassed it) [10:43] <larsemil> JEEB: cool. Must specify target arch and OS when cross-compiling it says now. [10:43] <larsemil> should it not get that self? [10:43] <JEEB> I have no idea what you are doing :P [10:44] <larsemil> solved it now [10:45] <larsemil> hope it does what it is supposed to later. ;) [10:45] <JEEB> also, personally I would only use/enable things I really need [10:46] <JEEB> there's a big chance that on a *nix environment you'll get enable-pthreads by default [10:46] <JEEB> win32/64 have their own threading implementation by default [13:40] <Shimmy> Any please help me urgently: https://ffmpeg.org/trac/ffmpeg/ticket/747 [16:29] <Renard> Hello people [16:29] <Renard> I'm trying to livecast my desktop on mips.tv, which requires the use of rtmp [16:30] <Renard> for this, i'm trying a combination of ffmpeg + x11grab + rtmp:// streaming [16:30] <Renard> So far, I've come up with a syntax like this one: [16:30] <Renard> ffmpeg -re -f x11grab -r 25 -s 800x600 -i :0.0 -f flv rtmp://mips.tv/live/grejghioghre?key=b2e65b2e&id=103899 [16:30] <Renard> (random channel) [16:31] <Renard> which does not work [16:31] <Renard> Anyone could help me adjust it? [16:31] <Renard> (also, add audio) [16:48] <Slackyman> hi [16:51] <Slackyman> I have a LICENSE-related question: I noticed that the OpenSSL license is not compatible with GPLv3 so have I to chose another SLL lib to redistribute my GPLed ffmpeg? [17:17] <StaRetji> Hi folks, I have problem with a picture quality coming from dreambox stream, looks interlaced [17:17] <StaRetji> what would be correct sintax for deinterlaced in latest git ffmpeg? [17:17] <StaRetji> thx [17:39] <Joelito> hi all, question I'm trying to so a simple enconding using this: ffmpeg -i "file.flv" -avcodec copy -ab 128k "out.mp3" [17:40] <Joelito> seems that the flag "-ab" is not use with "-acodec copy" [17:40] <sacarasc> No. [17:40] <sacarasc> You're asking it to copy it exactly without reencoding. [17:40] <Joelito> oww [17:43] <Joelito> can ffmpeg change or "normalize" the bitrate? [17:44] <sacarasc> You could remove -acodec copy [17:56] <Joelito> k, thanks for the help sacarasc [18:21] <StaRetji> folks, how to deinterlace input video with latest ffmpeg? [18:22] <StaRetji> appreciate help, thx ;) [18:23] <JEEB> -vf yadif [18:23] <JEEB> the -deinterlace setting used a very crappy thing, so it was removed because it probably did more bad than good] [18:24] <StaRetji> -vf yadif just after -i path-to-the-source-video? [18:24] <sh_t> hi everyone. i compiled x264 with --prefix=/opt/x264 and i'm wondering how i can get ffmpeg to recognize thats where the x264 libraries are installed when compiling it. there doesn't seem to be an option for specifying the root for x264. any suggestions on this? [18:24] <StaRetji> btw, JEEB thx for reply :) [18:24] <JEEB> after -i yes [18:24] <StaRetji> JEEB thx one more time! [18:25] <JEEB> sh_t, --extra-cflags"-I/opt/x264/include" and --extra-ldflags="-L/opt/x264/lib" [18:26] <JEEB> (or use pkg-config and set PKG_CONFIG_PATH=/opt/x264/lib/pkgconfig/ before calling ./configure) [18:30] <sh_t> JEEB: thank you kind sir [18:34] <andrew_> If I am decoding raw h264 and converting them to JPG images, then do I need to have x264 installed, or will the built in h264 decoder handle this? [18:34] <JEEB> x264 does not decode [18:34] <JEEB> it's an encoder [18:35] <andrew_> ok, so decoding h264 doesn't require x264. I would only require it for encoding h264? [18:36] <JEEB> yes [18:36] <andrew_> Thanks [18:59] <maujhsn> LjL How have you been? [19:06] <maujhsn> ubitux How are you these days? [19:07] <ubitux> fine now i guess [19:22] <sh_t> JEEB: regarding compilation of ffmpeg where x264 is in a different root.. those --extra-cflags were meant to be used in ./configure, right? it's still not seeing it [19:23] <JEEB> extra-cflags for setting another include folder, and extra-ldflags for setting another library folder [19:23] <JEEB> and then you of course have to have --enable-gpl and --enable-libx264 set [19:23] <JEEB> and of course have x264 build the libraries :) [19:45] <TACPILOT> dealing with footage that was deinterlaced in an external app .. [19:46] <TACPILOT> ffmpeg -i reveals this [19:46] <TACPILOT> Seems stream 0 codec frame rate differs from container frame rate: 119.88 (2997/25) -> 59.94 (60000/1001) [19:46] <TACPILOT> how can I change the container value to 59.94 without changing the actual frame rate ?? [20:12] <Diogo> i i have this code, what is the best way to restreming a rtmp? [20:12] <Diogo> Command: /ffmpeg -i "rtmp://ORIGIN" -re -isync -vcodec libx264 -b 600k -acodec libfaac -ab 128000 -ar 32000 -f flv "rtmp://DESTINY" [20:12] <Diogo> this code work but the player stop after some time.. [20:13] <Diogo> this is possible to add a buffer? [20:13] <Diogo> thanks.. [20:21] <Diogo> anyone..please [20:27] <burek> Diogo, you could move -re before -i [20:27] <burek> but you don't actually need it [20:27] <burek> -re is used when you are reading input from file [20:27] <burek> ffmpeg tries to read it as fast as possible (to speed up encoding process) [20:28] <burek> so if you need to "play" that input stream somewhere, like online stream or something, then you need to slow down the input with -re (real time reading of input) [20:28] <burek> if you are getting your input from some other live streaming source, it is already "real time" :) [20:29] <burek> and also, you can copy/paste the output of that command to pastebin.com, so we can investigate [20:33] <Diogo> http://pastebin.com/bQMi45zk [20:34] <Diogo> this is possible the rtmp:// do a buffer and stream from buffer.. [20:34] <Diogo> to rtmp server.. [20:35] <burek> Received signal 2: terminating. [20:35] <burek> did you stop it? [20:37] <burek> well, ffmpeg already is using some buffering [20:37] <burek> try reading the documentation to find out the exact parameter you need: http://ffmpeg.org/ffmpeg.html [20:38] <burek> You can also try using fifo filter: http://ffmpeg.org/ffmpeg.html#fifo [20:38] <burek> with: -vf "[in] fifo [out]" [20:38] <Diogo> yes [20:39] <Diogo> ok thanks for your help [20:40] <burek> :beer: :) [20:40] <Diogo> and for stream a file to a rtmp:// .. [20:40] <Diogo> its the same method?? [20:45] <burek> yes, just add -re before -i [20:45] <burek> to slow down the reading of that input file, to be real time [20:46] <Diogo> but i need slow that real time... [20:54] <burek> Diogo, slow down the input? why? [20:57] <burek> anyway, you can use setpts filter to slow down your input: http://ffmpeg.org/ffmpeg.html#setpts [20:58] <roxlu> hi [20:58] <roxlu> how can I use ffmpeg to create a quicktime with alpha support of a set of png's ? [21:02] <fantastic001> hi, i am trying to compile ffmpeg but i get this error [21:03] <fantastic001> ./libavcodec/dsputil.h:452:47: error: expected ;, , or ) before v1 [21:03] <fantastic001> can anyone help me ? [21:03] <fantastic001> version 0.10.2 [21:11] <luc4_mac> Hi! I'm trying to rotate a h264 video but it seems that after some time the encoding stops. Video is 6:40 and now I'm stuck at 1:12. Any idea why? [21:26] <DelphiWorld> hey all :) [21:27] <cbsrobot> luc4_mac: with so little information given, I'd say it because of cosmic rays [21:28] <DelphiWorld> ffmpeg -formats [21:28] <DelphiWorld> DE amr 3GPP AMR file format [21:28] <DelphiWorld> can AMR be encoded so ? [21:29] <luc4_mac> cbrrobot: I can provide more. This is the line I used: ffmpeg -i <inputfile> -metadata author="" -vf "transpose=1" -vcodec libx264 -crf 20 -r 25 <outfile>. [21:29] <luc4_mac> sorry, mispelled cbsrobot [21:30] <cbsrobot> DelphiWorld: looks like you can [21:30] <DelphiWorld> cbreak: so amr is free? [21:30] <DelphiWorld> hmmm [21:30] <cbreak> hmm? [21:31] <cbsrobot> DelphiWorld: did you compile with --enable-nonfree ? [21:31] <DelphiWorld> cbsrobot: ask debian:) [21:31] <cbsrobot> cbreak: that was a random ping ... :) [21:32] <DelphiWorld> cbreak: LOL. wanted to hack you but you're awake;) [21:32] <cbreak> good luck with that. [21:32] <cbreak> my IP address is 127.0.0.1 [21:32] <DelphiWorld> cbsrobot: but my program is complaining can't find encoder [21:33] <DelphiWorld> cbreak: just a joke, i wanted to talk to cbsrobot [21:33] <DelphiWorld> cbsrobot: http://pastebin.com/vs6rxaHU [21:35] <cbsrobot> cbreak: I hacked you ... I've the same IP ! [21:36] <cbreak> how strange. [21:36] <TACPILOT> cbsrobot needs to send a virus to 127.0.0.1 [21:36] <DelphiWorld> cbreak: me i have 239.100.0.0 [21:36] <cbreak> I think you are not alone [21:37] <cbsrobot> localhost is evil - lets DOS it [21:37] <TACPILOT> lol [21:38] <eviljames> Good thing you guys don't know my top secret password, ********* [21:39] <sacarasc> Your password is potatoes. [21:39] <sacarasc> We know this already. [21:39] <eviljames> HOW ? [21:40] <eviljames> this is horrible /o\ [21:40] <cbsrobot> mine is also: secret [21:40] <cbsrobot> damn nooooo - now it's public [21:40] <andrew_> Does anyone know anything about compiling ffmpeg as a binary for android, rather than the library? [21:41] <DelphiWorld> cbsrobot: please did you see my conf? [21:42] <DelphiWorld> new rev http://pastebin.com/VFcZxPYf [21:43] <cbsrobot> DelphiWorld: what config is that ? [21:44] <DelphiWorld> cbsrobot: MumyuDVB [21:44] <DelphiWorld> cbsrobot: but at least check for me the codec naming if it's correct [21:45] <cbsrobot> transcode_audio_codec=libfaac ?? [21:45] <DelphiWorld> cbsrobot: yes acording to the doc to transcode it to AAC [21:45] <cbsrobot> true [21:46] <DelphiWorld> cbsrobot: but look like, is not finding it. would be good my brother if you give me the Debian Multimedia repository. [21:46] <cbsrobot> but please use the commandline ffmpeg - I do not know MumyuDVB [21:46] Action: DelphiWorld is blind, folks [21:48] <DelphiWorld> cbsrobot: ok give me ffmpeg idea's to confirm if that formats exist [22:21] <vinhdizzo> Hi, I compiled ffmpeg on ubuntu per these instructions https://ffmpeg.org/trac/ffmpeg/wiki/UbuntuCompilationGuide. My ffmpeg command can't find preset files. Can anyone help? ffmpeg -y -f alsa -ac 2 -i pulse -f x11grab -r 15 -s $(xwininfo -root | grep 'geometry' | awk '{print $2;}') -i :0.0 -vcodec libx264 -vpre veryfast -crf 22 -acodec libmp3lame -ar 44100 -ab 128k -threads 0 /tmp/screencast_$DATE-$TIME.mp4 [22:21] <vinhdizzo> or lossless_ultrafast [22:22] <cbsrobot> vinhdizzo: they are obsolet [22:23] <vinhdizzo> hmmm [22:23] <vinhdizzo> are they replaced with anything else? [22:25] <vinhdizzo> well, i actually don't even know what the preset files are used for [22:26] <cbsrobot> hehe [22:26] <cbsrobot> try [22:26] <cbsrobot> ffmpeg -y -f alsa -ac 2 -i pulse -f x11grab -r 15 -s $(xwininfo -root | grep 'geometry' | awk '{print $2;}') -i :0.0 -c:v libx264 -preset ultrafast -crf 22 -c:a libmp3lame -ar 44100 -ab 128k -threads 0 /tmp/screencast_$DATE-$TIME.mp4 [22:26] <cbsrobot> ^untested [22:28] <vinhdizzo> cbsrobot: works, but file size seems to be bigger than what it used to be [22:29] <cbsrobot> use the crf option to tweak the filesize/quality [22:29] <cbsrobot> lower is better qualitiy & bigger files [22:30] <vinhdizzo> (compared to lossless_ultrafast) [22:31] <vinhdizzo> cbsrobot: is the difference between your command and my command the replacement of -vcodec and -acodec with -c:v and -c:a, respectively? [22:32] <cbsrobot> well I removed the -vrpe aswell [22:32] <vinhdizzo> -vrpe with -preset? [22:32] <cbsrobot> -c:v, -c:a is just the way I wrote it [22:32] <cbsrobot> ah yeah [22:33] <vinhdizzo> where can i find the list of values i can pass after -preset? [22:33] <cbsrobot> http://mewiki.project357.com/wiki/X264_Settings [22:33] <vinhdizzo> lossless_ultrafast doesn't seem to wrok [22:43] <praedo> why do i get errors saying ac eob marker is absent? [22:43] <praedo> source is ntsc dv [22:48] <ckb> Seems stream 0 codec frame rate differs from container frame rate: 59.94 (5994/100) -> 29.97 (30000/1001) [22:48] <ckb> Anyone tell me why this version of ffmpeg is giving me this error? (FFmpeg version SVN-r0.5.1-4:0.5.1-1ubuntu1.3) [22:48] <ckb> Should I use another version? [22:56] <cbsrobot> ckb yes [22:56] <ckb> 0.10.2? [22:57] <ckb> How should I install ffmpeg on Ubuntu? [22:57] <cbsrobot> http://goo.gl/9l5Nl [22:58] <TACPILOT> how to get actual frame count ? [22:58] <ckb> gracias [22:58] <ckb> framerate * duration? [22:59] <cbreak> play it with mplayer -benchmark -noframedrop -vo null or so and look at the output :) [22:59] <TACPILOT> will take a look [23:00] <cbsrobot> TACPILOT: ffprobe -show_streams input.mkv [23:02] <praedo> why do i get errors saying ac eob marker is absent? [23:02] <praedo> source is ntsc dv [23:04] <TACPILOT> cbsrobot: thats good to know will take a look [23:06] <TACPILOT> wondering why start time is such "start_time=0.497889" ?? [23:06] <TACPILOT> footage is avchd [23:06] <TACPILOT> interlaced [23:13] <TACPILOT> tnx cbreak and cbsrobot both have useful info, but not seeing frame count [23:14] <cbreak> yeah, it has framerate and exact duration [23:14] <cbreak> I think there was some option that made it spit out frame count but I can't find it right now [23:15] <cbsrobot> nb_frames=..... [23:15] <mp4man> hey! [23:15] <TACPILOT> I'll dig in the mans to see what I can find tnx [23:16] <cbsrobot> ffprobe -show_streams input.mkv 2>&1 | grep nb_frames [23:18] <lghtbng> Good evening; I converted a movie but the first input VOB is completely out of sync (a/v). Any idea how to change the command line so that this works for most DVD scenarios? here is the line: cat VTS*.VOB | ffmpeg -i - -c:v libxvid -c:a vorbis -c:b 192k movie.mkv [23:19] <cbsrobot> lghtbng: try the mkfifo route [23:19] <mp4man> I was wondering whether some of you would have an idea about ahy "convert -resize 2240x1000 pic.png" gives me a 2240x999 png ? [23:20] <mp4man> and if I ask for a 2240x998 it gives me a 2239x998 :/ [23:20] <mp4man> but 2238x998 is ok ! [23:20] <TACPILOT> cbsrobot: blank return [23:20] <cbsrobot> TACPILOT: what kind of file ? [23:20] <cbsrobot> latest git ? [23:22] <TACPILOT> .mts [23:22] <TACPILOT> avchd footage [23:22] <lghtbng> cbsrobot: do you really think that shall cure the problem? [23:23] <cbsrobot> lghtbng: I had to do it once and it worked for me .... [23:23] <cbsrobot> wait ... I try to remember [23:24] <cbsrobot> it was not quality important stuff as it was aimed to be on the internet [23:25] <cbsrobot> maybe I did it like this: [23:25] <cbsrobot> cat VTS*.VOB | ffmpeg -i - -c:v copy -an movie.vob [23:26] <cbsrobot> then extract all audio from the vobs -> wav [23:26] <lghtbng> cbsrobot: I'll try that fifo way, thanks for the hint [23:26] <cbsrobot> and then you can cat the wavs again [23:27] <cbsrobot> and merge the vob and the wav together and encode it to xvid [23:27] <cbsrobot> I do not remember exactly what I did [23:27] <cbsrobot> but lately I tried the mkfifo route and it was really nice [23:28] <cbsrobot> only writing the correct syntax was annoying [23:28] <lghtbng> cbsrobot: what??? is ffmpeg really designed for such cripple ways (split a/v and merge afterwards...)? [23:29] <cbsrobot> lghtbng: it depend on your input files [23:29] <cbsrobot> I also needed to cut and and crop the files ... so [23:30] <cbsrobot> normally I'd say you can cat vobs, but if it does not work, try the other way [23:32] <lghtbng> cbsrobot: it's mpeg2 (standard European DVD). So I am wondering that it's so unintuitive [23:40] <guest12345> does anybody have experience setting up a ffserver (Ubuntu 12.04)? My feed1.ffm file never gets any larger than 4K and the MaxFileSize is set to 2M. The only error message I have to go on is:Feed '/mnt/ramdisk/feed1.ffm' stream number does not match registered feed [00:00] --- Sat Apr 28 2012

1 0

ffmpeg-devel.log.20120426
by burek 27 Apr '12

27 Apr '12

[00:35] <canbal> hi, does anyone know if there are any plans or working progress on supporting H.264/MVC decoding in ffmpeg? [00:36] <michaelni> why do you ask ? [00:37] <michaelni> do you want to work on that ? :) [00:46] <canbal> i work on H.264/MVC and Video+Depth compression in my research [00:46] <canbal> and I rely on ffmpeg constantly to do frame extraction etc [00:47] <canbal> i recently encountered a consumer level camera that encodes MVC and i tried to extract frames from it [00:47] <canbal> ffmpeg obviously doesn't support it, so i was wondering if i could be any help for the implementation [00:47] <canbal> i don't have the time to take the lead, but i would like to contribute if there is any working progress or any intentions [00:49] <gnafu> Ooh, which camera? [00:52] <canbal> http://amzn.to/IqAGL7 [00:52] <canbal> sony td20 3d camera [00:53] <gnafu> Cool, thanks. [00:53] <canbal> np, so any ideas if anyone is working on this and looking for some help? [00:53] <michaelni> canbal, there definitly is interrest in a MVC implementation and i had considered to work on it myself [00:53] <michaelni> but iam too busy :( [00:54] <michaelni> with other open source things ... [00:56] <canbal> i see [00:56] <canbal> i am quite busy too - working part time and doing a phd at the same time [00:56] <canbal> that's why i can't dare to start on my own [01:08] <Compn> canbal : ask on the list [01:08] <Compn> i think there are some interested parties who may want to join up [01:09] <canbal> ok i will do that [01:09] <canbal> thank you for the info [01:09] <Compn> at least michael said he would mentor , in our failed bid for summer of code project :) [01:10] <Compn> so there will be help, if you want to start [01:10] <Compn> i think a good place to start is just to start a discussion on how to work the api [01:10] <Compn> e.g. how to demux it and pass to decoder [01:12] <Compn> seems like a difficult thing to do [01:12] <Compn> well maybe not [01:12] <Compn> wonder how optimizing would go , can motion vectors be reused? :P [01:31] <canbal> for decode process, the MVC stream macroblocks have their own motion vectors [01:32] <canbal> the only difference is they may refer to the first view's decoded frames as reference picture, and that is called disparity compensation [01:32] <canbal> i don't quite understand by the optimization and reusing the motion vectors though? do you mean for encoding the second view? [01:51] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r36a9dd2f4d 10ffmpeg/libswresample/swresample.c: [01:51] <CIA-17> ffmpeg: swr: align options a little [01:51] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [01:51] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r20fc610439 10ffmpeg/libswresample/swresample.c: [01:51] <CIA-17> ffmpeg: swr: Capitalize help texts for AVOptions [01:51] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [01:51] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * ra522e03754 10ffmpeg/libswresample/swresample.c: [01:51] <CIA-17> ffmpeg: swr: Add longer more descriptive alternative AVOptions. [01:51] <CIA-17> ffmpeg: They match avresample where it supports the same fields. [01:51] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [01:51] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rda9587952a 10ffmpeg/libswresample/ (swresample.c swresample_internal.h): [01:51] <CIA-17> ffmpeg: swr: pass resample parameters from the user to the resampler. [01:51] <CIA-17> ffmpeg: doxy and names stolen from avr [01:51] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [01:51] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r23282c2222 10ffmpeg/libswresample/swresample.c: [01:51] <CIA-17> ffmpeg: swr: set AV_OPT_FLAG_AUDIO_PARAM for the AVOptions [01:51] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [04:42] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * ra4b58fd9e8 10ffmpeg/libavcodec/ffv1.c: [04:42] <CIA-17> ffmpeg: ffv1: add examples of supported slices counts [04:42] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [05:22] <taqattack> Anyone know what does fd in "RTMP_SendPacket: fd=5, size=374" mean? [05:23] <Compn> file descriptor ? [05:23] Action: Compn hasnt a clue [05:41] <taqattack> :) [07:32] <ubitux> michaelni: xbmc ppl were annoyed by the libavresample/libswresample state too [07:32] <ubitux> i guess the message was/will be relayed anyway [10:09] <ubitux> should i --disable-libavresample for the old yasm box? [10:21] <TimNich> Not sure I get the significance of the libavresample merge. Is it just there, but unused, or have any calls been switched to use it? [10:41] <michaelni> ubitux, dunno, probably yes [10:42] <michaelni> TimNich, its unused [10:42] <michaelni> or at least should be [10:54] <TimNich> Thnx [11:11] <kierank> mvc isn't even worth a soc project [13:05] <RobertNagy> is there any filter for splitting an interlaced frame into two separate half height frames? [13:06] <RobertNagy> wrong channel [15:51] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rdd63e27574 10ffmpeg/libswresample/swresample_test.c: [15:51] <CIA-17> ffmpeg: swr-test: add sorted formats table from avr [15:51] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [15:51] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r030d6b65ae 10ffmpeg/libswresample/swresample_test.c: [15:51] <CIA-17> ffmpeg: swr-test: indent [15:51] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [15:51] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rb1519102f1 10ffmpeg/libswresample/swresample_test.c: [15:51] <CIA-17> ffmpeg: swr-test: add rates tables containing a new set of sample rates. [15:51] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [15:51] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r69f3af9590 10ffmpeg/libswresample/swresample_test.c: [15:51] <CIA-17> ffmpeg: swr-test: remove 0 terminator in layouts for consistency [15:51] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [15:51] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r7461e9ae45 10ffmpeg/libswresample/swresample_test.c: [15:51] <CIA-17> ffmpeg: swr-test: use formats array [15:51] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [15:51] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r08809e3085 10ffmpeg/libswresample/swresample_test.c: [15:51] <CIA-17> ffmpeg: swr-test: improve output [15:51] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [15:51] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r5347b9da98 10ffmpeg/libswresample/swresample_test.c: [15:51] <CIA-17> ffmpeg: swr-test: print channel layout, instead of just channel counts [15:51] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [15:51] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r27de7a8eb3 10ffmpeg/libswresample/swresample_test.c: [15:51] <CIA-17> ffmpeg: swr-test: remove unneeded hack to test planar formats. [15:51] <CIA-17> ffmpeg: planar formats are now chosen like any other parameter. [15:51] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [15:51] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * re6c460e3ea 10ffmpeg/libswresample/swresample_test.c: [15:51] <CIA-17> ffmpeg: swr-test: Use randomized collision free hypercube corner enumeration to select tests. [15:51] <CIA-17> ffmpeg: This way the user can specify how many or few tests should run while [15:51] <CIA-17> ffmpeg: still providing good coverage over the whole parameter set. [15:51] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [15:56] <ubitux> michaelni: 44100 is a fairly common frequency, wouldn't it be wise to test it? [15:57] <TimNich> ubitux: seconded! [16:04] <kierank> the world should use 48000 khz :) [16:15] <michaelni> ubitux, feel free to add it [16:15] <ubitux> ok :) [16:18] <TimNich> kierank: 96000 surely... [16:19] <kierank> entering audiophile realms imho there [16:19] <kierank> and i meant 48khz above of course [19:23] <CIA-17> ffmpeg: 03Matthieu Bouron 07master * r5473a98e45 10ffmpeg/libavfilter/af_amerge.c: af_amerge: return errors from subfilters when requesting a frame [20:39] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r56f89bae52 10ffmpeg/libswresample/swresample_test.c: [20:39] <CIA-17> ffmpeg: swr-test: improve rounding in set() [20:39] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [20:39] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r73be22f526 10ffmpeg/libswresample/swresample_test.c: [20:39] <CIA-17> ffmpeg: swr-test: fix u8 zero point. [20:39] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [20:39] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rf1f5b39866 10ffmpeg/libswresample/swresample_test.c: [20:39] <CIA-17> ffmpeg: swr-test: support fabrices audiogen like avr [20:39] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [22:59] <Compn> anyone invest in the stock market ? [23:00] <Compn> i'm going to get me some facebook when it goes ipo [23:00] <Compn> even tho i loath the company, its policies and its website.... i think they are a fairly sound business [23:00] <Compn> and seems popular [23:03] <iive> everybody is going to buy. this is no brainer. The real trick is to know when to sell. [00:00] --- Fri Apr 27 2012

1 0

ffmpeg.log.20120426
by burek 27 Apr '12

27 Apr '12

[00:00] <JEEB> anyways, if you're interested in working with libavcodec code or x264's code to embed MVC support, feel free to hop onto #x264dev or #ffmpeg-devel [00:01] <canbal> ok thanks for the guide, i will go on there and see if anyone else is interested in this [00:01] <canbal> my interest started when i realized some consumer level cameras started to encode with MVC [00:03] <canbal> and i rely on ffmpeg to extract frames from videos usually, and i couldn't use it in this scenario [00:08] <DaKaZ> all, I have an interesting audio question that I can't seem to ask correctly in google to find out IF it can be done. Given a sound card with 8 outputs (7.1 audio but outputs as 4 stereo headphone jacks), could I run 4 discrete instances of ffmpeg and decode a different audio file to each output pair? If this is possible, could you point me toward the right documentation on the cli parameters? [01:08] <Yeti218> i just noticed my Galaxy Nexus cannot play the mp4 files with libx264 codec that i generated [01:19] <Yeti218> maybe it is the level 4 that is the problem .. it could play the level3/high version [01:37] <Yeti218> it is not the level as i just made a 3.1 level version and that one also cannot be played. i am stumped [01:40] <Yeti218> i found another difference .. the new one has Standard: PAL [01:40] <Mavrik-> check profile instead of level. [01:41] <JEEB> Yeti218, pastebin your command line for creation of those files, and ffprobe output_file.mp4 of those two files [01:42] <dashcloud> is there a simple way to convert MPEG2-TS files to DVD-compliant video? [01:42] <JEEB> and/or a MediaInfo output of those two files [01:47] <Yeti218> http://pastebin.com/N0NLt0ri [01:48] <Yeti218> it has the (big) working stats an then the (small) not working one [01:50] <Yeti218> dashcloud, i use convertxtodvd to make dvd's [01:50] <Yeti218> but i dont think it is free [01:51] <aphid> "-target dvd" should take care of it, no? [01:52] <Yeti218> JEEB do you need more information? [01:53] <dashcloud> okay- I'm trying that [01:54] Action: JEEB is trying to find the algorithm for the reference frame for a level [01:55] <JEEB> ok, found it [01:57] <Yeti218> did i forget a parameter when encoding? [01:57] <JEEB> I'm just checking some things [01:58] <Yeti218> ok [01:59] <JEEB> funny enough you are within level 3.1 with that reference frame-wise [02:02] <Yeti218> maybe the codecs do differ even though they have the same name? [02:02] <JEEB> your source is from vimeo, guess what they use :P [02:04] <JEEB> try two things: one where you set level to 3.1, and one where you do -x264opts ref=4 [02:04] <JEEB> (with level 4) [02:04] <Yeti218> ref = reframes? [02:04] <Yeti218> k, will do [02:06] <JEEB> if my guess goes OK, the decoder just doesn't like the amount of reference frames >_> (albeit they _are_ within spec in this case, even for level 3.1 -- level 3 would need the number to be dropped to 8) [02:09] <Yeti218> bad option '-ref': '4' [02:09] <JEEB> ... [02:09] <JEEB> I told you how to set it [02:09] <JEEB> -x264opts ref=4 [02:10] Action: JEEB slaps Yeti218 for not listening [02:10] <Yeti218> i have: -profile:v high -level 4 -x264opts -ref=4 [02:10] <Yeti218> ah ref instead of -ref [02:10] <JEEB> you can basically set raw libx264 settings with -x264opts [02:10] <JEEB> so you have opt=setting:opt=setting [02:10] <JEEB> etc. etc. [02:11] <JEEB> in this case we only set one setting [02:11] <JEEB> thus only ref=4 [02:11] <Yeti218> i see, it is almost done [02:12] <Yeti218> hurray [02:13] <Yeti218> it plays .. takes like 10 secs while it seems to prepare for the job, but then it plays [02:13] <JEEB> so yeah, it doesn't like 16 reference frames [02:13] <JEEB> (even if that is within the specified level) [02:14] <Yeti218> indeed [02:14] <Yeti218> tomorrow i will test on an iphone 3gs [02:20] <Yeti218> Thank you for the help JEEB! I have gained a knowledge level [02:20] <JEEB> and I have no idea why I'm still awake [02:27] <Yeti218> it is 2:26 am here .. in 4,5 hrs my kids will wake crying and demanding attention [02:27] <JEEB> add another hour here [02:33] <Yeti218> You are from eastern europe? [05:22] <taqattack> Anyone know what does fd in "RTMP_SendPacket: fd=5, size=374" mean? [11:15] <fgro> Hi folks. Running the following command "ffmpeg -v 10 -vframes 1 -i blocking_video.mp4 -vcodec pam -an -f rawvideo -y output". I expect ffmpeg to process only one frame (1). On my gentoo box with ffmpeg 0.7.8 this works just fine. However, on a mac box with ffmpeg 0.9 (complete version info here https://gist.github.com/2497950 ) every frame gets processed, leading to heavy load. What has changed between the versions? Or is it a bug? [11:15] <fgro> same for debian ffmpeg 0.10.2 [11:21] <fgro> ok guys got it https://ffmpeg.org/trac/ffmpeg/ticket/898 [11:21] <fgro> damn, why does the order of parameters matter...mmm [11:29] <TimNich> Its not unreasonable for the order of parameters to matter if the system needs to be able to decide to which file(s) the parameter should apply.... [11:31] <fgro> TimNich: true [12:27] <Yeti218> when trying to make a video from 1 image i get the error: [libx264 @ 0x2e1e9c0] width not divisible by 2 (1379x1772) [12:27] <Yeti218> width is indeed an odd number but hy is this a problem? [12:28] <JEEB> because with 4:2:0 YCbCr (also colloquially called "YUV") has the Chroma planes saved as 2x2 blocks :P (one value for a 2x2 area) [12:29] <Yeti218> so best is to convert the image to even numbers? [12:30] <Yeti218> i think ffmpeg has a scale option, trying to figure out if that will solve it [12:46] <Yeti218> ok, i got a good video from an image now .. but it has 1 frame/sec (which is ok), but it also lasts 1 sec. Is there an option to enlengthen the video to say 40 secs? [12:48] <Yeti218> A silly way would be to make 40 images, rename them 001 to 040 and then put them in movie with 1 frame per second (-r 1) [12:48] <Yeti218> But i assume there is a smarter way when you only have 1 image to consider to determine the length [12:55] <Yeti218> i now found something that should work: ffmpeg -loop_input -shortest -y -i test.png -i test.mp3 -acodec copy -vcodec libx264 test.mp4 [12:56] <Yeti218> But that gives errors: 1) Unrecognized option 'loop_input' 2) Failed to set value '-shortest' for option 'loop_input' [13:01] <Yeti218> k, got it now .. when i use -loop 1 it works .. -loop_input is deprecated [13:06] <RobertNagy> is there any filter for splitting an interlaced frame into two separate half height frames? [13:08] <RobertNagy> basically what I want to do is: interlaced frame -> split into two half height frames -> scale both -> reinterlace [13:34] <esby> hmmm, are the x264 presets files gone ? (or were there renamed ?) i only see vpx*.presets and some ipod related presets in the source but nothing else? [13:35] <JEEB> you now use libx264's internal presets [13:35] <JEEB> http://mewiki.project357.com/wiki/X264_Settings#preset [13:37] <esby> verthx [13:37] <esby> thx [13:37] <Mavrik> we should really staple this with 15-line size letters on join of this channel :\ [13:37] <esby> sorry :/ [13:42] <Mavrik> not your fault, documentation really isn't explicit enough about that thange [13:42] <Mavrik> and there's alot of obsolete tutorials out there [13:47] <burek> it might be a good idea to start versioning major changes [13:55] <malinens> hi guys. I have problems with deinterlacing videos. In old good days I used -deinterlace option and it worked fine but now with yadif I get occasional artifacts :/ [14:00] <Yeti218> i just noticed that my galaxy Nexus wont play video's with large dimensions .. like 2700x2500 [14:00] <Yeti218> i assumed it would just scale to fit it when displaying [14:03] <JEEB> Yeti218, regarding hardware decoding you should always keep to up to 720p on mobile, unless you are absolutely, positively sure 1080p works. Anything over 1920x1080 is bound to fail. Newest nvidia ASICs are told to technically work with up to 4K (width) content, but I don't think any drivers support this -- and this is PC hardware, not mobile. [14:05] <nutz_> hi all [14:06] <nutz_> i'm tyring to transcode flac-files to mp3. everything works pretty well, except a few metadata-variables dont get transfered =(. e.g. the year. i get from the input file: DATE : 2006, and for the output-file it says: TDRL : 2006 and even TRACKTOTAL : 10, but both wont show up in the id3-tags in easytag [14:08] <nutz_> okay, specifying "-id3v2_version 3" helped the year, not the tracktotal though =( [14:17] <Yeti218> Is there a way to scale the video output so that it only scales when necessary and keeps the aspect ratio? [14:18] <esby> JEEB: it's not possible anymore to use a different preset for first pass and second pass ? [14:19] <Mavrik> Yeti218, you'll probably have to write an expression for scale filter [14:19] <Mavrik> esby, the "fastfirstpass" presets are chosen automatically [14:19] <Mavrik> anything else doesn't make any sense really [14:19] <JEEB> esby, x264 will use fast first pass settings in case you are not using the placebo preset [14:19] <JEEB> which is "everything up to eleven" [14:19] <esby> Mavrik: k [14:20] <esby> well i tried to use veryfast for first pass and slow for second pass, i m getting a [libx264 @ 0x23d4b80] different weightp setting than first pass (2 vs 1) [14:20] <Mavrik> of course. [14:21] <Mavrik> you have to use same presets for both passes [14:21] <esby> k [14:21] <detly> can I specify down to the frame where ffmpeg should start and stop reading input? I can only find CLI options for time, not frames [14:22] <detly> for start, I mean... the "-frames" option seems to allow *stopping* after a certain number of frames [14:31] <b_jonas> detly: could you figure out the fps and just use the time option? -ss I think [14:32] <detly> b_jonas: maybe, but I thought I'd check to see if there's a simpler way [14:32] <b_jonas> I don't know [14:32] <detly> I could also just use avidemux to do the cutting [14:33] <detly> and then ffmpeg for the transcoding, etc [14:37] <Yeti218> weird i have videos that sometimes play and sometimes don't play on my phone .. no idea what it is as i dont see a stacktrace, but i have a feeling it might be available memory related [14:45] <pespin> HI, I need to use hw accelerated decoding for h.264 in android. I saw there's support for it using libstagefright [14:46] <pespin> I fear I need that the apps run with root privilegies to be able to access the hw dev node though, is this correct? [14:47] <pespin> as I guess the layers go like app -> ffmpeg -> libstagefright -> OMXCore -> /dev/msm_vidc_dec [14:47] <pespin> which in turn has these rights: [14:47] <pespin> crw-rw---- system audio 247, 0 2012-04-23 18:51 msm_vidc_dec [14:48] <pespin> does someone now if stagefright is running as a different process with root/system privilegies? [15:23] <hivemind> q [15:24] <hivemind> how i can convert ALL files in catalog from m4a to mp3? [15:25] <hi117> lame [15:50] <hivemind> hi117, Warning: unsupported audio format [15:52] <hivemind> it's ALAC [16:21] <detly> I'm trying to convert video + stereo audio to video + mono, but the output always comes out stereo. I'm using this command line, including "-ac 1": http://paste.pocoo.org/show/587485/ ...am I missing something? [17:44] <Yeti218> i used -ac 1 yesterday and it made my sound mono [17:47] <Yeti218> how did you check stereo? with ffprobe? [18:01] <Yeti218> is there a reason why an mp4 with libx264 video codec and aac audio codec will not stream? [18:04] <JEEB> A) The format, by default, is heavily reliant on an index to make it play'able B) To create the index, the whole file has to be completed encoding-wise, thus it is often put in the end [18:05] <JEEB> you could start playing it semi-right-away if the index was in the beginning, or, if you were using the movie fragments feature of the container [18:06] <JEEB> putting the index into the beginning works for most things, and I have no idea how many things actually support the movie fragments feature (although it would let you play a clip that is being live-encoded f.ex.) [18:06] <JEEB> you can use the qt-faststart app to put the index into the beginning of the file after it has been encoded [18:07] <Yeti218> i will try that! [18:08] <taqattack> JEEB: Do u know what the difference between "RTMP_SendPacket: fd=436, size=308" and "RTMP_SendPacket: fd=5, size=308" is by chance? [18:08] <JEEB> nope [18:09] <Yeti218> hmm i installed ffmpeg from source (for 264 reasons) and it seems qt-faststart is not avaiable .. it tells me to apt-get ffmpeg .. [18:13] <JEEB> Yeti218, you have to compile it separately, it's in one of the folders [18:15] <Yeti218> ah yes the tools folder [18:15] <Yeti218> is omitting acodec the same as -acodec copy? [18:16] <JEEB> no [18:17] <JEEB> nowadays it might make a more sane'ish selection depending on your container of choice, but never ever do expect ffmpeg to do something you might like if you haven't set it [18:37] <Yeti218> i found a log of my video errors [18:37] <Yeti218> it says: GraphicBufferAlloc::createGraphicBuffer(w=1536, h=1872) failed (Out of memory) [18:37] <Yeti218> however that is bigger than the size of my actual video [18:49] <mp4man> Hi there [18:50] <mp4man> I'm trying to encode a video from a series of PNG files, here is the command I'm using : ffmpeg -r 24 -b 18000k -y -qscale 1 -sameq -i test%06d.png current054.mp4 The movie is produced, but the colors are altered (a bit darker). Could someone help me to figure out which argument would preserve the colors? [18:50] <mp4man> tahnks ! :) [18:51] <Mavrik_> colors are darker probably because of color space conversion [18:52] <mp4man> here is the original colors : http://i.imgur.com/w97d8.png [18:53] <mp4man> here is what I obtain with the previous command : http://i.imgur.com/PpM52.png [18:53] <Mavrik_> that actually looks the same on my screen :P [18:54] <mp4man> no kidding [18:54] <Mavrik_> mp4man: it's normal and expected that red's darken when you to RGB->YUV color conversion [18:54] <mp4man> the first one is not more orange? more brighter? [18:54] <Mavrik_> actually, RGB -> YUV420 [18:54] <mp4man> here is what I have in my terminal [18:54] <mp4man> http://pastebin.com/WqcJ7as6 [18:54] <Mavrik_> yeah, I see it now [18:55] <Mavrik_> yes, your input is bgra and output is yuv420p [18:55] <mp4man> I have another software, called "framed" o my mac, which only takes JPEGs and produces MOVs. with it the colors are better... but I don't want jpgs and mov :-/ [18:55] <Mavrik_> mp4man: sadly, there's not good way of preserving colors in such output video [18:56] <Mavrik_> mp4man: unless you switch to a format which is YUV444 or RGB [18:56] <Mavrik_> but those are rare and unsupported [18:56] <mp4man> ha :-/ [18:57] <mp4man> see, I don't really know which format I want... I want to produce movies that are "portable" and that have the same colors as the pngs [18:57] <mp4man> I don't care the saize of the files [18:57] <Mavrik_> well... most consumer widely supported video formats demand yuv420 [18:57] <Mavrik_> which is where you lose your colors [18:59] <mp4man> do you have a solution, even not really well supported, that would preseve my colors? [19:00] <Mavrik_> mp4man: I've had some success with huffyuv with RGB video [19:00] <Mavrik_> and x264 encoded yuv444 video (but I don't think you can encode that with ffmpeg yet) [19:00] <mp4man> i've tried this already : ffmpeg -r 24 -b 18000k -y -vcodec huffyuv -i test%06d.png current054.mp4 [19:00] <mp4man> But it does not work and produces errors : [19:00] <mp4man> [huffyuv @ 0x102054000] get_buffer() failed [huffyuv @ 0x102054000] [IMGUTILS @ 0x7fff5fbfde80] Picture size 0x0 is invalid [19:01] <Mavrik_> put huffyuv in .avi container. [19:01] <Mavrik_> and I think you need so specify image size [19:01] <mp4man> ah [19:02] <mp4man> sorry for my naive questions, I admit I know nothing about encoding videos etc. [19:04] <mp4man> is there a unit for the frame size? using -s 1615×585 (pixels) does not work [19:11] <mp4man> Mavrik_: are you gone? :) [19:13] <DelphiWorld> hey [19:13] <DelphiWorld> can someone help me fix a software that use ffmpeg [19:13] <DelphiWorld> MumuDVB [19:21] <mp4man> Could someone indicate me how I can create a video from PNG files that preserves the colors ? thx [19:29] <mp4man> no one ? :( [19:49] <JEEB> mp4man, if you can limit yourself to very up-to-date libavcodec-based players you can set output colorspace to rgb with vcodec x264 [19:49] <JEEB> that would create H.264-in-4:4:4 H.264 [19:49] <JEEB> argh [19:49] <JEEB> RGB-in-4:4:4 [19:50] <JEEB> otherwise just match the conversion matrices when converting from RGB to YCbCr and back [19:50] <JEEB> because that discrepancy kind of looks like a mismatch of those [19:50] <mp4man> What argument should I add to this command ffmpeg -r 24 -b 18000k -qscale 1 -y -i test%06d.png current054.mp4 [19:51] <JEEB> that is one mess of a command line [19:51] <mp4man> lol [19:51] <JEEB> (ffmpeg uses the BT.601 matrix, just FYI) [19:51] <mp4man> its my copy/paste_I_dont_get_nothing command.... sorry [19:51] <JEEB> (for RGB<->YCbCr (also colloquially called 'YUV') conversion) [19:51] <JEEB> anyways [19:52] <JEEB> ffmpeg -r 24 -i test%06d.png -c:v libx264 -crf 18 out.mp4 would probably be a good base from which begin [19:53] <JEEB> -r sets the rate you want ffmpeg to read the pictures in [19:53] <mp4man> yep [19:53] <mp4man> no -b 18000k ? :) [19:53] <mp4man> I've put this to have the best quality [19:53] <JEEB> then you set the input, any setting after that gets set as a setting for the encoder (before -i = decoder settings) [19:53] <JEEB> then we set the codec for encoding, libx264 (which does H.264) [19:53] <mp4man> and if I get rid of the -qscale 1, I have a lot of pixels, that's just horrible [19:54] <JEEB> then we set the rate control mode for x264, crf [19:54] <JEEB> and we set that to 18 [19:54] <JEEB> (crf = constant rate factor, closest thing we currently have to "constant quality") [19:54] <JEEB> (lower = more bitrate used, higher = less bitrate used) [19:54] <JEEB> anyways, that should be pretty good looking as a start [19:55] <Mavrik_> JEEB: hmm, -vcodec libx264 will force YUV conversion iirc [19:55] <JEEB> yes [19:55] <mp4man> here is what your command gives me : http://pastebin.com/JNGhzVDV [19:55] <JEEB> defaults to 420 [19:55] <JEEB> just like x264 cli [19:55] <JEEB> (and libx264's default) [19:56] <JEEB> okk [19:56] <JEEB> you have a lol old ffmpeg [19:56] <mp4man> :-/ [19:56] <JEEB> switch c:v to vcodec [19:56] <JEEB> also, oh man [19:56] <JEEB> not even 0.8 [19:56] <JEEB> can you build a newer one, together with a newer x264? [19:56] <mp4man> hum [19:57] <mp4man> I think I got this one one the official website [19:57] <mp4man> for mac [19:57] <mp4man> ah no [19:57] <mp4man> from macports [19:58] <JEEB> would recommend getting a newer one generally to make sure you have the easiest possible access to libx264 functionality from ffmpeg [19:58] <mp4man> oki [19:59] <mp4man> I'll try... But if I remeber well, I used macport because I couldnt install the latest version for some reason [20:01] <mp4man> What do you recommand : http://ffmpeg.org/download.html for mac os 10.6.8 [20:01] <JEEB> just grab the current git HEAD from the git repository? [20:02] <mp4man> Dont know what git is [20:02] <JEEB> it's a revision control software [20:03] <cbreak> git is great [20:03] <cbreak> I had some problems compiling HEAD on OS X a week ago or so [20:03] <JEEB> git clone git://source.ffmpeg.org/ffmpeg.git would get the current source code of the given git repository into a folder called 'ffmpeg' [20:03] <mp4man> I have to get that to get ffmpeg? [20:03] <cbreak> but v0.10.2 worked fine [20:03] <cbreak> if you have XCode 4.2 or so then you have git [20:03] <cbreak> I think it comes with xcode from 4.0 on [20:03] <cbreak> and if you want to compile ffmpeg, you need xcode anyway :) [20:04] <cbreak> (unless you want to bootstrap your own gcc/clang which is a pain) [20:05] <mp4man> :-/ [20:07] <mp4man> ok I'm gitting [20:07] <JEEB> also you'll probably have to grab the current libx264 [20:07] <mp4man> ah [20:07] <JEEB> git clone git://git.videolan.org/x264.git [20:07] <mp4man> JEEB: your my hero [20:08] <JEEB> also, do you have yasm installed? [20:08] <mp4man> yasm? [20:08] <cbreak> it's an assembler [20:08] <mp4man> seems so [20:08] <JEEB> check if you get anything when you write yasm --version onto the command line? [20:08] <cbreak> it does not come with OS X, you don't have it unless you installed it yourself [20:08] <JEEB> and if yes, which version do you have? [20:08] <mp4man> yasm 1.1.0.2352 [20:08] <JEEB> oh jolly good [20:08] <mp4man> juil. 0211 [20:08] <JEEB> you don't have to install that then :D [20:08] <mp4man> oh yeah:) [20:09] <JEEB> so yeah, wait for ffmpeg's sources to come around, then grab x264, and begin from x264 [20:09] <mp4man> x264 is being downloaded [20:09] <mp4man> done [20:09] <JEEB> great [20:10] <mp4man> now comes the compilation? :-( [20:10] <JEEB> I guess you will want to install this stuff into your home dir [20:10] <mp4man> hum [20:10] <JEEB> ./configure --prefix=/home/Username/ownapps --enable-static or something [20:10] <mp4man> static? [20:10] <JEEB> yes [20:10] <JEEB> easier to deal with to be honest [20:11] <cbreak> unless you want lgpl closed source apps :) [20:11] <cbreak> (which you won't get anyway with x264) [20:11] <cbreak> what's the reason to install it into your ~/ ? [20:12] <mp4man> I'm doing it in my /Applications/bin/x264 [20:12] <mp4man> make in progress [20:12] <cbreak> that's even weirder [20:13] <cbreak> you won't even have write rights there unless you're admin [20:13] <mp4man> ok make install done [20:13] <cbreak> and it's not in $PATH either :) [20:13] <mp4man> cbreak: it is [20:14] <mp4man> JEEB: is there something in particular to do to indicate ffmpeg where is x264? [20:17] <mp4man> make in progress ... [20:17] <JEEB> --extra-cflags="-I/where/you/installed/include" and --extra-ldflags="-L/where/you/installed/lib" [20:17] <JEEB> (and you need to --enable-libx264 when configuring ffmpeg) [20:17] <mp4man> arf [20:17] <mp4man> ok [20:18] <Yeti218> is it your work that you know so much about ffmpeg? [20:18] <JEEB> I've not had a job for three years [20:18] <JEEB> I'm a jobless student that has less than 10 euros on his bank account :P [20:19] <mp4man> argh [20:19] <mp4man> so JEEB , is this ok : ./configure --extra-cflags="/Applications/bin/x264/include" --extra-ldflags="-L/Applications/bin/x264/lib" --enable-libx264 [20:20] <JEEB> uhhh [20:20] <JEEB> where did the -I go ? [20:20] <JEEB> from cflags [20:20] <mp4man> oh yes [20:21] <Yeti218> with so much knowledge about it stuff and video encoding in particular it should be doable to get a job (if you actually want one ofc) [20:21] <mp4man> libx264 is gpl and --enable-gpl is not specified. [20:21] <JEEB> yes, you need to specify that of course [20:21] <JEEB> Yeti218, I would welcome a job :P [20:21] <mp4man> of course :) [20:21] <mp4man> what is your domain? [20:21] <JEEB> ? [20:22] <mp4man> area of expertise [20:22] <JEEB> oh [20:23] <mp4man> :) [20:23] <mp4man> compiling... lots of warnings btw [20:23] <JEEB> x264 usage, C, Java, Python, video encoding generally (IVTC/deint, avisynth usage etc.), Finnish/Russian native, can speak English freely and can write/read Japanese very well (used to speak it freely as well, but haven't lived there for a couple of years now). [20:23] <JEEB> warnings are OK [20:24] <mp4man> ok :) [20:24] <mp4man> Iam a scientist [20:25] <JEEB> I should really write a resume on the internets [20:25] <Mavrik_> JEEB: linkedin profile? :) [20:25] <mp4man> r u looking for in finland? [20:26] <JEEB> Never got to registering there, to be honest > LinkedIn [20:27] <mp4man> (still compiling....) [20:27] <Mavrik_> at least around here, that's the place where people go look for your CV (if they're not a huge old enterprise) :) [20:27] <JEEB> ah [20:27] <JEEB> oh well [20:27] <JEEB> will have to register after I finish these university projects [20:27] <mp4man> finished [20:28] <JEEB> great [20:29] <mp4man> shit I forgot to put the prefi [20:29] <mp4man> x [20:29] <mp4man> can I just move it from /usr/local/bin to where I want? [20:29] <JEEB> it should be static by default so yeah [20:31] <mp4man> ok [20:31] <mp4man> seems to be operational [20:32] <mp4man> so : ffmpeg -r 24 -i test%06d.png -c:v libx264 -crf 18 out.mp4 again ? [20:32] <JEEB> yes, that should create 4:2:0 YCbCr ('YUV') with the BT.601 matrix [20:32] <mp4man> here is what I got : http://pastebin.com/nHJZwfCy [20:32] <JEEB> also, I would generally recommend to make sure that whatever you are previewing your output with uses the correct matrix :) [20:32] <mp4man> :-/ [20:33] <JEEB> yes, not divisible by two [20:33] <mp4man> is there a way to bypass this? [20:33] <JEEB> no [20:33] <mp4man> :-/ [20:33] <JEEB> you'll just have to resize [20:34] <mp4man> hum [20:34] <mp4man> ok [20:34] <JEEB> 4:2:0 means that the Chroma (Cb,Cr) planes of the video have one value per 2x2 block [20:34] <mp4man> would you know the argument of convert that allows me to resize my images? [20:34] <mp4man> otherwise I'll just google it dont woryr [20:35] <mp4man> -resize apparently [20:36] <JEEB> also, just a quick question -- with how wide of a variety of players etc. are you planning to play this thing back? [20:36] <mp4man> well I don't really know [20:37] <mp4man> I use VLC, quicktime [20:37] <mp4man> on macs mostly [20:37] <JEEB> ok, that rules out RGB in 4:4:4 :) [20:37] <mp4man> ah? [20:37] <JEEB> (QuickTime) [20:37] <JEEB> so yeah, 4:2:0 YCrCb it shall be [20:37] <mp4man> ok [20:37] <JEEB> which is what you are creating atm [20:37] <mp4man> that's a bit chineese to me [20:39] <mp4man> ok... resizing the 2000 PNGs... that's gonna take a while [20:39] <JEEB> 4:2:0 is a weird way of saying "Luma plane has one value for every pixel, and chroma planes have one value for a 2x2 block" [20:39] <JEEB> and YCbCr is "YUV" [20:40] <JEEB> 4:2:0 YCbCr is the format 90%+ of all video is [20:40] <mp4man> and its going to preserve my colors? [20:41] <JEEB> if you use the same conversion matrix from RGB to YCbCr and back [20:41] <JEEB> ffmpeg uses the BT.601/bt470bg one for conversions [20:42] <mp4man> ok colors are still a bit different [20:42] <JEEB> probably due to a conversion mismatch [20:42] <JEEB> try checking in quicktime? [20:42] <mp4man> and the white is black [20:42] <mp4man> that's in quicktime [20:43] <mp4man> VLC plays like one frame and then black screen [20:43] <JEEB> lolwut [20:43] <JEEB> how old is your VLC? [20:43] <JEEB> anyways, let's try setting the color matrix in the stream [20:43] <mp4man> http://pastebin.com/sGR29pBx [20:43] <mp4man> Version 1.1.12 The Luggage (Intel 64bit) [20:43] <JEEB> that's old [20:44] <mp4man> argh [20:44] <JEEB> go get the current one [20:44] <mp4man> lol Im running [20:44] <JEEB> at least you won't have to compile it :P [20:44] <JEEB> -x264opts colormatrix=bt470bg [20:44] <JEEB> try re-encoding with this [20:45] <mp4man> lol dmg this time [20:46] <mp4man> http://pastebin.com/xzJMThe2 [20:47] <mp4man> Version 2.0.1 Twoflower (Intel 64bit [20:47] <JEEB> mp4man, set settings for x264 AFTER -i [20:47] <JEEB> before -i = decoder settings [20:47] <JEEB> after -i = encoder settings [20:48] <mp4man> ah ok sorry [20:48] <mp4man> ok [20:48] <mp4man> it's better I have all the frames [20:49] <mp4man> and the orange seems nice [20:49] <JEEB> anyways, with colormatrix set correctly you should have the players that don't fail showing it right [20:49] <mp4man> although the white is still black [20:49] <JEEB> is the "white" actually see-through? [20:50] <JEEB> because that one png file you linked was such [20:50] <mp4man> ah yes [20:50] <JEEB> there might be an ffmpeg setting to make see-through portions white [20:52] <mp4man> ok [21:00] <mp4man> JEEB: is it in the --help doc you think? [21:00] <JEEB> no idea [21:01] <mp4man> oki ;) [21:09] <mp4man> ok -alpha off in convert [21:14] <mp4man> JEEB: thanks so much [21:14] <JEEB> np [21:14] <mp4man> it works now [21:15] <mp4man> my colors are nice [21:15] <JEEB> yes, you just had to specify which color matrix you were using for the encode :) [21:16] <JEEB> (ffmpeg uses BT.601 aka bt470bg) [21:16] <Mavrik> JEEB: what difference does that make - does that mean it changes the encoder or is it just a flag in the stream? [21:17] <mp4man> JEEB: Im affraid I dont really understand what it means [21:17] <JEEB> Mavrik, just the flag in the stream [21:17] <JEEB> ffmpeg uses BT.601 so we set a matching flag [21:17] <JEEB> (for RGB<->YCbCr) [21:18] <Mavrik> ah, so that probably means the decoder has to check that flag and be able to use the correspondent reverse transform [21:18] <Mavrik> if it doesn't your colors are still crap... right? [21:18] <JEEB> unless the player guesses by your resolution :) [21:18] <Mavrik> are the matrices resolution-dependent? [21:18] <JEEB> nope [21:19] <JEEB> but usually for SD BT.601 is used [21:19] <JEEB> and for HD, BT.709 [21:19] <Mavrik> ah... well since I almost always got bad colors when doing RGB->YUV with ffmpeg, I guess the players suck at guessing :P [21:19] <JEEB> some players don't guess at all and just use BT.709 or BT.601 [21:20] <Mavrik> but it seems you just solved one of my major problems lately, thanks :) [21:20] <mp4man> ok so JEEB you said my videos wont work with quicktime right? [21:20] <JEEB> no, these should work just fine [21:21] <mp4man> hum [21:21] <JEEB> since they are 4:2:0 YCbCr [21:21] <mp4man> its all blakc [21:21] <JEEB> splendid [21:21] <JEEB> you're welcome to ask why it happens from Apple :P [21:22] <mp4man> argh [21:22] <Mavrik> :D [21:23] <mp4man> colors are not even the same on my laptop with VLC [21:23] <Mavrik> are your screens calibrated even? :) [21:23] <JEEB> your stream is fine [21:23] <JEEB> every discrepancy now would be because of derping players [21:23] <JEEB> and naturally different screens will have different colors [21:24] <JEEB> (unless you calibrate) [21:25] <mp4man> ok its the same colors as the png on my laptop [21:25] <mp4man> so I guess its the screen [21:25] <Mavrik> yeah, go find nearest graphic designer and annoy the hell out of him until he lends you his calibrator or few days :) [21:25] <Mavrik> (that's what I did on my Workstation which had two different screens with different colors) [21:26] <mp4man> ok [21:26] <Yuhjn> hey guys. I've been asked to develop an app using adobe AIR and does some minor video editing. I'm using ffmpeg to do the heavy lifting and I've got everything working on windows. My problem is mac... the version of ffmpeg.dmg for mac doesnt have the non-free codec used to generate .mov files, and the experimental one fails. I'm trying to compile the mac version from source using the right [21:26] <Yuhjn> libraries but still get failures [21:26] <mp4man> as long as the pngs look the same [21:26] <Yuhjn> wondering if someone who's farmilliar with compiling and the enable options can give me a little help plz [21:26] <mp4man> thats fine by me [21:27] <JEEB> Yuhjn, you can't give out nonfree versions of ffmpeg :) [21:27] <JEEB> they can't be distributed [21:28] <mp4man> JEEB: I wonder, is there a way to do the very same movie from PS files instead ot PNGs? [21:28] <Yuhjn> they can be licensed i assume [21:28] <JEEB> no [21:28] <mp4man> my ps files have a much better quality [21:28] <JEEB> mp4man, most probably not -- but I don't know every format ffmpeg supports :) [21:28] <Yuhjn> ok well if that's the case then how does the guy who posts the windows builds get away with it? [21:29] <Yuhjn> i took a windows ffmpeg off the shelf and it works fine [21:29] <Yuhjn> im not even sure i need the nonfree codec to be honest, kind of a newbie with video [21:29] <Mavrik> Yuhjn: he doesn't have non-free codecs in [21:30] <Mavrik> only GPL ones [21:30] <Yuhjn> ok then, so my origional question was wrong [21:30] <Yuhjn> I'm back to "why does it work on win and not on mac" [21:30] <JEEB> for AAC there's the GPLv3 compatible libvo-aacenc, and in newer revisions even the internal experimental aac encoder should work [21:32] <Mavrik> Yuhjn: for that, I suggest you do what ChanServ said - give us a pastebin of your command line AND the output error you get. [21:32] <Yuhjn> when i run the same thing on mac i get "codec is experimental but experimental codes are not enabled, see -strict -2... then when i try to use that flag I get "too many bits per frame requested [21:32] <Yuhjn> ok I'll do that... one sec [21:33] <JEEB> sounds like something old [21:33] <Mavrik> sounds like you're using integrated AAC codec instead of libfaac/libvo_aacenc [21:33] <JEEB> but we'll see with the pastebin how old ffmpeg that is [21:33] <Mavrik> (which is horribly broken) [21:33] <JEEB> it's no longer broken nowadays [21:33] <JEEB> just not really good [21:33] <Yuhjn> march 17th [21:33] <Yuhjn> 2012 [21:34] <JEEB> -strict experimental was it ? [21:34] <Mavrik> JEEB: really? I tested it like three or four months ago and it created completely broken AAC output :\ [21:34] <JEEB> the last time I had completely broken output was in january 2011 or so [21:34] <Yuhjn> -strict -2 [21:34] <JEEB> no [21:35] <JEEB> it's supposed to be -strict experimental [21:35] <Yuhjn> oh ok let me try that [21:35] <Yuhjn> the error says -strict -2 but let me try [21:35] <Yuhjn> same thign, too many bits per frame requested [21:35] <JEEB> also remember that settings for encoder(s) go AFTER -i [21:36] <JEEB> anyways, where's that pastebin :P [21:36] <JEEB> we're not espers [21:36] <Yuhjn> yes one sec.. hehe [21:36] <Yuhjn> btw thx for the help guys, appreciated [21:38] <mp4man> JEEB: http://www.sendspace.com/file/lr637t thx to you [21:38] <Yuhjn> http://pastebin.com/A6inCSYk [21:38] <Yeti218> i created a video in portrait mode but every player i watch it with on my HTC Desire shows it in landscape mode .. any suggestion why that happens? [21:38] <Yeti218> on my galaxy Nexus i can watch it in whatever orientation i have the phone [21:39] <Mavrik> um... I think desire plays all videos in landscape mode [21:41] <Yuhjn> I can pastebin the successful windows execution too if that would help... all works great on win [21:41] <Yuhjn> without -strict [21:42] <Yeti218> Mavrik, it seems that way, but isnt that kinda silly? When the video resolution is clearly portrait? [21:43] <Mavrik> yeah, probably :P [21:43] <Yeti218> the video gets shown much smaller than is strictly necesssary [21:43] <Mavrik> I don't watch many videos on my Desire (and I have it hacked to oblivion), but I do remember them playing in landscape mode always [21:43] <Mavrik> with the exception of Youtube [21:43] <Yeti218> only thing i can think of is to crop the view but then i would lose quite some information [21:45] <Yuhjn> incase you missed it: :) http://pastebin.com/A6inCSYk [21:47] <Mavrik> Yuhjn: it's quite possible the AAC encoder can't handle mono audio - try "-ac 2" and see if turning it to stereo works [21:47] <Mavrik> Yuhjn: or you can recompile ffmpeg with libvo_aacenc like the windows builds are [21:47] <Yuhjn> oh shit i think that worked [21:48] <Yuhjn> but woooo it's slow on mac [21:48] <Yuhjn> still working [21:49] <mp4man> ok that's unfortunate it doesn't work in quicktime :-/ [21:49] <mp4man> 'cause I think that's what keynote is using :-/ [21:50] <Yuhjn> the resulting .mov is not working right... framerate is really slow and it's acting very funny [21:52] <Yuhjn> hrm... [21:52] <Yuhjn> one of my avi's works but the resulting mov is all screwed up, and the other avi gives me too many bits per frame again [21:52] <Yuhjn> ok so plan B, try to recompile with the same enable flags as windows i guess [21:53] <Yuhjn> this is frustrating :) [21:54] <Yuhjn> i really really appreciate the help though guys, dont know what I'd do without it [21:54] <Yuhjn> probably not get paid is my guess :) [21:54] <Yuhjn> fail [21:57] <mp4man> seems that no matter what I do, any video wont play in Quicktime [21:58] <mp4man> :-/ [21:58] <mp4man> with the new version of ffmpeg [21:58] <Yuhjn> ok so I'm trying to configure the compile... I'm using the exact same flags that my current mac build has, plus the --enable-libvo-aacenc .... the configure gives me libmp3lame not found.... [21:59] <Yuhjn> if it's not found, that means it's not in the code, which means it's not a free lib? [22:00] <mp4man> anyone already had this issue with ffmpeg and quicktime? [22:16] <gebbione> hi I am trying to add audio and video together but i cannot get ffmpeg to do it ... it gives me several errors [22:17] <gebbione> ffmpeg -i prelisten.mp3 -i Mamasutra2.mpeg Mamasutra2_audio.mpeg -vcodec mpeg2video -r 25 -ar 22050 -ab 46k [22:18] <gebbione> but i get lots of errors [22:18] <Yuhjn> you're listing two mpeg's in a row? dont think you can do that... think you have to use concat: protocol (newbie myself so not 100%) [22:19] <gebbione> i think my problem is more to do with the video i have [22:19] <gebbione> it gives me errors while creating the file [22:19] <gebbione> like mpeg2video @ 0x23a6600] end mismatch left=72 108E22 [22:20] <gebbione> and i have no clue on how to get it to work [22:22] <mp4man> ok, adding :-pix_fmt yuv420p in the command line makes it playable with quicktime [22:22] <mp4man> but then bad colors are back :-/ [22:23] <juanmabc> tm [22:25] <mp4man> JEEB: does it sound normal to you? [22:33] <mp4man> I've found this on the internet : "I don't think QuickTime supports H264 High 4:2:2. Use -pix_fmt yuv420p if you need a QuickTime compatible file." [22:37] <JEEB> I don't know why ffmpeg starts using something not 420p by default o_O [22:37] <JEEB> and yes, it only supports 420p [22:38] <JEEB> can you pastebin me what ffmpeg gives out when you encoded the last time? [22:40] <mp4man> http://pastebin.com/Y3keUcZQ this wont play in quicktime [22:41] <mp4man> but this : http://pastebin.com/r96AbFA3 will [22:41] <mp4man> although the colors are bad again [22:41] <Yuhjn> can ffmpeg fade audio in and out? [22:44] <JEEB> mp4man, lol [22:44] <JEEB> it selected 4:4:4 by default o_O [22:44] <mp4man> lol ? :) [22:44] <JEEB> I wonder if that's the new default [22:44] <mp4man> ah [22:45] <JEEB> (in case your source is closest to it) [22:45] <JEEB> libx264's default is 4:2:0 [22:45] <JEEB> I guess ffmpeg now just sees you have RGB as source [22:45] <JEEB> and thus encodes 4:4:4 by default? :D [22:46] <JEEB> also, the colors should be fine, you just lose chroma fidelity [22:46] <mp4man> so is there a way to make it work in quicktime and get my colors back good? [22:46] <JEEB> 4:4:4 -> chroma planes are one value = one sample [22:47] <JEEB> mp4man, your colors should be fine, if there are any problems that's because of the chroma subsampling [22:47] <JEEB> 4:2:2 -> chroma planes are one value = two samples on one line [22:47] <mp4man> should be fine with the pix_fmt yuv420p option? [22:47] <JEEB> 4:2:0 -> chroma planes are one value = a 2x2 area of samples [22:48] <JEEB> mp4man, I'm not sure if you understand what I mean with "fine" [22:48] <mp4man> Im not sure either [22:48] <JEEB> the colors should be as good as possible given the fact that you are converting to 4:2:0 YCbCr ('YUV') [22:48] <mp4man> ah [22:49] <mp4man> so no way I can have the same colors as I have in VLC without this pix_fmt yuv420p option? [22:49] <JEEB> yuv420p (4:2:0 YCbCr) is the only format most decoders take in :P [22:49] <JEEB> such as quicktime, and many others [22:50] <JEEB> flash surprisingly supports 4:2:2 [22:50] <JEEB> but not 4:4:4 [22:50] <mp4man> do you think about any other pixel format in the pix_fmts list that would be better and work in quicktime? [22:50] <JEEB> there are better pix_fmts, but they won't work in quicktime [22:51] <JEEB> quicktime (and 90%+ of all H.264 decoders) will want 4:2:0 YCbCr [22:51] <mp4man> only? [22:51] <JEEB> yes [22:51] <mp4man> I have seen this : http://developer.apple.com/library/mac/#qa/qa1501/_index.html [22:51] <mp4man> not very recent though since its for QT7 and I have 10 [22:51] <JEEB> yes, they support other colorspaces overall [22:51] <JEEB> but they don't support H.264's compression of other colorspaces [22:51] <JEEB> capisci? [22:52] <mp4man> lol not really but if you say its impossible [22:52] <JEEB> ok [22:52] <JEEB> H.264 has different profiles [22:52] <JEEB> different ways of compressing stuff [22:53] <mp4man> what I dont get is that... using the software "framed" I have a good .mov quality and color [22:53] <mp4man> but it requires jpgs, which I dont like [22:53] <mp4man> but the colors are nice and it is working in quicktime [22:54] <mp4man> so I don't really understand how that's posible given what you've just said [22:54] <JEEB> ffprobe working.mov [22:54] <JEEB> pastebin that [22:55] <mp4man> workin.mov being the movie created with framed? [22:55] <JEEB> also, how does the 4:2:0 video with the correct colormatrix look in VLC? [22:55] <JEEB> yes [22:55] <mp4man> let me answer the second question first [22:55] <mp4man> I have to look for a movie that were created with framed [22:56] <mp4man> the color does look better (brither) with VLC [22:56] <mp4man> but not as red as the PNG [22:56] <cbsrobot> whats framed ? [22:58] <cbsrobot> ah - image sequence to movie ... [22:59] <mp4man> http://i.imgur.com/IPYev.png [22:59] <mp4man> top to bottom : PNG, quictkime, VLC [22:59] <mp4man> video done with : ffmpeg -r 24 -y -i test%06d.png -pix_fmt yuvj420p -c:v libx264 -crf 18 -x264opts colormatrix=bt470bg out.mp4 [23:00] <mp4man> JEEB: do you see the difference? [23:01] <mp4man> the PNG is the most red, in quicktime the video is awfull, in VLC its better but not as red as the original PNG [23:01] <cbsrobot> mp4man: where did you get these 48bit png from ? [23:02] <mp4man> I created them from an IDL code [23:02] <mp4man> (IDL = Interactive Data Language, as far as I know) [23:02] <mp4man> http://idlastro.gsfc.nasa.gov/idl_html_help/WRITE_PNG.html [23:03] <mp4man> JEEB: regarding your other question, ffprobe working.mov does not work, it says : [mov,mp4,m4a,3gp,3g2,mj2 @ 0x102051000] moov atom not found jz040.mov: Operation not permitted [23:04] <cbsrobot> and why do you use colormatrix=bt470bg ? [23:04] <mp4man> ask JEEB :) [23:09] <JEEB> cbsrobot, because that's what ffmpeg uses [23:09] <JEEB> for RGB<->YCbCr conversions [23:09] <cbsrobot> I just read the backlog [23:10] <JEEB> <mp4man> the PNG is the most red, in quicktime the video is awfull, in VLC its better but not as red as the original PNG <- ok, so I'll just say that it's because of the chroma subsampling [23:10] <JEEB> of course, if you do mov instead of mp4 you could just stick PNGs into mov >_> [23:11] <JEEB> -c:v copy [23:11] <mp4man> ? [23:11] <mp4man> and it would be ok? [23:11] <JEEB> ffmpeg -r 24 -y -i test%06d.png -c:v copy out.mov [23:11] <JEEB> could end up huge tho [23:11] <JEEB> but it should work in Quicktime, I guess? [23:12] <mp4man> Im trying [23:12] <mp4man> you think that's what framed is doing? [23:12] <mp4man> because it gives very large files [23:12] <mp4man> like 500MB [23:12] <JEEB> if it needs jpeg I'm just guess it puts that into mov :P [23:12] <JEEB> *guessing [23:13] <mp4man> ok [23:13] <mp4man> 800MB [23:13] <mp4man> wow :) [23:13] <mp4man> color is still less red than the PNG [23:13] <mp4man> (in QT) [23:13] <mp4man> although much better than before [23:13] <mp4man> but 800MB wow [23:14] <JEEB> well, it's just your PNGs in a mov container with a frame rate :P [23:14] <mp4man> yeah [23:14] <JEEB> and if Quicktime changes the colors, it's its fault [23:14] <mp4man> I see [23:15] <mp4man> and so with the previous command [23:15] <cbsrobot> mp4man: can you shre one png ? [23:15] <mp4man> you said it was because ot the chroma subsampling [23:15] <cbsrobot> *share [23:15] <mp4man> is their a way to improve htat? [23:15] <JEEB> chroma subsampling and probably some QuickTime failed added to it as well [23:15] <JEEB> and no [23:15] <mp4man> cbsrobot: sure [23:16] <JEEB> chroma subsampling is chroma subsampling [23:16] <JEEB> Deal With It [23:16] <mp4man> :) [23:16] <mp4man> ok [23:16] <JEEB> all DVDs, all TV stations, all blu-ray have this [23:16] <mp4man> so I guess for day to day movies I'll just do the mp4 thing [23:16] <mp4man> and for big occasions I'll do the mov then [23:17] <JEEB> pretty much all video we have around us that is YCbCr, is 4:2:0 [23:17] <mp4man> cbsrobot: http://i.imgur.com/yH3y8.jpg [23:17] <JEEB> I only happen to have something better than that because I got access to some masters via various ways [23:17] <mp4man> cbsrobot: actually sorry it was not created from IDL directly [23:18] <mp4man> cbsrobot: I created a postscript with IDL, then I used "convert" ot obtain the PNG from the PS [23:18] <HampshireGuy> good evening everyonw [23:18] <cbsrobot> post ps or png then [23:18] <HampshireGuy> everyone even :) [23:18] <cbsrobot> hi [23:18] <mp4man> cbsrobot: sorry? [23:18] <cbsrobot> post a postscript or a png file then [23:19] <mp4man> here is the png : http://i.imgur.com/yH3y8.jpg [23:19] <mp4man> ah sorry [23:19] <cbsrobot> jpg ? [23:19] <mp4man> stipid website [23:19] <cbsrobot> lol [23:19] <mp4man> it converts the file [23:19] <HampshireGuy> does anyone have a moment to look a command line I have put together that isn't working for me at the moment? [23:19] <cbsrobot> most of them are [23:19] <mp4man> do you know how I could send it to you? [23:19] <mp4man> without converting it [23:20] <VooDooNOFX_> MP4_maniac, imageshack won't convert it. [23:20] <HampshireGuy> I'm trying to build a standard command line to convert files into the right format for my daughters VTech Innotab :) [23:20] <mp4man> cbsrobot: here you go : http://www.sendspace.com/filegroup/%2FWsKO8uqNCkTcXM38tglyQ [23:20] <mp4man> cbsrobot: ps and png together [23:20] <cbsrobot> page not found [23:20] <mp4man> damn [23:21] <cbsrobot> I take the jpg [23:21] <mp4man> http://www.sendspace.com/file/q17fqy [23:21] <mp4man> png [23:21] <cbsrobot> you had the same problem with jpg aswell [23:21] <mp4man> http://www.sendspace.com/file/ggqzyt ps [23:21] <cbsrobot> thanks [23:22] <mp4man> he ps was converted to png with : convert -quality 100 -alpha off -resize $1x$2 $f ${f%.*}.png [23:29] <Mista_D> can't I use /dev/urandom any more to encode noise? `ffmpeg -ar 48000 -acodec pcm_s16le -i /dev/urandom -ab 128k -f mp2 -t 10 noise.mp2` [23:30] <Tjoppen> try -f s16le [23:31] <Mista_D> Tjoppen: thanks! it works. [23:32] <mp4man> cbsrobot: why did u wnat the files btw? [23:39] <cbsrobot> ffmpeg -loop 1 -t 2 -i ~/Downloads/test000000.png -c:v libx264 -x264opts colormatrix=bt470bg -pix_fmt yuv420p Desktop/test.mp4 [23:39] <cbsrobot> gave me :http://bayimg.com/dAOeiaADE [23:39] <cbsrobot> doesnt look that bad - does it ? [23:40] <cbsrobot> left being png in preview, right mp4 in qt [23:40] <mp4man> cbsrobot: sorry my employer's admin blabla does not allow me to go on your website [23:40] <mp4man> could you upload it to sendspace? [23:40] <cbsrobot> hahaha [23:41] <cbsrobot> tell him its mission critical [23:41] <cbsrobot> hehe [23:41] <cbsrobot> http://imgur.com/BaGan [23:41] <cbsrobot> and dont spend the rest of your day there ... [23:41] <mp4man> what's on the right and left? [23:41] <cbsrobot> left being png in preview, right mp4 in qt [23:42] <cbsrobot> a bit of a loss in the reds ... [23:42] <mp4man> that's funny my png look more red on my screen [23:43] <mp4man> although not on my laptop [23:43] <mp4man> ok it appears I'm fucking flies, as we say in french :) [23:43] <mp4man> flyes [23:43] <mp4man> flies [23:44] <mp4man> enculer les mouches [23:45] <mp4man> cbsrobot: and if the red are less pronounced, the images look also a bit darker [23:45] <mp4man> dont you think? [23:46] <mp4man> (flies again) [23:47] <cbsrobot> encule ce qui tu veuxmais pas une mouche stp [23:47] <cbsrobot> -ce [23:47] <cbsrobot> hehe [23:48] <mp4man> ahah [23:48] <mp4man> oue c degueu [23:49] <Yuhjn> is there another audio code i can use other than AAC build in? [23:49] <Yuhjn> codec [23:51] <Yuhjn> I'm really confused... I have an .avi file and I can extract the audio from it into mp3, but then when I try to put that mp3 into an mov file I get that experimental codec -strict -2 problem [23:51] <Yuhjn> how can it get the audio out, but then not be able to write it back in? [23:52] <mp4man> ok guyes [23:52] <mp4man> guys [23:52] <mp4man> thks a loot for your help [23:52] <cbsrobot> np [23:52] <mp4man> cbsrobot: and JEEB, I understand things better now [23:52] <mp4man> time for me to go home ;) [23:54] <mp4man> cheers [00:00] --- Fri Apr 27 2012

1 0

ffmpeg-devel.log.20120425
by burek 26 Apr '12

26 Apr '12

[00:13] <CIA-17> ffmpeg: 03Yusuke Nakamura 07master * r462a5b7839 10ffmpeg/libavformat/isom.c: [00:13] <CIA-17> ffmpeg: isom: Support more DTS codec identifiers. [00:13] <CIA-17> ffmpeg: DTS LBR identifier ('dtse') is not included since libavcodec doesn't support it yet. [00:13] <CIA-17> ffmpeg: Signed-off-by: Derek Buitenhuis <derek.buitenhuis(a)gmail.com> [00:13] <CIA-17> ffmpeg: 03Anton Khirnov 07master * r8c4022aceb 10ffmpeg/avconv.c: avconv: fix a segfault on -c copy with -filter_complex. [00:13] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r8f89954550 10ffmpeg/: [00:13] <CIA-17> ffmpeg: Merge remote-tracking branch 'qatar/master' [00:13] <CIA-17> ffmpeg: * qatar/master: [00:13] <CIA-17> ffmpeg: avconv: fix a segfault on -c copy with -filter_complex. [00:13] <CIA-17> ffmpeg: isom: Support more DTS codec identifiers. [00:13] <CIA-17> ffmpeg: Merged-by: Michael Niedermayer <michaelni(a)gmx.at> [03:20] <Compn> http://forums.anandtech.com/showthread.php?t=2240545 2tb $110 [04:16] <Zeranoe> I'm trying to come up with a patch to allow FFmpeg to install without .lib files. Currently the make install will fail if a shared build can't make the .lib files. Can anyone help me out on where I should be looking to fix this? [04:38] <ohsix> aren't the import library and the dll's generated at the same time [04:39] <Zeranoe> .lib files are generated from lib.exe, and when cross compiling wine needs to be installed. make install should not fail to install FFmpeg if the .lib files arent built [04:41] <ohsix> if they aren't import libraries then what are you "install"ing [04:43] <Zeranoe> ohsix: They are import libs [12:18] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r16524d1be0 10ffmpeg/libavcodec/ffv1.c: [12:18] <CIA-17> ffmpeg: ffv1: include solution hint in the slice error message [12:18] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [12:18] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r0c27854532 10ffmpeg/libavcodec/ffv1.c: [12:18] <CIA-17> ffmpeg: ffv1: Always store slice size when error reobustness is enabled. [12:18] <CIA-17> ffmpeg: This simplifies finding slices within a damaged bitstream. [12:18] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [12:18] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rc24ca7b88e 10ffmpeg/libavcodec/ffv1.c: [12:18] <CIA-17> ffmpeg: ffv1: add 1 status byte to slices in in case crcs are stored too. [12:18] <CIA-17> ffmpeg: This will allow storing of information about corrected and uncorrectable [12:18] <CIA-17> ffmpeg: errors. [12:18] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [15:14] <ubitux> http://mattrichardson.com/Descriptive-Camera/ we need to write a video filter to do this ;) [15:15] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r3ed6917ab8 10ffmpeg/libavcodec/ffv1.c: [15:15] <CIA-17> ffmpeg: ffv1: store slice size with all slices in 1.3 [15:15] <CIA-17> ffmpeg: This simplifies handling by removing a special case. [15:15] <CIA-17> ffmpeg: Its also needed to make the next change possible. [15:15] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [15:15] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r60217b5b9c 10ffmpeg/libavcodec/ffv1.c: [15:15] <CIA-17> ffmpeg: ffv1: Remove slice count field with 1.3 and just count slices. [15:15] <CIA-17> ffmpeg: This field was problematic because in case of damaged slices it can be [15:15] <CIA-17> ffmpeg: lost [15:15] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [15:16] <iive> i don't understand the description of the above change. [15:16] <iive> isn't slice count like mpeg2, aka indicating the number of row? [16:02] <michaelni> iive, slice count simply was the count of slices like 4 if there are 4 slices in a frame [16:03] <iive> is there something indicating where the slice starts (in the picture)? [16:05] <michaelni> yes each slice stores its x and y position as well as its size [16:29] <ubitux> does anyone knows if it makes sense to have in a mov/mp4 the first entry of the stss (sync samples table) a value of 1? [16:29] <ubitux> basically, if the first frames can be non-key frames [16:30] <ubitux> (if the value is 3, the first two frames won't be key frames) [16:30] <ubitux> a value *different* than 1 sorry [16:32] <ubitux> it seems that ffplay doesn't display anything until the first keyframes [16:32] <ubitux> while mplayer is trying to display it anyway [16:32] <ubitux> (this reminds me of the encode failure issue) [16:45] <iive> ubitux: in theory you can have P frame that consists only of intra-blocks. but it may be something different. [16:45] <iive> there is h.264 mode where no key frames are inserted. A value is given that ensures that after decoding X number of frames, there will be no missing information in the decoded picture. [16:46] <iive> aka, the partial updates would refresh everything. [16:48] <ubitux> the samples tables don't give much hints about the kind of frames/samples [16:49] <ubitux> i was testing the nginx mp4 streaming plugin [16:49] <ubitux> which basically remux at a given time the video [16:50] <ubitux> looking at the output given, it seems to split without caring much about the key frames [16:50] <ubitux> the file looks correct, but the sync samples don't start at 1 (first frame is a sync/key) [16:51] <MP4_maniac> I [16:51] <ubitux> and the standard doesn't seem to be explicit about this [16:54] <MP4_maniac> if the file is splitted, stss indexes of the file might not start at 1 [16:54] <ubitux> yes that's exactly what seems to happen [16:54] <ubitux> but then, the players behave quite differently [17:01] <MP4_maniac> h264 in mp4 with gradual decoder refresh should be specified with roll recovery grouping which indicates how many frames to be needed to decode correctly. if the file has that feature, demuxer should follow. [17:05] <ubitux> sample grouping would solve the issue? mmh.. [17:06] <MP4_maniac> i'm writing the patch for that feature https://github.com/VFR-maniac/libav/commit/bc6ef68208337d87cb52d53267103645… [17:07] <MP4_maniac> i need someone's help for this [17:08] <ubitux> interesting [17:09] <MP4_maniac> this patch is for only demuxing [17:09] <ubitux> what is this "gradual decoder refresh" all about? [17:11] <MP4_maniac> about h264, recovery point sei [17:15] <ubitux> ok, thx [17:15] <MP4_maniac> note: for h264 open-gop (recovery_frame_cnt == 0), use random access point grouping instead of this [17:20] <MP4_maniac> gpac already implements roll recovery grouping for their muxer, however they misunderstand the definition of recovery_frame_cnt and/or frame_num :P [17:42] <mkoch> any VDA experts around? [17:58] <michaelni> mkoch, try to send the author a mail maybe if you still have some problems ... [17:58] <michaelni> waiting here is maying not that practocal [17:58] <michaelni> dunno [17:58] <mkoch> michaelni: ok, I will try, thanks [18:24] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r3f4c0dad16 10ffmpeg/libavcodec/ffv1.c: [18:24] <CIA-17> ffmpeg: ffv1: move init_slice_state() into the decoder threads. [18:24] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [18:24] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * rcab15f9db4 10ffmpeg/libavcodec/ffv1.c: [18:24] <CIA-17> ffmpeg: ffv1: remove clear_state(), it has become unused [18:24] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [18:24] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r68a9194bb6 10ffmpeg/libavcodec/ffv1.c: [18:24] <CIA-17> ffmpeg: ffv1: indent [18:24] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [18:24] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r83bf9fb6ec 10ffmpeg/libavcodec/ffv1.c: [18:24] <CIA-17> ffmpeg: ffv1: move clear_state() to clear_slice_state() in encoding threads [18:24] <CIA-17> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni(a)gmx.at> [21:39] <j-b> I can HAZ eac3 in iso? [21:40] <s3r[]> what does "pts" stand for? [21:40] <s3r[]> in the source [21:40] <nevcairiel> presentation time stamp [21:40] <nevcairiel> usually, anyway [21:41] <s3r[]> thanks [21:43] <s3r[]> I've been slowly going through it for 2 days now [21:46] <s3r[]> So I'm reading through the developer documentation, and for any changes/bug fixes, I should state/post them here and let it be reviewed? [21:46] <s3r[]> unless the ffmpeg-devel mentioned in the docs is a mailing list and not the irc [21:47] <JEEB> for patches and such, ffmpeg-devel the mailing list indeed [21:47] <JEEB> for bugs, there's the trac issue tracker [21:54] <jdstrand> hi, someone reported a security issue against libav in Ubuntu, and it looks like it also affects ffmpeg. what is the email address for the security contact for ffmpeg? [22:09] <iive> michaelni ? [22:11] <Compn> jdstrand : you can send security mail to Michael Niedermayer <michaelni(a)gmx.at> [22:11] <Compn> since i dont know of any security mailing list, there is ffmpeg-legal but who knows what goes on there :D [22:12] <Compn> jdstrand : or if you prefer, you can make bugs on http://ffmpeg.org/trac/ffmpeg [22:12] <Compn> our bug tracker [22:12] <jdstrand> Compn: thanks. I am trying to coordinate this among various people, so will stick with the email address [22:12] <jdstrand> (for now) [23:35] <CIA-17> ffmpeg: 03Justin Ruggles 07master * rc8af852b97 10ffmpeg/ (27 files in 5 dirs): [23:35] <CIA-17> ffmpeg: Add libavresample [23:35] <CIA-17> ffmpeg: This is a new library for audio sample format, channel layout, and sample rate [23:35] <CIA-17> ffmpeg: conversion. [23:35] <CIA-17> ffmpeg: 03Justin Ruggles 07master * rc5671aeb77 10ffmpeg/tests/ (lavf-regression.sh ref/lavf/dv_fmt ref/seek/lavf_dv): [23:35] <CIA-17> ffmpeg: FATE: avoid channel mixing in lavf-dv_fmt [23:35] <CIA-17> ffmpeg: This partially reverts acb1730218f1c614dc8ca3ba45d9de1e05059515 [23:35] <CIA-17> ffmpeg: which would only have needed to change the checksums if channel mixing had [23:35] <CIA-17> ffmpeg: been properly avoided. This changes the output file size reference and the [23:35] <CIA-17> ffmpeg: seek test reference back to the previous values. [23:35] <CIA-17> ffmpeg: 03Justin Ruggles 07master * rbcb82fe1f4 10ffmpeg/ (avconv.c cmdutils.c): avconv: use libavresample [23:35] <CIA-17> ffmpeg: 03Michael Niedermayer 07master * r3ead79eaa3 10ffmpeg/: (log message trimmed) [23:35] <CIA-17> ffmpeg: Merge remote-tracking branch 'qatar/master' [23:35] <CIA-17> ffmpeg: * qatar/master: [00:00] --- Thu Apr 26 2012

1 0

ffmpeg.log.20120425
by burek 26 Apr '12

26 Apr '12

[02:13] <bencc> I want to let users publish live camera stream from both Flash and Html5 and to publish it live to both Flash and Html5 [02:13] <bencc> erlyvideo let me do this for Flash [02:13] <bencc> can I send the live stream to ffmpeg and transcode it, send it back to erlyvideo and send it to the users? [02:15] <phunyguy_> Is decoding 7.1 channel 24bit dts-hd ma on the horizon for ffmpeg and getting the raw data out forencoding to flac as 8ch 24bit? [04:40] <eli_> hello [04:45] <phunyguy_> Hai [04:46] <eli_> hi phunyguy :D [05:03] <gustavo_cm> Hello there! I find here http://blog.pcode.nl/2010/10/17/encoding-webm-using-ffmpeg/ a ffmpeg preset to libvpx. Does anyone know if it's possible to use a libvpx preset like libx264-max? Where could I find such information? [05:11] <gustavo_cm> Well, I found this: http://www.webmproject.org/tools/encoder-parameters/#10-sample-command-lines -- how to pass these parameters to ffmpeg? [09:33] <cbreak-work> what's the time base of the value I get with best_effort_timestamp? [09:34] <cbreak-work> when I call av_opt_ptr(avcodec_get_frame_class(), _frameVideo, "best_effort_timestamp"); [09:34] <cbreak-work> _frameVideo is an AVFrame* [09:38] <cbreak-work> for some videos I tried, the associated AVCodecContext has a working seeming time base [09:43] <cbreak-work> seems it's in the video stream [09:44] <cbreak-work> the one in the codec is not valid in this case. weird. [15:24] <andrew_> Is there a way to have ffmpeg output a non-seekable mp4 format? [15:25] <andrew_> Basically, I need to pipe the output of ffmpeg to a video player, and need to get ffmpeg to output it appropriately [15:26] <JEEB> I don't think you can make a completely non-seekable mp4 file, but there's the fragments feature in the container that lets one have some data without getting the full index [15:26] <JEEB> not sure if ffmpeg supports fragments [15:27] <andrew_> Ok, I did see that in the MP4 muxer [15:27] <andrew_> in the documentation [15:27] <JEEB> of course you'll also have to make sure the player side likes fragments [15:27] <andrew_> yea, thats annoying [15:27] <JEEB> otherwise it'll just put a trollface on you [15:28] <JEEB> *on a trollface for you [15:28] <andrew_> Lol, would you happen to know if Android's MediaPlayer class supports it? [15:29] <JEEB> no idea [15:29] <andrew_> ok, thought it would be worth asking [15:30] <andrew_> but, at least, ffplay should support it, right [15:30] <andrew_> but, at least, ffplay should support it, right? [15:31] <JEEB> everything that uses a recent enough libavformat/-codec should support it [15:32] <andrew_> ok, last compiled last week should be recent enough I imagine then [15:33] <JEEB> yeah [17:00] <Yeti218> Newbie question: i did the not option part of the oneiric ffmpeg install on ffmpeg.org [17:01] <Yeti218> however at the end the executables are not in my path, the hash commands fail and when i try to encode a file with h264 encoding it says Unknown encoder 'h264' [17:02] <Yeti218> Any idea why it doesn't recognize -vcodec h264? [17:05] <JEEB> because ffmpeg has no H.264 encoder of its own? [17:05] <JEEB> the only way to encode H.264 is to compile libx264, and then ffmpeg with it [17:11] <Yeti218> I understand, but i followed https://ffmpeg.org/trac/ffmpeg/wiki/UbuntuCompilationGuide [17:11] <Yeti218> it installs x264 [17:12] <Yeti218> I didnt get any errors, just zillions of warnings, but i am used to that when installing from source [17:14] <JEEB> well, then if you have libx264 support compiled into ffmpeg, then it's -vcodec (or -c:v) libx264 [17:14] <JEEB> not vcodec h264 [17:14] <JEEB> because h264 is only a decoder [17:15] <Yeti218> ah that could be it, let me check [17:16] <Yeti218> it seems to be working, it is running [17:17] <JEEB> now you'll probably want to learn about some things relative to x264 encoding [17:17] <JEEB> there's a -preset:v setting that can set the libx264 preset (which is basically the setting to set compression/speed) [17:17] <Yeti218> it worked! [17:17] <JEEB> http://mewiki.project357.com/wiki/X264_Settings#preset [17:18] <Yeti218> what i am interested in is .. converting videos into videos that are better suited for smartphones .. filesize-wise and resolution-wise [17:19] <Yeti218> i have decided to go for mp4 container and h264 encoding [17:20] <JEEB> also, the default is crf 23 if you don't set any rate control mode for the video codec. crf is the closest thing to "constant quality" we currently have, and it works so that higher numbers mean less bitrate used, while lower numbers mean more bitrate used (the idea is to use the highest number that still looks good for you) [17:20] <JEEB> (in other words, 1pass crf if you care about quality, 2pass bitrate-based encoding if you care more about file size) [17:20] <Yeti218> ok, so i should experiment with that number and try to get it as high as possible? [17:20] <JEEB> (the option for crf in ffmpeg is -crf btw) [17:21] <JEEB> and it goes up to 51 + then there are the emergency modes that also do denoise (which are used in cases when you have set bitrate limitations and otherwise a frame wouldn't fit the buffer) [17:21] <JEEB> (zero is lossless encoding, which won't work with hardware decoders) [17:21] <JEEB> also, for devices you usually have to set your -profile:v and -level [17:22] <JEEB> (Depending on what your device can handle) [17:22] <Yeti218> that is a lot of information for me to process as i am a bit new to ffmpeg [17:22] <Yeti218> i am converting videos for Android phones, but there is a large range of Android phones... [17:22] <JEEB> well, I just told you that crf goes from zero to 51, and that you will also have to set the H.264 profile and levle [17:22] <JEEB> *level [17:22] <JEEB> yes [17:23] <JEEB> very wide [17:23] <JEEB> f.ex. my first android phone could only take baseline profile + level 3 [17:23] <JEEB> my current phone can take up to high profile (default) level 4.1 or so [17:24] <JEEB> so basically you'll have to set your own limits depending on the handsets you want to support [17:24] <JEEB> baseline + level 3 tends to work at many places [17:26] <andrew_> Does Android support more than baseline profile H.264? I was under the impression that it didn't, but your's can? [17:26] <JEEB> it completely depends on the phone [17:26] <JEEB> most phones from middle of 2010+ usually can support more than baseline [17:27] <Yeti218> that is solid information.. as i only have 2 phones to test it on (galaxy Nexus and htc desire) and both are kinda new so i would not test it correctly without knowing this [17:27] <JEEB> (but of course there's a LOT of phones that don't, and are still on sale most probably) [17:27] <andrew_> so it all depends on the hardware? No software decoding support? [17:27] <JEEB> I don't think many phone makers make the CPU do the decoding [17:27] <Yeti218> i dont need super quality but i also dont want super big files .. if i manage to reach that goal for older phones i am happy [17:28] <JEEB> there might be a decoder in the Android source code, but I don't remember. Probably not, it just probably tells you to bolt the mediaplayer API into your hardware decoder's drivers [17:28] <JEEB> Yeti218, well if you support the older phones you'll be on a resolution of 320x180 or something [17:28] <Yeti218> gtg, will be back this evening after some experimenting [17:29] <JEEB> anyways, I'd recommend testing out with the default crf 23 and gradually going up or down from there depending on if it looks good or not [17:29] <Yeti218> i was planning to make the resolution big enough for newer models and playing it by default player .. which will know how to downscale it is that is needed [17:29] <JEEB> many phones will just herp a derp at bigger resolutions [17:29] <JEEB> I think my old phone did, too [17:30] <JEEB> also I think my old phone's decoder chip was flawed [17:30] <Yeti218> well that would explain something too [17:30] <Yeti218> :) [17:30] <JEEB> it either failed inloop deblocking, or weighted prediction [17:30] <JEEB> don't remember [17:30] <JEEB> never cared to test it for the exact problem [17:31] <Yeti218> gtg, i will start out with this knowledge and test it out .. (atm my videos are 35mb and i have to download them first - way too big if you ask me) [17:31] <Yeti218> my goal is somewhere < 5 mb if i can do it [17:32] <JEEB> depends on the length of the clip, frame rate, complexity of the content as well as the resolution (and of course the crf value) [17:33] <Yeti218> now i really have to go else my kids will stand alone at the street waiting .. [17:33] <Yeti218> thx for the info, you were really helpful [18:24] <juanmabc> is there some audio to text, ala youtube testing transliteration module, to make subtitles on the fly? [18:30] <burek> juanmabc, I don't think so, but you can always just extract the audio and pass it through the speech recognition software [18:57] <andrew_> Hi again. Im still trying to get android to read a piped m4 stream, but having issues [18:58] <andrew_> I tried fragmenting the mp4 using flags in the MP4 muxer documentation for ffmpeg, but Android starts playing them, then just sits there [18:58] <andrew_> Nothing reports any trouble, I just see a black screen [18:59] <andrew_> I don't know if the issue is that I did it wrong or if Android just doesn't support MP4 fragmented files [19:00] <andrew_> Does anyone have a suggestion of how to write have ffmpeg format the piped file in such a way that it it would be readable at the begining? [19:04] <taqattack> Okay. How come I can't use RTMP to stream videos with 1000k bitrate but as soon as I change it to 900k, it starts to stream? [19:20] <Yeti218> i am a bit overwhelmed by the 6500 lines of man file of ffmpeg ... [20:50] <Yeti218> Does any of you have settings i could use for compressing (too large) videos into a format that is supported by both Android and iPhone? [20:52] <JEEB> you need to decide which devices you are trying to support first of all [20:52] <JEEB> f.ex. for iphone there's the pre-3GS stuff (baseline only AFAIK), and 3GS and further that could take high profile, level 4.1 just fine [20:53] <JEEB> (just that itunes would not let you transfer the files over if you tried to do it via that way because of their dumb lockdowns) [20:53] <JEEB> with android there is no such "clear" line as there are over 9000 various devices [20:54] <JEEB> but there too, you will most probably want to decide if you want to start supporting the low-end as well, or only the high-end devices [20:54] <JEEB> think of profiles are "feature sets" for compression [20:57] <Yeti218> ok [20:57] <Yeti218> suppose i do not want to support any pre-3GS [20:58] <Yeti218> i have tried a few ffmpeg options and managed to downsize the file from 34mb to 12mb and i do not see much difference [20:58] <Yeti218> but i have a feeling i can get a lot lower than 12mb still [20:59] <JEEB> well, if you don't want to support pre-3GS then it's easy >_> the only problem would be stemming from the fact of how you are going to distribute those clips [20:59] <Yeti218> the only change i made to th source was: -vcodec libx264 -r 13 [20:59] <Yeti218> i have them all in a zip and my app downloads them to the phone and then unzips them [21:00] <JEEB> ok, so you won't be limited by itunes's crappy limitations [21:00] <Yeti218> no no, dear god no [21:00] <Mavrik-> 13 fps videos could look choppy [21:00] <Yeti218> the app will be in the itunes store, but the videos will be downloaded when installed [21:01] <JEEB> first of all, 3GS and latter can take up to 720p (was it?) in high profile, level 4.1, but you will probably want to resize to some comfy-for-you level first [21:01] <JEEB> frame rate limiting I would do only if absolutely needed [21:03] <Yeti218> atm i have several videos in the zip and i would like to reduce the size, but the videos should still be nice to look at [21:03] <Yeti218> i am unsure what size is realistic [21:05] <b_jonas> Yeti218: in that case, just try multiple bitrates, and see which one you like [21:06] <b_jonas> make sure you lower the resolution to the res of the phone displays [21:06] <JEEB> first make sure you can output the video size that you are aiming for (resize!), unless you already have that [21:06] <JEEB> after that, find your highest crf value that still looks good [21:06] <JEEB> after that, set the preset setting of libx264 to the highest you can bear speed-wise [21:07] <Yeti218> the source i have: Video: mpeg2video (Main), yuv420p, 720x576 [SAR 16:15 DAR 4:3], 8000 kb/s, 25 fps, 25 tbr, 90k tbn, 50 tbc [21:07] <Yeti218> Audio: mp2, 48000 Hz, stereo, s16, 224 kb/s [21:08] <JEEB> (note: crf value is not supposed to be exactly the same thing between presets as the compression algorithms that calculate these values start differing, but it should be relatively close) [21:08] <JEEB> well, I don't think mobile devices like anamorphic? [21:08] <JEEB> audio would have to be re-encoded to AAC [21:10] <Yeti218> i am now trying: -crf 30 -vcodec libx264, does that make sense? [21:10] <Yeti218> that means i do not change dimensions and sound [21:12] <JEEB> do -an then to not copy the original audio track over, or do tell me which aac encoders did you compile with? [21:13] <JEEB> also, yes -- it makes sense, just make sure to check the result? [21:14] <Yeti218> i think i lost connection [21:15] <Yeti218> i see horizontal lines but i think that is a problem of the source not the compression i did [21:16] <JEEB> <JEEB> do -an then to not copy the original audio track over, or do tell me which aac encoders did you compile with? [21:16] <JEEB> <JEEB> also, yes -- it makes sense, just make sure to check the result? [21:16] <JEEB> ok, it could be that you are dealing with interlaced content [21:16] <JEEB> add -vf yadif [21:16] <JEEB> vf = video filter [21:16] <JEEB> yadif = deinterlacing video filter [21:17] <Yeti218> i think you mean this: Stream #0:1 -> #0:1 (mp2 -> libfaac) [21:19] <Yeti218> i just checked the result and it seems ok but when there is movement i see horizontal lines (glitches) .. i would have to check on the phone to see how bad it is there (last try i did -crf 35) [21:19] <Yeti218> yes, interlaced content .. i think that is what i am seeing [21:20] <Yeti218> Wow! not only did the filesize get smaller (no idea why), but it also looks a lot better [21:21] <Yeti218> it now went from 34mb -> 2.8mb and no indication of interlaced anymore [21:23] <Yeti218> is this de-interlace option a problem for when i convert videos that are not interlaced? [21:24] <JEEB> yes [21:24] <Yeti218> i would like to script processing them so 1 set of parameters would be ideal [21:25] <JEEB> you can use ffprobe or something to check if the bitstream is tagged interlaced or not :P [21:26] <JEEB> of course, telecined content is usually encoded interlaced, too -- but that's a whole separate bowl of derp [21:27] <Yeti218> i am afraid a lot of my content will have a tv source [21:27] <JEEB> ...telecined has nothing to do with whether or not it's from TV :P [21:28] <JEEB> anyways, yadif shouldn't rape even telecined content too badly (in the best case it'll just leave you alone with the extra frames) [21:28] <JEEB> so yeah, just set something to check if the bit stream in the input file is interlaced or not :P [21:28] <Yeti218> ah ok, i am not a native English speaking person so i kinda guessed telecined meant : broadcasted on tv [21:29] <JEEB> I can give you a link to a long write-up regarding this all, but I think it'd just make you go wee even more :P [21:29] <Yeti218> i just did ffprobe but i didnt see any indication that it was interlaced .. not obvisouly that is [21:29] <Yeti218> i try to absord just as much information as i need for this .. i should be spending my time writing the app [21:30] <Yeti218> the moment i know what works on (most) devices i will tell others to deliver me videos in that format [21:32] <JEEB> if tbc is 2x fps it's probably interlaced :P [21:32] <JEEB> would be my guess [21:32] <JEEB> because yes, it doesn't seem to straightforwardly give you the answer [21:32] <Yeti218> i just searched for some indications .. it seems it is not very easy to spot [21:33] <Yeti218> tbc is indeed 2 x tbs [21:33] <JEEB> it shouldn't be hard to make ffprobe to output if the stream is marked interlaced or not, not sure why they don't give you that information (Ž [21:34] <JEEB> mediainfo would be one lib/app to let you get that info [21:34] <Yeti218> mediainfo is a windows application right? [21:35] <JEEB> not only [21:35] <JEEB> also a library :P [21:37] <taqattack> What is the proper way to use ffmpeg to stream to RTMP? [21:37] <Yeti218> i just checked it with mediainfo -> it is interlaced [21:38] <JEEB> well, d'uh [21:38] <JEEB> that was pretty obvious [21:38] <JEEB> but yes, you can use it to check if something is flagged interlaced or not [21:38] <aphid> rtmp is an abomination [21:39] <taqattack> Yes but I need to use some protocol to stream [21:39] <Yeti218> ok, i just checked -crf 40 and it has gotten pretty bad [21:40] <JEEB> yes, you should add/remove it in smaller steps :P [21:41] <JEEB> (and it probably got pretty small as well) [21:43] <Yeti218> i changed changed another setting .. -ac 1 it went to mono and it decreased a bit again [21:43] <Yeti218> i assume that you wont notice the difference on a smart phone [21:44] <Yeti218> 2.8mb -> 2.6mb [21:44] <JEEB> unless you use headphones, yeah [21:44] <Yeti218> hmm true [21:44] <Yeti218> is there a codec that you recommend for audio? [21:45] <JEEB> you can't really use much else but AAC or HE-AAC for the devices you're aiming at [21:45] <JEEB> if you really want to minimize, use HE-AAC with libaacplus [21:48] <Yeti218> i think there is more to win on the video side of the file [21:48] <Yeti218> i am rather happy with the result so far .. but i havent done anything conscious yet with the profile and level we talked about earlier [21:49] <JEEB> if you are doing stuff for 3GS and latter you don't have to care about profile, it can stay at high (default) [21:49] <JEEB> as for level, you can just set it at 4 [21:49] <JEEB> also, you did play with the presets, right? [21:50] <JEEB> because those control the compression vs speed thing in libx264 [21:50] <JEEB> http://mewiki.project357.com/wiki/X264_Settings#preset [21:50] <Yeti218> -vprofile baseline -level 1 [21:51] <JEEB> lol [21:51] <JEEB> I don't think you'll be fitting in level 1 :P [21:51] <JEEB> also, not vprofile [21:51] <Yeti218> that was an example of the paramaters :) [21:51] <JEEB> it's -profile:v baseline [21:51] <Yeti218> oh [21:52] <JEEB> (in cases of presets and profiles you might want to append :v to the end as AAC encoders love to try and apply the profile f.ex. and then fail) [21:52] <Yeti218> k, so simply if i set it to slow it will compress better right? [21:53] <Yeti218> i read something about the problem with profile and AAC [21:53] <jdstrand> hi, someone reported a security issue against libav in Ubuntu, and it looks like it also affects ffmpeg. what is the email address for the security contact for ffmpeg? [21:54] <JEEB> yes (in the end of the line preset-wise [slowest ones] for the same crf you might get a wee bit bigger size, but that just means that the algorithms "see" more detail) [21:54] <jdstrand> hmmm, that is probably better for #ffmpeg-devel [21:55] <Yeti218> preset slow -> it got a little bit smaller [21:55] <JEEB> yes, that's the general way of how it goes [21:59] <Yeti218> how much does the profile and level influence the bytesize? [21:59] <JEEB> the more limited a profile, the less features can be used to compress a clip [22:00] <JEEB> level is just a flag [22:01] <Yeti218> ah, as in the targeted device does not have the decode chip to process it? [22:02] <Yeti218> hmm the high profile got 1/6 smaller than the baseline [22:03] <JEEB> (also the slower the preset the more features from higher profiles get used) [22:03] <Yeti218> i am testing veryslow atm to see what the result is [22:04] <Yeti218> btw are you a user of ffmpeg or one of its developers? [22:05] <JEEB> I have a commit in ffmpeg, but it's a very simple one, so you can't really count me as a "developer" -- I'm mostly a user :P [22:06] <Yeti218> heh heh [22:19] <Yeti218> maybe something too far away from the functionality from ffmpeg .. but would it be possible to create a video from a foto + mp3? [22:24] <taqattack> What's xuggler. How is it related to Ffmpeg [22:27] <Yeti218> hmm i found a little glitch in ffmpeg [22:27] <Yeti218> if i use as input -i "bla, (sdf).mp4" it cannot find the file to process [22:28] <Yeti218> even though the OS accepts this as a valid filename to copy etc [22:45] <JEEB> Yeti218, you have to select a container that is supported by the thing you are encoding (reply to your damn pm) [22:46] <Yeti218> i just re-used the container the original was in [22:46] <JEEB> so you think an iPhone would support mpeg-ps? [22:46] <JEEB> sure, use it if you want :P [22:46] <Yeti218> i might have been lucky [22:46] <JEEB> mp4 is the format you want [22:46] <JEEB> end of story [22:46] Action: JEEB sighs [22:51] <aphid> maybe someday apple will support webm [22:53] <JEEB> ahaha [22:53] <JEEB> maybe after there will be actual ASICs for VP8 [22:53] <JEEB> actually on board actual hardware [22:55] <aphid> there's momentum on the hardware stuff [22:55] <JEEB> let's just say that I'm not holding my breath [22:55] <aphid> ...it's almost as if companies don't like paying into patent cartels.. preposterous I know [23:15] <Yeti218> the problem with patents is that there is always someone who has patented "use a thing with another thing" and then you are screwed [23:17] <aphid> you're not screwed if it's for something obvious [23:18] <aphid> but it ultimately just comes down to who has more lawyers. mpegla and the vp8 pool they're still in the process of forming (more than a year after they called for patent submissions) have plenty of them.. so does google and their pool. [23:46] <Yeti218> are all x264 parameters, parameters of the ffmpeg command as well? [23:46] <JEEB> what? [23:47] <Yeti218> x264 has -tune, does that mean ffmpeg has it as well? [23:47] <JEEB> first of all [23:47] <JEEB> x264 has --tune [23:47] <JEEB> but yes [23:47] <JEEB> ffmpeg has -tune [23:47] <JEEB> it has -preset, -profile, -level and -tune [23:48] <JEEB> almost everything else you will most probably have to set via the x264opts interface [23:48] <Yeti218> i am fiddling a bit with -tune now to see if i can notice any difference [23:49] <canbal> hi, is there any working progress or any plans on supporting H.264/MVC encoding and/or decoding in ffmpeg? [23:50] <JEEB> ffmpeg doesn't handle MVC to begin with [23:50] <JEEB> s/MVC/H.264/ [23:50] <JEEB> (encoding) [23:50] <JEEB> x264 had an MVC GSoC last year [23:50] <JEEB> but the student has since vanished into thin air [23:51] <canbal> i see, so there are no plans to add the support to ffmpeg? and x264 somewhat supports it, but not entirely? [23:51] <JEEB> there was also a patch provided by a company that corp. licensed x264 for MVC encoding, but that was not merged into the mainline because the student IIRC was still around then and they were similar [23:51] <JEEB> (the patches, I mean) [23:52] <JEEB> at the moment mainline x264 doesn't support it at all [23:52] <JEEB> the patches for MVC probably lie somewhere on pastebins, but neither of them was officially implemented [23:52] <canbal> ok, what you mean by ffmpeg doesn't support h.264 encoding, it supports mpeg4 though [23:52] <JEEB> it doesn't have a H.264 encoder? [23:53] <JEEB> it only has an interface for usage of libx264 [23:53] <canbal> oh i see [23:53] <canbal> sorry for the confusion [23:53] <JEEB> as for decoding, I don't remember if anyone worked on MVC [23:54] <canbal> ok, and there are no plans for that either? i work on H.264/MVC related research [23:54] <JEEB> also, x264 can use frame packing for multiple views, and IIRC the compression benefit for MVC wasn't even that big with the stuff that was implemented [23:54] <JEEB> *from MVC [23:55] <canbal> and i am trying to see to what extent it is supported (dev version) and if i can be any help [23:55] <JEEB> dunno about plans, I've not seen any [23:55] <JEEB> people are welcome to try their luck [23:55] <canbal> :) [23:55] <JEEB> (with both x264 and the GSoC project that never got merged, as well as libavcodec) [23:57] <JEEB> does the JM H.264 reference decoder even support MVC yet? [23:57] <canbal> yes it does now [23:58] <canbal> only works for stereo though [23:58] <canbal> they have a hard limitation on number of views [23:58] <canbal> i personally work on JMVC [23:58] <canbal> which is the reference SW for MVC [23:58] <JEEB> ok, so it's still separate from JM [23:59] <canbal> i guess you may say that [23:59] <canbal> but JM also started to support 3D [23:59] <canbal> and i think it does have full support except for number of views supported [00:00] --- Thu Apr 26 2012

1 0