Ffmpeg-devel-irc

Download

ffmpeg-devel-irc@ffmpeg.org

March 2014

1 participants
62 discussions

ffmpeg-devel.log.20140315
by burek 16 Mar '14

16 Mar '14

[00:06] <llogan> shit. forgot a period. OCD OCD OCD. [00:50] <gueb> how can I configure ffmpeg to use -pg? [00:52] <gueb> -pg is a flag for profiling with gcc [00:54] <BBB_> gueb: --extra-cflags="-pg" [00:54] <gueb> thanks [00:54] <BBB_> gueb: also, user questions in #ffmpeg-users please [00:55] <gueb> profiling is developing, isn't it? [00:55] <BBB_> passing a flag to configure isn't really [00:55] <wm4> does gprof really have any worth? [00:56] <gueb> wm4, what do you recommend for profiling? [00:56] <wm4> on linux, maybe the perf utility [00:56] <wm4> or counting cycles like ffmpeg devs seem to love it [00:56] <wm4> BBB should know a lot about this [00:56] <BBB_> I've used perf, it's pretty good [00:57] <BBB_> kcachegrind was quite useful, in combination with valgrind, but that's long ago [01:05] <iive> i've done gprof on mplayer on my previous athlon tbird. the speed up was quite miniature, about 1%. [01:05] <iive> and I don't consider that cpu to be modern. :) [01:09] <BBB_> gueb: anyway, perf, really - don't use gprof [01:09] <gueb> ok [01:09] <gueb> I have a performance problem [01:09] <gueb> when I compile the code externally [01:09] <gueb> it does some work in less than 5 ms [01:10] <gueb> when doing inside ffmpeg it uses about 40 ms [01:10] <gueb> :S [01:11] <gueb> any advice? [01:12] <BBB_> gueb: I don't know what code you're talking about so it's really hard to say [01:12] <gueb> it is a part of my new codec [01:12] <BBB_> also 5ms is a weird metric, we typically talk about cycles for performance-sensitive stuff (unless this is sth. like network io or so) [01:13] <gueb> I know that there is a performance problem because the file is processed very slow [01:13] <BBB_> I'd try perf and see where it's spending its time [01:13] <gueb> with some other tests outside ffmpeg it was fast [01:14] <BBB_> I don't have enough context to answer your question, I'd need things like code, compile flags, sample files, etc. - probably a spec of the codec would be useful although not necessary [01:14] <BBB_> maybe you compiled ffmpeg with -O0? [01:14] <gueb> I do not know, I just added the module to Makefile [01:14] <gueb> OBJ-... += foo.c [01:17] <iive> gueb: is the code complete codec, or just some routine? [01:17] <gueb> some routine [01:19] <iive> when you do synthetic benchmark, you usually totally ignore the cache pollution. [01:29] <cone-797> ffmpeg.git 03Michael Niedermayer 07master:2b8d28439b6e: avcodec/h264_cabac: move the arm unchecked_bitstream reader special case closer to where the issue is [01:29] <cone-797> ffmpeg.git 03Michael Niedermayer 07master:41d08ca57564: avcodec/arm/cabac: fix inline cabac reader with the UNCHECKED bitstream reader [01:58] <gueb> solved, the problem was a spurious initialization that I had to put to avoid an absurd and annoying warning [01:59] <gueb> warning that only appeared with -O3 [02:09] <cone-797> ffmpeg.git 03Peter Ross 07master:d1bb17940dd2: avcodec/adpcm: ADPCM_IMA_DK3 packets are padded to 16-bit packet boundary [02:33] <rcombs> can I get lavf/lavc to give me MKV-format ASS instead of text-file-format? [02:35] <wm4> rcombs: at what point? lavf should give you always the mkv format [02:35] <wm4> but lavc output not [02:35] <wm4> (and that's annoying) [02:35] <rcombs> AFAICT matroskadec spits out the wrong thing [02:35] <wm4> makes me wonder, how does ffmpeg.c do srt-to-ass and then mux into mkv [02:36] <wm4> rcombs: used to [02:36] <wm4> and Libav probably still does [02:40] Action: rcombs stares upon Dialogue: 0,0:00:39.93,0:00:46.02,op trans,,0,0,0,,{}Both our infinite dreams and our sketched out future\r\n [02:41] <rcombs> oh, damn, that's my fault [02:44] <rcombs> erm, wait, no it's not [02:45] <rcombs> how do I force AV_CODEC_ID_ASS over AV_CODEC_ID_SSA in lavf? [02:45] <wm4> oh maybe it's still not bumped [02:46] <wm4> it's connected to an api bump [02:46] <rcombs> yeah, it is [02:46] <rcombs> but that seems to imply it'd be overrideable at runtime [02:56] <cone-797> ffmpeg.git 03Stephen Hutchinson 07release/2.2:2b9ee7d5b901: doc: Add section about AviSynth support [02:56] <cone-797> ffmpeg.git 03Michael Niedermayer 07release/2.2:29df24252aa9: Merge commit '2b9ee7d5b901e0d7ba617511e4ed31d3043894d3' into release/2.2 [02:56] <cone-797> ffmpeg.git 03Stephen Hutchinson 07release/2.2:c4f5f4dbd303: doc/general.texi: Adjust the notes on AviSynth [02:56] <cone-797> ffmpeg.git 03Matt Oliver 07release/2.2:bf08665e2ede: Fix modplug linkage on Windows. [02:56] <cone-797> ffmpeg.git 03Carl Eugen Hoyos 07release/2.2:3caa6a5a5744: Revert "Allow stream-copying grayscale mov files." [02:56] <cone-797> ffmpeg.git 03Michael Niedermayer 07release/2.2:46c2dba20e09: doc/texi2pod: fix encoding type [02:56] <cone-797> ffmpeg.git 03Michael Niedermayer 07release/2.2:242df26b4440: mvformat/movenc: fix IMX [02:56] <cone-797> ffmpeg.git 03Michael Niedermayer 07release/2.2:70e3cc282bee: avutil/timestamp: Warn about missing __STDC_FORMAT_MACROS for C++ use [02:56] <cone-797> ffmpeg.git 03Michael Niedermayer 07release/2.2:125bea15d1f6: avcodec/libx265: fill headers in extradata [03:22] <cone-797> ffmpeg.git 03Michael Niedermayer 07release/2.2:6ba07e9948d5: update for 2.2-rc2 [04:26] <cone-797> ffmpeg.git 03Peter Ross 07fatal: ambiguous argument 'refs/tags/n2.2-rc2': unknown revision or path not in the working tree. [04:26] <cone-797> Use '--' to separate paths from revisions [04:26] <cone-797> refs/tags/n2.2-rc2:HEAD: avcodec/adpcm: ADPCM_IMA_DK3 packets are padded to 16-bit packet boundary [04:58] <cone-797> ffmpeg.git 03Peter Ross 07master:dadf668df20e: avformat/mov: decode compilation metadata [04:58] <cone-797> ffmpeg.git 03Peter Ross 07master:ac9bfb1fafde: avformat/movenc: encode compilation metadata [04:58] <cone-797> ffmpeg.git 03Peter Ross 07master:847d8af59af5: avformat/id3v2: decode compilation metadata [05:32] <gueb> good night [06:58] <yanrong> Is there anyone know the difference between ff_fft_calc_interleave_xx and ff_fft_calc_xx? When to use respectively, thanks. [15:52] <cone-678> ffmpeg.git 03Michael Niedermayer 07master:da89572004f6: avcodec: Move STRIDE_ALIGN to internal.h [15:52] <cone-678> ffmpeg.git 03Michael Niedermayer 07master:06a3185e3861: avcodec/mpegvideo_enc: dont use direct mode for unaligned input [17:27] <jnvsor> With regards to bug 3192, the summary is wrong, can I change it or should I leave that to a dev? [17:30] <Compn> jnvsor : change it :) [17:30] <Compn> if you have a dev account on the bug [17:30] <Compn> not sure if users can change it [17:30] <Compn> you might face carl in mortal kombat, but thats always fun to watch [17:31] <jnvsor> Compn: Cool :) [18:15] <wm4> interplayvideo is broken [18:16] <wm4> block soup on descent 2 movies [19:58] <kierank> these guys are hillarious [19:59] <kierank> they put all their software in a hidden folder [19:59] <JEEB> lol [20:02] <kierank> it's literally every media software under the sun in that hidden folder [20:02] <kierank> from jm to realmedia [20:03] <JEEB> sounds like a payware version of SUPER, with a hidden directory :D [20:09] <kierank> basically yes [20:41] <J_Darnley> kierank: Thet should make it more secret by making it a system file too [20:41] <J_Darnley> uh folder in this case [22:23] <cone-678> ffmpeg.git 03Michael Niedermayer 07master:8803b970ef98: ffmpeg_opt: Check ist is valid before using it in audio_channels_map init [22:58] <cone-678> ffmpeg.git 03Michael Niedermayer 07master:6c47a4e97248: swscale/x86/swscale: fix missing xmm clobbers in yuv2yuvX_sse3() [00:00] --- Sun Mar 16 2014

1 0

ffmpeg.log.20140315
by burek 16 Mar '14

16 Mar '14

[00:03] <llogan> relaxed: damn it, you made me procrastinate again. [00:04] Action: llogan goes back to boring, paid work [00:05] <JEEB> at least you have one ;) [00:07] <relaxed> you're no longer the "unfortunately i've been neglecting that particular guide lately" guy. [00:07] <relaxed> ...and now I have a wiki account. I see what you did there. [00:08] <llogan> there's still the PATH goofiness. ~/.bash_profile was from me being lazy, but it, as users oft say "doesn't work" for everyone [00:09] <llogan> but i don't want to throw it into /usr/local because some users didn't have su powers but still have enough deps to compile, etc, IIRCAFAIK. [00:11] <llogan> relaxed: did trac bitch to you about spam? [00:11] <llogan> during registration [00:11] <relaxed> a little bit [00:12] <relaxed> /usr/local prefix is bad, in my opinion [00:12] <relaxed> I'll look at the $PATH stuff later [00:13] <llogan> yes, i also want to avoid it. especially on Debian and spawn since of conflicting packages in repo. [00:14] <llogan> and deps that expect certain shit. although for just the binary i don't know how many dependent packages use that (KDenlive) now how they use it (via PATH or full path). [00:14] <llogan> s/now/nor [00:15] <llogan> i add the ~/.manpath to CentOS guide later too [01:34] <gueb> hello [02:11] <ItsMeLenny> ohaiguyz, i need a little help here, i'm trying to replace the audio stream in a video with an audio file, but the command line i have so far is just adding an extra audio stream [02:11] <ItsMeLenny> -i ./MVI_2936.MOV -i ./MVI_2936_audio.flac -map 0 -map 1 -c:v libx264 -s 640x480 -aspect 16:9 ./Output.mov [02:12] <ItsMeLenny> http://pastie.org/8920449 [02:13] <klaxa> that's neither the complete command, nor the console log [02:13] <ItsMeLenny> oh right [02:14] <ItsMeLenny> http://pastie.org/8920453 [02:14] <llogan> what's avconv? [02:14] <ItsMeLenny> ffmpeg fork [02:14] <llogan> try #libav [02:14] <ItsMeLenny> it's the same command [02:14] <sacarasc> But different programmes. [02:15] <ItsMeLenny> theres nobody ever talking in libav [02:15] <llogan> forks and third party products are not supported here [02:15] <ItsMeLenny> and thats why i didnt paste that start of it in the first place [02:15] <ItsMeLenny> i'm not asking for support [02:15] <llogan> or use ffmpeg [02:16] <ItsMeLenny> ok, i am using ffmpeg now, the exact same question still remains [02:17] <sacarasc> ItsMeLenny: I think it is, and I suck at map, so who knows... -map 0.0:0.0 -map 1.0:0.1 [02:17] <klaxa> i'd rather go with -map 0:v -map 1:a [02:17] <ItsMeLenny> llogan, paste bin for what? i am asking what the command is to replace an audio in a video with a different audio file [02:17] <klaxa> map the video of the first input, the audio of the second input [02:17] <klaxa> and maybe copy codecs, re-encoding seems pointless [02:17] <llogan> ItsMeLenny: your fake ffmpeg that you're probably using is not supported here. [02:18] <ItsMeLenny> lol, thanks sacarasc [02:18] <ItsMeLenny> i went with klaxa 's, it worked like a charm, thanks very much [02:19] <klaxa> ItsMeLenny: do consider switching to ffmpeg though [02:20] <llogan> or next time ask in the empty, echoing halls of #libav is you want to continue using that [02:20] <ItsMeLenny> klaxa, yeah i grab the nightly binaries usually, but avconv is in ubuntu by default, also theres some things which avconv does that ffmpeg doesnt and vice versa [02:21] <klaxa> i haven't heard many good things about avconv, i compile ff-- oh well... [02:21] <klaxa> this may be generalizing, but damn those ubuntu folks [02:22] <llogan> s/those folks/a particular maintainer [02:22] <klaxa> well apparently the usebase doesn't care enough either [02:22] <klaxa> *userbase [02:25] <mmint> The userbase probably just doesn't know [02:26] <mmint> I (debian user) didn't know the story for almost a year [02:26] <klaxa> at least this particular user knows, but refuses th coorperate [02:26] <klaxa> *to [02:27] <klaxa> i run debian on my servers too, at some point the ffmpeg binary said it was deprecated, so i asked, was pointed to the stackoverflow comment and started compiling ffmpeg from source [05:32] <gueb> good night [14:40] <sinhayash_> Hello everybody! [14:40] <sinhayash_> I am new to ffmpeg. [14:41] <sinhayash_> I am trying to convert an array of png images to video using ffmpeg in javascript in firefox browser [14:41] <sinhayash_> What command do I need to call to achieve this? [14:41] <sinhayash_> I am trying with a web worker. [14:56] <klaxa> sinhayash_, i don't think you can do that in javascript with ffmpeg [15:01] <sinhayash_> klaxa https://www.webrtc-experiment.com/ffmpeg/merging-wav-and-webm-into-mp4.html [15:01] <sinhayash_> https://github.com/bgrins/videoconverter.js [15:05] <sinhayash_> klaxa: Is there something I am missing out? [16:19] <eth2wlan> Hi [16:19] <eth2wlan> I have a question [16:19] <eth2wlan> (of course) [16:19] <jarainf> We all do~ [16:19] <eth2wlan> ok, I'm going to explain [16:20] <eth2wlan> I've a 720p «H264 - MPEG-4 AVC (part 10) (H264)» [16:20] <eth2wlan> (this is what VLC says) [16:21] <jarainf> eth2wlan, please post the ffprobe output :3 [16:21] <eth2wlan> yes [16:21] <eth2wlan> ("please hold the line") [16:22] <eth2wlan> 1) Duration: 00:01:59.86, start: 0.000000, bitrate: 1953 kb/s [16:23] <eth2wlan> 2) Stream #0:0: Video: h264 (Baseline) (H264 / 0x34363248), yuv420p, 1280x720, 600 fps, 600 tbr, 600 tbn, 1200 tbc [16:23] <eth2wlan> This is what ffprobe says [16:23] <eth2wlan> Ok [16:23] <jarainf> eth2wlan, so what are you trying to do? [16:24] <eth2wlan> jarainf, justo to improve compression with as-high-as-possible quality [16:25] <eth2wlan> Al my attempt results in an bigger and/or worse video file [16:25] <klaxa|web> eth2wlan: see: https://trac.ffmpeg.org/wiki/x264EncodingGuide [16:26] <eth2wlan> Note: it is a fixed videocamera video file. For a long time, nothing change in the picture. [16:26] <klaxa|web> well what have you tried so far? the encoding guide should be comprehensive enough for a novice to encode to sufficient quality [16:27] <eth2wlan> Last but not least: the ffmpeg job must be faster than the movie duration [16:27] <klaxa|web> that may or may not be possible [16:27] <klaxa|web> what hardware are you encoding on? [16:27] <eth2wlan> For example, for a 60 minutes video, ffmpej must recompress it in less than 60 minutes [16:28] <klaxa|web> i'll ask again, what hardware are you encoding on? [16:28] <eth2wlan> The hardware is an old Acer Extensa 5220 [16:29] <eth2wlan> Googling, it is a Celeron M 530 @ 1.74GHz [16:30] <klaxa|web> i highly doubt you will be able to encode with a sufficient quality, faster than real-time on that hardware [16:30] <eth2wlan> Yes, I'm afraid that is not possible to [16:31] <eth2wlan> So, this is the real question [16:32] <eth2wlan> ("it is possible to compress faster than movie lenght with sufficient quality") [16:33] <jarainf> that was a statement. [16:33] <klaxa|web> apart from that: short answer for your specific usecase: probably not [16:34] <klaxa|web> there is too little information about the video content (not even the encoding and measuered "quality", but perceived complexity) and your hardware is old and weak [16:37] <eth2wlan> Ok, probably I can't have what I need. I just asked here before of deepen (and waste of time, if is not possible) in the documentation [16:38] <eth2wlan> Thank you very much [16:40] <eth2wlan> Sorry, I've forgotten an important thing. [16:40] <eth2wlan> The greetings. :-) [16:40] <eth2wlan> Bye! [21:03] <sajjadg> How can I detect the audio codec/format inside a container (mkv) and extract is without encoding (copy) to another container(e.g. mka)? [21:06] <sajjadg> avconv -i part2.mkv -acodec copy part2.mka works [21:06] <sajjadg> but but ogg doesn't! [21:07] <sacarasc> #libav is the channel for avconv. It could be an avconv bug. [21:07] <sacarasc> (Also, ogg supports video.) [21:07] <sacarasc> Try -an [21:07] <sajjadg> Stream #0.1: Audio: vorbis, 48000 Hz, stereo, fltp (default) [21:07] <sajjadg> I'm the enemy now :P [21:22] <wodim> hello, what would you say are the most safe audio and video codecs you think are supported by all home dvd players? [21:24] <sacarasc> MPEG2 and AC3. [21:24] <wodim> alright, let's see [21:24] <wodim> and what container? [21:24] <wodim> avi? [21:24] <sacarasc> No, make an actual DVD. [21:25] <wodim> it's for a home dvd player with a usb socket [21:25] <wodim> so it has to be a file [21:25] <wodim> i guess an avi is ok [21:25] <sacarasc> You said ALL home DVD players. [21:25] <wodim> heh, well... my fault [21:25] <sacarasc> My DVD player can't do anything except actual DVDs. :p [21:25] <wodim> do you change your mind now that you have this new information? :p [21:26] <sacarasc> You could look up what your DVD player can do, but you probably want MPEG4 ASP and MP3 in AVI, then. [21:26] <wodim> ok, let's see [21:39] <relaxed> wodim: you'll need -vtag XVID if you're using ffmpeg's mpeg4 encoder. [21:39] <wodim> alright alright [21:40] <wodim> -acodec libmp3lame -vcodec mpeg4 [21:40] <wodim> how right does this sound? [21:43] <relaxed> ffmpeg -i input -c:v mpeg4 -q:v 3 -vtag XVID -c:a libmp3lame -b:a 192k -ac 2 output.avi [21:44] <wodim> ok. [21:44] <wodim> but the size of the new movie is worrying me. [21:45] <wodim> is it because this codec is shit? [21:45] <sacarasc> Do you want it to be a specific size? [21:45] <wodim> i would like it to look good [21:46] <wodim> but it's like 200MB for 10 minutes [21:46] <wodim> while the original file is 650MB for 1h30 and looks perfectly [21:46] <wodim> haha [21:46] <sacarasc> 650MB for 90 minutes will not look perfect. :( [21:46] <wodim> yes, you're right [21:47] <wodim> let's say "very good" [21:47] <sacarasc> What is the original file? [21:47] <wodim> "h264 mpeg-4 avc" [21:47] <wodim> if that says something to you [21:48] <sacarasc> Yeah, AVC is a LOT better at compression than ASP. [21:49] <wodim> 500MB for 18 minutes [21:49] <wodim> haha [21:49] <wodim> god [21:49] <wodim> i think this dvd player won't be able to read that much of the USB drive at once [21:49] <wodim> it will jam [21:49] <relaxed> use 2 pass encoding if you need a specific size [21:49] <wodim> i will see if the player can do this [21:50] <relaxed> https://trac.ffmpeg.org/wiki/x264EncodingGuide [21:50] <wodim> no, i'm not recoding in h264, but mpeg4 [21:50] <relaxed> look at the "Two-Pass" section and apply that to your mpeg4 ecndoing command [21:50] <wodim> ah [21:50] <relaxed> encoding* [21:54] <wodim> can subtitles be embedded in avi files? [21:55] <wodim> alright, with mpeg4 i get audio but no sound [21:55] <wodim> uh [21:56] <wodim> oops [21:56] <wodim> audio but no video [21:56] <wodim> heh [21:57] <relaxed> try -vtag DIVX [22:14] <wodim> can ffmpeg take advantage of multicore systems? [22:27] <klaxa> depends on the codec, usually yes [22:27] <klaxa> as far as i know at least [00:00] --- Sun Mar 16 2014

1 0

ffmpeg-devel.log.20140314
by burek 15 Mar '14

15 Mar '14

[00:07] <cone-365> ffmpeg.git 03Diego Biurrun 07master:831a1180785a: Update dsputil- and SIMD-related comments to match reality more closely [00:07] <cone-365> ffmpeg.git 03Diego Biurrun 07master:edd1f833fa14: x86: h264_idct_10_bit: Use proper type in function prototype comments [00:07] <cone-365> ffmpeg.git 03Diego Biurrun 07master:55519926ef85: x86: Make function prototype comments in assembly code consistent [00:07] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:1c788eaca9a0: Merge commit '831a1180785a786272cdcefb71566a770bfb879e' [00:07] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:a9b1936a4e46: Merge commit 'edd1f833fa145eb9c5026877c699ebe6efca00a0' [00:07] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:4104eb44e629: Merge commit '55519926ef855c671d084ccc151056de9e3d3a77' [00:16] <cone-365> ffmpeg.git 03Diego Biurrun 07master:0a8f91b1fbc1: dsputil: Remove disabled avg_no_rnd_qpel leftovers [00:16] <cone-365> ffmpeg.git 03Diego Biurrun 07master:8199bb7be0db: dsputil.h: Clean up comments [00:16] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:ea2bd139b75d: Merge commit '0a8f91b1fbc1b627d8db8c734c812beb32102936' [00:16] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:153557dfcb10: Merge commit '8199bb7be0db423bcbbd252d092386634fcf1d68' [00:24] <cone-365> ffmpeg.git 03Diego Biurrun 07master:503dda7d4f5b: dsputil.h: K&R formatting cosmetics [00:24] <gueb> hello again [00:24] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:db94b144e95f: Merge commit '503dda7d4f5b4d76eed0239e9da66993a182b009' [00:25] <gueb> wm4, are you there? [00:36] <cone-365> ffmpeg.git 03Diego Biurrun 07master:bb12c2e9c00e: dsputil: K&R formatting cosmetics [00:37] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:dd7c2e9bce0c: Merge commit 'bb12c2e9c00e3d3c393885849b8fa31cd849dc92' [00:44] <wm4> gueb: sort of... [00:44] <gueb> :D [00:45] <gueb> hello [00:45] <cone-365> ffmpeg.git 03Diego Biurrun 07master:635ec127d4bc: dsputil: Merge variable declarations and initializations [00:45] <cone-365> ffmpeg.git 03Diego Biurrun 07master:05563ccacc98: dsputil: cosmetics: Lose camelCase on ff_cropTbl and ff_squareTbl names [00:45] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:16f619f57c07: Merge commit '635ec127d4bc0c49ae0ac4a50acd3da1ca8c2150' [00:45] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:6016b8329bca: Merge commit '05563ccacc98fd185affdbf8cbaf094caf36b852' [00:45] <gueb> what do you think that is easier, making my own muxer or modifying an existing one? [00:46] <nevcairiel> for simple audio codecs just define a TwoCC RIFF code and shove it into existing containers, many support that [00:47] <wm4> yeah, there's a big lookup table somewhere [00:47] <nevcairiel> avformat/riff.c -> ff_codec_wav_tags [00:48] <gueb> thanks [00:48] <gueb> wow, 0x1234 is not taken :P [00:50] <nevcairiel> the "official" codes are just sequentially numbered from 1 upwards, although there are some more modern codecs with higher numbers, mostly AAC related [00:50] <Compn> does the world need yet another audio codec ? :P [00:51] <cone-365> ffmpeg.git 03Diego Biurrun 07master:acd2b8e42df2: rnd_avg.h: K&R formatting cosmetics [00:51] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:3a27f968f990: Merge commit 'acd2b8e42df25d46440583f93c12e5f285aebb31' [00:52] <gueb> flac's tag is F1AC [00:52] <gueb> :O [00:52] Action: wm4 wonders if 0xBEEF is taken [00:52] <nevcairiel> any 16-bit number will do [00:53] <nevcairiel> as long as its not taken [00:53] <gueb> no, BEEF neither [00:54] <gueb> it works :D [00:57] <cone-365> ffmpeg.git 03Diego Biurrun 07master:945673f792eb: dsputil_template: K&R formatting cosmetics [00:57] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:fdf4eb7911b3: Merge commit '945673f792eb4141860ed66664156df685f7b7f0' [00:58] <nevcairiel> anyway, with a riff code defined, putting it into wav, avi or even mkv should work [00:58] <nevcairiel> probably a bunch more [00:58] <gueb> yes, I am trying it with matroska [01:08] <cone-365> ffmpeg.git 03Diego Biurrun 07master:6a8b35dc88b4: dsputilenc_mmx: Merge two assignment blocks with identical conditions [01:08] <cone-365> ffmpeg.git 03Diego Biurrun 07master:38675229a879: dsputil_mmx: K&R formatting cosmetics [01:08] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:55f53f6c29ac: Merge commit '6a8b35dc88b4a1a452f192fbbf53ae7f59bc3f23' [01:08] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:d926c4b240ed: Merge commit '38675229a879aa5258a8c71891fc8cbf74cf139f' [01:18] <cone-365> ffmpeg.git 03Diego Biurrun 07master:a36947c167d7: dsputilenc_mmx: K&R formatting cosmetics [01:18] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:b7a5f5dc6662: Merge commit 'a36947c167d7278b891453083b57dc56b7a7f5c5' [01:25] <cone-365> ffmpeg.git 03Diego Biurrun 07master:f8bbebecfd7e: x86: motion_est: K&R formatting cosmetics [01:25] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:925ce6faf4dd: Merge commit 'f8bbebecfd7ea3dceb7c96f931beca33f80a3490' [01:42] <cone-365> ffmpeg.git 03Diego Biurrun 07master:4cb4680c1087: x86: dsputil_x86.h: K&R formatting cosmetics [01:42] <cone-365> ffmpeg.git 03Diego Biurrun 07master:441b09391571: x86: dsputil_init: K&R formatting cosmetics [01:42] <cone-365> ffmpeg.git 03Diego Biurrun 07master:db3f61a04f1f: x86: dsputil_init: Drop some unnecessary parentheses [01:42] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:236874a57192: Merge commit '4cb4680c1087a2cd13d4b0c9167a2eb3147f99d8' [01:42] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:27cab16ce719: Merge commit '441b093915717afa7d24be34bdab2a4911b30a57' [01:42] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:5dd97d5809ad: Merge commit 'db3f61a04f1f66746660f921bb2780ddf1141f3b' [01:47] <Compn> i thought we were going to do one giant k&r commit and get it over with [01:47] <Compn> did that plan fail ? [01:48] <cone-365> ffmpeg.git 03Diego Biurrun 07master:671c8785596e: bfin: dsputil: Remove unused global variable [01:48] <cone-365> ffmpeg.git 03Diego Biurrun 07master:14efbe694ac5: bfin: dsputil: Drop broken vsad bits [01:48] <cone-365> ffmpeg.git 03Diego Biurrun 07master:f5f099766816: bfin: hpeldsp: Drop broken put_pixels_*_xy2_nornd bits [01:48] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:0f95856834d0: Merge commit '671c8785596ea4a45817ba302289004ff456630a' [01:49] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:c686cf23aa86: Merge commit '14efbe694ac5beff98f5c5e652c166603f1738cf' [01:49] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:2b02c1a698e3: Merge commit 'f5f099766816aab3ad805c63e85f12cb664dc9a8' [01:56] <cone-365> ffmpeg.git 03Diego Biurrun 07master:1e33095b83ad: bfin: dsputil: Drop broken and trivial profiling code [01:56] <cone-365> ffmpeg.git 03Diego Biurrun 07master:a3e7562621bd: bfin: dsputil: K&R formatting cosmetics [01:56] <cone-365> ffmpeg.git 03Diego Biurrun 07master:e99af2a3b166: bfin: Refactor duplicated l1_text/l1_data_B macros [01:56] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:a4369dbb8fda: Merge commit '1e33095b83addd8d24bc16bd08aa734fa161d515' [01:56] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:edf426135005: Merge commit 'a3e7562621bd96ce909e07149c419b6ba85cc54f' [01:56] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:91a61c4c156d: Merge commit 'e99af2a3b1660b4f328335149980064692097cad' [02:04] <cone-365> ffmpeg.git 03Diego Biurrun 07master:da785231ea0b: bfin: Refactor duplicated assembly-related macros [02:04] <cone-365> ffmpeg.git 03Diego Biurrun 07master:80173c71bab3: bfin: hpeldsp: K&R formatting cosmetics [02:04] <cone-365> ffmpeg.git 03Diego Biurrun 07master:bbf0ef57e9d2: bfin: Move all assembly declarations only used within one .c file there [02:04] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:3141b04c7fd4: Merge commit 'da785231ea0b82b5c2526babbb2871c935b21a87' [02:04] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:cbf190f7aef0: Merge commit '80173c71bab3b8817c148837bdc920557018f428' [02:04] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:c49797637ff7: Merge commit 'bbf0ef57e9d2d7bb2995f664dd74c8229bca25f5' [02:10] <gueb> how is managed aligned memory? [02:10] <gueb> is there a function for request aligned memory? [02:10] <wm4> av_malloc [02:10] <gueb> oh [02:10] <gueb> so simple! [02:12] <gueb> I am adapting a code for using matrices [02:12] <cone-365> ffmpeg.git 03Diego Biurrun 07master:13ae97b210f3: bfin: hpeldsp: "nornd" ---> "no_rnd" function name suffix for consistency [02:12] <cone-365> ffmpeg.git 03Diego Biurrun 07master:a7cc4ac9d99b: bfin: dsputil: Unconditionally compile hpel pixels code [02:12] <cone-365> ffmpeg.git 03Diego Biurrun 07master:929ec39ec449: bfin: hpeldsp: Split no_rnd pixel operations off into a separate file [02:12] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:866e11b33b29: Merge commit '13ae97b210f35d77290f4edb0d6846269271729a' [02:12] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:26f4a509820d: Merge commit 'a7cc4ac9d99bd586c8939f4312f8c9e56d5f5730' [02:12] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:be60e20498c7: Merge commit '929ec39ec4497827f7530f2aaa31646ef35462ee' [02:13] <gueb> and I would like to align rows, so I need to put some padding [02:13] <gueb> at the end of each row [02:13] <gueb> is there any way to know the align size? [02:14] <gueb> a macro, perhaps [02:14] <wm4> well, you can look what libavutil/mem.c does [02:15] <gueb> mmmm [02:16] <gueb> alignement is 16 unless AVX is present, then is 32 [02:16] <gueb> thanks [02:18] <cone-365> ffmpeg.git 03Diego Biurrun 07master:eb2b8bce39c9: bfin: Use more sensible file names [02:18] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:f3d3e3fa4e10: Merge commit 'eb2b8bce39c93d16a02656238248c6cf697f447f' [03:20] <Zeranoe> lol http://ffmpeg.zeranoe.com/forum/viewtopic.php?f=17&t=1851 [03:26] <Compn> llogan : bypassing an opportunity to bikeshed about number of developers in git repo list on ffmpeg.org ? you're no fun! [03:27] <michaelni> Zeranoe, i dont think a default of 1.4center + 0.5 surround would work with all kinds of 5.1 input [03:27] <michaelni> s/work/sound good/ [03:34] <Compn> michaelni : want to run some double blind tests? :p [03:38] <michaelni> cant hurt if you have volunteers, diverse input material and someone to organize it [04:21] <llogan> Zeranoe: lol indeed. much anger in that one. [04:27] <llogan> Compn: next bikeshed is site redesign. go ape. [04:41] <gueb> good night [04:43] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:dded5ed9c5eb: avcodec/libx265: fill headers in extradata [04:59] <Zeranoe> Hats off to FFmpeg, ffplay was able to handle 4k at 60fps while VLC and MPC-HC chocked. Proabably multithreaded decoding [05:34] <Zeranoe> Does the output bitrate effect encoding speed? [06:05] <anshul> yes encoding speed directly propotional bitrate [06:20] <Zeranoe> anshul: Lower bitrate=faster speeds? [06:21] <anshul> higher birate = faster speeds [06:23] <anshul> since algorithm of encoders time complexity increase when lower bitrate [06:23] <anshul> lower bitRate = more compression [06:23] <anshul> morecompression = high timecomplexity [06:24] <anshul> high timecomplexity = slow speed [06:24] <anshul> hence lower bitrate = slow speed [06:25] <Zeranoe> I'm looking to do some benchmarking of a few encoders. Is there a fall off point where the higher bitrate will start to cause IO problems? [06:26] <Zeranoe> I'm looking for some data for the fastest realtime encoder FFmpeg has. [06:28] <anshul> when bitrate is high then look for swap read write when encoder start using swap then they go slow because of time taken to read write from swap is very high. [06:29] <anshul> so the benchmark for higher bitrate might depend on ram size used and specification of PC [06:29] <anshul> It might also depend on number of application already running on that PC [06:30] <anshul> sp bench mark for dedicated server and personal pc differs a lot [06:31] <Zeranoe> will benchmark results be helpful to other people, or not because of what you just said [06:34] <anshul> benchmark will help in choosing the encoder. [06:49] <anshul> so things wil help if we get time , bitrate , ramsize graph of different encoder [06:54] <Zeranoe> wow vp9 is slow [07:00] <Zeranoe> I think mjpeg is the fastest realtime encoder [07:39] <ubitux> https://blog.mozilla.org/research/2014/03/05/introducing-the-mozjpeg-projec… [07:40] <ubitux> seems it's talking about you pengvado (jpgcrush) [07:41] <ubitux> > So far as we know, no production encoder has this functionality built in [07:41] <ubitux> meh [09:21] <RT|AO> ubitux: mozjpeg breaks in many situations now, at least I found 2. 1 for RGB, another 1 for CMYK/YCCK. [09:22] Action: RT|AO is thinking about writing his own JPEG stripper in C. [09:23] <anshul> I was trying to add closed caption in ffmpeg, I found that some frames startbyte are fc, fe fd which means that they contain EIA-608 and EIA -708 [09:24] <anshul> is it really possible that one stream contain both standurd of closed caption [10:23] <kierank> anshul: yes [10:23] <kierank> For backwards compat [10:24] <kierank> Closed captions are complicated [10:26] <anshul> In ffmpeg we have two codec 608 and 708, if select anything on basis of startbyte it may lead to wrong result [10:31] <anshul> I think selection to be done per frame basis. [10:36] <kierank> Yes it is allowed to change [10:36] <kierank> And it does in reality [11:44] <BBB> omg there is a question on the libvpx mailing list about how to calculate bitrate, the answers people come up with is amazing [11:45] <av500> send it to phoronix [11:46] <av500> codec-devel? [11:46] <ubitux> BBB: hey [11:47] <ubitux> BBB: filter_8tap_2d_fn() in x86/vp9dsp_init [11:47] <ubitux> iirc plepere suggested to make a single merged asm version [11:48] <ubitux> do you think it's worth the effort? [12:00] <cone-797> ffmpeg.git 03Vittorio Giovara 07master:2e708f17083f: ljpeg: fix duplicated pixel format entry [12:00] <cone-797> ffmpeg.git 03Michael Niedermayer 07master:0c402810faf6: Merge commit '2e708f17083fc41f10161f2406d57ec154c75468' [12:15] <cone-797> ffmpeg.git 03Luca Barbato 07master:afdf94689c4e: aac: K&R formatting cosmetics [12:15] <cone-797> ffmpeg.git 03Michael Niedermayer 07master:b5bdd04f0818: Merge remote-tracking branch 'qatar/master' [12:23] <BBB> ubitux: yes! [12:24] <BBB> ubitux: if you're in for that, I'll happily review... not sure how much gain to expect, but a few percent might be possible [12:25] <ubitux> i don't think i'm in currently [12:25] <BBB> av500: yes [12:25] <ubitux> i was just wondering again about that while looking at namespacing the asm symbols [12:25] <BBB> yeah they're both good ideas [12:25] <ubitux> (for consistency i had to rename those into ff_vp9_* [12:25] <ubitux> ) [12:25] <ubitux> (consistency for making it work actually) [12:33] <BBB> did plepere post a new patch on the ml? [12:34] <BBB> I don't see any [12:55] <plepere> BBB : I didn't give a new version [12:56] <plepere> BBB : I've added plenty of stuff, and mraulet wants to review it before me pushing it. [12:57] <BBB> plepere: ok [12:57] <BBB> plepere: when you've got something you'd like me to look at, poke me [12:57] <BBB> plepere: I thought it was pretty good last time, so hopefully after next round it can go in [12:57] <plepere> :D [12:58] <plepere> do you want to see the 2d_8tap I've done ? [12:58] <BBB> yeah [12:58] <plepere> mraulet suggests not using it and just using a C wrapper, though. :( [12:58] <BBB> I hope it doesn't have 7 movdqas per row :-p [12:58] <BBB> but I don't know how else to do it [12:58] <JEEB> :) [12:59] <plepere> well I'm doing shifting registers, so I've got that going for me. :) [13:00] <plepere> BBB : can I send you a patch with everything I've done, so you've got an intermediate version of what I'm probably going to submit ? [13:00] <BBB> sure, but I don't mind waiting another few days [13:01] <BBB> do whatever is faster for you [13:01] <BBB> getting it in is more important than me looking at it 2 days early or so ;) [13:01] <plepere> if I can have some input during the week-end, I'm all for it. :) [13:02] <BBB> ok [13:16] <plepere> BBB : ok, sent. :) [13:17] <BBB> \o/ [13:18] <BBB> will look over the weekend [13:18] <plepere> thanks a bunch. [13:18] <BBB> happy to, let's get it in, maybe I'll want to write some assembly myself later on [13:19] <BBB> I mean, we can mux it into mp4 and x265 is not dog-slow anymore right? so it's actually like something that can be used for real now [13:19] <plepere> BTW, unrelated : I'm starting to look for my next job for the end of the year. Do you know any company interested in a low-level programmer ? [13:19] <BBB> google hires people like that [13:20] <JEEB> too bad google jobs usually mean that one has to relocate and that usually means visas and that usually means that one has to already have papers out of the uni \o/ [13:20] <BBB> email me privately I can list a few companies, also tell me specifically what you want to do (or if you don't know yet), there's a lot of sectors where such skills are useful but they're not all related to multimedia or even in pure engineering firms [13:21] <BBB> although typically out of uni you'll want a pure engineering firms, the other jobs require a few years experience [13:21] <BBB> (most of the time - not always) [13:21] <plepere> JEEB : I want to relocate. :D [13:21] Action: JEEB just got a call from intel [13:22] <Compn> JEEB : are they inviting you to join the cabal of evil ? [13:22] <BBB> work time now, bbl [13:22] <JEEB> Compn, yes -- they want to check if I am as people have told for two hours on tuesday [13:22] <plepere> BBB : x264 is the encoder. :p but yeah, openHEVC is starting to have interesting performance with all this assembly. [13:23] <JEEB> x265 is pretty fast now, preset placebo and refs/bframes maxed is now ~0.4fps with 720p on a i5 3570K [13:24] <plepere> I think once I've got the sao and idct done, we'll be able to decode ra HD @ 50fps on a single core. [13:25] <saigono> 0.4fps is pretty fast? No way [13:25] <saigono> doesn't seem that perfomance increased since summer or autumn '13 [13:26] <JEEB> yeah, but that at least gives you better quality than x264 maxed out :P [13:26] <JEEB> and it is fast when compared to libvpx and VP9 [13:26] <JEEB> you can go faster, but I'm not sure how fast you can go until you start losing the edge on x264 [13:26] <JEEB> anyways, 0.4fps is something I can already kind of use if I have the need to [13:27] <saigono> well, it's still far away from realtime encoding :) [13:28] <JEEB> you missed the "preset placebo" part? [13:28] <saigono> oh [13:28] <saigono> yep [13:29] <plepere> hmmm. [13:30] Action: plepere wonders if he should do a script to convert all his "holiday videos" to h264. [13:45] <mraulet> BBB: as you have the patch and you are on mac, you would probably see the same segfault than me. if you have any idea where it comes from. it would be good to know. [15:00] <yanrong> I'd like to contribute a new fft kernel to FFmpeg, which is optimized for Power platform, single precision float point format. The kernel will be written in c language and will be optimized with PowerPC VSX built-in functions, so that one can use GCC compiler options to switch between big endian and little endian cases without extra code. Is there any comments on this contribution proposal? Appreciate your feedback. [15:01] <av500> powerpc still a thing? [15:02] <Compn> do we want BE/LE runtime code ? [15:03] <Compn> or are mac people used to making those dual-binary programs ? [15:03] <Compn> for ppc and intel-mac [15:03] <Compn> yanrong : sounds good :) [15:04] <yanrong> Thanks. I noticed the current fft kernel is written in assembler code. There maybe some problems when run on little endian Power platforms (e.g. IBM power8), so I want to do some contribution. [16:36] <plepere> ubitux. I don't understand something. [16:37] <plepere> I've doce C wrappers so when doing a filter on a block width of 12 and plus, it calls multiple times the functions for the block width 4 or 8. [16:37] <plepere> *done [16:38] <plepere> I've got a function, it works with the widths 12 and more, but not on the core 4 and 8 functions. [16:38] <plepere> even though the large widths, using the smaller ones are correct ! [17:01] <michaelni> yanrong, how does it perform speedwise compared to existing code ? [17:27] <cone-797> ffmpeg.git 03Michael Niedermayer 07master:2c9e5cb1a698: avcodec/arm/fft_fixed_neon: reduce diff by 2 spaces to libav [17:27] <cone-797> ffmpeg.git 03Peter Ross 07master:7380201451a2: avcodec/adpcm: squelch 'mismatch in coded sample count' warning for AV_CODEC_ID_ADPCM_EA_R2/3 [18:21] <cone-797> ffmpeg.git 03James Almer 07master:722b48bc1256: configure: add missing avx2 support check [18:34] <cone-797> ffmpeg.git 03Diego Biurrun 07master:36b822b8be7f: arm: dsputil: Drop restrict keyword from add_pixels_clamped_armv6 prototype [18:34] <cone-797> ffmpeg.git 03Diego Biurrun 07master:07ea8a7e943e: configure: Group and sort function/struct/type checks [18:34] <cone-797> ffmpeg.git 03Michael Niedermayer 07master:5e5d8ace8a7f: Merge commit '36b822b8be7f9ecd6f9d87acaa786b128a873cd9' [18:34] <cone-797> ffmpeg.git 03Michael Niedermayer 07master:2b94189e961b: Merge commit '07ea8a7e943ea6a7bdcd736e2fc3ed357dc15c57' [18:47] <cone-797> ffmpeg.git 03Diego Biurrun 07master:d0adf55c05be: configure: Split some logical blocks off from CONFIG_LIST [18:47] <cone-797> ffmpeg.git 03Michael Niedermayer 07master:332f1698559e: Merge commit 'd0adf55c05beaf80420747dded2d1884bff377ec' [19:15] <cone-797> ffmpeg.git 03Diego Biurrun 07master:7a650caf6436: configure: Split some logical blocks off from HAVE_LIST [19:15] <cone-797> ffmpeg.git 03Michael Niedermayer 07master:56a498bbba3f: Merge commit '7a650caf6436ddfe417bdcb5ce520c96c779af7b' [19:24] <cone-797> ffmpeg.git 03Vittorio Giovara 07master:999d31d354f1: codec_desc: K&R formatting cosmetics [19:24] <cone-797> ffmpeg.git 03Michael Niedermayer 07master:11e66d20d51f: Merge commit '999d31d354f1f0616a09691913659ac0abdcc899' [19:25] <gueb> hello! [19:33] <cone-797> ffmpeg.git 03Vittorio Giovara 07master:d42ea06798b8: codec_desc: add missing .long_name [19:33] <cone-797> ffmpeg.git 03Michael Niedermayer 07master:4d10b93bd196: Merge commit 'd42ea06798b83d87397449ec6aad40cb5ea46e43' [19:33] <michaelni> hello [19:46] <cone-797> ffmpeg.git 03Vittorio Giovara 07master:f20977921be5: codec_desc: group image codecs together [19:46] <cone-797> ffmpeg.git 03Michael Niedermayer 07master:a8a84fa8b2cd: Merge remote-tracking branch 'qatar/master' [21:51] <gueb> what should I do to manage endianness? [21:51] <gueb> for example [21:52] <gueb> I have audio samples s16le [21:52] <gueb> on x86 I can use it without problem [21:52] <gueb> but on another architectures... [21:52] <JEEB> look at the AV_RLXX/AV_WLXX [21:52] <JEEB> functions [21:53] <JEEB> you can read and write something from/to specific endianness [21:53] <JEEB> or well, macros [21:53] <JEEB> not functions [21:54] <gueb> oh, nice [21:57] <gueb> what is N? [21:57] <JEEB> native [21:58] <JEEB> as in, not specific ÖP [21:58] <JEEB> :P [21:58] <gueb> :S [23:02] <gueb> I am puzzled :S [23:02] <gueb> I am trying to read from frame->data [23:02] <gueb> but I get segmentation fault [23:03] <gueb> I am using two channels [23:03] <gueb> when I read from frame->data[1][0] it breaks [23:03] <gueb> where is the second channel? [23:04] <wm4> gueb: that depends entirely on the audio format [23:04] <wm4> data[1] is used only with non-interleaved format (called planar in ffmpeg) [23:04] <gueb> oh [23:04] <wm4> for normal (interleaved) formats, all data is packed in data[0] [23:05] <gueb> thanks, wm4 [23:06] Action: Compn wonders if libav is asking any other devels to 'reunite' [23:06] <wm4> Compn: what where [23:06] <Compn> wm4 : in private of course [23:06] <Compn> libav way, always private [23:07] <wm4> no [23:07] <wm4> currently they're backporting ffmpeg stuff [23:07] <wm4> because debian possibly wanting to switch to ffmpeg makes them nervous [23:08] <Compn> wm4 : do you troll in their channel ? :P [23:08] <gueb> :O [23:08] Action: Compn wonders who is trolling there [23:08] <wm4> wherever I am I must also troll [23:08] <Compn> :) [23:08] <gueb> I discovered libav just some days ago [23:09] <Compn> gueb : we used to be one project :P [23:09] <gueb> I know, have read about that [23:10] <gueb> ego's programmers strikes back :P [23:13] Action: Compn goes afk [23:14] <gueb> bless valgrind [23:31] <cone-797> ffmpeg.git 03Michael Niedermayer 07master:f6122ed375ab: configure: remove unused identifer that was added by cleanups [23:31] <cone-797> ffmpeg.git 03Michael Niedermayer 07master:c17dea05f746: configure: factorize mips arch extensions out [23:31] <cone-797> ffmpeg.git 03Yu Xiaolei 07master:fbe9ae482d51: avfilter: make avfilter_graph_get_filter use const string name [00:00] --- Sat Mar 15 2014

1 0

ffmpeg.log.20140314
by burek 15 Mar '14

15 Mar '14

[00:12] <stonie_> hello all.. i have a mxf mpeg file with 8 mono audio channels... i need it in stereo.. to have channels 1 and 3 as left and right. how can i do that? [00:18] <llogan> stonie_: probably pan=stereo:FL=FL:FR=FR [00:19] <llogan> oh..1 and 3 [00:19] <llogan> you'll ahve to modify that example [00:19] <llogan> pan=stereo:c0=c1:c1=c2 [00:19] <stonie_> can i define the channel numbers? [00:20] <stonie_> ah ok thanks ill try that soon [00:20] <llogan> see "man ffmpeg-filters" if i'm wrong. [00:20] <stonie_> aye tyvm [01:13] <wacomcito> hi :) [01:14] <wacomcito> my cron daemon don't run a ffmpeg command line. no error, no warning... but if i run the script in bash works perfectly. [01:15] <wacomcito> it's ffmpeg ungry with cron? [01:15] <wacomcito> xDDD [02:01] <relaxed> wacomcito: how are you running it? [02:02] <wacomcito> fron a /bin/sh with root [02:02] <wacomcito> and i all this script from cron [02:02] <wacomcito> call [02:03] <wacomcito> but don't work. [02:03] <wacomcito> ffmpeg not launched, no error, no warnings.. [02:03] <wacomcito> command line: 02 2 * * * /usr/local/bin/ffmpeg -i /var/www/gup/gup-template/procesando/Insiduos_2.avi /dev/null 2>&1 [02:05] <wacomcito> crontab line. [02:07] <wacomcito> syslog -> CRON[13966]: (root) CMD (/usr/local/bin/ffmpeg -i /var/www/gup/gup-template/procesando/Insiduos_2.avi /dev/null 2>&1) [02:07] <wacomcito> no more info... [02:07] <wacomcito> but no ffmpeg proccess. [02:08] <wacomcito> [relaxed] any ideas [02:08] <wacomcito> ? [02:25] <another> wacomcito: check if /dev/null is a video file now [02:26] <wacomcito> xDDDDDD [02:27] <another> the way i experienced /dev/null is, you don't want to overwrite it [02:27] <another> ah forget that [02:28] <wacomcito> xDDD [02:28] <wacomcito> mmm [02:29] <wacomcito> a bulb shine over my head.! [02:29] <wacomcito> thanks another [02:30] <another> ? hokay. [02:47] <wacomcito> well the problem was the space between words in filename. [02:48] <wacomcito> i trate this with " in route file... [02:48] <wacomcito> but not works. [02:49] <wacomcito> may be cron sh enviroment? [03:40] <relaxed> wacomcito: put the command in a script or use "ffmpeg -i input .... output </dev/null" [03:42] <relaxed> because, I think, ffmpeg expects keyboard input [04:26] <llogan> wacomcito: use -y as a global option and ffmpeg wil need -f as an output option to tell it the output format for /dev/null [04:41] <gueb> good night [06:14] <anshul> what is the purpose of writing movie in naming input -i "movie=test.avi[out0]" [06:55] <relaxed> anshul: I think it's used for ffplay [06:57] <anshul> i am using it like this ffmpeg -f lavfi -i "movie=/home/anshul/work/Project/cc_ext/test_5sec.ts[out0+subcc]" -map v frame%08d.png -map s -c copy -f rawvideo subcc.bin [06:58] <anshul> but i If i dont add movie i get following error [lavfi @ 0xa714d60] No such filter: '/home/anshul/work/Project/cc_ext/test_5sec.ts' [07:00] <relaxed> are you hard subbing? [07:02] <relaxed> "-vf movie=input" was used before -filter_complex was created. The latter allows use to use multiple inputs instead of specifying them in the filter chain, example: ffmpeg -i input1.mkv -i input2.mkv ... [07:04] <relaxed> whereas before you used: ffmpeg -i input1.mkv -vf "movie=input2.mkv,..." [07:06] <anshul> do you mean that movie is used when we want to specify two input [07:07] <relaxed> yes, but -filter_complex is the recommended way now. [07:07] <anshul> I am not trying to hard subbing. [07:08] <anshul> ffprobe -filter_complex "movie=/home/anshul/work/Project/cc_ext/test_5sec.ts[out0+subcc]" [07:08] <anshul> the above command does not work [07:08] <anshul> it says Failed to set value 'movie=/home/anshul/work/Project/cc_ext/test_5sec.ts[out0+subcc]' for option 'filter_complex': Option not found [07:09] <relaxed> explain what you're trying to do [07:12] <anshul> http://pastebin.com/wPAy8hNz [07:12] <anshul> Actually i was trying to find out whether the stream contain closed_caption or not [07:13] <relaxed> try, ffprobe -show-streams input [07:14] <anshul> Using libav I was able to do it successfully, one day nicolas made patch for me so that using ffprobe i can check whether stream contain closed caption or not [07:14] <relaxed> er, -show_streams [07:15] <relaxed> ffprobe and a small amount of scripting would work [07:15] <anshul> there will be no useful output of that since closed_caption are interlaced inside the video [07:15] <anshul> mean those subtitle are there in video streams [07:16] <relaxed> Stream #0:1: Subtitle: cea_708 <-- closed caption subs [07:17] <relaxed> so ffprobe or ffmpeg will show them [07:17] <anshul> see output for same video http://pastebin.com/TKBp6W7G [07:18] <anshul> cea_708 is shown when i use -f lavfi -i "movie=/home/anshul/work/Project/cc_ext/test_5sec.ts[out0+subcc]" [07:20] <anshul> My goal is to understand what i am doing, just now i am running the command given to me by nicolas [07:20] <relaxed> what about -show_format? [07:21] <relaxed> can you put this video up somewhere? [07:23] <anshul> http://pastebin.com/9LU3NwdJ show_format [07:24] <anshul> video alone might not help you, you have to also patch the latest git of yours and compile your source code [07:26] <relaxed> humor me [07:26] <SirCmpwn> I'm using ffmpeg to encode videos for playback in the browser [07:26] <SirCmpwn> however, there's a weird decoding issue sometimes [07:27] <SirCmpwn> really weird interlacing shows up, but only when the video's not in full screen (@1920x1080) [07:27] <SirCmpwn> example: https://mediacru.sh/ZUkpscG0k5jj [07:28] <anshul> https://docs.google.com/nonceSigner?nonce=9huu8vt278722&continue=https://do… [07:29] <SirCmpwn> how about a URL that doesn't need wrapping to display fully [07:31] <anshul> https://drive.google.com/file/d/0B4sMK9Wz9oX3cDRHdko1eldTczQ/edit?usp=shari… [07:33] <anshul> patch 1 https://drive.google.com/file/d/0B4sMK9Wz9oX3MXpNbmdSbzVKTHc/edit?usp=shari… [07:33] <anshul> patch 2 https://drive.google.com/file/d/0B4sMK9Wz9oX3aEc3blgxd0FvNXc/edit?usp=shari… [07:34] <anshul> relaxed, i have uploaded all three things [07:36] <anshul> SirCmppwn I think search something similar in track, if nothing found then you should place your bug report with ffmpeg command [07:37] <anshul> SirCmppwn, I am unable to find plugin for your media file in browser [07:38] <anshul> Is behaviour same when you try to play that video using ffplay [07:40] <anshul> relaxed, what is the use of out0 in input [08:07] <relaxed> anshul: They're labels. Have you looked at http://ccextractor.sourceforge.net/ ? [08:09] <anshul> yes, I am trying to intigrate ffmpeg in cc_extractor [08:11] <anshul> relaxed, what is the use of labels? [08:12] <anshul> What purpose is the out0 solving [10:47] <benimnetz> hi everybody. i am seeking a bit of assistance in compiling ffmpeg 2.1.4 on centos 6.5 [10:50] <benimnetz> i largely followed this guide: https://trac.ffmpeg.org/wiki/CentosCompilationGuide [10:53] <relaxed> what's the problem? [10:54] <relaxed> benimnetz: ^^ [10:58] <benimnetz> the i only need a subset of ffmpegs function so i only --enable-opus --enable-libx264 [10:59] <benimnetz> sorry was preparing a pastebin [10:59] <benimnetz> configure gives me ERROR: libx264 not found [11:00] <benimnetz> here a pastebin regarding my problem: http://pastebin.com/DHE39G5p [11:08] <axys^> is vp9 multithreaded yet ? [11:08] <JEEB> depends on if you're talking about decoding or encoding [11:08] <axys^> encoding [11:08] <JEEB> no [11:09] <JEEB> libvpx is still single-threaded with vp9 encoding [11:09] <axys^> also while encoding the fps cosntantly drops [11:09] <axys^> I get from 30fps to 0.1fps [11:09] <relaxed> benimnetz: did you install libx264 git using $HORNET_PREFIX ? [11:10] <JEEB> axys^, yes - it's slow as hell [11:11] <spaam> still slow? =/ [11:12] <benimnetz> zes [11:12] <benimnetz> yes i did [11:14] <benimnetz> for example pkg-config --cflags x264 gives -I/home/bgeese2s-local/hornet_dist/include [11:14] <relaxed> pastebin.com the bottom of config.log [11:17] <benimnetz> http://pastebin.com/vkfUtyEZ [11:18] <relaxed> add -ldl to ldflags [11:18] <relaxed> --enable-fdflags"-ldl -L...." [11:19] <benimnetz> give me second [11:19] <relaxed> s/fdflags/ldflags/ [11:21] <benimnetz> Unknown option "--enable-ldflags" do you mean --extra-ldflags ? [11:21] <relaxed> yes, sorry [11:22] <benimnetz> ok cool, that seems to help! [11:22] <relaxed> llogan: https://trac.ffmpeg.org/wiki/CentosCompilationGuide needs --extra-ldflags="-ldl ..." for libx264. [11:23] <relaxed> oh, I see you have it as --extra-libs="-ldl" [11:23] <axys^> I wonder if there is some active development on multithreading the vp9 encoder [11:24] <relaxed> llogan: maybe a blurb in the x264 section mentioning --extra-libs="-ldl" is required for it. [11:25] <JEEB> axys^, #vp8 would be the channel for that, but so far it seems like multithreading is not on their list of things to do [11:26] <benimnetz> thanks relaxed for the fast help, in the pastebin you can see the commands i used [11:27] <axys^> isn't youtube already streaming vp9 videos ? [11:28] <JEEB> yes, but for them multithreading is just an extra :P Also I think only some clips get encoded as VP9 [11:28] <JEEB> also it's not streaming [11:28] <JEEB> it's just reading already encoded data via http [11:28] <JEEB> because VP9 definitely isn't usable for streaming purposes yet [11:30] <axys^> hm, with all the fuss around 4k TVs I would have expected them to focus more on this aspect [11:31] <axys^> I mean it's almost impossible to encode a 480p video, 4k is out of the question [11:32] <JEEB> well, at least the HEVC implementations seem OK, preset placebo at 720p gives me around 0.4fps. Which is already much faster than libvpx's vp9 [11:32] <JEEB> although neither libvpx's VP9 or libx265's HEVC still have as much visual optimizations as libx264 [14:39] <theholyduck> JEEB, have they gotten aroudn to adding mulithtreading to libvpx vp9 yet? [14:39] <theholyduck> cause the last time i tried, it hasnt [14:40] <theholyduck> lol 14 hours to encode a 3000 frame video [14:41] <theholyduck> means it did 3.5 frames a Minute :P [14:41] <spaam> shit [14:41] <spaam> that is really slow [14:41] <spaam> when did they release vp9 ? in november+ [14:44] <ubitux> 14 hours for 3k frames [14:44] <ubitux> that's actually pretty good [14:44] <ubitux> it took several days for an HD encode of 5k frames last time [14:45] <sacarasc> That is crazy. [15:46] <onefix_work> How can I reencode an MPEG-TS stream without having audio sync problems when I get dropouts? [15:48] <klaxa> what? [15:51] <onefix_work> I have an MPEG-TS stream (from a TV source) that will occasionally have a garbled frame. If I play it in realtime, I see garbage in the video sometimes, but it's always in sync. I would like to reencode it to an MPEG4 while keeping the audio in sync...probably keep it from dropping frames [15:53] <jarainf> onefix_work, seems like you want to ivtc the transport stream. Use -vf "fieldmatch, decimate" to do so. [15:53] <onefix_work> OK, thanks\ [15:54] <jarainf> If there are still some interlaced frames you can try adding yadif=deint=1 between fieldmatch and decimate [15:54] <klaxa> oh that might actually be what he means [15:54] <klaxa> i thought something like transmission errors in the payload [15:57] <onefix_work> right [18:04] <albator> hello [18:04] <albator> i m facing a strange issue. [18:06] <albator> so I wanted to setup nginx php auth [18:07] <albator> and it worked fine using open caster or flash media encoder [18:07] <albator> when I put the same string in the rtmp url , but using ffmpeg [18:07] <albator> it basically shutdown the whole server [18:07] <albator> ffmpeg -i file -c copy - f flv rtmp-server/path&user=xx&pass=xxx [18:08] <albator> without the auth & , no issue ofc [18:18] <Wzee> Hello, I'm hoping someone is able to fill me in on how I can demux a live sliding .m3u8. Starting from the first segment inside the playlist? I've tried ffmpeg -ss 0 -i http://localhost:8888/master.m3u8 -ss 0 -f image2 -vsync 0 path/%02d.jpg [18:32] <llogan> relaxed: feel free to add it. unfortunately i've been neglecting that particular guide lately. [22:01] <albator> absolutely no hint on my issue? :( [22:49] <relaxed> albator: no, but did you single quote the url? [22:50] <albator> nop [22:50] <relaxed> llogan: NO U [23:16] <albator> works with " " yes [23:16] <relaxed> my work here is done [23:40] <relaxed> llogan: done [23:52] <llogan> relaxed: thanks. i made it into a "fancy note". maybe i'll do the same for the rest of the libraries sometime and de-noteize them. [00:00] --- Sat Mar 15 2014

1 0

ffmpeg-devel.log.20140313
by burek 14 Mar '14

14 Mar '14

[00:14] <wm4> Layout indicates a different number of channels than actually present [00:14] <wm4> well fuck [00:14] <wm4> that's what I get for trying to make the code work with both ffmpeg and libav [01:00] <llogan> baptiste: why is ffmbc licensed as GPL v2? why is there no public vcs? [01:53] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:72d44f15834a: mvformat/movenc: fix IMX [03:58] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:12ce58bebdff: doc/texi2pod: fix encoding type [04:58] <cone-365> ffmpeg.git 03James Almer 07master:7d7487e85c06: x86/float_dsp: add ff_vector_{fmul_add, fmac_scalar}_fma3 [07:49] <cone-365> ffmpeg.git 03Michael Niedermayer 07release/1.2:fb44a3e4a3e3: avcodec/vorbisdec: use the stored previous window type only when the actual previous is not known [07:49] <cone-365> ffmpeg.git 03Michael Niedermayer 07release/1.2:f31682f2f2b1: avformat/flvdec: discard inconsistent timestamps [07:49] <cone-365> ffmpeg.git 03Michael Niedermayer 07release/2.1:a9753049b54e: avcodec/vorbisdec: use the stored previous window type only when the actual previous is not known [07:49] <cone-365> ffmpeg.git 03Michael Niedermayer 07release/2.1:4362a222e2d2: avformat/flvdec: discard inconsistent timestamps [07:49] <cone-365> ffmpeg.git 03Michael Niedermayer 07release/2.1:3712b398e9e9: avidec: calculate missing bitrates from index [07:49] <cone-365> ffmpeg.git 03Michael Niedermayer 07release/2.1:3e73bea05f71: avformat/avidec: Check required demuxing buffer sizes in guess_ni_flag() [07:49] <cone-365> ffmpeg.git 03Michael Niedermayer 07release/2.1:c7a854af1901: mvformat/movenc: fix IMX [08:20] <anshul> this command is not working at my place ./ffmpeg_g f lavfi -i "movie=test.ts[out0+subcc]" -map v frame%08d.png -map s -c copy -f rawvideo subcc.bin [08:20] <anshul> it saying no such file movie=test.ts[out0+subcc] [10:58] <Compn> anshul : you are missing a - in front of "f lavfi" [10:58] <Compn> not sure if thats a copy paste error or if that is your problem, of course [11:13] <durandal_1707> llogan: i like your EncodeforYouTube page, exp. musicfilter.jpg maybe i will get motivation to write yet another avf filter :) [11:20] <anshul> Compn Thanks,I really missed - , it was my bad, it was copy paste mistake from document to command line [11:24] <Compn> anshul : no problem :) but there is an actual bug, i think ffmpeg error message could be improved upon [11:24] <Compn> anshul : aka if ffmpeg sees 'f' it could throw a warning [14:34] <cone-365> ffmpeg.git 03Luca Barbato 07master:81498ceb5b2d: avconv: Match stream id [14:34] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:bb6de15d927d: Merge remote-tracking branch 'qatar/master' [14:41] <ubitux> lol [14:42] <ubitux> let's make the exact same thing with a different separator [15:04] <Compn> is timothy gu on irc ? [15:04] <Compn> is that daemon404 ? [15:07] <cone-365> ffmpeg.git 03Carl Eugen Hoyos 07master:aac9e5675935: Fix hardware detection on aix. [15:07] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:f385a099c199: Merge remote-tracking branch 'cehoyos/master' [15:18] <plepere> ubitux : does a bus error:10 tell you anything ? mraulet gets it when testing my code on mac. :/ [15:33] <ubitux> plepere: bus error = seg fault [15:33] <ubitux> i can't guess. [15:33] <plepere> ok thanks [16:55] <ubitux> bingo, diego pushed the commit with the spaces before casts [16:56] <wm4> how terrible [16:56] <ubitux> wm4: several people asked him not to [16:56] <ubitux> wm4: it makes no sense, and it's now inconsistent and less readable [16:57] <ubitux> i love how review is basically pointless [16:58] <ubitux> also fun to see him introduce his own way of indenting asm code [16:58] <ubitux> oh well.. [17:14] <ubitux> any idea how it's possible to "ldr r0, =foobar" with AS on apple? [17:15] <ubitux> (foobar being a label, i want the address) [17:16] <ubitux> i'm doing a movw r0,#:lower15:foobar + movt r0,#:upper16:foobar [17:16] <ubitux> but it sucks really hard [17:16] <ubitux> i don't know if gas-pp is doing some magic to that ldr r0,=foobar syntax, i'm not able to try that [17:30] <ubitux> > The iOS assembler does not support the =258014 pseudo-immediate form. You will need to either replace these with loads from a constant pool, or split them into mov + movt [17:30] <ubitux> ok.. [17:42] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:8b02dfd37cb3: avutil/timestamp: Warn about missing __STDC_FORMAT_MACROS for C++ use [18:28] <llogan> Compn: Daemon404 is Derek Buitenhuis. Timothy is Timothy_Gu, IIRC, but doesn't seem to be here often. [18:31] <llogan> bwheatenhise? how is that pronounced anyway? [18:57] <ubitux> https://lists.libav.org/pipermail/libav-devel/2014-March/057429.html T_T [18:58] <ubitux> git blame killer [19:00] <av500> alwys blame libav :) [19:01] <ubitux> that's actually where you want git blame to work [19:01] <ubitux> like "when was the support for such id was added" [19:02] <ubitux> and blame -w won't help you anymore [19:02] <ubitux> BBB: ping [19:08] <llogan> kierank: can you confirm that emotion fixed their violation? [19:09] <kierank> I can do when I get home [19:10] <llogan> thanks. no hurry. [19:11] <llogan> or i guess i can stop being a lazy ass and start my Windows VM... [19:25] <wm4> ubitux: lol what's the point [19:25] <ubitux> generating fake activity [19:34] <gueb> hello! [19:36] <gueb> I am trying to add a dummy codec, but ffmpeg says that "Codec 0x73736d77 is not in the full list." [19:36] <gueb> which did I wrong? [19:37] <gueb> what [19:37] <gueb> :S [19:37] <ubitux> can you show a diff? [19:39] <gueb> I put an entry in enum AVCodecID {...} [19:39] <gueb> another one in avcodec_register_all() [19:39] <gueb> and in the Makefile [19:39] <gueb> that was all [19:39] <ubitux> what about libavcodec/codec_desc.c? [19:39] <gueb> :O [19:42] <gueb> :D [19:42] <gueb> it works! [19:43] <gueb> thank you! [19:45] <gueb> it is curious [19:46] <gueb> when the ff_foo_decoder was almost empty [19:46] <gueb> ffmpeg -codecs did not show D flag [19:47] <gueb> with .init = (void *)-1, .decode = (void *)-1, ... [19:47] <gueb> it shows that [19:47] <gueb> it seems that it checks for null pointers [19:47] <ubitux> AVCodec are shared between encoders and decoders [19:48] <ubitux> if .decode callback is present it's a decoder [19:48] <ubitux> (i guess) [19:48] <gueb> now let's to fill the encoder [19:49] <gueb> I have read that a frame can contain some packers, or vice versa [19:50] <gueb> wow, I have forgotten it in few minutes! [19:50] <gueb> what is a frame and a packet? [19:50] <ubitux> see http://ffmpeg.org/ffmpeg.html#Detailed-description [19:52] <gueb> so a packet has frames [19:52] <gueb> can a packet have more than 1 frame of the same stream? [19:52] <gueb> source? [19:54] <gueb> for example, 3 packets of the same audio track [19:54] <gueb> not packets [19:54] <gueb> frames [19:54] <gueb> 1 packet with 3 frames of the same audio track [19:54] <gueb> is it posible? [19:55] Action: gueb has the feeling of being annoying ubitux [19:55] <gueb> sorry [19:57] <ubitux> i'm not annoyed [19:57] <gueb> :) [19:57] <gueb> I ask too much [19:57] <ubitux> generally you have 1+ packet(s) for 1 frame [19:57] <ubitux> most of the time 1 [19:59] <wm4> gueb: for video, 1 frame 1 packet is required [19:59] <wm4> for audio, there can be multiple frames per packet [19:59] <gueb> I want to make an audio codec [20:00] <ubitux> ah sorry my bad multiple frames per packets yeah [20:00] <ubitux> the packet need to be consumed [20:01] <ubitux> http://git.videolan.org/?p=ffmpeg.git;a=blob;f=doc/examples/demuxing_decodi… [20:01] <gueb> is it a decision of ffmpeg or the programmer of the codec? [20:01] <ubitux> basically this [20:09] <gueb> from av_read_frame() documentation [20:09] <gueb> "For audio, it contains an integer number of frames if each [20:09] <gueb> * frame has a known fixed size (e.g. PCM or ADPCM data). If the audio frames [20:09] <gueb> * have a variable size (e.g. MPEG audio), then it contains one frame. [20:09] <gueb> talking about pkt->buf [20:10] <gueb> I am going to dinner [20:13] <lindenle> Hi all. I was wondering if there is a standard set of tricks to make demuxing and muxing work faster on android phones. I have adde hwaccels and passed libx264 the fastdecode parameter...but thought there might be some expertise in this area somewhere online. [20:24] <lindenle> Another question how do i set the "tune" or "preset" in libx264 through libavcodec? [20:26] <Compn> in ffmpeg or in a 3rd party application ? [20:26] <Compn> btw in ffmpeg, h264 decoding is done with libavcodec. x264 is an encoder only [21:02] <gueb> then, for encode [21:02] <gueb> I have to get data from AVFrame to put it in AVPacket? [21:06] <cone-365> ffmpeg.git 03Reinhard Tartler 07master:fb9a5bdc516b: Prepare for 11_alpha1 Release [21:06] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:a47dfdfa7dcf: Merge commit 'fb9a5bdc516b75b4ba1e36ca2223dec46571dea3' [21:16] <cone-365> ffmpeg.git 03Diego Biurrun 07master:88b240646f57: build: Drop stray mpc8 decoder dependency on dsputil [21:16] <cone-365> ffmpeg.git 03Diego Biurrun 07master:7d7be44f3e8b: build: Drop stray PNG decoder dependency on dsputil [21:16] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:4b50de88a4c0: Merge commit '88b240646f57a62299c6c4c4d7554f0a80c3a29a' [21:16] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:ecf16f039def: Merge commit '7d7be44f3e8bc14224540a477ef0607f18964421' [21:20] <mankan> Hi all. I've been tweaking x265 encoding to include vps/sps/pps before every keyframe and I wonder what would be suitable config parameter to enable this mode? [21:22] <cone-365> ffmpeg.git 03Diego Biurrun 07master:9ecf1b62f37e: build: Drop stray MLP decoder dependency on dsputil [21:22] <cone-365> ffmpeg.git 03Diego Biurrun 07master:6cd76e5412d7: build: Drop stray MJPEG encoder dependency on dsputil [21:22] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:07dbdcf9d0d4: Merge commit '9ecf1b62f37ef9cbdf2d7300a052cd704c4ba731' [21:22] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:2631d155baa5: Merge commit '6cd76e5412d794dbd1cc6ab0e4c9f2ee53cf028f' [21:33] <cone-365> ffmpeg.git 03Diego Biurrun 07master:0e29c5e1285e: build: Drop stray LPC dependency on dsputil [21:33] <cone-365> ffmpeg.git 03Diego Biurrun 07master:cd832f032f94: ffv1: Drop bogus dependency on dsputil [21:33] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:981aa023387e: Merge commit '0e29c5e1285eeb5b4842d96d00dfe06898cd87e5' [21:33] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:daded4f89a77: Merge commit 'cd832f032f945a57a1f1dd385942c2fadc7ce4e5' [21:46] <cone-365> ffmpeg.git 03Diego Biurrun 07master:b5f112d82093: configure: Add explicit dsputil dependency to mpegvideoenc [21:46] <cone-365> ffmpeg.git 03Diego Biurrun 07master:bbb64356cca6: build: Record dependency of eatqi decoder on dsputil [21:46] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:96b02460fce5: Merge commit 'b5f112d820934ade66e809ef30bfcfc8210929f8' [21:46] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:cbfc9046e1c7: Merge commit 'bbb64356cca65d7a2a33f59b2267ee58f28a061a' [21:55] <gueb> what is the attribute "av_cold" for? [22:10] <ubitux> gueb: http://gcc.gnu.org/onlinedocs/gcc/Function-Attributes.html [22:10] <ubitux> look for "cold" [22:11] <gueb> oh [22:14] <cone-365> ffmpeg.git 03Diego Biurrun 07master:d6096a674225: Remove all SH4 architecture optimizations [22:14] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:7ac1e27f0f06: Merge commit 'd6096a67422534918405abb46dafbbac4608cbc3' [22:18] <gueb> which container should I use with a new codec? [22:19] <gueb> [matroska @ 0x30ed1c0] No wav codec tag found for codec foo [22:19] <gueb> :S [22:19] <gueb> have I to create a container too? [22:20] <gueb> avi does not complaint [22:20] <gueb> but ffmpeg says that the codec parameters are incorrect [22:21] <ubitux> are you trying to design a new codec? [22:21] <gueb> yes [22:22] <ubitux> audio only? [22:22] <gueb> yes [22:23] <ubitux> not sure nut would be relevant for audio only [22:23] <ubitux> you make your own dummy demuxer [22:23] <ubitux> or just start with raw [22:24] <wm4> it should be pretty simple to add a codec to matroska [22:24] <wm4> or avi, for that matter [22:24] <gueb> I am going to watch TV now, would you explain me that later, please? [22:25] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:fb1b70c1ed50: libavcodec/sh4: add note about the removial [22:25] <gueb> thanks for all [22:25] <gueb> see you later [22:36] <cone-365> ffmpeg.git 03Diego Biurrun 07master:b4dd424d96f0: Remove all SPARC architecture optimizations [22:36] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:3fc2362aef5a: Merge commit 'b4dd424d96f09f9bafb88e47f37df65dc4529143' [22:54] <cone-365> ffmpeg.git 03Diego Biurrun 07master:08dba0e1c3df: x86: mpegvideoenc: Remove some remnants of the long-gone libmpeg2 IDCT [22:54] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:b7cffc81b7ed: avcodec/sparc: add note about the removial [22:54] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:8ccf21839ad8: Merge commit '08dba0e1c3df3b71270ebce4527bffa155b91b97' [23:06] <cone-365> ffmpeg.git 03Diego Biurrun 07master:49676eb7301e: dsputil: Remove prototypes for nonexisting optimization functions [23:06] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:1306359ea9cd: Merge commit '49676eb7301e775d08bdbba5380159b106ee258f' [23:25] <cone-365> ffmpeg.git 03Diego Biurrun 07master:9103185bd116: dsputil: Drop ff_check_alignment() function [23:25] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:f5e6b49cd23b: Merge commit '9103185bd116930f90b847090e66a64fa9971ce2' [23:40] <cone-365> ffmpeg.git 03Diego Biurrun 07master:ed9625eb62be: dsputil: Move intreadwrite.h #include from header file to .c files [23:40] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:5dce8e56c61f: Merge commit 'ed9625eb62be1e1c44cecdd73ea0d80077a15d48' [23:59] <cone-365> ffmpeg.git 03Diego Biurrun 07master:fd9e2221bd3b: ppc: Add some missing headers [23:59] <cone-365> ffmpeg.git 03Diego Biurrun 07master:17608f6ee3d2: x86: Add some more missing headers [23:59] <cone-365> ffmpeg.git 03Diego Biurrun 07master:d1184b8110b4: arm: dsputil: Add a bunch of missing #includes [23:59] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:e52d97a7aceb: Merge commit 'fd9e2221bd3b2c5d62605134b62efa26eb9ef0f7' [23:59] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:d61e1156beae: Merge commit '17608f6ee3d2088cdb8d1e704276d8b34f01160d' [23:59] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:be879af217e8: Merge commit 'd1184b8110b4847013bf25222e6809eb3462913c' [00:00] --- Fri Mar 14 2014

1 0

ffmpeg.log.20140313
by burek 14 Mar '14

14 Mar '14

[09:18] <termos> When using libav across multiple threads, is there a list of any pitfalls I should look out for? [09:56] <k0bi> is it posible to extract img and audio from rtmp stream and build it on site Synchrony [10:20] <kvz> Can I tell ffmpeg to discard streams it doesn't understand? I have a file with 3 streams, 2 of which it doesn't understand. I would like it continue with the one it does. Here's the paste: http://pastebin.com/DhGgFkbL [11:44] <relaxed> kvz: ffmpeg -i input -map 0:v:0 ... [11:46] <kvz> relaxed: I'm looking for a syntax where I can leave out streams based on codec. I think the proposed syntax hardcodes by stream index & type = video? [11:47] <kvz> relaxed: because sometimes people throw files at the same command, that have a perfectly sane video #0:1 stream. and by applying -map 0:v:0 now that one is lost. Correct? [11:51] <relaxed> you would have to script it [12:16] <kvz> relaxed: ok, thank you! [13:16] <steinchen> hello all.. i have a mxf file with 8 mono audio channels... i need it in stereo.. have channels 1 and 3 to left and right stereo [15:59] <asten> Hi. I have a question on h264 decoding process. Is it a good place to ask? [15:59] <average> asten: well.. ask the question [15:59] <average> asten: if someone knows, and they have time to answer you, and if they want to answer you.. if if if if if , then you'll get an answer [16:00] <asten> http://stackoverflow.com/questions/22382040/h264-cavlc-decoding-coeff-token [16:51] <srcreigh> hello [16:58] <srcreigh> I have a question about configure flags. [16:58] <srcreigh> I'm trying to compile FFmpeg minimally so that I can do h264 encoding to mp4. [16:58] <srcreigh> Here's my current configuration: http://pastebin.com/dz12JL36 [16:59] <sacarasc> You disabled ffmpeg? [17:00] <srcreigh> I'm using av* libs in C code directly. [17:00] <srcreigh> With that configuration, avformat_alloc_output_context2(&outputCtx, NULL, "mp4", filename); results in a null output context. [17:01] <srcreigh> Sorry, I should have said "I'm trying to compile av* libraries to do h264 mp4 encoding" [17:02] <Zeranoe> Is there any encoder faster than mjpeg [17:04] <Zeranoe> I feel like there are no good status for encoder speed or even quailty with modern hardward and modern codecs. I might need to fill that void [17:07] <srcreigh> Update: I just compiled FFmpeg without the --disable-everything flag, and it worked! [17:07] <srcreigh> Well, the code ran, Which is good. I still would like to minimize the build. [17:11] <srcreigh> I just need to figure out which ./configure flags to pass in after --disable-everything to enable h264 encoding and mp4 muxing. [17:44] <jarainf> srcreigh, why don't you use libx264 to encode whatever you are encoding? [17:44] <jarainf> as in: not using ffmpeg at all [17:46] <srcreigh> jarainf: I haven't looked at using libx264 directly. FFmpeg's muxing.c is a really great example for encoding as far as I can tell. [17:46] <sacarasc> Because they want the MP4 muxing. [17:47] <srcreigh> jarainf: also what sacarasc just said [17:47] <srcreigh> and my project already has ffmpeg, so why not continue using it if it supports libx264 [17:48] <srcreigh> My question is really about mp4 muxing, I'm pretty sure that --enable-libx264 will give me h264 encoding. [17:48] <srcreigh> --enable-muxer=mp4 doesn't seem to be enough for avformat to give me an mp4 output context for whatever reason. [17:49] <Zeranoe> srcreigh: Your stripping the libs? [17:49] <srcreigh> Zeranoe: yeah [17:52] <Zeranoe> In my experience, it isn't easy to single out a single component to be compiled [17:53] <srcreigh> Well, I've already done it for mp4/h264 demuxing and decoding. as you can see from the pastebin I sent here [17:53] <srcreigh> (http://pastebin.com/dz12JL36) [18:01] <srcreigh> I wonder if it's possible to make some kind of documentation for lib stripping. [18:03] <srcreigh> Or get some logs that will say, like, "[avformat_alloc_output_context] OPENING MUXER foo" [18:37] <llogan> Zeranoe: rawvideo, libx264, huffyuv, mpeg2video is faster. probably a bunch of others too. [18:37] <llogan> mpeg4 [18:37] <Zeranoe> it seems that for very low latency it preforms well, better than x264 -crf 0 [18:38] <llogan> i assumed encoding speed [18:38] <Zeranoe> yes [18:52] <srcreigh> test [19:20] <Zeranoe> How is does ffmpeg encode to gsm? I tried -c:a libgsm -f gsm file.gsm, and -c:a libgsm -f gsm out.wav [19:24] <geqsaw> hi, I want to grap a video from my webcam, but my cpu is not fast enought to encode it with x264 in real time, and withoud encoding I get huge files. Which codec should I use to get relatively small videos without too much cpu load? [19:25] <srcreigh> geqsaw: When I tried a bunch of different encoding formats yesterday, flv seemed to be very fast [19:26] <llogan> flv is a container format. do you mean the flv1 encoder? that's old and shitty. [19:26] <Zeranoe> ffmpegs first [19:27] <llogan> geqsaw: use a pastebin site to show the complete output of: ffmpeg -f v4l2 -list_formats all -i /dev/video0 [19:27] <llogan> assuming /dev/video0 is your actual input name [19:27] <Zeranoe> assuming your on Linux [19:29] <srcreigh> well, whatever encoder you get from ffmpeg guessing an encoder by filename for "foo.flv" [19:29] <srcreigh> was about an order of magnitude faster than foo.mp4 [19:29] <srcreigh> (I was playing around with muxing.c yesterday) [19:29] <llogan> the defalt encoder can depend on how your ffmpeg was configured [19:29] <Mavrik> geqsaw, you can control CPU use/quality tradeoff by using presets [19:29] <llogan> for example, ffmpeg will use libx264 for flv, and if that is not available then it will use flv1 [19:30] <geqsaw> http://pastebin.com/Ksh99E84 [19:30] <srcreigh> llogan: huh, that's interesting. I had libx264. I wonder why the mp4 format was so much slower than flv. [19:30] <llogan> i'm not sure. i'm also only familiar with the cli tool. [19:31] <llogan> geqsaw: i was hoping your device would already provide H.264 video so you can simply stream copy, but i guess not. [19:32] <Zeranoe> llogan: Do you know if there can be a IO bottleneck with the steam size of rawvideo? [19:32] <Zeranoe> or is it not *that* big [19:33] <llogan> you can test with: ffmpeg -i input -map 0:v:0 -c:v rawvideo -f null - [19:33] <Zeranoe> and look for dropped frames? [19:34] <llogan> geqsaw: what kind of hardware do you have that can't encode 640x480 (or smaller) fast enough? [19:34] <llogan> Zeranoe: you can look at the encoding speed with null, and then compare it to a local file output and see if there is much of a difference [19:34] <geqsaw> llogan: I have a intel i5 @ 1.8 Ghz dual core [19:37] <Zeranoe> whats the container for rawvideo [19:38] <Zeranoe> and is rawvideo bascially just the decoded stream? [19:39] <geqsaw> I do not have a abnormal output, only after a few seconds the cpu is at 100% and ffmpeg is freezing for a couple of seconds... [19:41] <llogan> then without any command and output i can only recommend to "use git head" for x264 and ffmpeg [19:45] <geqsaw> llogan: with "git head" you mean the latest version of ffmpeg from github? [19:46] <llogan> the latest from git://source.ffmpeg.org [19:46] <llogan> and make sure to use yasm [20:55] <lifeofgu_> is there any trick to compile ffmpeg for intel atom cpus? [20:55] <lifeofgu_> on intel i7 ffmpeg is using all cores/threads, on intel atom it is only using one core/thread [21:23] <mankan> Hi all. I've been tweaking x265 encoding to include vps/sps/pps before every keyframe and I wonder what would be suitable config parameter to enable this mode? [21:48] <gers4302> I'm muxing multiple audio / video files. Everything seems to work fine with the output in VLC, but in WMP or QuickTime, it messes up. I'm getting encoded packets and using av_write_frame to write it back out. If the audio files doesn't start for 10 seconds, WMP will just start playing the audio right away. Is there a proper way to write frames that are offset using av_write_frame? [00:00] --- Fri Mar 14 2014

1 0

ffmpeg-devel.log.20140312
by burek 13 Mar '14

13 Mar '14

[02:20] <cone-924> ffmpeg.git 03Luca Barbato 07master:db9d39b4b5e5: avformat: Report the duration analysis reached [02:20] <cone-924> ffmpeg.git 03Michael Niedermayer 07master:85e1368f57a6: Merge commit 'db9d39b4b5e5a3c20aeecf787ddeadd88f4906cf' [02:28] <cone-924> ffmpeg.git 03Luca Barbato 07master:390acbea0697: configure: Provide --pkg-config-flags [02:28] <cone-924> ffmpeg.git 03Michael Niedermayer 07master:3a05d7a9e583: Merge commit '390acbea0697a60300f249602dbf701e04274693' [02:41] <cone-924> ffmpeg.git 03Luca Barbato 07master:55a215ba63d9: http: Return meaningful error codes [02:41] <cone-924> ffmpeg.git 03Michael Niedermayer 07master:03fd80dcb1f8: Merge commit '55a215ba63d9fa79cd7ee265ee2e777ee86b200c' [02:44] <cone-924> ffmpeg.git 03Luca Barbato 07master:78b21c1d7177: http: Drop doxy comments [02:44] <cone-924> ffmpeg.git 03Michael Niedermayer 07master:b752d02f42bb: Merge commit '78b21c1d7177e1d61ad3c9225f67699da089aa7c' [02:50] <cone-924> ffmpeg.git 03Luca Barbato 07master:7a2fddb44801: http: K&R formatting cosmetics [02:50] <cone-924> ffmpeg.git 03Michael Niedermayer 07master:df41cbee85c8: Merge commit '7a2fddb4480121712df560cf619c1c3566cae3ff' [02:58] <cone-924> ffmpeg.git 03Luca Barbato 07master:4ff99ab3d7d5: http: Refactor process_line [02:58] <cone-924> ffmpeg.git 03Michael Niedermayer 07master:21d4c571fabe: Merge commit '4ff99ab3d7d5576e99e6b8a411b4a44500ed88fa' [03:25] <cone-924> ffmpeg.git 03Luca Barbato 07master:8075c3d8bb1f: http: Add support reading ICY metadata [03:25] <cone-924> ffmpeg.git 03Michael Niedermayer 07master:ca2369cdee0d: Merge commit '8075c3d8bb1f6aade0cc7c5c40db9bc1bcd84cab' [03:29] <Guest52893> hello, just to try this irc. [03:38] <cone-924> ffmpeg.git 03Michael Niedermayer 07master:e58c85b06868: http: Export Content-Type information [03:38] <cone-924> ffmpeg.git 03Michael Niedermayer 07master:c03af3ac1c4c: Merge commit 'e58c85b0686892960042232e51c77168b264838a' [03:40] <yanrong> s [03:42] <hightallyht> hi yanrong [03:55] <cone-924> ffmpeg.git 03Clément BSsch 07master:ddfc98906373: http: Support setting custom User-Agent [03:55] <cone-924> ffmpeg.git 03Michael Niedermayer 07master:f36da16ede52: Merge commit 'ddfc98906373d1f17f6205cedd14c68d7a75995f' [04:28] <cone-924> ffmpeg.git 03Anssi Hannula 07master:2ec33d271272: http: Add support for selecting a request range [04:28] <cone-924> ffmpeg.git 03Michael Niedermayer 07master:76a939d0e58c: Merge commit '2ec33d27127251bbc45e1f88e60691ad59cf2319' [04:37] <cone-924> ffmpeg.git 03Anssi Hannula 07master:ab76d9f628ad: http: Always allow no-op seek [04:37] <cone-924> ffmpeg.git 03Michael Niedermayer 07master:1cc9468289b2: Merge commit 'ab76d9f628ad46e1d3bbf26c5bf1f87083f239ab' [04:51] <cone-924> ffmpeg.git 03Alessandro Ghedini 07master:fe568b3d27ca: http: Improve options descriptions [04:52] <cone-924> ffmpeg.git 03Michael Niedermayer 07master:0d487654ea85: Merge commit 'fe568b3d27ca2c5cca3878b2a7a3a968e605aec4' [04:58] <cone-924> ffmpeg.git 03Clément BSsch 07master:2572d07c1f0a: http: Allow setting a Content-Type for POST requests [04:58] <cone-924> ffmpeg.git 03Michael Niedermayer 07master:cd25412f5984: Merge remote-tracking branch 'qatar/master' [05:18] <cone-924> ffmpeg.git 03Michael Niedermayer 07master:1f36ebf63a8d: avformat: revert %c changes from d92024f18fa3d69937cb2575f3a8bf973df02430 [09:10] <ubitux> > FreeType can now use the HarfBuzz library [09:10] <ubitux> mmh [09:10] <av500> ah harfbuzz [09:10] <av500> arabic all the things! [09:11] <ubitux> http://sourceforge.net/projects/freetype/files/freetype2/2.5.3/ [12:32] <cone-365> ffmpeg.git 03Carl Eugen Hoyos 07master:54bbe3e2a645: Revert "Allow stream-copying grayscale mov files." [12:32] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:3ddf76ee070d: Merge remote-tracking branch 'cehoyos/master' [13:01] <J_Darnley> If git send-email says "5.1.0 <y> invalid address" as its last line, is that en error from the smtp server? [13:05] <J_Darnley> Nevermind. The emails were sent when I tried again [13:19] <cone-365> ffmpeg.git 03Stephen Hutchinson 07master:5336cd6374c7: doc/general.texi: Adjust the notes on AviSynth [13:19] <cone-365> ffmpeg.git 03Bryce McLeod 07master:581957cd8618: avformat/movenc: rescale when copying duration to timecode track [13:21] Action: durandal11707 awesome, now just to add mt frame audio encoder [13:22] <durandal11707> wm4: is tak seeking really that slow? [13:22] <wm4> durandal11707: not slow, but imprecise [13:22] <wm4> I think it'll even output wrong timestamps after seeking [13:23] <durandal11707> that is because duration is usually set in header every ~15th frame [13:24] <wm4> it's probably hopeless to try to seek correctly, and the only way is creating an index in advance [13:24] <durandal11707> but it actually should never change, so maybe i need to set flag or something... [13:25] <durandal11707> wrong timestamps do not happen always... [13:32] <durandal11707> it just seems that parser context is lost when seeking [13:33] <durandal11707> so, creating index in advance by some way is right solution [13:33] <nevcairiel> the parser is flushed, not sure how much of the data survives that? [13:33] <durandal11707> but lavf should already do that... [13:33] <nevcairiel> it doesn't do that in advance, only in areas you already played [13:34] <durandal11707> but for areas not played it should parse all packets up until it finds such timestamp [13:35] <durandal11707> so perhaps index does not have duration set by parser... [13:36] <durandal11707> so parser is run once to seek and again to play... or i'm speaking shit [13:39] <durandal11707> but anyway, mpv should for duration of frame use decoder samples and not stuff reported by lavf/whatever [13:44] <wm4> what's that about duration? [13:44] <wm4> I never use packet/frame duration, except for subtitles [13:50] <cone-365> ffmpeg.git 03Fred Rothganger 07master:3f3229cd109f: avformat: extracting NTP timestamp from RTCP [13:51] <ubitux> wm4: speaking of subtitles, i don't know if it's a bug in ffmpeg, but when 2 events appear at the same time i have the 2 one appear on top of the first [13:52] <ubitux> is that due to libass? mpv internal? bug in ffmpeg? [13:52] <ubitux> iirc we sort them by timestamps, then by position, so it shouldn't have happened [13:52] <ubitux> (it was a .srt file) [13:52] <ubitux> s/2 one/2nd one/ [13:53] <wm4> probably libass [13:53] <wm4> seems logical that the 2nd appears top, because well, the 1st takes the normal place, and the 2nd has to put where there's still place [13:54] <ubitux> yeah right [13:54] <ubitux> couldn't it be changed? [13:55] <ubitux> i mean, it's extremely confusing when it happens :) [13:55] <wm4> one could manually merge events that happen at the same time [13:55] <wm4> with what format did this happen? [13:55] <ubitux> srt [13:56] <ubitux> wm4: it's hard to merge properly [13:56] <ubitux> it would work with easy formats such as srt [13:56] <wm4> I don't think I really want to touch the libass collision logic [13:56] <ubitux> but it won't with all the other [13:56] <wm4> and it'd probably be wrong for ASS [13:57] <wm4> usually srt uses multi line events for multi line events [13:57] <ubitux> could you insert them backward for event at the ssame time? [13:57] <wm4> I don't think srt is even supposed to be overlapping (but then it's a wishy washy half-standard) [13:57] <wm4> yes [13:58] <wm4> ubitux: can I get a sample? [13:58] <ubitux> sure [13:59] <ubitux> wm4: just the srt? [13:59] <wm4> yes [14:04] <ubitux> wm4: http://b.pkh.me/Shokuzai.E05.720p.HDTV.x264.AAC-YYeTs.srt [14:04] <ubitux> grep -- '-->' Shokuzai.E05.720p.HDTV.x264.AAC-YYeTs.srt > a [14:04] <ubitux> grep -- '-->' Shokuzai.E05.720p.HDTV.x264.AAC-YYeTs.srt|uniq > b [14:05] <ubitux> diff -u a b [14:05] <ubitux> you'll get the different ts where that happens [14:05] <wm4> so they're just successive; maybe libavformat could do this automatically [14:06] <ubitux> as i said, yes we could merge them [14:07] <ubitux> but that won't be a generic solution for all the formats [14:07] <wm4> that is why srt could set a flag or call a function to do it [14:07] <ubitux> it will happen in others, where it's not obviously possible to merge [14:07] <ubitux> why don't you just swap the order of rendering? [14:08] <ubitux> i'm pretty sure i can found (or craft since i'm evil) for a format where we can't merge them [14:08] <ubitux> and where the order would matter [14:14] <wm4> that's why I said to merge in srt only [14:14] <wm4> and swapping the order of rendering? I'm just passing packets to the decoder [14:14] <bove> Has anyone been looking at integrating >8bit depths to vf_lut.c? [14:22] <durandal11707> wm4: when i seek tak in mpv it shows ??? for some short time, because packets are missing duration [14:25] <wm4> durandal11707: did you mean ffplay (not mpv) and packet timestamps (not duration)? [14:26] <ubitux> wm4: i suppose that if the subtitles decoders were outputing valid "ass" (so properly "ass ordered") that would help? [14:26] <wm4> ubitux: probably [14:26] <ubitux> oh well that ass insanity [14:26] <ubitux> i just forgot i don't want to touch it [14:27] <ubitux> bove: http://trac.ffmpeg.org/ticket/2475 [14:27] <wm4> but really, the best way would probably be merging them in the demuxer, since srt normally does not have overlapping events, and other subtitle renderers/converters will also have problems with that [14:28] <ubitux> i guess i'll add a method to lavf/subtitles.c [14:28] <ubitux> to auto merge on request [14:28] <ubitux> for "compatible" formats [14:29] <ubitux> wm4: now, what if the duration is different? [14:30] <bove> ubitux: Thanks. I'll put it on my todo list then ... [14:30] <wm4> ubitux: don't merge [14:30] <ubitux> wm4: and so i'll have the exact problem [14:30] <ubitux> +same [14:37] <ubitux> well that will work for simple case anyway [14:40] <ubitux> that's actually pretty trivial to do [14:41] <wm4> also send a hitman to the ones who produced the subs [14:41] <ubitux> :) [14:42] <ubitux> i've seen much worse [14:43] <ubitux> ok so not even compiled but i guess something like that would do the trick http://pastie.org/8911390 [14:44] <wm4> that's all? [14:44] <ubitux> with a ';' at the end tyes [14:45] <ubitux> we already have the merge logic available [14:45] <wm4> sounds like this is the way to go [14:45] <ubitux> until i found samples with different duration& [14:45] <ubitux> or misordered [14:51] <durandal11707> wm4: mpv, not ffplay. and timestamps displays ??? when seeking in mpv [14:53] <wm4> I don't remember having code that prints "???", but I don't have a tak sample right now [14:53] <wm4> though IMO the first after a seek really should have a timestamp [14:53] <ubitux> wm4: i guess it's missing a \n though.. [14:54] <wm4> does ffmpeg have a coding style document? I'm finding not anything obvious in doc/ [14:54] <durandal11707> wm4: but you had one? [14:54] <durandal11707> wm4: send code to libav ml [14:55] <ubitux> wm4: developers.{texi,html} [14:55] <durandal11707> general coding style should be: keep coding style already present in file - do not add your own [14:55] <wm4> ubitux: thanks [14:55] <ubitux> wm4: http://ffmpeg.org/developer.html [14:55] Action: wm4 going to steal it [14:57] <durandal11707> last time i compiled mpv (before configure change/removal/whatever) it displayed ??? or nothing iirc [14:57] <wm4> durandal11707: sample? (also this should probably go into the mpv channel) [14:59] <durandal11707> but iirc you complained about tak here, so you had sample or? [15:03] <ubitux> wm4: actually, it will probably break stuff [15:03] <ubitux> guess what: for ass tag used in srt :D [15:04] <ubitux> if you have a srt even with something like "{\a9}bla\nfoo" and then merge it with the following which has no alignment [15:04] <ubitux> but that's insane anyway [15:10] <wm4> I guess you'Re right [15:11] <ubitux> anyway, ugly hack that works for me http://pastie.org/8911440 [15:29] <wm4> hm [15:30] <wm4> if we don't find a better and more correct way, we should probably go with this... but fact is that some subtitle formats seem to want one ordering, while other formats prefer something different [15:30] <wm4> I'd be inclined to say "fuck it" and go with this hack [15:30] <ubitux> that is only for srt [15:30] <ubitux> ordered ones [15:30] <ubitux> and with matching durations [15:30] <ubitux> pretty limited case but work for the use case i had [15:32] <wm4> there's no spec for srt, so you can only guess [15:33] <ubitux> there is no spec so everyone generate shit and call it srt :) [15:34] <wm4> e.g. the mplayer code doesn't allow overlaps at all, and it just throws away the first line [15:34] <wm4> and it seemed to work for them for years [15:44] <ubitux> (w & ~15) + ((w & 15) ? 16 : 0) [15:44] <ubitux> anything simpler in mind? [15:45] <nevcairiel> isnt that just FFALIN(w, 16)? [15:45] <nevcairiel> ALIGN* [15:45] <ubitux> mmh indeed [15:46] <ubitux> so (w + 15) & ~15 heh. [16:00] <Eventh> :q [16:22] <plepere> in ASM, is there a fast way to copy from an adress to another ? I have function that just does dst[x]=src[x] [16:23] <thardin> rep movsd or somesuch? [16:24] <thardin> copies 4 bytes at a time [16:26] <ubitux> plepere: read 128b, write 128b? [16:39] <nevcairiel> yeah if its aligned you can just use movdqa [16:44] <iive> just use memcpy. I don't think that ffmpeg have its own accelerated memcpy, but gcc might produce simd code for it, given the right alignment of glibc, gcc, options, planets and stars... [16:45] <nevcairiel> yeah if its just C code, use memcpy, many c runtimes have vectorized variants for it [16:45] <nevcairiel> if its fixed size, 64 or 128 byte or something, you can also use AV_COPY64 macros [16:52] <wm4> glibc does have simd code [16:53] <wm4> but according to some tests, it's not even much faster than a naive rep movsd loop [16:53] <wm4> (on never CPUs) [16:54] <nevcairiel> i only know that even a manual movdqa loop isnt faster then memcpy, on msvcrt anyway [17:07] <jnvsor> I noticed that ffmpeg supports flac but it's not in the codec docs or the `./configure -h` - is it just baked in by default? [17:07] <JEEB> there is a libavcodec encoder for flac, yes [17:08] <av500> how does it sound? [17:08] <nevcairiel> configure -h only lists external libraries [17:08] <av500> as bad as the AAC one? [17:08] <kierank> av500: sounds lossy [17:08] <kierank> we need to get neil young in [17:08] <kierank> he'll sort it out [17:08] <av500> +1 [17:08] <JEEB> :D [17:08] <jnvsor> What, flac? [17:08] <nevcairiel> don't listen to them trolls [17:09] <nevcairiel> they're just making fun of pono [17:09] <av500> flac is not good enough, you also need zero feedback [17:09] <nevcairiel> he found quite a bunch of fools to pay up though [17:09] <nevcairiel> 1.5mil now [17:09] <av500> sure [17:10] <av500> at that rate, he can also make a ubuntu phone soon [17:10] <nevcairiel> what i find more baffeling though is the design of that player [17:10] <av500> the design I like [17:10] <jnvsor> I'm confused, I think I've stumbled into a conversation halfway through [17:11] <av500> no [17:11] <wm4> pono? [17:11] <av500> https://www.kickstarter.com/projects/1003614822/ponomusic-where-your-soul-r… [17:12] <nevcairiel> my boss seems to like the initiative, but maybe thats because he is about the same age as neil not-so-young [17:12] <jnvsor> So flac is lossless and pono is a triangle ipod? Gotcha [17:12] <av500> the age when your hearing maxes out at 12khz :) [17:13] <wm4> pretty annoying shape and colore [17:13] <av500> beauty is in the eye of the beholder [17:14] <av500> pono is a flacPod [17:14] <wm4> what if my eye is the most objective eye? [17:14] <av500> wm4: then stick a pono into it [17:14] <wm4> ow [17:14] <nevcairiel> just get the black one, not so annoying color [17:14] <nevcairiel> didnt i read something about their own special blend of audio format earlier [17:14] <nevcairiel> kickstarter claims they use flac [17:14] <av500> yes [17:15] <av500> yes, I think they changed rethoric [17:15] <av500> maybe they realised ppl will call out the BS [17:15] <wm4> so not yet another lossless audio codec? :( [17:17] <av500> https://plus.google.com/u/0/+VladimirPantelic/posts/J3aM2xYHaCb [17:17] <nevcairiel> but they must be throwing some drm on top of it, you know, cant have flac highres audio without protection! [17:17] <av500> this link to the older article about "special sauce" [17:17] <av500> nevcairiel: seems DRM is gone now too [17:17] <nevcairiel> the RIAA will not let that happen :p [17:17] <av500> they do [17:18] <av500> you get music sans DRM these days, no? [17:19] <nevcairiel> i suppose it depends how this service is setup, buying individual albums vs. subscription access to the entire library [17:20] <wm4> ah lol there's talk about flac etc. not being good enough [17:21] <ubitux> let's rar+zip it [17:21] <wm4> repeat 10 times for better compression [17:23] <av500> yes [17:23] <av500> its neil young! [17:23] <av500> just believe! [17:24] <thardin> feels like hipster bait [17:24] <ubitux> so they are able to reach the same size as flac but at 192kHz? [17:24] <ubitux> (basically a codec decoding flac and upsampling) [17:24] <nevcairiel> flac with sbr? [17:26] <thardin> I'm not seeing much technical information scrubbing through the video. just feels [17:29] <thardin> maybe they'll get more nicely mastered tracks [21:03] <cone-365> ffmpeg.git 03Martin Storsjö 07master:e77a2ea95058: http: Declare more parameters as const where possible [21:03] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:ad341b3bc65e: update doc/APIchanges [21:03] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:7e6c6c45ec7c: Merge commit 'e77a2ea9505863e50bf013706f66bf8b7325e524' [21:36] <cone-365> ffmpeg.git 03Diego Biurrun 07master:b3c6ee199e75: configure: Group toolchain options together in help output [21:36] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:aa175983a196: Merge commit 'b3c6ee199e75bbad2908253f11e871500dd38531' [21:43] <cone-365> ffmpeg.git 03Vittorio Giovara 07master:b4e355c89e23: copy_block: K&R formatting cosmetics [21:43] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:990f956ce66b: Merge commit 'b4e355c89e23664b8dac26936bf3fa7e7bc2110f' [21:52] <cone-365> ffmpeg.git 03Luca Barbato 07master:aa807425395c: configure: Support older version of openjpeg1 [21:52] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:58b8d268d971: Merge commit 'aa807425395caa17a85ed2833133278e8bd44a76' [22:15] <cone-365> ffmpeg.git 03Janne Grunau 07master:5a7f382a5d33: armv6: vp8: use explicit labels in motion compensation asm [22:15] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:e5920425b02b: Merge commit '5a7f382a5d33d9a26890affe6c8c5070a48dfc22' [23:21] <cone-365> ffmpeg.git 03Martin Storsjö 07master:d15c536123a4: doc: Point to the correct, actually maintained gas-preprocessor repo [23:21] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:835f41376da1: Merge commit 'd15c536123a44362ace6299c391a492c90b83fc7' [23:21] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:675a66a93bf8: doc: switch github urls to https [23:21] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:98eb99d20bf5: tools/build_libstagefright: switch git urls to https [23:21] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:b5c6b23acd6f: doc/platform: mention that yuvis gas-preprocessor is currently missing some changes [23:30] <cone-365> ffmpeg.git 03Alessandro Ghedini 07master:6998a9f4c4e0: http: Properly initialize icy headers string [23:30] <cone-365> ffmpeg.git 03Michael Niedermayer 07master:11ed7ec092d3: Merge remote-tracking branch 'qatar/master' [23:38] <wm4> what does all_channel_counts on buffersink do again? Libav doesn't have it [23:39] <nevcairiel> its apparently a boolean flag that tells it to accept any input channel counts [23:39] <nevcairiel> instead of just the list of supported ones [23:39] <wm4> and what if neither is set? [23:40] <nevcairiel> chaos ensues? [23:40] <wm4> in that case it seems not to call ff_set_common_channel_layouts(), whatever that means in the end [23:40] <nevcairiel> never used lavfi for audio [00:00] --- Thu Mar 13 2014

1 0

ffmpeg.log.20140312
by burek 13 Mar '14

13 Mar '14

[03:04] <hightall> hello every [03:04] <hightall> hi [03:04] <hightall> any one? [03:05] <hightall> is this ffmpeg channel? [03:05] <hightall> somebody help me [03:05] <sacarasc> hightall: Most IRC communication is asynchronous, that means you shouldn't wait for an answer like that, but just ask your actual question. [03:06] <hightall> how can i compile ffmpeg with vsz [03:07] <hightall> how can I compile ffmpeg with vsa [03:10] <sacarasc> hightallyht: What is vsa? [03:11] <hightallyht> it is a gcc parameter [04:15] <hightall> hello [04:16] <hightall> q_bit hello [06:45] <metaphysician> .ffpreset files should contain option=value pairs. How do I add the -vn (disable video recording) flag to a .ffpreset? [10:54] <jarainf> Is it possible to use a x11grab in a complex filtergraph? [10:54] <jarainf> I'm currently stuck with: [10:54] <jarainf> Stream specifier 'in' in filtergraph description [in]setpts=PTS-STARTPTS[x11grab] matches no streams. [11:17] <Aiena> is it pssible to tell ffmpeg to grab a specific window only for screen capture ? [11:18] <Aiena> e.g. if I specifiy the pid to it ? [11:20] <Aiena> i mean fo it to capture actuin within a specific open application only [11:20] <Aiena> no matter where it moves [12:29] <pespin> Hi, I'm trying to use ffmpeg to encode a real time streaming with h264 (x64 encoder). I am using "zerolatency" option, and as far as I see now I get 1 frame delay latency [12:30] <pespin> I need to drop the latency to 0 frames. [12:30] <pespin> I think it's related to the fact that ffmepg sets x264's "b_vfr_input" property to 1, which adds that 1 frame of extra latency [12:31] <pespin> I can't find the way of disabling it through ffmpeg API though :/ [12:31] <pespin> any hints appreciated :) [12:40] <Aiena> pespin no idea [12:40] <Aiena> I am seeking help too [12:40] <Aiena> I CANT;T GET -QSCAL 0 TO WORK [12:40] <Aiena> *-qscale 0 [12:45] <Aiena> pespin do you know which video codec is best for encoding with windows ? [12:46] <pespin> Aiena: what does "best" mean in this case? :P [12:46] <Aiena> pespin which is not too lossy [12:46] <Aiena> msmpegv2 wokrs everywhere ? [12:47] <Aiena> pespin or a decent codec to encode with and preserve quality when encoding into avi's [12:47] <Aiena> ffv1 seems to not play on windows [12:49] <Aiena> pespin okmpeg leads to too much graniness [12:54] <pespin> Aiena: I guess most people use h264 or vp8 [12:58] <Aiena> thanks pespin [12:58] <Aiena> pespin what is the best container for windows and linux ? [12:58] <Aiena> sometihng which works out of the box with WMP and can be encoded in linux ? [12:59] <pespin> Aiena: no idea about containers [12:59] <Aiena> OK thanks [12:59] <Aiena> pespin what format do you normally encode in ? [12:59] <Aiena> .mpg ? [12:59] <Aiena> and also what bitrate do you normally use ? [13:02] <pespin> Aiena: I don't do video files :) [13:02] <Aiena> pespin OK sorry to bother you [13:02] <Aiena> But thanls for the h264 tip [13:02] <Aiena> really helped a lot [13:02] <pespin> np! [13:03] <Aiena> this channle is so quiet with so many people [13:03] <Aiena> I just can;t fathom why (: [13:03] <Aiena> you were one of the few who responded :P [14:16] <Obi-Wan_> hello guys [14:16] <Obi-Wan_> I was looking for someone to help [14:19] <zap0> i helped, once. [14:19] <Obi-Wan_> I have an MPEG-TS with an AVC/H.264 video and an AAC audio [14:19] <Obi-Wan_> sorry but I am a newbie [14:20] <Obi-Wan_> I would like to convert boath audio and video into one .MPEG file [14:20] <Obi-Wan_> thank you, zap0 this can be the second time :) [14:21] <zap0> sounds like a newbie/common problem. i would bet all my monies, and all my neighbours monies that there is an example of this on the examples page. [14:22] <Obi-Wan_> also in case of this input file? [14:22] <Obi-Wan_> http://pastebin.com/5ibTc76H [14:24] <Obi-Wan_> what I should looking for, zap0? (and where, please)? [14:27] <jarainf> Obi-Wan_, do you want to ictv it? [14:29] <Aiena> HI zap0 [14:29] <Obi-Wan_> sorry, what do you mean with ictv ? [14:29] <Obi-Wan_> (thank you jarainf for you reply) [14:30] <Obi-Wan_> hello Aiena [14:30] <Aiena> Hi [14:30] <jarainf> do you want to deinterlace it and decimate it to 23.976 fps? [14:30] <Obi-Wan_> no, no [14:31] <Obi-Wan_> I simply need to generate a .MPEG file [14:31] <Obi-Wan_> I will rip the result with avidemux later [14:31] <Aiena> jarainf do you know which codec and video contaner is best across all platforms ? [14:31] <jarainf> Aiena, I would recommend matroska (.mkv) [14:31] <Aiena> and vcodec ? [14:32] <jarainf> x264 [14:32] <Aiena> ffv1 does n ot work well on windows [14:32] <Aiena> Ok jarainf how would I boots the quality of the x264 video ? [14:32] <Aiena> boost [14:32] <Aiena> also it complains about data not in sync leading to frame loss [14:32] <jarainf> Aiena, what do you mean by boost? [14:33] <Aiena> I mean increase the quality what bitrate should I use ? [14:33] <Aiena> reduce presense of artifacts due to compression [14:33] <Obi-Wan_> jarainf, of course I would like to leave the quality the same [14:34] <jarainf> Obi-Wan_, try ffmpeg -i input -c copy -f mpeg output.mpeg [14:34] <blight> hi guys [14:34] <Obi-Wan_> hi blight [14:34] <blight> does anybody know if there is a lossless audio format which i can put into mp4 together with h264 videoß [14:34] <jarainf> Aiena, if you already have artifacts in your source there is no way to "boost" its quality by simply reencoding it [14:35] <blight> i know mkv works with h264+flac but thats not supported by the app with which i want to open the video [14:35] <Aiena> jarainf the source is the x11grab itself [14:35] <jarainf> Ah [14:35] <Aiena> i.e. a grab of my desktop so the source is clean [14:36] <Aiena> i just altered the script to capture a specific window area by using data from xwinfo [14:36] <jarainf> Aiena, how does your current command look like? [14:37] <Aiena> ffmpeg -f x11grab -y -r 24 -s $WIN_GEO -i :0.0+$WIN_XY -vcodec h264 $NAME.mp4 [14:37] <Aiena> WIN_XY is the offset and WIN_GEO is the size here [14:37] <Aiena> $NAME is name of file specified in the shell jarainf [14:37] <Obi-Wan_> jarainf, sorry to bother you again ... would you be so kind to tell me if I can reach my goal with ffmpeg? [14:38] <jarainf> Obi-Wan_, did you try my command? [14:38] <Aiena> jarainf it is designed to capture video with no audio [14:38] <Obi-Wan_> no, sorry ... I thought you where talking with someone else [14:38] <jarainf> Aiena, so you are currently having problems with artifacts? [14:38] <Obi-Wan_> :/ [14:39] <Obi-Wan_> now I got it [14:39] <Aiena> jarainf yes a bit I just need to increase the bit rate a bit more [14:39] <Aiena> but not sure how or if its desirable [14:39] <jarainf> You should try first by setting presets [14:40] <Aiena> presets like ? [14:40] <JEEB> if you are not trying to hit a specific file size over time (average bit rate over the whole file), then use the crf rate control (which is the default one for libx264). -crf 23 is the default, and higher uses less rate and lower uses more rate. Use the highest that still looks good [14:41] <JEEB> as for presets, those control speed vs compression settings [14:41] <jarainf> Aiena, ffmpeg -y -i input -c:v h264 -preset -tune -f mp4 /dev/null [14:41] <jarainf> I don't know if output to /dev/null is simulated in windows [14:41] <JEEB> it's NUL on windows [14:41] <jarainf> But this should give you a list of available presets [14:41] <JEEB> and no, it will most probably just tell you that -tune is not a valid preset :P [14:42] <jarainf> Ah, thank you JEEB [14:42] <JEEB> if you are doing realtime encoding you probably want to pick a preset that's still fast enough [14:42] <JEEB> also personally I wouldn't do a "final encode" live [14:42] <JEEB> because realtime encoding puts constraints on you [14:42] <JEEB> personally I would just use crf zero (or quantizer zero) to losslessly encode the clip [14:42] <JEEB> and then encode with a slower preset afterwards for final distro [14:44] <JEEB> http://mewiki.project357.com/wiki/X264_Settings#preset [14:44] <JEEB> list of presets, btw [14:47] <Aiena> JEEB so what would you recommend I just dump raw frames and then encode more slowly later ? also how would I do this two phase operration I read the guides inline but ffmpeg is complex and there are so many ways of doing something that it is confusing [14:47] <Aiena> also they don;t explain what they are doing very well often [14:47] <JEEB> not raw frames [14:47] <JEEB> lossless compression is compression [14:47] <JEEB> it's just that you get what you pushed into the encoder when you decode it :P [14:47] <Obi-Wan_> jarainf, thank you this is the result > http://pastebin.com/3U6Ky2a0 < [14:48] <Obi-Wan_> but unfortunately I cannot see any video [14:48] <Obi-Wan_> :( [14:48] <JEEB> Aiena, you already have most of the stuff up until the input done, and it's not really that hard after that [14:48] <Obi-Wan_> my input is available here > http://www.hdpvrcapture.com/hdpvrcapturev3/samples/20131211_102729-1920x108… < [14:48] <JEEB> -c:v libx264 -crf 0 -preset preset_that_is_fast_enough_for_you [14:48] <JEEB> that should give you lossless 4:2:0 YCbCr [14:49] <Aiena> ok I have to google the -crf 0 option [14:49] <JEEB> and then you tweak the crf and preset to your liking after you have captured [14:49] <JEEB> crf is constant rate factor [14:49] <JEEB> zero with 8bit x264 is a special value, sets lossless coding [14:49] <JEEB> same as setting the quantizer to zero [14:49] <JEEB> higher bit depth x264 has that value go negative, so you want to use quantizer zero in those cases [14:50] <Aiena> ok [14:50] <JEEB> (otherwise the CRF value range goes bonkers, by adjusting where the "zero" is one can use similar CRF values between bit depths) [14:50] <JEEB> (for lossy that is) [14:51] <Aiena> JEEB "ffmpeg -f x11grab -y -r 24 -s $WIN_GEO -i :0.0+$WIN_XY -vcodec h264 -crf 0 $NAME.mp4" would be correct ? where win_geo is the size and win_xy is the offset ? [14:51] <Obi-Wan_> jarainf, I examined the result with mediainfo [14:51] <JEEB> yes, that would use the "medium" preset (default for x264) and c:v/vcodec h264 just happens to work, the actual encoder is called "libx264" [14:52] <Obi-Wan_> it seems that the output is againg AVC and not MPEG-2 [14:52] <JEEB> just that you only have libx264 as a AVC/H.264 encoder, so it probably auto-picks by that [14:52] <JEEB> when you set "h264" as the encoder [14:52] <Aiena> JEEB so I should replace h264 with libx264? [14:52] <JEEB> do as you wish, I don't see another H.264 encoder picking pace by now :P [14:53] <JEEB> but libx264 gets less weird looks [14:53] <Aiena> Ok [14:53] <jarainf> Obi-Wan_, ohgod [14:53] <jarainf> I'm sorry [14:53] <JEEB> because it sets the encoder specifically, and many people helping probably aren't going to be expecting that you are using the auto-mapping [14:53] <jarainf> I misread the mediainfo [14:54] <Aiena> JEEB Ok [14:54] <Obi-Wan_> jarainf, it's my fault [14:54] <Obi-Wan_> probably I was not clear enough [14:54] <Obi-Wan_> :( [14:54] <jarainf> Na, it was [14:54] <jarainf> it was mine* [14:55] <Obi-Wan_> no problem [14:55] <Obi-Wan_> I will patiently wait [14:55] <Obi-Wan_> :) [14:57] <jarainf> ffmpeg -i input -vcodec mpeg2video -acodec copy -f mpeg output.mpeg should do the trick, but you'll get artifacts from reencoding [14:58] <Obi-Wan_> what do you mean with get artifacts from reencoding? [14:58] <JEEB> the default for most video encoders is 200kbps [14:58] <JEEB> and mpeg-1 feature set [14:59] <Aiena> JEEB how would I specify "libx264-lossless_fast" preset ? woud I say -preset libx264-lossless_fast.ffpreset $NAME.mp4 ? [14:59] <JEEB> ugh, no [14:59] <Aiena> then [14:59] <JEEB> that command line is short enough by itself, isn't it? [14:59] <JEEB> also the setting for setting the derpy ffpreset files is -vpre [15:00] <JEEB> not -preset [15:00] <Aiena> Ok [15:00] <JEEB> -preset is a x264-specific setting that sets the libx264 preset [15:00] <Aiena> ok a bit confusing sigh [15:00] <JEEB> like -preset fast, which sets the 'fast' preset from the list of presets that can be easily seen @ http://mewiki.project357.com/wiki/X264_Settings#preset [15:01] <Aiena> i ingnored it and just specified libx264 FILENAME [15:01] <JEEB> yes, default preset is 'medium' [15:01] <Aiena> ok [15:02] <jarainf> Obi-Wan_, -acodec libmp3lame [15:02] <Aiena> but slower the better in this aspect for quality [15:02] <blight> so lame! [15:02] <blight> -acodec flac ! [15:02] <blight> :p [15:02] <JEEB> Aiena, only with lossy, lossless coding just has you compress the data more or less [15:02] <jarainf> Obi-Wan_, you might need that, I'm not sure if .mpeg supports aac [15:02] <JEEB> if you're doing live encoding you most probably will want to use a fast'ish preset [15:02] <Aiena> JEEBok so since I set crf 0 the preset is irrelevant [15:03] <Aiena> JEEB ok so since I set crf 0 the preset is irrelevant [15:03] <JEEB> not irrelevant, it sets the amount of CPU power used for compression [15:03] <Aiena> Ok [15:03] <JEEB> but it is irrelevant for quality, yes [15:03] <JEEB> you just get a bigger or a smaller file [15:04] <Aiena> JEEB how do I make ffmpeg use both cores (i have 2 cores 4 threads but hyperthreading does not make much of a difference) [15:04] <JEEB> libx264 should be defaulting to as many threads as you have CPUs (including virtual ones) by default as long as your build isn't 2-3 years old [15:04] <Aiena> No its latest [15:05] <Aiena> JEEB I am using 2.1.3 [15:05] <Aiena> or recent enough [15:05] <Aiena> yeah it seems up to date [15:07] <jarainf> 8 [15:07] <Obi-Wan_> jarainf, the conversion works fine [15:07] <Obi-Wan_> but the quality is very low [15:08] <Obi-Wan_> is it possible to keep the same quality? [15:08] <Aiena> JEEB I get extremelty small file sizes with h264 (670kb for like 10 seconds) but the quality is not as good with crf 0 I can't figure out why [15:09] <jarainf> Obi-Wan_, you can set the videobitrate with -b:v [15:09] <JEEB> the only difference would be the colorspace, libx264's encoder defaults to 4:2:0 for compatibility reasons [15:09] <jarainf> But as far as I know mpeg2 does not support lossless reencoding [15:09] <JEEB> Aiena, you can see what you're comparing against [15:09] <JEEB> by comparing the logs of whatever you encoded first [15:09] <JEEB> and whatever you just encoded [15:09] <Aiena> JEEB I am comparing against the output of the ffv1 codec [15:09] <JEEB> well look what colorspace it used :P [15:09] <Obi-Wan_> jarainf, and what about mpeg1? [15:10] <JEEB> it should tell you if it used rgb, or if it used yuvXXXp or whatever [15:10] <jarainf> it's worse [15:12] <Obi-Wan_> jarainf, do you think that -qscale 1 is a good option? [15:12] <JEEB> Aiena, if you're indeed comparing ffv1 encoding in RGB and libx264 in yuv420p, then you can just switch from the libx264 encoder to the "libx264rgb" one, which is pretty much the same, it just defaults to RGB instead of 4:2:0 YCbCr [15:13] <jarainf> Obi-Wan_, you should try yourself [15:14] <jarainf> I would go for 2 or 3 though [15:15] <Obi-Wan_> -qscale 1 is the maximum quality I can reach, right? [15:15] <jarainf> yes. [15:15] <JEEB> Aiena, do note though that for final distribution, if you want your stuff to be widely supported, 4:4:4 modes are a no-go (RGB and 4:4:4 YCbCr are the things encoded with the 4:4:4 mode), as it's not supported by any commercial decoder [15:16] <JEEB> (and that is why the non-rgb libx264 "encoder" defaults to 4:2:0 YCbCr) [15:17] <Obi-Wan_> thank you jarainf [15:17] <Obi-Wan_> I have also found these options [15:17] <Obi-Wan_> -pix_fmt yuv422p -qscale 1 -qmin 1 -intra -an [15:17] <Obi-Wan_> do they make sense for you to improve quality? [15:20] <jarainf> they probably do [15:20] <jarainf> while -an disables audio [15:20] <jarainf> -intra does give you only I-Frames though [15:20] <jarainf> afaik [15:20] <jarainf> but dont quote me on that [15:21] <Obi-Wan_> ok :) [15:24] <Obi-Wan_> jarainf, just one couple of things [15:25] <Obi-Wan_> > ffmpeg -i [INPUT] -vcodec mpeg2video -qscale 1 -acodec mp2 -b 320k -f mpeg output.mpeg < [15:26] <Obi-Wan_> does it make sens for you? [15:27] <jarainf> Uh [15:27] <jarainf> -b sets your bitrate [15:27] <jarainf> so you'd set the bitrate to 320k [15:28] <Obi-Wan_> it's constant bitrate encoding, right? [15:29] <JEEB> just -b is ambiguous [15:29] <JEEB> use -b:a for audio [15:29] <JEEB> and -qscale:v for video [15:31] <jarainf> what JEEB said [15:35] <Obi-Wan_> thank you JEEB and jarainf [15:35] <Obi-Wan_> I also have this error: [15:35] <Obi-Wan_> [h264 @ 03029800] Cannot use next picture in error concealment [15:35] <Obi-Wan_> [h264 @ 03029800] concealing 110 DC, 110 AC, 110 MV errors in P frame [15:35] <Obi-Wan_> after a lots of frames processed with no errors [15:37] <Obi-Wan_> this is the complete result [15:37] <Obi-Wan_> http://pastebin.com/76xwtmuw [15:40] <jarainf> A p-frame is a predicted frame [15:41] <jarainf> It's basically a diff from a frame presented before this one [15:41] <Obi-Wan_> so, it's an info, a warning or an error? [15:41] <jarainf> I'm only guessing here, but due to the nature of encodes it might be the case that the i-frame on which this p-frame relies may not be in the transportstream [15:42] <jarainf> the encode should still be fine [15:42] <jarainf> it's no fatal error :3 [15:42] <Aiena> I get tihs message [15:42] <Aiena> [swscaler @ 0x74f180] deprecated pixel format used, make sure you did set range correctly [15:42] <Aiena> No pixel format specified, yuv444p for H.264 encoding chosen. [15:42] <Aiena> Use -pix_fmt yuv420p for compatibility with outdated media players. [15:42] <Aiena> How should I address it ? [15:42] <Obi-Wan_> jarainf, and what about this warning? [15:43] <Obi-Wan_> [mpeg @ 0303f020] VBV buffer size not set, muxing may fail [15:43] <Obi-Wan_> should I set the VBV buffer? [15:43] <jarainf> Aiena, you probably want to use -pix_fmt yuv420 [15:44] <JEEB> Aiena, post command and full log :P [15:44] <JEEB> on pastebin [15:44] <Aiena> ok [15:44] <JEEB> and link your pastebin here [15:45] <Aiena> JEEB http://pastebin.com/7Dcq9Mmr [15:45] <Obi-Wan_> jarainf, and the other strange thing is that VLC is not able to detect the duration of the output :( [15:45] <Obi-Wan_> I am not sure everything went fine :'( [15:45] <JEEB> Aiena, that doesn't contain the actual ffmpeg command :P [15:45] <jarainf> Huh. [15:45] <Aiena> JEEB can I pm you the script ? [15:45] <JEEB> no [15:46] <Aiena> Ok [15:46] <JEEB> just note what the hell you're doing there with ffmpeg and that's it [15:46] <jarainf> Obi-Wan_, VBV stands for Video Buffering Verifier [15:46] <jarainf> It basically ensures, that an mpeg encode can be properly decoded [15:47] <Obi-Wan_> well, thank you [15:47] <Aiena> Ok [15:47] <Aiena> one sec [15:47] <Obi-Wan_> so, what I can do to solve this warning? [15:47] <jarainf> I don't really get what a verified buffering verifier buffer should be, though. [15:48] <JEEB> jarainf, Obi-Wan_ -- basically mpeg-ts and probably mpeg-ps are made for bandwidth limited muxing (you set VBV variables and it muxes accordingly, and adds extra padding when needed to hit the set maxrate over bufsize) [15:48] <JEEB> basically it's saying that he hasn't set that and thus depending on the muxing mode it might fail [15:48] <JEEB> that's more or less what it means [15:48] <JEEB> if the output works fine, then that's fine [15:48] <jarainf> Ah, that makes sense. [15:49] <Obi-Wan_> JEEB: YEAH, thank you ... but the result is strange [15:49] <Obi-Wan_> VPL is not able to detect the lenght of the file, for example [15:49] <Obi-Wan_> *VLC not VPL sorry [15:49] <JEEB> well given the container [15:49] <Aiena> JEEB http://pastebin.com/tqA5ff02 should do it [15:50] <JEEB> any length is guesswork :P [15:50] <JEEB> Aiena, surprising that it picks 4:4:4 YCbCr by default [15:50] <Obi-Wan_> JEEB, sorry I did not understand [15:50] <JEEB> you probably want RGB, though [15:51] <Aiena> JEEB but RGB will work well on windows too ? [15:51] <Aiena> h264RGB ? [15:51] <Aiena> also files are ultra light this makes text rendering is specific cases not sharp enough [15:51] <JEEB> Aiena, well didn't we just goddamn go over that it's better to capture stuff in lossless and then encode it with better compression for final distribution or whatever [15:52] <JEEB> also way to go ignoring the crf recommendation [15:52] <Aiena> JEEB we did but you said crf 0 was lossless [15:52] <JEEB> yes [15:52] <JEEB> you are using -b:v 1000k [15:53] <JEEB> so unless you posted the wrong information (I JUST NEEDED THE FUCKING FFMPEG LINE FOR FUCK'S SAKE), you are just trying to play with me [15:53] <JEEB> Obi-Wan_, the mpeg-ps container has no length information in it, so the only way to guess the length is when the mux is of constant rate and thus you can just divide the file size with the mux rate [15:53] <Aiena> JEEB got disconnected what was the last line I mentioned ? [15:54] <jarainf> <Aiena> BEEJ we did but you said crf 0 was lossless [15:54] <JEEB> Aiena, I'm sorry, find someone else to waste time because you clearly are ignoring any recommendations. [15:56] <Aiena> JEEB alright I am sorry I did try your recommendations I did exactly as told except that I used -b:v to try to understand how that option worked [15:56] <Aiena> I am sorry if I ghurt your feelings in any way. [15:57] <Obi-Wan_> JEEB, so a MPEG .TS container should be better? [15:57] <JEEB> you didn't, it's just aggravating when people clearly ignore when you're trying to guide them to a good way of doing things [15:57] <JEEB> Obi-Wan_, > implying that has it either [15:58] <jarainf> Obi-Wan_, wasn't the whole point to have an mpeg container? [15:58] <JEEB> although what players do to guess the length in those containers can depend, so everything depends on what the flying derp you want :D [15:59] <Obi-Wan_> OK! [15:59] <Obi-Wan_> when I will be at home I will try to see if audio and video are in sync [15:59] <Aiena> JEEB I took everything you told me into consideration I was just experimenting at the end sorry. Jee what is the difference beween -qp 0 and -crf 0 [15:59] <jarainf> Obi-Wan_, btw, why do you need to reencode the ts? [16:00] <Obi-Wan_> jarainf, the issue is that I have an Hauppage HD PVR recorder [16:00] <JEEB> Aiena, quantizer zero works with all bit depths in x264 (8,9,10) while crf 0 only works with 8bit [16:00] <Obi-Wan_> it records in AVC/H.264 and this is not well handled by AVIdemux [16:00] <Aiena> Ok and how do I set the bit depth for the h264 codec ? [16:00] <JEEB> you can't on runtime [16:00] <JEEB> it's a build-time thing [16:00] <Obi-Wan_> so I need to convert H.264 to MPEG [16:01] <Obi-Wan_> and so to be handled by AVIdemux without issues [16:01] <Aiena> JEEB Ok thank you for all your help [16:01] <Aiena> I will ask you if I am unsure of something. [16:01] <jarainf> Obi-Wan_, what do you need AVIdemux for? [16:01] <Obi-Wan_> does it make sens for you, jarainf? [16:01] <Obi-Wan_> I use it to encode the video in Xvid/MP3 [16:01] <Obi-Wan_> after applying filters [16:02] <Obi-Wan_> such as deinterlacing [16:02] <Obi-Wan_> cropping [16:02] <Obi-Wan_> resizing [16:02] <Obi-Wan_> and so on [16:02] <jnvsor> Is flac a good lossless audio codec? [16:02] <jarainf> you can do all that in ffmpeg too [16:02] <Obi-Wan_> Yes, you are right [16:03] <Obi-Wan_> but, as you know I am a newbie [16:03] <Obi-Wan_> and not very confident with command lines tools [16:03] <Obi-Wan_> last question, please [16:03] <Obi-Wan_> how can I transform from 30fps to standard PAL 25 fps? [16:04] <Dark-knight> anybody want to play a free fps with me? [16:07] <Dark-knight> im done re-muxing all my videos [16:07] <Aiena> JEEB I let my freind on windows test the h264 encoded stream it is not playable on windows with WMP. I read somewhere that windows sometimes does not play a strem if audio is not present or somethig how do I put a dummy silet audio track ? [16:07] <jarainf> Obi-Wan_, -vf=fps decimates the stream to a default of 25 fps [16:08] <Aiena> I encoded it with the regular -crf 0 option [16:08] <JEEB> well, d'uh [16:08] <jarainf> But I'm not sure if that's enough to be PAL conform or how it handles with interlaced frames [16:08] <JEEB> seriously, I went over this already [16:08] <Aiena> JEEB let me look up [16:08] <Aiena> one sec [16:09] <JEEB> 1) lossless is not supported in commercial decoders 2) anything but 4:2:0 -||- [16:09] <Obi-Wan_> jarainf, -vf=fps seems to be not a valid options [16:09] <JEEB> so for the file for distribution you will want to use the non-rgb encoder and crf higher than 0 [16:10] <Aiena> Ok I did not know about crf 0 not being supported I knew about RGB not being supported [16:10] <Aiena> from you [16:10] <Aiena> JEEB they don;t mention it https://trac.ffmpeg.org/wiki/x264EncodingGuide either [16:11] <jarainf> Obi-Wan_, -vf "fps=fps=25" [16:11] <jarainf> or -vf "fps" [16:11] <JEEB> 23 is the default, lower makes x264 do less error, higher makes x264 do more error. thus for lossy you pick the highest value that still looks good [16:11] <Aiena> Ok I am testing with 18 [16:12] <Aiena> -crf 18 [16:15] <Aiena> JEEB even with -crf 18 it is not working on windows [16:15] <Aiena> JEEB can I try adding a dummy audio channel ? [16:16] <Aiena> how would I do that [16:16] <Aiena> JEEB also is mp4 an ok container for h264 video ? [16:16] <Aiena> or is it a container issue ? [16:17] <Obi-Wan_> thenk you once more, jarainf [16:17] <JEEB> Aiena, show current line and log [16:17] <Obi-Wan_> -vf "fps=fps=25" is fine [16:17] <JEEB> pastebin [16:17] <Aiena> JEEB one moment [16:18] <Obi-Wan_> since I know that the max bitrate is 18.8 Mbps [16:18] <Obi-Wan_> do you know if I can set this [16:18] <Obi-Wan_> ad a cbr for the output [16:19] <Obi-Wan_> I am not sure but in this way I should reach the max quality [16:19] <jarainf> Obi-Wan_, -vf fps does the same, -vf=fps just doesn't work because of syntax [16:19] <jarainf> I'm a bit tired. [16:19] <jarainf> Obi-Wan_, setting a constant bit rate does not improve quality [16:19] <jarainf> bit rate is not directly proportional to quality [16:21] <Aiena> JEEB HERE http://pastebin.com/4NzJTXVq [16:21] <Aiena> *here [16:21] <JEEB> interesting [16:22] <JEEB> I would think that version would default to 4:2:0 [16:22] <JEEB> -pix_fmt yuv420p then after -i [16:22] <Aiena> Ah "4:4:4 8-bit" [16:22] <Aiena> ok adding [16:24] <Obi-Wan_> jarainf: I'm very sorry to make you tired ... I will be silent, now. ;-) [16:24] <Aiena> JEEB should there be a space between " -pix_fmt yuv420p" and ":0.0+$WIN_XY" like " -i -pix_fmt yuv420p :0.0+$WIN_XY " or should it be with no space like " -i -pix_fmt yuv420p:0.0+$WIN_XY " ? [16:24] <jarainf> Obi-Wan_, it's not your fault. After all it's rather me giving you a hard time. [16:25] <JEEB> Aiena, what do you think ":0.0+$WIN_XY" is? [16:25] <Aiena> that is offset [16:25] <JEEB> isn't it a parameter for -i [16:25] <Aiena> oops [16:25] <JEEB> thus, if it's a parameter, why are you adding something in between the setting and the parameter :P [16:26] <JEEB> I did say "after -i" but that means, after setting the input [16:26] <Aiena> yeah my mistake [16:26] <Aiena> sorry [16:26] <Aiena> thaks for pointing it out [16:28] <Aiena> JEEB "ffmpeg -f x11grab -y -r 24 -s $WIN_GEO -i :0.0+$WIN_XY -pix_fmt yuv420p -vcodec libx264 -crf 18 $NAME.mp4" seems ok to you ? [16:28] <JEEB> yes [16:28] <Aiena> OK thanks encoding a new test video [16:29] <Obi-Wan_> jarainf, I promise this to you ... just the real last thing :) [16:30] <Obi-Wan_> do you know if there is a loseless codec alternative to MPEG? [16:30] <jarainf> You can do lossless encodings in x264 [16:31] <jarainf> But you wont be able to fit them into an .mpeg :3 [16:31] <jarainf> but x264 works fine with .mp4 [16:31] <Obi-Wan_> :( unfortunately it's not managed by avidemux :( [16:32] <jarainf> Obi-Wan_, I think it would suit your needs if you encode tranports-streams in ffmpeg [16:32] <jangle> Greetings all. What parts of the library are responsible for parsing an h264 annex b file and passing the data to a decoder? [16:33] <jarainf> Obi-Wan_, ffmpeg -i transportstream -vf "fieldmatch, yadif=deint=1, decimate, scale=width:heigth" -c:v libx264 -c:a copy -preset slow output.mp4 [16:33] <Obi-Wan_> jarainf, do you mean what we have done now, or other encodings? [16:34] <Obi-Wan_> but I already have a .mp4 file [16:34] <jarainf> This will deinterlace your input and decimate it to 23.976 fps [16:34] <jarainf> It will also scale it to width:height [16:34] <jarainf> which you would need to specify then [16:35] <jarainf> Obi-Wan_, is it interlaced? [16:35] <Obi-Wan_> sometimes yes [16:35] <Obi-Wan_> sometimes not [16:36] <Obi-Wan_> and what about from my source to .VOB? [16:36] <Obi-Wan_> is it better from wyk or not? [16:36] <Obi-Wan_> compared with the MPEG conversion we have just done? [16:38] <jarainf> I'm sorry but I don't really get why you need to reencode the stream [16:39] <jarainf> Your recorder supports x264 and avidemux seems to support mp4 and x264 [16:40] <jarainf> Obi-Wan_, all these are just container formats, it really depends on the reencode you do [16:41] <jarainf> as in the codec you use [16:41] <Obi-Wan_> AVIdemux is NOT able to REALLY support [16:41] <Obi-Wan_> AVC/H.264 codec [16:41] <Obi-Wan_> at the end of the conversion audio and video goes out of sync [16:42] <Obi-Wan_> the only option I have is using Xmedia Recode [16:42] <jarainf> Uhm [16:42] <jarainf> I doubt that [16:42] <jarainf> since it uses libx264 and ffmpeg [16:42] <Obi-Wan_> Do you doubt what? [16:42] <Obi-Wan_> I have tried [16:43] <jarainf> Obi-Wan_, does it do that only for interlaced streams? [16:43] <jarainf> >While Avidemux uses built-in libavcodec from FFmpeg for H.264 decoding, it needs an additional (external) library for H.264 encoding. Therefore Avidemux uses x264. [16:43] <Obi-Wan_> no, I don't think so [16:43] <Obi-Wan_> but I am not 100% sure [16:43] <jarainf> you should test that [16:43] <jarainf> If it does it only for interlaced streams [16:44] <jarainf> Then it's probably related to your attempt of ivtcing [16:44] <Obi-Wan_> no, sorry [16:44] <Obi-Wan_> the real issue [16:44] <Obi-Wan_> I have found is from 3-5 seconds of silence [16:44] <Obi-Wan_> at the beginning of the converted video [16:45] <jarainf> Obi-Wan_, does it also do that for long videos? [16:45] <jarainf> or is there a longer period of silence? [16:50] <jnvsor> I notice niether the ./configure nor the documentation mention flac - is it just baked in and assumed to work? [16:51] <Obi-Wan_> I do not really know, jarainf [16:51] <Obi-Wan_> but the other issue is that AVIdemux 2.6 does not support XviD open libraries anymore [16:53] <jarainf> Obi-Wan_, after processing the files with avidemux, what do you use them for? [16:55] <Obi-Wan_> sharing, playing on TV and so on [16:55] <Obi-Wan_> now I have to leave or I will loose the train [16:56] <jarainf> Obi-Wan_, yeah, see you [16:56] <Obi-Wan_> thank you very much for your support (also to JEEB) [16:56] <jarainf> You should try to get rid of avidemux [16:56] <jarainf> If it's really the fault of avidemux [16:56] <Obi-Wan_> I will try [16:56] <Obi-Wan_> and I will probably come back to you [16:56] <jarainf> feel free to do so [16:57] <Obi-Wan_> but I think it's an avidemux issue [16:57] <Obi-Wan_> since with Xmedia Recode [16:57] <Obi-Wan_> everything is fine [16:57] <Obi-Wan_> :) [16:57] <jarainf> honestly, I really can't imagine it being an avidemux issue [16:57] <Obi-Wan_> bye bye talk to you next time [17:00] <Aiena> JEEB when using -pix_fmt yuv420p the height of the screen should be divisible by 2 ? [17:01] <JEEB> the encoded picture should be, yes [17:01] <JEEB> due to chroma subsampling [17:01] <Aiena> OK that is why the encoder fails [17:01] <Aiena> hmm [17:01] <Aiena> I do not know what to do in this case [17:01] <JEEB> you can use -vf scale to resize the picture [17:02] <JEEB> after -i [17:02] <JEEB> http://ffmpeg.org/ffmpeg-all.html [17:02] <JEEB> ctrl+F here and you should see both options and examples [17:04] <Aiena> hmm there is a 0vf frames and a -vf filtergraph [17:04] <Aiena> the thing is many of these technical terms are alien to me. [17:04] <Aiena> But let me read more [17:05] <JEEB> uhh what [17:05] <JEEB> just ctrl+F for "-vf scale" [17:05] <JEEB> the scale video filter [17:06] <Aiena> OK thank you JEEB [17:11] <Aiena> JEEB is there a way to make h264 encode at a 4000k bit trate constantly [17:11] <Aiena> or is bitrate limited by the crf factor adncannot go beyond that ? [17:12] <Aiena> JEEB "ffmpeg -f x11grab -y -r 24 -s $WIN_GEO -i :0.0+$WIN_XY -pix_fmt yuv420p -vf scale=1280:720 -vcodec libx264 -b:v 4000k -minrate 4000k -maxrate 4000k -bufsize 1835k $NAME.mp4" is my current line. [17:12] <Aiena> I tried to follow a documentation example [17:13] <JEEB> why the flying fuck do people use -minrate [17:13] <JEEB> are they fucking dumb [17:13] <JEEB> anyways, why the fuck would you want fucking pseudo-CBR [17:14] <JEEB> I've once again lost what the flying fuck you're actually trying to do [17:16] <Aiena> JEEB ok I was looking at the video and I can't get the text to be sharper. I don;t mind the file being larger. Currently with gh264 enocidng I get files sizes of 400kb for 10 seconds [17:16] <JEEB> why was your fucking first idea to switch all to something else instead of just using the goddamn tool I had told you [17:16] <JEEB> which was the CRF value [17:17] <Aiena> JEEB with CRF too the text is blurry [17:17] <JEEB> THAT IS BULLSHIT [17:17] <JEEB> I mean [17:17] <Aiena> I personally think it is sharp enough [17:17] <JEEB> CRF does the exact same thing [17:17] <JEEB> it just sets a rate factor and encodes to it [17:17] <Aiena> yeah I read and re read the documentation [17:18] <JEEB> doing 2pass encoding for a bit rate and CRF value for the same file size give the same fucking quality [17:18] <Aiena> and I was convinced that everything yopiu said was right [17:18] <JEEB> so as fucking long as you're in the same bit rate spectrum [17:18] <JEEB> then you are getting the same quality [17:18] <Aiena> ok [17:18] <JEEB> so you are either comparing apples to oranges (completely different rates) [17:18] <JEEB> or you have no idea what 4:2:0 YCbCr and why it looks different to RGB [17:19] <Aiena> Probably thee other person viewing the video expects something else [17:19] <JEEB> asdf [17:19] <JEEB> stop going into conclusions [17:19] <JEEB> for fuck's sake [17:19] <JEEB> there are two alternatives [17:20] <Aiena> Ok [17:20] <JEEB> the average bit rate of the results (The CRF value you used vs ABR bit rate mode) is different [17:20] <JEEB> OR [17:20] <JEEB> one of the clips is 4:2:0 and the other isn't [17:20] <JEEB> these are the ONLY TWO REASONS WHY YOU WOULD HAVE A DIFFERENCE [17:21] <JEEB> and if you see a difference and they're both 4:2:0 then just fucking adjust the CRF value accordingly goddamnit [17:21] <JEEB> instead of saying fuckdumb things like "with CRF the text is blurry" [17:21] <Aiena> Ok [17:21] <Aiena> JEEB so the real answer is to switch to an RGB codec [17:22] <JEEB> well, what the fuck was even compared [17:22] <jarainf> Wat [17:22] <JEEB> I'm not an esper [17:22] <Aiena> JEEB ffv1 versus CRF [17:22] <JEEB> I'm trying to tell you the two possible sources of differences [17:22] <Aiena> *libx264 sorry [17:23] <JEEB> what!? [17:23] <Aiena> ffv1 vs libx264 [17:23] <Aiena> ffv1 as you mentioned is an RGB one [17:23] <JEEB> did you make sure that you're comparing the same pixel format output from both? [17:23] <Aiena> let me rerender one moment [17:23] <JEEB> if you're comparing RGB to 4:2:0 YCbCr then you're going to see the difference :P [17:24] <JEEB> because you lose information when converting one to another [17:24] <Aiena> JEEB is there a good RGB cross paltform format assuming the video is going to palyed obnly on PC's ? [17:24] <JEEB> Aiena, "no" [17:24] <JEEB> at least not for vanilla installations of systems [17:25] <JEEB> which is what I thought you were aiming for [17:25] <Aiena> Even with Windows media player installed [17:25] <JEEB> that is part of "vanilla installation" for me :P [17:25] <JEEB> and I still don't know your final use case [17:26] <JEEB> as in, I don't know if you really want to start looking for something very dumb to get RGB (and literally no compression), or if you can just live with 4:2:0 and use something normal and generally supported [17:26] <Aiena> JEEB so basically flv's are also YCbcR (luminance and chrominance) like telivision ? [17:26] <JEEB> flv is a container [17:27] <Aiena> Ok [17:27] <JEEB> but yes, most if not all video in FLV container is 4:2:0 [17:27] <JEEB> YCbCr [17:27] <JEEB> heck, most video all around is YCbCr, and in 99%+ of all cases 4:2:0 [17:27] <JEEB> (in production you generally can see 4:2:2 if interlacing is touched) [17:28] <JEEB> (but end user stuff is pretty much wholly 4:2:0) [17:29] <Aiena> Ok [17:29] <Aiena> then I do not know my freind really expected [17:29] <Aiena> (sigh) [17:33] <Venti^> Aiena: maybe try turning off true type? then chroma subsampling shouldn't matter [17:33] <Aiena> Venti^ that is a good Idea but I don;t know how to do that on my linux system [17:33] <Aiena> let me google it [17:33] <JEEB> ? truetype, what? [17:34] <Venti^> font smoothing for lcd's [17:34] <JEEB> nothing is used in the ffmpeg command that would have /anything/ to do with that [17:34] <JEEB> oh [17:34] <JEEB> so to make the source content different? [17:34] <JEEB> so it might look better with chroma subsampling [17:34] <JEEB> you kind of noted it in a way that just brought up question marks all over in my brain [17:35] <JEEB> because freetype in ffmpeg generally means the lulzy ASS subtitle rendering filter [17:35] <JEEB> which definitely isn't used here [17:35] <Aiena> Venti^ actually its freetype [17:36] <Aiena> perhaps font rasterisatyion has something to do with it [17:36] <Aiena> rasterisation [17:36] <JEEB> also one old trick was to first resize to 2x width/height and then convert to 4:2:0 YCbCr from RGB [17:36] <JEEB> the video becomes twice as big, but at least you're kind of unaffected by chroma subsampling [17:37] <JEEB> (you have to use nearest neighbor resizing, of course) [17:37] <Venti^> JEEB: I couldn't think of any other reason why chroma subsampling would matter for text blurriness [17:37] <JEEB> Venti^, well you generally lose contrast after the RGB->YCbCr conversion plus the subsampling [17:38] <JEEB> I'm not so sure it can be fully mitigated by disabling font smoothing [17:38] <Aiena> JEEB probably that is the issue but there is no real way around it [17:38] <Aiena> even with -crf 0 [17:38] <rsdrsdrsd> Is there a universal method to etrieve the height of a video with ffprobe or ffmpeg? With ffprobe I get the info back and can parse it, but sometimes the video stream is on a different index [17:38] <JEEB> Aiena, there is the "workaround" I noted :P [17:38] <JEEB> 2x nearest neighbor resize [17:39] <Aiena> upscale or downscale ? [17:39] <Aiena> with -vf ? [17:39] <JEEB> upscale, naturally [17:39] <Aiena> Ok [17:39] <JEEB> also make sure with -v debug that it actually first resizes and then converts [17:40] <JEEB> because it becomes useless if it first converts and then resizes that [17:40] <Aiena> shoud i put -v debug after -vf=*:* ? [17:40] <JEEB> you can put it pretty much everywhere, I usually put it in the input options' side (before -i) [17:40] <JEEB> and it will spam a lot [17:41] <Aiena> ok can I pastebin output for you to look at ? [17:41] <Aiena> JEEB someone suggested something like this online -level 3.1 what does that mean ? [17:42] <Aiena> along with libx264 ? [17:42] <JEEB> it sets a level and is completely unrelated to this problem [17:42] <Aiena> ok [17:42] <JEEB> levels are more or less flags that tell the decoder that it will need X amount of memory to decode the stream [17:43] <JEEB> just like profiles are things that tell the decoder which features it will need to decode the stream :P [17:44] <JEEB> scale=w=2*iw:h=2*ih -sws_flags neighbor [17:44] <JEEB> this should do 2x upscale with nearest neighbor [17:44] <Aiena> JEEB I thought "-vf scale=2560:1440 " [17:45] <Aiena> for 720p [17:45] <JEEB> yes, you can set the actual numbers as well, but it's just simpler to tell it "2x input width/height please" [17:45] <Aiena> my screen is 1366x768 so cannot get larger than that and I need multiples of 2 [17:46] <Aiena> JEEB actually I tghink so this will be make any resolutioin a mutiple of 2 automatically [17:46] <Aiena> so that will sove the other problem of enforcing a fixed size [17:46] <Aiena> solve [17:46] <JEEB> yes, multiplying by two kind of does that :P [17:47] <Aiena> JEEB so I shoud replace -vf scale=2560"1440 with "-vf scale=w=2*iw:h=2*ih -sws_flags neighbor" [17:47] <JEEB> basically yeah [17:47] <Aiena> Ok trying [17:48] <Aiena> thank you so much for your patience I know that it taking too much time to wrap my mind around certain ffmpeg concepts [17:48] <JEEB> oh, you might also need the full_chroma_inp and full_chroma_int flags [17:49] <JEEB> -sws_flags neighbor+full_chroma_inp+full_chroma_int [17:50] <JEEB> sets nearest neighbor, and sets some "use chroma better" flags that I'm still not sure what they do exactly :P [17:50] <Aiena> ok I will try but quality has improved [17:50] <Aiena> pretty neat thanks JEEB [17:51] <JEEB> now to just make sure it doesn't first convert the pixel format and then resize [17:51] <JEEB> but I guess if it looks better then it most probably isn't doing that [17:51] <JEEB> -v debug should show what kind of filter chains it makes internally I think? [17:51] <Aiena> Ooh forgot to add the -v debug flag [17:51] <Aiena> one sec [17:51] <Aiena> thanks for reminding [17:52] <JEEB> but yeah, since 4:2:0 basically means that chroma has one value per 2x2 area, the only thing to kind of mitigate its damage is to upscale 2x [17:52] <JEEB> of course even that is not perfect [17:54] <Aiena> JEEB debug output http://pastebin.com/Mjwj9rCy [17:54] <Aiena> is it doing what it is supposed to do ? [17:55] <Aiena> because I cannot interepret that output. It has a lot in it I don't know what to look fo [17:57] <JEEB> meh, swscale is so silent [17:57] <JEEB> [Parsed_scale_0 @ 0x2583700] w:1366 h:741 fmt:bgr0 sar:0/1 -> w:2732 h:1482 fmt:yuv420p sar:0/1 flags:0x6010 [17:57] <JEEB> this is the input and output of your resizing [17:57] <JEEB> (and other things) [17:58] <JEEB> but yeah, if it looks better there's a big chance it actually resizes first [17:58] <Aiena> ok so the output does not indicate which operation happened first [17:58] <Aiena> JEEB will upscaling 4x be better ? [17:58] <JEEB> no [17:59] <Aiena> Ok [17:59] <JEEB> "4:2:0 basically means that chroma has one value per 2x2 area" [17:59] <JEEB> that's where the 2x upscale thing comes from :P [18:00] <Aiena> JEEB what if I use lanczos instead of neighbor ? [18:00] <Aiena> for upscaling ? [18:00] <Aiena> i know lanczos is better for downscaling usually [18:00] <JEEB> well then you are no longer making the process less reversable'ish :P [18:00] <JEEB> uhh [18:00] <JEEB> remove the less [18:00] <JEEB> basically you can reverse nearest neighbor scaling rather easily [18:01] <Aiena> Ok [18:04] <Aiena> JEEB suppose I encode at 2x scale and then downscale later again the text will become blurry right ? [18:04] <Aiena> Downscale with ffmpeg to the actual resolution [18:05] <klaxa> depends on the algorithm i think [18:05] <JEEB> Aiena, well the idea was to get around chroma subsampling by having all details be at least of the size 2x2... [18:06] <Aiena> OK so resizing the output post rendering will not matter [18:06] <JEEB> "post rendering"? [18:06] <JEEB> you mean during playback? [18:06] <JEEB> yes, the idea is that you can then have the playback side enjoy it in any way it wants [18:06] <Aiena> i mean after rendering/encoding from the screen into the original file [18:07] <JEEB> since stuff generally gets converted to RGB first, before being scaled [18:07] <JEEB> so resizing during playback is just fine [18:07] <Aiena> Ok but resizing withh ffmpeg itself is not ? [18:08] <Aiena> with ffmpeg the source mpg [18:08] <JEEB> but wasn't your idea to create a file that was playable on "vanilla" systems? [18:09] <JEEB> I really am having difficulty trying to parse what you're trying to say [18:09] <Aiena> jeeb isnt mp4 just a container ? [18:09] <JEEB> yes? [18:09] <Aiena> so if the video in it is h264 it should play on WMP right [18:09] <JEEB> if it's 4:2:0 and not lossless, yes [18:10] <Aiena> ok let me look at my log [18:12] <JEEB> anyways, as I still didn't understand your ffmpeg/resizing question, here's a generic answer: after you have the 2x 4:2:0 YCbCr content, you cannot downscale that in the same pixel format if you do not want to lose the effect you gain by upscaling it to begin with (not having details that are smaller than 2x2) [18:13] <JEEB> you can make the playback window smaller during playback because at that point the picture has already been converted to RGB in most cases, and thus you have no subsampling [18:19] <Aiena> Ok [18:19] <Aiena> SO downscaling with ffmpeg will in effect make the output the same as if it was upscaled and rendered normally [18:19] <Aiena> thanks JEEB [18:23] <jangle> I guess my question now boils down to, when packing nals from h264 into an AVPacket structure, does avcodec_decode_video2 accept "chunked" decoding, where I should load one nal at a time into avpacket and then make the decode call? Or does the AVPacket need to be filled with a full frame worth of nals before the decode, and if so, should they be packed with startcodes or just be sequential raw nal payloads? [20:09] <kingsob> I am trying to invert a video, but getting an error "Error setting option pixel_aspect to value -1/1." --> http://paste.ubuntu.com/7080930/ [20:36] <kingsob> it seems ffmpeg can't do anything with my video that has the SAR -1:1 DAR -71:40 [20:44] <Aiena> kingsob well I can;t help you there [21:03] <kingsob> I was able to use h264_mp4toannexb to extract the bit stream, then encode back to h264 [21:03] <kingsob> but that doesn't seem ideal [21:33] <Snaggle> I'm building ffmpeg-head, and although I don't explicitly enable the jack indev, -ljack is still added to EXTRALIBS. According to the configure output, jack is not in the enabled indevs either. This lazy linking to -ljack causes runtime failures if the jack library is removed (since it was not expected to be linked to). Is this a bug in the configure jack check (that's when -ljack first seems to be added into the EXTRALIBS [21:33] <Snaggle> stream) ? [21:35] <klaxa> can you pastebin your config.log? [21:38] <Snaggle> klaxa: http://pastebin.com/8Ccn4EUA [21:39] <Snaggle> oops. that's actually for ffmpeg-2.1.4, but same thing happens with git-head [21:39] <klaxa> it's listed as an indev device [21:39] <klaxa> can you explicitly disable it? [21:43] <Snaggle> klaxa: yes, i can explicitly disable it via configure parameters. But I would have thought that if the "check_func sem_timedwait" function failed (which is part of the "enabled jack_indev" test) failed, then -ljack wouldn't be added to EXTRAFLAGS. But it seems that just passing the first check_lib2 test adds -ljack to EXTRAFLAGS. config.h and config.mak both have the jack_indev as disabled. [21:44] <klaxa> that's weird, i think jack isn't tested for functionality, but presence, so if it is present it is added to the build [21:44] <klaxa> like every other library [21:46] <Snaggle> right. I think since the check is this: enabled jack_indev && check_lib2 jack/jack.h jack_client_open -ljack && check_func sem_timedwait, that passing check_lib2, -ljack is added, and then when check_func fails, the jack indev is disabled, but -ljack from the previous test function is not removed from EXTRAFLAGS. I'm going to see if bracketing it fixes [22:04] <DX099> hello, is there a way for ffmeg to be able to write metadata to aac files ? [22:06] <jnvsor> ffmpeg -i file.aac xxx -c copy file.aac replace xxx with stuff from the docs: http://www.ffmpeg.org/ffmpeg-all.html#AC_002d3-Metadata [22:07] <DX099> jnvsor, I'm actually converting from flac. More like "ffmpeg -i myfile.flac -c:a libfdk_aac -vbr 4 -map_metadata 0 output.aac" [22:08] <DX099> but output.aac doesn't have any metadata even though ffmpeg displays them [22:12] <jnvsor> does ffprobe display the metadata? [22:13] <DX099> yes, there's no problem for original .flac file metadata, they're displayed correctly by ffprobe and at convert time by ffmpeg [22:23] <DX099> http://trac.ffmpeg.org/wiki/AACEncodingGuide#Metadata , according to this , there should be no problem setting the metadata to .aac but it just doesn't work... [00:00] --- Thu Mar 13 2014

1 0

ffmpeg-devel.log.20140311
by burek 12 Mar '14

12 Mar '14

[00:06] <cone-589> ffmpeg.git 03Michael Niedermayer 07master:77e9123fe5d6: lavf: always use av_free [00:06] <cone-589> ffmpeg.git 03Michael Niedermayer 07master:cbe92a93026a: Merge commit '77e9123fe5d64b0960158de6e1713d3c6c7878a7' [00:09] <mark4o> lindenle: ffmpeg -help encoder=aac will show the supported sample formats [00:09] <lindenle> mark4o: thanks! [00:53] <cone-589> ffmpeg.git 03Diego Biurrun 07master:7caf48e03658: configure: miscellaneous cosmetics [00:53] <cone-589> ffmpeg.git 03Michael Niedermayer 07master:4da8dc64be50: Merge remote-tracking branch 'qatar/master' [01:09] <cone-589> ffmpeg.git 03Carl Eugen Hoyos 07master:d68ac92dc2ae: Replace an incorrect av_free() in movenc.c with av_freep(). [01:09] <cone-589> ffmpeg.git 03Carl Eugen Hoyos 07master:596841515600: Define ff_log2_run[] in libavcodec/internal.h. [01:09] <cone-589> ffmpeg.git 03Michael Niedermayer 07master:47754244463c: Merge remote-tracking branch 'cehoyos/master' [01:10] <cone-589> ffmpeg.git 03Carl Eugen Hoyos 07master:d68ac92dc2ae: Replace an incorrect av_free() in movenc.c with av_freep(). [01:11] <cone-589> ffmpeg.git 03Carl Eugen Hoyos 07master:596841515600: Define ff_log2_run[] in libavcodec/internal.h. [01:11] <cone-589> ffmpeg.git 03Michael Niedermayer 07master:47754244463c: Merge remote-tracking branch 'cehoyos/master' [01:25] <BBB> ubitux: yes, fpel is uber-trivial [02:31] <lindenle> In general how do i know what all i have to freee when using ffmpeg libraries? should i free wverything i alloc? [02:31] <lindenle> i ask b/c i am looking at the muxing example and the pkt is never freed...so I am confused ;) [02:37] <cone-589> ffmpeg.git 03Michael Niedermayer 07master:9a05e8a38502: doc/APIchanges: fill in dates [03:25] <cone-589> ffmpeg.git 03Michael Niedermayer 07master:e893440a285f: configure: make it possible to disable xlib dependancy [04:49] <cone-589> ffmpeg.git 03Michael Niedermayer 07master:e161c1bbfce9: avcodec/mjpegdec: support adob transform 2 [10:11] <ubitux> michaelni: careful with "avformat: more correct printf format specifiers" if it gets applied [10:11] <ubitux> there are a few problems with it [10:12] <ubitux> like going from signed to unsigned [10:12] <ubitux> and misc (probably harmless) nonsense [10:13] <ubitux> such as: [10:13] <ubitux> - av_log(s, AV_LOG_WARNING, "skipping CAF chunk: %08X (%c%c%c%c)\n", [10:13] <ubitux> + av_log(s, AV_LOG_WARNING, [10:13] <ubitux> + "skipping CAF chunk: %08"PRIX32" (%"PRIu8"%"PRIu8"%"PRIu8"%"PRIu8")\n", [10:13] <ubitux> not sure that's relevant [10:15] <nevcairiel> X is not standard? [10:17] <ubitux> how so? [10:17] <nevcairiel> %08"PRIX32" [10:17] <nevcairiel> i figured at least X would be a standard printf thing [11:01] <cone-924> ffmpeg.git 03Janne Grunau 07release/2.2:a1ab3300c83a: arm: hpeldsp: prevent overreads in armv6 asm [11:01] <cone-924> ffmpeg.git 03Reinhard Tartler 07release/2.2:4b476e6aa4b8: configure: enable PIC on s390(x) [11:01] <cone-924> ffmpeg.git 03Reinhard Tartler 07release/2.2:d5254230068e: Revert "Add libx265 encoder" [11:01] <cone-924> ffmpeg.git 03Reinhard Tartler 07release/2.2:72a58c077245: Update default FATE URL for release/10 [11:01] <cone-924> ffmpeg.git 03Lukasz Marek 07release/2.2:3171e2360aae: Revert "lavu/buffer: add release function" [11:01] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:123981930f0a: Merge commit 'a1ab3300c83a16c2d5f5d29c51393668b9d92667' into release/2.2 [11:01] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:ddd3301badee: Merge commit '4b476e6aa4b830f919cf3c67ba2caa039ff285b9' into release/2.2 [11:01] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:3d05625136b6: Merge commit 'd5254230068e196a2496618c0d89cdfbc41f7478' into release/2.2 [11:01] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:b82860caa7d8: Merge commit '72a58c0772450993d375c6cf4b187a068f5bc765' into release/2.2 [11:12] <cone-924> ffmpeg.git 03Janne Grunau 07release/2.2:23af29e8825a: arm: hpeldsp: fix put_pixels8_y2_{,no_rnd_}armv6 [11:12] <cone-924> ffmpeg.git 03Derek Buitenhuis 07release/2.2:a643a47d41f4: fic: Properly handle skip frames [11:12] <cone-924> ffmpeg.git 03Keiji Costantini 07release/2.2:db67b7c31b6f: rv10: Forward error from rv10_decode_packet [11:12] <cone-924> ffmpeg.git 03John Stebbins 07release/2.2:daa5a988e2ec: matroskaenc: allow override of "writing application" tag [11:12] <cone-924> ffmpeg.git 03John Stebbins 07release/2.2:c3861e14ceac: movenc: allow override of "writing application" tag [11:12] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:27e6b4a3ffec: Merge commit '23af29e8825ac112877b9ac0572ef11e5f0539f2' into release/2.2 [11:13] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:bc3648d4b476: Merge commit 'a643a47d41f4924b66fce339e4b82aaee20825be' into release/2.2 [11:13] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:46f8d838b37c: Merge commit 'db67b7c31b6fdd3747e2b5328945ad2091533698' into release/2.2 [11:13] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:66030e8133d1: Merge commit 'daa5a988e2ec8275ad8b724ea68f78306c271ae7' into release/2.2 [11:13] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:a475755b3ad6: Merge commit 'c3861e14ceace7ee69820091871173b4abcae311' into release/2.2 [11:39] <cone-924> ffmpeg.git 03Pierre Lejeune 07release/2.2:f2693e98b449: build: Use pkg-config for openjpeg [11:39] <cone-924> ffmpeg.git 03Janne Grunau 07release/2.2:5b933be089ab: arm: vp3: remove incorrect const in ff_vp3_idct_dc_add_neon declaration [11:39] <cone-924> ffmpeg.git 03Janne Grunau 07release/2.2:0ede7b534483: float_dsp: fix errors in documentation [11:39] <cone-924> ffmpeg.git 03Reinhard Tartler 07release/2.2:e4cbd0d6e5a7: changelog: Cleanups and prepare for v10_beta2 [11:39] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:5aa4b29bbefc: hevc: Use get_bits_long() in decode_vui() [11:39] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:d79cb6947e4a: hevc: use av_mallocz() for allocating tab_ipm [11:39] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:80122a3af371: Merge commit 'f2693e98b449592ec0ed4979220814bf54e60a16' into release/2.2 [11:39] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:cbabbe8220b3: Merge commit '5b933be089ab2657eb754ebf5b804ae43badf13d' into release/2.2 [11:39] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:3e4b95784793: Merge commit '0ede7b534483c5c90f404a8f11f776d2f2da4e7e' into release/2.2 [11:39] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:82c96b5ad88e: Merge commit 'e4cbd0d6e5a7b3b850d72f4f4ef0124b27dbdcbd' into release/2.2 [11:39] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:b12c5cbbb21c: Merge commit '5aa4b29bbefc06fc2bbcb52af7a14393a1bcf504' into release/2.2 [11:39] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:830c3058ff46: Merge commit 'd79cb6947e4a9c42ac20925dd920d3a0910d9a26' into release/2.2 [11:54] <durandal_1707> nice, looks like there is nothing left to reindent so they started to copy "useless" codecs from irellevant project [11:57] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:fa6b99d351ed: hevc: Do not turn 32bit timebases into negative numbers [11:57] <cone-924> ffmpeg.git 03Vittorio Giovara 07release/2.2:ca2c9d6b9bfa: hevc: make pps/sps ids unsigned where necessary [11:57] <cone-924> ffmpeg.git 03Martin Storsjö 07release/2.2:1c1e252cd1cb: movenc: Add a fallback fragmentation method for plain mp4 as well [11:57] <cone-924> ffmpeg.git 03Tim Walker 07release/2.2:ea3309eba715: movenc: use 'hev1' tag for HEVC in MODE_MOV. [11:57] <cone-924> ffmpeg.git 03Tim Walker 07release/2.2:c761379825ff: movenc: write hvcC tag for HEVC. [11:57] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:03ae616b194e: Merge commit 'fa6b99d351ed483766a875054676a56fd8459774' into release/2.2 [11:57] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:0adde39e0488: Merge commit 'ca2c9d6b9bfadb64e1502594fdf745a391699890' into release/2.2 [11:57] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:89a9c84ebb12: Merge commit '1c1e252cd1cbd5f59fe118c49f6d7207dbdfdbd4' into release/2.2 [11:57] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:3aee1fa5b650: Merge commit 'ea3309eba715e83027e8ece4a226e39a4bf2a6ce' into release/2.2 [11:57] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:3faebed6fa29: Merge commit 'c761379825ff0bf9dd191e244c4b2f7697fb2b3c' into release/2.2 [11:57] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:95ddd2227baa: avformat: fix hevcs use of golomb from avformat [12:14] <cone-924> ffmpeg.git 03Tim Walker 07release/2.2:eaa79b79b25a: movenc: enable Annex B to MP4 conversion for HEVC tracks. [12:14] <cone-924> ffmpeg.git 03Tim Walker 07release/2.2:eabefe83f40a: movenc: allow muxing HEVC in MODE_MP4. [12:14] <cone-924> ffmpeg.git 03Baptiste Coudurier 07release/2.2:7940306a47df: movdec: handle 0x7fff langcode as macintosh per the specs [12:14] <cone-924> ffmpeg.git 03Mark Himsley 07release/2.2:d37fac6dbbdd: isom: lpcm in mov default to big endian [12:14] <cone-924> ffmpeg.git 03Janne Grunau 07release/2.2:45acc228a6d5: doc: fix a couple of typos in frame.h [12:14] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:2642ad9f551d: Merge commit 'eaa79b79b25ac0ceaf44fe575a3ae724b87285b2' into release/2.2 [12:14] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:186e0ff067c4: avformat/hevc: Make return codes consistent and more flexible [12:14] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:30099413ecf6: Merge commit 'eabefe83f40a65d0f0c2a9a0521f6d96c3932545' into release/2.2 [12:15] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:ebe356bf1cc9: avformat/hevc: fix mix of av_malloc() with free() [12:15] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:bb116e6ba39f: Merge commit 'd37fac6dbbdddb76225aa691b83ffd9a0c7dae6b' into release/2.2 [12:15] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:ff1e98220568: Merge commit '45acc228a6d5f1e7d6c5ce6da63b293bd5eda57d' into release/2.2 [13:12] <cone-924> ffmpeg.git 03Vittorio Giovara 07release/2.2:6230de03aad9: vf_frei0r: refactor library loading from env variable [13:12] <cone-924> ffmpeg.git 03Vittorio Giovara 07release/2.2:bd4ad1a1d52b: vf_frei0r: fix missing end of line character [13:12] <cone-924> ffmpeg.git 03Vittorio Giovara 07release/2.2:416847d19593: vf_frei0r: prevent a segfault when filter parameters are not set [13:12] <cone-924> ffmpeg.git 03Vittorio Giovara 07release/2.2:227cfc1f10a9: vf_frei0r: adjust error messages [13:12] <cone-924> ffmpeg.git 03Luca Barbato 07release/2.2:6d7ab09788bd: golomb: Add a get_se_golomb_long [13:12] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:4a1e7a6fb7b8: Merge commit '6230de03aad9f26d5843afb913d196622e0b5b98' into release/2.2 [13:12] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:3bfb7a2537cf: Merge commit 'bd4ad1a1d52b8882df016826b8bdcf7b1009cb97' into release/2.2 [13:12] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:35e63f35b0e3: Merge commit '416847d19593e87ee1704c26a9a638fd6b0d977c' into release/2.2 [13:12] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:f67e75b5dc04: Merge commit '227cfc1f10a940c88ad3742ec805c07b6a5e7abb' into release/2.2 [13:12] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:7f8804296da9: Merge commit '6d7ab09788bdafffb3f3fc4f7feb262eb8cdf0b1' into release/2.2 [13:22] <cone-924> ffmpeg.git 03Luca Barbato 07release/2.2:b37b83214ae3: hevc: Use get_se_golomb_long [13:22] <cone-924> ffmpeg.git 03Luca Barbato 07release/2.2:2c5e1d0933fa: configure: Use the right pkgconf file for openjpeg [13:22] <cone-924> ffmpeg.git 03Tim Walker 07release/2.2:6d56bc9a6d85: lavf: simplify ff_hevc_annexb2mp4_buf [13:22] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:3533a850e766: lavf: always use av_free [13:22] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:ad8bf22086e2: Merge commit 'b37b83214ae3a462df1e8d3cc765ddbd2bfc73aa' into release/2.2 [13:22] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:bb40f8f5e22c: Merge commit '2c5e1d0933facc20c6926a788cce05d3e6cad149' into release/2.2 [13:22] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:5227eac5b040: Merge commit '6d56bc9a6d853a33fe53ab63db580c4facaba420' into release/2.2 [13:22] <cone-924> ffmpeg.git 03Michael Niedermayer 07release/2.2:f800cacada79: Merge remote-tracking branch 'qatar/release/10' into release/2.2 [13:56] <cone-924> ffmpeg.git 03Matt Oliver 07master:99b48fd448fa: Fix modplug linkage on Windows. [13:56] <cone-924> ffmpeg.git 03Anshul Maheshwari 07master:61f96be08a66: Documented av_dump_format [15:45] <cone-924> ffmpeg.git 03Carl Eugen Hoyos 07master:666749f66945: Decode png images without iend chunk. [15:45] <cone-924> ffmpeg.git 03Michael Niedermayer 07master:52b6db848d41: Merge remote-tracking branch 'cehoyos/master' [16:46] <plepere> ubitux, can you explain to me how you use pmulhrsw to do (result + offset) >> shift ? [16:46] <plepere> please. :) [16:48] <ubitux> iirc pmulhrsw is (A*B + (1<<14)) >> 15 [16:49] <ubitux> 1<<14 being for the rounding [16:49] <ubitux> pmulhrsw m0, m5 ; (x*2048 + (1<<14))>>15 <=> (x+8)>>4 [17:03] <plepere> ok [17:04] <plepere> so if I want to do result + (1<<5) >> 6 [17:05] <plepere> i do result<<9 with the multiplication [17:06] <plepere> but then the second shift is too much... [17:06] <ubitux> what's result? [17:06] <ubitux> (x*1024 + (1<<14))>>15 <=> (x+16)>>5 [17:07] <ubitux> (x*2048 + (1<<14))>>15 <=> (x+8)>>4 [17:07] <plepere> I need to do (x+32)>>6 [17:07] <ubitux> so i suppose... you want (x*512 + (1<<14))>>15 <=> (x+32)>>6 [17:07] <plepere> so basically 512 [17:07] <plepere> ok great. :D [17:08] <plepere> thanks [17:08] <ubitux> and guess what [17:08] <ubitux> <<9 is *(1<<9) [17:08] <ubitux> which is 512 [17:09] <plepere> so I guess I can skip on the operation ? [17:09] <ubitux> yes [17:09] <ubitux> pmulhrsw x, [pw_512] [17:12] <plepere> hmmm [17:14] <plepere> well it doesn't work. :p [17:15] <ubitux> too bad [17:15] <ubitux> >>> (1234+32)>>6 [17:15] <ubitux> 19 [17:15] <ubitux> >>> (1234*512 + (1<<14)) >> 15 [17:15] <ubitux> 19 [17:16] <ubitux> works for me [17:16] <plepere> I believe you. :) [17:18] Last message repeated 1 time(s). [17:50] <plepere> ok it works. [17:50] <plepere> the asm was right. it was all the rest that wasn't. [17:50] <plepere> :) [17:51] <plepere> thank you very much, ubitux [17:51] <ubitux> your welcome [17:53] <ubitux> oh fuck diego is doing that space after cast insanity [18:40] <llogan> michaelni: i set BadContent as read only if, somehow, they get smart and try to delete it [19:05] <llogan> whats the max frame size for MPEG-4 part 2 video, level 1? [19:09] <wm4> [libav-devel] [PATCH 001/132] [19:09] <wm4> fun [19:11] <ubitux> wm4: fun except the spaces after cast all over [19:11] <wm4> as expected, the main problem are cosmetic issues [19:12] <ubitux> Diego was asked to change it, but i'm pretty sure he won't, or only partially [19:12] <ubitux> that was a really great idea to stack that much patches, ofc now for cosmetic issues.. [19:13] <ubitux> and knowing that the main reason for stacking that much patches was just a dick contest with Anton TEP makes me feel really sad [19:14] <ubitux> apart from that they are intersesting changes hidden in the haystack [19:15] <ubitux> 3-4 on the first ~35 afaict [19:26] <wm4> ubitux: lol yes, they seem to love spamming large patch sets [19:26] <wm4> repeatedly spamming them seems also to be the only thing that makes the libav-devel ML as active as the ffmpeg one [19:26] <wm4> (ffmpeg has more discussion) [19:37] <cone-924> ffmpeg.git 03Michael Niedermayer 07master:01000064c951: avidec: calculate missing bitrates from index [19:37] <cone-924> ffmpeg.git 03Michael Niedermayer 07master:5d75730c58f7: avformat/avidec: Check required demuxing buffer sizes in guess_ni_flag() [20:56] <cone-924> ffmpeg.git 03Michael Niedermayer 07master:5171ae781a24: avcodec/vorbisdec: use the stored previous window type only when the actual previous is not known [21:12] <cone-924> ffmpeg.git 03Diego Biurrun 07master:d92024f18fa3: lavf: more correct printf format specifiers [21:12] <cone-924> ffmpeg.git 03Michael Niedermayer 07master:28ee7757f5b0: Merge commit 'd92024f18fa3d69937cb2575f3a8bf973df02430' [21:17] <cone-924> ffmpeg.git 03Luca Barbato 07master:5926603ad65e: configure: Document --pkg-config option [21:17] <cone-924> ffmpeg.git 03Michael Niedermayer 07master:a1b92569db65: Merge commit '5926603ad65e5fd4fefe8f0d72c381059f71cd59' [21:24] <michaelni> ubitux, about signed/unsigned fee free to revert/fix hunks if some are wrong [21:24] <ubitux> yeah i'll do that [21:26] <cone-924> ffmpeg.git 03Stephen Hutchinson 07master:908836e20743: doc: Add section about AviSynth support [21:26] <cone-924> ffmpeg.git 03Michael Niedermayer 07master:50a3c823023b: Merge remote-tracking branch 'qatar/master' [21:57] <Case> michaelni: thanks, it seems like you fixed all the bugs I reported [22:26] <lotharkript__> what should be the correct behavior for PTS/DTS from a M2ts file when the first PTS is really closed to the 33bit max? (for example, 1<<33 - 10). Should it be -10 or 1<<33 - 10 ? [22:26] <lotharkript__> for example, if I generate a file like this: ffmpeg -f s16le -ac 1 -i /dev/zero -af asetpts="PTS+95400/TB" -y test.ts [22:26] <lotharkript__> ffprobe will tell me the first pts is -42 s [22:26] <lotharkript__> but if i do, ffmpeg -f s16le -ac 1 -i /dev/zero -af asetpts="PTS+47200/TB" -y test.ts [22:26] <lotharkript__> the first PTS is positive and the PTS will never be negative and go back to 0, even after generating a 24h video. [00:00] --- Wed Mar 12 2014

1 0

ffmpeg.log.20140311
by burek 12 Mar '14

12 Mar '14

[00:43] <ALbSpirit> hello i'm using ffplay version 2.1 on ubuntu and i'm trying to play a file with ffplay. I have the video but don't have any voice on that can someone help me please. Here is the link http://pastebin.com/hq0DCfJF [02:20] <savid> How can I preserve EXIF metadata when re-encoding a video? [02:45] <Hello71> videos have EXIF? [07:55] <amigojapan> hi [07:55] <amigojapan> why is my resulting video distorted like this? do you notice how the background gets distorted in the middle of the video? http://www.youtube.com/watch?v=RP3SF_Fi1gA&feature=youtu.be [07:55] <amigojapan> The original images don't seem to be distorted. [07:55] <amigojapan> comman d line:d:\ffmpeg\bin\ffmpeg.exe -r 7 -i img%d.png -r 24 o.avi [07:55] <amigojapan> pastebin of ffmpeg output http://pastebin.com/McL2E4Kd [07:58] <relaxed> amigojapan: see if it happens with o.mkv [07:58] <amigojapan> ok relaxed , thanks [07:58] <relaxed> unnless there's some reason you need mpeg4 + avi [07:59] <amigojapan> relaxed: hmm, I dont really have a good reason.. I dont think, let me try it [08:00] <amigojapan> relaxed: thanks, that seems to have gotten rid of the distortions.... I will now go to modify my program, thanks again for each time you have helped me [08:02] <amigojapan> relaxed: I guess I can always convert to avi if I need it. [08:18] <kode54> what would cause a missing codec id = 86056 on an atrac3p containing file? [08:19] <kode54> other than, say, being an idiot and using libav [11:03] <JEEB> kode54, that comment is funny because that decoder came from Libav :P Anyways, that usually sounds like either you disabled the decoder or that you're getting the wrong library loaded [12:11] <wiretap11> anyone here using ffmpeg for DVB-S TS transcoding for streaming? [12:11] <wiretap11> i need someone with experience to help me tune my setup [12:11] <wiretap11> would love to know if there is a cheat sheet for this [12:28] <pron> wiretap11 gief me too if you find something intresting [12:28] <pron> :} [12:29] <wiretap11> :) [12:29] <wiretap11> its weird that there is very little online about this setup [12:29] <wiretap11> im sure there is a "correct" way to tackle these cases [12:29] <wiretap11> CPU vs quality samples [12:29] <pron> wiretap11 you want "live" streaming? [12:29] <wiretap11> yes [12:30] <pron> then i think i have seen some pointers in wowza forum [12:30] <pron> some time ago [12:30] <Mavrik> wiretap11, there are so many use-cases around that it's not really easy to write an all-encompasing giude [12:30] <wiretap11> I have a xeon server (4cpu with 6 cores each= total 24 cores) and I want to use ffmpeg to catpure a TS stream from http link (local) and transcode it to libx264 with low cpu usage and good quality (not more than 1000kbits). input is 720x576i and i want to keep the same resolution. my target is to encode as many streams as possible by utilising all the cpu [12:30] <Mavrik> also the user needs to have at least basinc understanding of what he wants and how things work. [12:30] <wiretap11> ffmpeg -i "http://127.0.0.1/stream.ts%E2%80%9C -acodec libmp3lame -ac 2 -ar 44100 -ab 128k -vcodec libx264 -crf 28.5 -r 25 -g 25 -vprofile High444 -level 3.1 -preset veryfast -f flv "rtmp://192.168.1.1/stream1" [12:30] <wiretap11> also i would like to know if deinterlacing is suggested and if it kills the cpu [12:30] <Mavrik> wiretap11, there is not such thing as low CPU usage and good quality [12:31] <Mavrik> pick one. [12:31] <Mavrik> also, encoding 4:4:4 is going to cause problems with most players [12:31] <Mavrik> (not to mention you're losing bitrate for questionable quality gain) [12:31] <pron> whats 4:4:4 ?:< [12:31] <Mavrik> http://en.wikipedia.org/wiki/Chroma_subsampling [12:32] <Mavrik> most players you meet in the wild only really support 4:2:0 [12:32] <JEEB> setting the profile does nothing by itself [12:32] <Mavrik> true :) [12:32] <JEEB> heck, I would be pretty sure it gets plain ignored [12:32] <JEEB> just like all those users who set -profile:v to high and then complain when the encoder has instead selected main [12:32] <Mavrik> hmm, haven't really tested it witout specifying a conversion... [12:33] <Mavrik> but ffmpeg might get smart and coerce pix_fmt :) [12:33] <wiretap11> my players do only yuv420p [12:33] <JEEB> because libx264 basically auto-selects the lowest possible one [12:33] <wiretap11> what about the -preset veryfast [12:33] <wiretap11> is that always evaluated? [12:33] <JEEB> yes [12:33] <Mavrik> yes. [12:33] <JEEB> and the setting _is_ evaluated with the profile, too [12:34] <wiretap11> i was once told that by using higher profiles i get better compression [12:34] <JEEB> libx264 just internally has a minimizer there [12:34] <JEEB> yes, if you actually use their features :P [12:34] <JEEB> just setting a profile means a flag [12:34] <wiretap11> since i dont understand the concept of the profile (i dont know the differences except the encapsulated resolution and frame rates) [12:34] <JEEB> and libx264 internally then minimizes that value according to your overall settings [12:34] <JEEB> profile is just the feature set available [12:35] <wiretap11> i do see on the other side of encoding that the profile is enforced [12:35] <wiretap11> or at least as text version string [12:35] <pron> JEEB max feature set libx264 can use? [12:35] <JEEB> wiretap11, libx264/ffmpeg should output the actually used profile somewhere in stderr during encoding [12:35] <JEEB> pron, what? [12:36] <wiretap11> JEEB: so if my input stream is a TS 576i SD video what would be the best option to encode it [12:36] <pron> JEEB nwm [12:36] <Mavrik> wiretap11, your best option is whatever your expectations are [12:36] <JEEB> you're streaming, so you're basically limited by what your consumers are limited by [12:37] <wiretap11> bitrate [12:37] <Mavrik> and requirements as well [12:37] <JEEB> on the encoding side you use the slowest profile that is still fast enough for your use case [12:37] <wiretap11> slowest is the 4.0 4.1 and below ? [12:37] <JEEB> you most probably will want to deinterlace if possible during the encoding, unless you want to make people enjoy interlaced content in flash :P [12:37] <JEEB> ugh, not profile, preset I mean [12:37] <wiretap11> ok [12:37] <JEEB> also you just used level numbers when talking about profiles [12:37] <wiretap11> well deinterlacing kills the cpu :( [12:38] <JEEB> then just make your flash clients handle deint [12:38] <wiretap11> yes its ok [12:38] <wiretap11> they handle it [12:38] <JEEB> also interlaced encoding /will/ be slower [12:38] <JEEB> just keep that in mind [12:38] <pron> and you want to choose whats the oldest consumer device you want to be able to hw decode that video [12:38] <JEEB> and weightp 2 for example isn't usable [12:38] <pron> imo [12:38] <wiretap11> yeah its fast than deinterlacing + encoding [12:39] <JEEB> pron, he is doing rtmp streaming so he is only limited by bandwidth and flash [12:39] <Mavrik> JEEB, weightp2 isn't usable? any reason? [12:39] <JEEB> Mavrik, "not implemented for interlaced coding" [12:39] <Mavrik> doh, that [12:39] <wiretap11> whats weightp2 ? [12:39] <Mavrik> thought I was missing something :) [12:39] <JEEB> IIRC it gives you a warning [12:39] <JEEB> if you try to use it [12:39] <JEEB> wiretap11, a setting used with slower presets [12:39] <wiretap11> ok [12:40] <pron> JEEB afaik old andoid devices can struggle wtih too high profile , while still able to grab rtmp [12:40] <JEEB> no they are not [12:40] <JEEB> unless they have flash player installed [12:40] <JEEB> you are speaking of rtsp [12:40] <pron> transport does nto metter [12:41] <JEEB> oh for fuck's sake [12:41] <JEEB> he is streaming for flash [12:41] <JEEB> he is doing rtmp [12:41] <JEEB> and android most fucking definitely does not support rtmp [12:41] <JEEB> at least not out-of-box [12:41] <JEEB> I know many android devices are limited, but that's just fucking out of fucking scope right now [12:41] <Mavrik> wiretap11, if you're streaming, using crf might not be a good option if your clients are on limited connection [12:42] Action: Mavrik hugs JEEB. [12:42] <JEEB> wiretap11, anyways, 1) pick the slowest preset that still gives you good enough quality 2) pick a CRF value that hits your target quality more or less (highest that still looks good, start with 23) 3) set VBV limits (-maxrate and -bufsize) according to your bandwidth [12:42] <JEEB> and bufsize should be matched on consumer's and encoder's side [12:42] <Mavrik> yes, VBV is certanly a must for streaming [12:42] <JEEB> and yes, BOTH maxrate and bufsize HAVE to be set, otherwise the thing doesn't work [12:42] <Mavrik> yep [12:43] <JEEB> if your flash thingamajig only can set buffering in seconds, it's <amount of seconds> * maxrate [12:43] <wiretap11> ok [12:43] <JEEB> VBV will make sure that your stream never goes over those limits (average maxrate over bufsize) [12:43] <wiretap11> i stream to wowza and nginx does this affect the settings? [12:44] <pron> ;D [12:44] <JEEB> only depends on if you're limited by anything regarding them on the bandwidth side of things [12:44] <wiretap11> no [12:44] <JEEB> ok [12:44] <wiretap11> u said 23)3 [12:44] <wiretap11> is that a typo ? [12:44] <JEEB> what [12:44] <JEEB> no [12:44] <wiretap11> crf [12:44] <JEEB> the 3 is part of 3) [12:44] <JEEB> 1) 2) 3) [12:45] <pron> why u so angry ?:D [12:45] <JEEB> the (highest that still looks good, start with 23) is part of 2) [12:45] <wiretap11> i lost u on the part [12:45] <JEEB> ok, fuck it [12:45] <JEEB> I'll separate the lines [12:45] <Mavrik> that reminds me [12:45] <wiretap11> setting is 28.5 now [12:45] <Mavrik> I really should submit the FLV patch for PTS wrap [12:45] <JEEB> 1) pick the slowest preset that still gives you good enough quality [12:45] <Mavrik> breaks streaming :) [12:45] <JEEB> 2) pick a CRF value that hits your target quality more or less (highest that still looks good, start with 23) [12:46] <wiretap11> ah :d [12:46] <JEEB> 3) set VBV limits (-maxrate and -bufsize) according to your bandwidth [12:46] <wiretap11> syntax :D [12:46] <wiretap11> is a bitch [12:46] <wiretap11> thx [12:47] <JEEB> always for any bandwidth limited scenario where you don't need to hit a specific average bit rate over the whole clip (as in, if the stuff doesn't need rate, then you don't need to use it), CRF + VBV + slowest preset you can take is what you need [12:48] <JEEB> in theory since flash supports 10bit H.264 if you have enough performance speed-wise you could just use a 10bit libx264 library. Gets somewhat better performance. The only negative part is that it kills any and all hardware acceleration so all consumers have to use their CPUs for decoding [12:48] <wiretap11> -acodec libmp3lame -ac 2 -ar 44100 -ab 128k -vcodec libx264 -crf 28.5 -r 25 -g 25 -maxrate 1000k -bufsize 5000k -preset veryfast -f flv URL [12:48] <JEEB> why do you set -r on the encoding side? [12:48] <wiretap11> to lower the frames [12:48] <wiretap11> from 50 to 25 [12:48] <JEEB> oh, so you're actually getting 50fps input? [12:48] <wiretap11> yes [12:48] <JEEB> as in, frames per sec [12:49] <JEEB> not fields per sec [12:49] <wiretap11> yes [12:49] <JEEB> ok [12:49] <JEEB> yeah, then it makes sense [12:49] <JEEB> you want to have one-second GOPs? [12:49] <wiretap11> i dont know [12:49] <JEEB> -g 25 makes the maximum GOP length one second [12:49] <wiretap11> i just matched the framerate [12:49] <wiretap11> yes [12:49] <wiretap11> i dont know if its good practice [12:49] <JEEB> you have a bufsize of 5 seconds right now [12:49] <wiretap11> yes [12:49] <JEEB> so I'm pretty sure you could make it a bit longer [12:50] <JEEB> say, 2sec [12:50] <wiretap11> so g 50 [12:50] <JEEB> yup [12:50] <wiretap11> ok ill test [12:51] <JEEB> but yeah, the basis for all bandwidth limited stuff is "-c:v libx264 -crf A -preset B -maxrate C -bufsize D" [12:52] <JEEB> then you add stuff on top of that to limit the stream for consumers that you have to support and might not support the stuff you would otherwise output by default [12:52] <JEEB> flash isn't exactly limited, so no such things needed really [12:52] <JEEB> Mavrik, so PTS wrapping is valid? [12:52] <JEEB> you poked Adobe about it? [12:52] <wiretap11> thanx JEEB [12:52] <wiretap11> I really appreciate the hints [12:53] <wiretap11> once i get this up i will do the same for HD streams [12:53] <wiretap11> which KILL the server :( [13:36] <theahindle> Howdy - is it possible to get a timecode from a frame? [13:37] <theahindle> Maybe timecode isn't the right word - I need to know when the video//frame was made [13:42] <koko_> hello [13:42] <koko_> how can i get Duration with ffmpeg [13:42] <koko_> ffmpeg -i song.mp3 [13:43] <koko_> ffmpeg -i song.mp3 | grep Duration ? [13:47] <steinchen> hello all.. i have a mxf file with 8 audio channels, and i need it in stereo.. to map channel 1 left and 2 right, the others can be dropped [14:06] <tim-kos> hey all. is it possible to force ffmpeg to copy over all meta data/id3tags? I have an mp3 file which contains Lyrics according to mediainfo. I encode it and after that the key is gone. I also tried -map_metadata g and -map_metadata:s:a 0:g, but no luck. Any pointers would be greatly appreciated. [14:09] <Mavrik> JEEB, technically standard does not consider or allow wrapping, however practical implementations for chunked streaming do [14:09] <Mavrik> (since it's the only thing possible) [14:10] <Mavrik> we have wrapping patch deployed to production and it works just fine [14:10] <JEEB> you should then poke adobe to add it to the spec [14:10] <Mavrik> mhm [14:10] <Mavrik> on TODO list somewhere :) [14:10] <JEEB> they after all have an e-mail there for comments etc [14:27] <theahindle> Is it possible to get the timecode of a frame using ffprobe or ffmpeg? using ffmpeg as: ./ffmpeg -i "rtmp://myserver/myapplication/mystreamname" -c copy -y -timelimit 10 -copyts ~/myfile.mp4 [14:29] <tim-kos> anybody know anything about my meta data question? [15:22] <TekniQue> I'm a bit puzzled about time stamps. I'm decoding audio and video with examples/demuxing_decoding.c. I've altered the code to print out the pkt_pts for each frame but I'm puzzled as to how I can make sure audio and video stay in sync when I process the output, when the timestamps aren't in the same format [15:23] <TekniQue> is there a way to convert the timestamps from whatever timebase they are into something real like nanoseconds [15:23] <TekniQue> because audio and video timestamps appear to be in different timebase [17:24] <Mavrik> TekniQue, well, each stream has it's own timebase [17:24] <Mavrik> if you divide PTS or DTS with time_base you get time in seconds :) [17:24] <Mavrik> there's a whole util class that deals with that [17:25] <Mavrik> av_mul_q or somethin [17:27] <TekniQue> Mavrik: the problem is, the time base of the audio as indicated is 1/44100 and the timebase of the video as indicated is 1/50 [17:27] <TekniQue> so the output of the two is completely different [17:28] <TekniQue> video_frame(cached) n:115791 coded_n:115789 pts:NOPTS size:345600 9223372036854775808 key 0 pktpts 2.31584e+06 (115792000 @ 1/50) [17:28] <TekniQue> audio_frame n:199470 nb_samples:1024 pts:NOPTS 4631.68 (204257280 @ 1/44100) [17:30] <TekniQue> but fwiw, the video stamps increment by 1000 per frame [17:30] <TekniQue> while the audio stamps increment by 1024 per frame [17:30] <TekniQue> as each audio frame appears to hold 1024 samples [17:34] <Mavrik> that's... completely normal :) [20:32] <nordlys> Hello. Is it possible to make encoding with -loglevel info reuse/update the same line for the encode process rather than make a new line for each update? [Windows] [20:33] <Maverick|MSG> is it possible to screencapture from a specific process id, rather than the entire desktop? [20:35] <llogan> Maverick|MSG: you can give the screen coordinates to capture a specific window. [20:35] <Maverick|MSG> let's assume I don't know those [20:36] <llogan> then use xwininfo and click on the desired window [20:44] <Maverick|MSG> llogan ffmpeg can't figure out the window dimensions automatically? [20:44] <Maverick|MSG> (and x/y position) [21:23] <ztx> hello, i was wondering if it is possible to copy streams directly from input to output without transcoding? [21:23] <ztx> specifically i'd like to copy the audio from one WMV file and the video from another WMV file, and put them into an output WMV [21:23] <ztx> i've checked the documentation but can't fathom how to do it [21:24] <average> ztx: that's because you haven't read it enough [21:24] <average> ztx: I don't mean to come across as a complete douchebag, but read-the-fine-manual :) [21:25] <ztx> you're probably right! but please could you suggest a hint of what to search for? [21:27] <average> 5359 questions tagged with [ffmpeg] on SO http://stackoverflow.com/questions/tagged/ffmpeg [21:27] <average> 938 tagged with [video-processing] http://stackoverflow.com/questions/tagged/video-processing [21:28] <average> ffmpeg-user mailing list http://ffmpeg.org/pipermail/ffmpeg-user/ [21:29] <average> ffmpeg-devel mailing list http://ffmpeg.org/pipermail/ffmpeg-devel/ [21:29] <average> also see /topic for some forums [21:30] <average> ztx: I know I'm an idiot, but I'd rather not give you the answer, since this is probably not the only question you'll have about ffmpeg [21:31] <average> ztx: so I'd like to guide you to places where it's likely that you'll find the answers(giving you the fishing rod instead of the fish so-to-say) [21:31] <ztx> thanks that actually helps a lot, i didn't think of SO and hadn't yet noticed there is such an active mailing list [21:32] <ztx> sorry for coming in and asking questions without checking all the available sources [21:32] <average> no problem [21:39] <ztx> ffmpeg -i AUDIOVIDEO.wmv -i JUSTVIDEO.wmv -map 1:0 -map 0:0 -codec copy OUTPUT.wmv [21:39] <ztx> did the trick [21:39] <ztx> wonderful software [22:02] <llogan> Maverick|MSG: no, but you can probably script it. [22:03] <Maverick|MSG> thanks llogan. but will ffmpeg capture the window's contents or just the area of the desktop based on the x/y and width/height? [22:04] <Maverick|MSG> basically: I want to capture what is in a certain window, even if that window is moved or another window overlaps it [22:22] <shevy> how does ffmpeg compare to sox in terms of speed, when i.e. converting a .mp3 into an .ogg file ? anyone knows? just a rough estimate [22:24] <Gregor> Is there an ffmpeg filter that can drop frames based on some external list of frames to be dropped? [22:24] <Gregor> (Rather than internal calculation) [22:26] <llogan> Maverick|MSG: just the specified area of the desktop unless you use -follow_mouse [22:32] <Maverick|MSG> darn, ok. thanks [00:00] --- Wed Mar 12 2014

1 0