Ffmpeg-devel-irc

Download

ffmpeg-devel-irc@ffmpeg.org

January 2013

1 participants
60 discussions

ffmpeg-devel.log.20130120
by burek 21 Jan '13

21 Jan '13

[01:12] <cone-175> ffmpeg.git 03Carl Eugen Hoyos 07release/1.1:90da0cb60e1c: The c99-to-c89 binaries are now hosted on videolan.org. (cherry picked from commit c29c7c1470f98f8e66752fb09c44ded625e9a991) [01:12] <cone-175> ffmpeg.git 03Michael Niedermayer 07release/1.1:bd531038e8af: init_get_bits8: check byte_size against being positive [01:12] <cone-175> ffmpeg.git 03Michael Niedermayer 07release/1.1:6f3bc92c294b: init_get_bits8: zero pointers & struct on error [01:12] <cone-175> ffmpeg.git 03Michael Niedermayer 07release/1.1:113592890332: init_get_bits: fix off by 1 error [01:12] <cone-175> ffmpeg.git 03Michael Niedermayer 07release/1.1:edc00dea020b: update for 1.1.1 [02:31] <cone-175> ffmpeg.git 03Michael Niedermayer 07fatal: ambiguous argument 'refs/tags/n1.1.1': unknown revision or path not in the working tree. [02:31] <cone-175> Use '--' to separate paths from revisions [02:31] <cone-175> refs/tags/n1.1.1:HEAD: sws: check that dither array is not null before memset [04:25] <cone-175> ffmpeg.git 03Michael Niedermayer 07master:0c6b0409af07: mpeg1enc: Disable threads for resolutions too large for multi-threading [04:40] <cone-175> ffmpeg.git 03Michael Niedermayer 07release/1.0:bca27b0480b6: swr: fix handling of timestamps that cause multiple drops or silence injections [04:40] <cone-175> ffmpeg.git 03Ronald S. Bultje 07release/1.0:ac6b2fecf473: h264: don't clobber mmco opcode tables for non-first slice headers. [04:40] <cone-175> ffmpeg.git 03Ronald S. Bultje 07release/1.0:d369cc83aa76: h264: add 3 pixels below for subpixel filter wait position. [04:40] <cone-175> ffmpeg.git 03Michael Niedermayer 07release/1.0:3e36cba430f7: mpeg12: Support decoding dimensions that are a multiple of 4096 [04:40] <cone-175> ffmpeg.git 03Michael Niedermayer 07release/1.0:5af41713046e: mpeg12enc: Correctly mask dimensions [04:40] <cone-175> ffmpeg.git 03Michael Niedermayer 07release/1.0:95f8701c32d8: mpeg12enc: check dimension validity [04:40] <cone-175> ffmpeg.git 03Michael Niedermayer 07release/1.0:ece68a0363ac: mpeg1enc: Disable threads for resolutions too large for multi-threading [05:07] <cone-175> ffmpeg.git 03Matthieu Bouron 07master:b3b456b2588f: lavf/id3v2: seek back to previous offset if header size is not matched [14:02] <cone-223> ffmpeg.git 03Clément BSsch 07master:40976257bc35: dvdsubdec: parse the size from the extradata [14:02] <cone-223> ffmpeg.git 03Xi Wang 07master:992b03183819: mpegvideo: fix loop condition in draw_line() [14:02] <cone-223> ffmpeg.git 03Ronald S. Bultje 07master:aeaf268e52fc: vp3: integrate clear_blocks with idct of previous block. [14:02] <cone-223> ffmpeg.git 03Michael Niedermayer 07master:cf061a9c3b86: Merge commit 'aeaf268e52fc11c1f64914a319e0edddf1346d6a' [14:19] <cone-223> ffmpeg.git 03Ronald S. Bultje 07master:fef906c77c09: Move vorbis_inverse_coupling from dsputil to vorbisdspcontext. [14:19] <cone-223> ffmpeg.git 03Michael Niedermayer 07master:c62cb1112ffc: Merge commit 'fef906c77c09940a2fdad155b2adc05080e17eda' [14:30] <cone-223> ffmpeg.git 03Ronald S. Bultje 07master:0e02b381b485: Remove put_no_rnd_pixels_l2 function pointer for w=16 from dsputil. [14:30] <cone-223> ffmpeg.git 03Michael Niedermayer 07master:28245fb466d8: Merge remote-tracking branch 'qatar/master' [15:11] <durandal_1707> is there way to silence ffmpeg configuaration output [15:17] <cbsrobot> durandal_1707: ./configure > /dev/null [15:17] <cbsrobot> or when using ffmpeg ? [15:17] <cbsrobot> ffmpeg -loglevel quiet [15:22] <cbsrobot> panic and fatal may work too [15:30] <cone-223> ffmpeg.git 03Michael Niedermayer 07master:60af6c3138dc: h264: Only apply error concealment if theres a frame [15:34] <michaelni> funman, all out of array reads with http://people.videolan.org/~funman/lol.flv should be fixed with ffmpeg [15:51] <durandal_1707> cbsrobot: quiet is tooo quiet for me [15:56] <durandal_1707> is there example how to mix 5 ending seconds of audio with 5 starting seconds of another audio? [17:19] <cone-223> ffmpeg.git 03Michael Niedermayer 07master:cc4a41727e29: swr: fix "may be used uninitialized in this function" warnings [17:19] <cone-223> ffmpeg.git 03Michael Niedermayer 07master:6136b989f658: vc1dec: add assert to impossible *_valid combination. [17:42] <funman> michaelni: ok, i didn't use valgrind but the file seemed to decode fine [17:43] <funman> michaelni: i still have my buggy muxer around if you want me to generate a few more files ;) [17:48] <michaelni> if they do out of array accesses or crash, sure ... [17:49] <funman> did my sample crash for you or you noticed the bug in valgrind? [17:52] <michaelni> valgrind [17:52] <michaelni> no crash [17:57] <cehoyos> The sample does not even crash on Windows which surprised me (crashes were reported for samples that valgrind only reported with an out-of-bounds read of 1). [17:59] <funman> hm the installed libraries are stripped? [17:59] <cbsrobot> cehoyos: thanks for the gdb hint &. [17:59] <funman> http://pastie.org/5736871 [18:00] <funman> #2 0x0000000006c657ff in av_strtod () from /home/fun/.local/ffmpeg/lib/libavutil.so.52 [18:01] <cone-223> ffmpeg.git 03Michael Niedermayer 07master:7cb27d216d1c: oggdec: Support chained streams, support replacing streams in multistream files. [18:01] <cone-223> ffmpeg.git 03Michael Niedermayer 07master:bc15fcb8cfa8: theora: support midstream reconfiguration [18:01] <cehoyos> cbsrobot: ?? [18:01] <cbsrobot> cehoyos: in some ticket I posted lately [18:02] <funman> michaelni: http://pastie.org/5736882 (not sure if false positive or glibc bug) [18:03] <michaelni> smells like glibc bug [18:07] <funman> => 0x0000000006c859bf <av_strtod+47>: movapd %xmm0,%xmm2 [18:08] <funman> xmm0 & xmm2 are all 0s [18:09] <cehoyos> ffplay -threads 1 chained_streams.ogg looks very different from mplayer chained_streams.ogg (ffmpeg/ffplay eats the part in the middle) [18:09] <michaelni> http://sourceware.org/bugzilla/show_bug.cgi?id=12424 [18:13] <michaelni> cehoyos, it doesnt eat it but it seems its displayed only for a short time [18:15] <funman> michaelni: thx [18:26] <cehoyos> michaelni: ffmpeg eats it (and ffplay plays it as fast as possible iiuc) [18:42] <michaelni> cehoyos, patch that fixes that on the ML [20:06] <durandal_1707> what happened with aac patch from madshi? [20:09] <durandal_1707> j-b: why is there bouny only for some codecs asked to be supported in vlc [20:09] <durandal_1707> *bounty [20:18] <kierank> durandal_1707: depends on demand [20:18] <kierank> from vlc users [20:19] <cone-223> ffmpeg.git 03Michael Niedermayer 07master:fc86f88b32a2: h264: fix use of uninitialized variable [20:19] <cone-223> ffmpeg.git 03Michael Niedermayer 07master:dcbe15813ed0: h264: print mismatching opcode details in check_opcodes() [20:19] <cone-223> ffmpeg.git 03Michael Niedermayer 07master:3fba3e79a3e4: h264/ff_generate_sliding_window_mmcos: fix use of uninitialized variable [20:23] <durandal_1707> kierank: i see nowhere demand list where people can vote, how is demand list made. my some random ad-hoc decision? [20:23] <kierank> it was done with agreement of the vlc developers [20:23] <kierank> based on people who request codec support [20:23] <kierank> it is the vlc foundations money so they can do what they like with it [20:29] <funman> durandal_1707: yeah we made the list ourselves [20:33] <durandal_1707> funman: pretty sparse one [20:34] <Daemon404> see what kierank said. [20:36] Action: durandal_1707 have already read what kierank said. [20:45] <kierank> durandal_1707: what exactly do you want on it [20:45] <funman> durandal_1707: if you have anything to suggest, before fosdem is the right time [20:46] <funman> +1 [20:46] <Daemon404> funman, some of those bounties are really easy [20:46] <Daemon404> notable "opus encoder wrapper" [20:46] <Daemon404> considering lavc has one iirc [20:48] <funman> Daemon404: feel free to claim it [20:48] <Daemon404> i have no intention of learning vlc's internals [20:48] <funman> you'll make easy bucks, users will be happy, world will be a better place [20:49] <Daemon404> last time i tried, the code i had to touch was a giant mess of spaghetti [20:49] <Daemon404> (i tried to add ARGB support) [20:49] <Daemon404> (colorspace stuff is accross N files in confusing ways) [20:49] <funman> well i started some work on opus but i stopped at ogg muxer [20:49] <j-b> durandal_1707: why would I care about anything else? [20:49] <Daemon404> funman, thats how it goes. everybody hates ogg. [20:50] <funman> anyway it's not so easy in the end if you consider everything [20:52] <durandal_1707> j-b: i really don't care [20:53] <Daemon404> funman, i really like how every foss media framework in existence duplicates teh same work [20:53] <Daemon404> it's really awesome [20:53] <Daemon404> and they all port eachothers stuff [20:55] <funman> Daemon404: yes it is not a perfect situation [20:55] <Daemon404> but it is hilarious. [20:55] <Daemon404> thats an upside [20:55] <funman> xine vlc ffmpeg libav gstreamer ? [20:56] <Daemon404> mplayer [20:56] <Daemon404> avisynth [20:56] <Daemon404> the latter being where everyone gets all their plugin code to prot from in teh first place [20:56] <Daemon404> or well i should note [20:56] <Daemon404> all /three/ mplayers [20:57] <Daemon404> mplayer, mplayer2, mpv [20:57] <Daemon404> \o/ [20:57] <wm4> hey I'm trying to use ffmpeg for everything [20:57] <wm4> (with rather mixed results...) [21:00] <mfg> michaelni saw your email... looking at it [21:02] <wm4> Daemon404: also there's a 4th mplayer that has some users, mplayer-ww (some chinese windows-only thing) [21:03] <durandal_1707> xine is dead, vlc is dying and gstreamer is going to die [21:03] <wm4> wat [21:03] <Daemon404> netcraft confirms bsd is dead [21:03] <durandal_1707> and indeed bsd is dead [21:04] <Daemon404> wm4, vlc being the single most downloaded foss project on sf.net == dead [21:04] <Daemon404> millions of user? poppycock! [21:06] <cone-223> ffmpeg.git 03Michael Niedermayer 07master:3317414fc11a: wtvdec: fix integer overflow [21:06] <funman> Daemon404: what do you mean? [21:06] <Daemon404> funman, about what? [21:06] <Daemon404> my trolling? [21:07] <funman> ah i see that durandal_1707 started the troll [21:07] <funman> sorry [21:07] <wm4> also someone on #mplayer is trying to use mplayer-rc1 [21:07] <wm4> which is from 2006 [21:07] <funman> durandal_1707: vlc thanks you well from its dying bed [21:07] <j-b> sparse list, of course, but noone wants to do any of those [21:08] <Daemon404> i heard j-b will buy you a car if you finish G2M [21:08] <j-b> except janne and Simon [21:08] <j-b> durandal_1707: I can buy you a car [21:08] <Daemon404> lol [21:09] <Daemon404> saste or ubitux -- ping on something ffprobe related [21:09] <funman> my car cost me one local resampling at the time [21:09] <funman> or about G2M + Voxware [21:09] <Daemon404> isnt voxware basically massive tables [21:09] <j-b> probably, it is [21:10] <j-b> but, anyway, people complained about no money sent to libav* development, especially from big companies. I did the bounties list for that. [21:10] <j-b> and when the first bounties will be done, new ones will take place [21:14] <Daemon404> yeah ok, whoever wrote sortdts should be shot [21:14] <Daemon404> daemon404@bbvm:~/dev/f/ffmpeg$ grep AVFMT_FLAG_SORT_DTS * -r [21:14] <Daemon404> libavformat/options_table.h:{"sortdts", "try to interleave outputted packets by dts", 0, AV_OPT_TYPE_CONST, {.i64 = AVFMT_FLAG_SORT_DTS }, INT_MIN, INT_MAX, D, "fflags"}, [21:14] <Daemon404> libavformat/avformat.h:#define AVFMT_FLAG_SORT_DTS 0x10000 ///< try to interleave outputted packets by dts (using this flag can slow demuxing down) [21:14] <Daemon404> libavformat/avidec.c: avi->non_interleaved |= guess_ni_flag(s) | (s->flags & AVFMT_FLAG_SORT_DTS); [21:14] <Daemon404> ^ it's a single hack for avi only [21:15] <Daemon404> but hey lets give it its own option and vague description [21:15] <michaelni> durandal_1707, if you want some work funded that has alot of users who would want it done we could put a news entry on the homepoage and maybe there would be donations for it ... [21:15] <Daemon404> michaelni, a blurb on vlc's homepage would be mroe effective [21:15] <Daemon404> tons more traffic [21:16] <Daemon404> oh my bad.. it doesn't actually do anything for avi either [21:16] <Daemon404> :< [21:16] <Daemon404> why does this exist [21:17] <michaelni> Daemon404, to annoy you ;) [21:17] <Daemon404> i was going to go add sortpts [21:17] <Daemon404> which would have been useful to me [21:17] <Daemon404> but... even sortdts doesn't exist O.o [21:17] <Daemon404> lies!!! [21:19] <j-b> videolan.org does 10 millions views per month [21:21] <michaelni> Daemon404, btw, a sortpts would not work, because packets cant be decoded in pts order if that differs from dts order [21:21] <durandal_1707> j-b: enjoy until it last [21:21] <Daemon404> michaelni, im using it for ffprobe's -show_packet [21:21] <Daemon404> nothing is decoded [21:21] <michaelni> ahh ok [21:22] Action: durandal_1707 breaking bad ... [21:23] <j-b> durandal_1707: why are you soo negative? did someone beat you with a stick this morning? [21:23] <Daemon404> michaelni, i might just sort it in my own script (luckily ffprobe has json output...) [21:24] <michaelni> yep, thats probably easiest, [21:25] <michaelni> would be interresting to have direct support in ffprobe or lavf for it though [21:26] <Daemon404> should be done at a higher level probably [21:26] <Daemon404> demuxer-specific sorting is semi evil [21:26] <Daemon404> and unmaintainable [21:27] <michaelni> for the ffprobe usecase yes [21:27] <michaelni> if you want it realtime and your avi is non interleaved ... [21:27] <michaelni> then non interleaved demuxing or 2 demuxers are needed [21:28] <Daemon404> well for pts, the latter use case just doesnt exist [21:28] <michaelni> true [21:29] <Daemon404> i dont adding sorting to my script(s) will be much overhead [21:29] <Daemon404> compared to C, anyway [21:29] <Daemon404> sorting is always implemenetd in C in the langs anyway [21:36] <cone-223> ffmpeg.git 03Xi Wang 07master:64b7e7dcaf21: cafdec: fix overflow checking in read_header() [21:49] <saste> michaelni, ok to remove CREDITS? [21:52] <michaelni> if no people are lost by pointing to git and [21:52] <michaelni> maybe put a note in it to point to git [21:52] <michaelni> btw, can filter->inputs[i] ever be null ? [21:52] <michaelni> [FFmpeg-devel] [PATCH] lavfi: fix null checking in avfilter_config_links() <-- [21:53] <saste> michaelni, no if i >= nb_inputs [21:53] <saste> Daemon404, pong? [21:53] <michaelni> saste, please reply then [21:53] <michaelni> i < nb_inputs i assume [21:54] <beastd> saste: >=? [21:54] <saste> michaelni, yes that's a precondition [21:54] <Daemon404> saste, already discussed [21:54] <Daemon404> was re: -sortdts [21:55] <saste> i was away ~2 days and now my inbox is a mess [21:57] <michaelni> saste, i can reply to the "lavfi: fix null checking in avfilter_config_links()" if you are busy with your inbox ? [21:58] <saste> i'll catch up today or tomorrow, but feel free to OK it if the patch is correct [22:06] <michaelni> saste, i think filter->inputs[i] can be NULL if functions are called in the wrong order [22:06] <michaelni> thus the patch should be ok [22:06] <saste> michaelni, lemme see the patch [22:07] <michaelni> saste: [FFmpeg-devel] [PATCH] lavfi: fix null checking in avfilter_config_links() [22:08] <michaelni> if the user forgets linking filters then the null check can be true [22:08] <michaelni> i think [22:40] <durandal_1707> so how to do what asyncts=first_pts=0 does? [22:42] <saste> durandal_1707, or use asyncts if it works for you [22:45] <durandal_1707> saste: asyncts needs libavresample which is not enabled by default [22:48] <Daemon404> as far as i know, only asyncts has that functionality [22:48] <Daemon404> i am specifically enabling avresample in ym builds so i can use it [22:49] <durandal_1707> what about asetpts=PTS-STARTPTS [22:49] <Daemon404> never tried [22:50] <Daemon404> it it relies on prior knowledge of pts info [22:50] <Daemon404> doesnt it [22:50] <Daemon404> cant be used pragmatically [22:50] <Daemon404> and afaik setpts wont actually trip or pad audio [22:50] <Daemon404> trim* [22:51] <durandal_1707> apad can pad it [22:51] <Daemon404> again, thats not a legitimate replacement [22:51] <Daemon404> more complex, requires prior knowledge [22:52] <Daemon404> why would i bother with that stuff if i can just enable asyncts and have it Just Work [22:53] <saste> stretching/squeezing/filling or trimming [22:53] <saste> this can be done with aresample, but documentation is (still) poor [22:53] <Daemon404> saste, do tell how to accomplish the same as asyncts=start_pts=0 [22:53] <Daemon404> i dotn see a way to do that [22:53] <saste> Daemon404, dunno [22:54] <Daemon404> anything that requries outside parsing or eval() is already failed in my mind [22:54] <Daemon404> compared to asyncts [22:54] <Daemon404> :P [22:54] <durandal_1707> it is trivial to port asyncts to lswr [22:56] <Daemon404> the question is do you want to deal with the merge conflicts when the actual author of asyncts updates its code? [22:56] <Daemon404> or do you want to fork it into a new filter [22:56] <saste> durandal_1707, min_comp, min_hard_comp, maybe async may help [22:56] <durandal_1707> dont be silly what new could be added to asyncts? [22:57] <durandal_1707> it is just usual boring maintaince work to port to W api [22:57] <Daemon404> bug fixes [22:57] <Daemon404> critical fixes actually went in not to logn ago [22:58] <Daemon404> to fix start_pts [22:58] <Daemon404> at my request [22:58] <Daemon404> =p [22:58] <Daemon404> [16:56] <@durandal_1707> dont be silly what new could be added to asyncts? <-- btw this is a horrible attitude for any dve [22:58] <Daemon404> only gonna lead to shitcode [22:58] <saste> keep asyncts, add features to aresample, document better aresample/resampler options [22:58] <durandal_1707> i disagree [22:59] <Daemon404> saste, ill gladly switch to aresample when i can a) have it do what i want b) figure out how to use it [22:59] <Daemon404> =[ [22:59] <Daemon404> =p [22:59] <michaelni> Daemon404, do you have a testcase for the features you need so i can take a look how to do it with aresample [22:59] <michaelni> ? [22:59] <Daemon404> i can trivially make some [22:59] <michaelni> please do [22:59] <durandal_1707> michaelni: another user asked same q on user ml [22:59] <Daemon404> basically i need [22:59] <Daemon404> if start pts != 0 [22:59] <durandal_1707> say few minutes ago [22:59] <Daemon404> pad or trim so it is 0 [22:59] <Daemon404> or rather [23:00] <Daemon404> until it is start_pts [23:00] <Daemon404> it's useful for audio sync stuff [23:01] <Daemon404> e.g. it can drop negatively pts'd audio samples [23:01] <Daemon404> or pad from 0 -> start [23:01] <Daemon404> with silence [23:01] <Daemon404> (used in conjunction with -copyts) [23:02] <saste> Daemon404, drop negative PTS => aselect [23:03] <saste> pad from the start: aresample=min_comp=0.001:min_hard_comp=0.100000 [23:03] <Daemon404> or i can stick with the simple asyncts [23:03] <Daemon404> which is not at all complex [23:03] <saste> than i agree that the aresample params should be documented better [23:03] <Compn> whats -vf fixpts do ? [23:03] <saste> Daemon404, sure [23:03] <Compn> i mean , not what you want, probably [23:03] <Daemon404> aresample=min_comp=0.001:min_hard_comp=0.100000 <-- thats ... what [23:03] <Daemon404> i cant even figure out what that means [23:04] <Daemon404> how how it owuld pad [23:04] <saste> you should use what works best for you [23:04] <Compn> Daemon404 : any chance for a sample file ? [23:04] <Daemon404> mux any file with an audio delay [23:04] <Daemon404> bam [23:04] <Daemon404> sample file [23:05] <saste> asetpts=PTS+X/TB [23:05] <Daemon404> saste, my only annoyance is the trolling whenever i mention asyncts as a solution to others :P [23:05] <saste> or something like that to add an initial delay [23:05] <Daemon404> "omg it uses avresample! hresey!" [23:06] <saste> OTOH we should improve what we have, since lavr is disabled by default [23:06] <saste> not just for silly pride [23:07] <michaelni> Daemon404, so i get no testcase ? [23:07] <Daemon404> michaelni, when i cazn be bothered to make one [23:07] <Daemon404> it's just a case of muxing e.g. an mp4 file with an audio delay (using gpac or l-smash) [23:07] <Daemon404> one positive delay test file, one negative delay test file [23:08] Action: Daemon404 is busy playing xbox and beign unproductive on a sunday. [23:08] <durandal_1707> porting is far simpler long term and avoids user questions.... [23:09] <Daemon404> and im trying to fix a different annoying problem right now [23:09] <Daemon404> some files (sup mpeg-ts) dont treat 0 as the defacto start point [23:09] <durandal_1707> how to kill that guy above ...? [23:10] Action: Daemon404 loves files where teh first dts=pts=<massive number> [23:10] <durandal_1707> cehoyos: you are not helpful to user questions at all [23:12] <cehoyos> durandal_1707: I am extremely happy if you take over! [23:12] <durandal_1707> user asked to simply translate asyncts=first_pts=0 reading asyncts code should provide enough knowledge to provide/seek/produce working help... [23:13] <durandal_1707> i really do not see point why you ask for full crap output and testcase and bla balllbal [23:13] <cone-223> ffmpeg.git 03Michael Niedermayer 07release/1.0:04c74a339ff7: update for FFmpeg 1.0.2 [23:13] <cone-223> ffmpeg.git 03Michael Niedermayer 07release/1.0:3e57b37ddb25: aacenc: Fix assignments in if() [23:13] <cone-223> ffmpeg.git 03Michael Niedermayer 07release/1.0:28ef8eaeefe3: libfaac: Fix assignments in if() [23:13] <cone-223> ffmpeg.git 03Michael Niedermayer 07release/1.0:6f2aa8adbf65: libfdk-aacenc: Fix assignments in if() [23:13] <cone-223> ffmpeg.git 03Michael Niedermayer 07release/1.0:3daa086b2eb4: libmp3lame: Fix assignments in if() [23:13] <cone-223> ffmpeg.git 03Michael Niedermayer 07release/1.0:57211d8a5f5d: libopencore-amr: Fix assignments in if() [23:13] <cone-223> ffmpeg.git 03Michael Niedermayer 07release/1.0:d7826b8b9b47: libspeexenc: Fix assignments in if() [23:13] <cone-223> ffmpeg.git 03Michael Niedermayer 07release/1.0:461907760fdf: libvo-aacenc: Fix assignments in if() [23:13] <cone-223> ffmpeg.git 03Michael Niedermayer 07release/1.0:2ff4257e642a: libvorbisenc: Fix assignments in if() [23:13] <cone-223> ffmpeg.git 03Michael Niedermayer 07release/1.0:1489e67072b7: nellymoserenc: Fix assignments in if() [23:13] <cone-223> ffmpeg.git 03Michael Niedermayer 07release/1.0:f00a9f00034c: ra144enc: Fix assignments in if() [23:13] <cone-223> ffmpeg.git 03Michael Niedermayer 07release/1.0:ee8d38191fbf: tiff: Fix assignments in if() [23:13] <cone-223> ffmpeg.git 03Michael Niedermayer 07release/1.0:dadf58908ceb: vsrc_testsrc: Fix assignments in if() [23:13] <durandal_1707> while user problem is quit obvious [23:13] <cone-223> ffmpeg.git 03Michael Niedermayer 07release/1.0:b006b2f4f54a: rtmpproto: Fix assignments in if() [23:13] <cehoyos> So to sum up: You cannot answer Daemon404's question how to translate a asyncts command line into aresample (I suspect there is nothing to translate in his case, because all our mp4 delay files play fine with -async 1), but I should guess what a user wants? [23:13] <durandal_1707> he want to change pts [23:14] <Daemon404> [17:13] <@durandal_1707> he want to change pts <-- that is nto an accurate description [23:14] <Daemon404> whatsoever [23:14] <Daemon404> changing the timestamps is NOT the same as padding/tirmmign audio [23:14] <Daemon404> based on timestamps [23:14] <Daemon404> they are two very distinct and different operations [23:17] <durandal_1707> so he want to pad audio so its pts becomes 0? [23:19] <Daemon404> it's first pts [23:19] <Daemon404> or if it is negative pts, drop them up intol 0 [23:19] <Daemon404> until* [23:19] <durandal_1707> michaelni: there is no equvivalet for avresample_available in swr? [23:20] <durandal_1707> i just wonder is swr capable doing of same filter without extra complicated code... [23:21] <durandal_1707> leaving now, i expect tomorrow when i woke up that this will be solved :evil [23:31] <cone-223> ffmpeg.git 03Carl Eugen Hoyos 07master:cb1c69c81d91: Clarify that the recommended audio squeezing/stretching filter is aresample. [23:42] <cone-223> ffmpeg.git 03Carl Eugen Hoyos 07fatal: ambiguous argument 'refs/tags/n1.0.2': unknown revision or path not in the working tree. [23:42] <cone-223> Use '--' to separate paths from revisions [23:42] <cone-223> refs/tags/n1.0.2:HEAD: Clarify that the recommended audio squeezing/stretching filter is aresample. [23:42] <Daemon404> lulz [23:42] <saste> and yes compensation docs need to be improved (for example: filled samples are silence-samples) [00:00] --- Mon Jan 21 2013

1 0

ffmpeg.log.20130120
by burek 21 Jan '13

21 Jan '13

[00:17] <cbsrobot> llogan: so just bussines as usual then [00:24] <Zeeflo> How are you guys this fine evening! [00:37] <sloof> llogan: What are you encoding that is taking so long? [00:38] <klaxa> 24 minute clip, 1080p -preset placebo :3 [00:40] <Zeeflo> WHAT+ [00:40] <Zeeflo> ? [00:40] <Zeeflo> is someone using placebo? [00:40] <Zeeflo> thats insane! Not to say the least.. [00:42] <klaxa> that was a joke :P [00:42] <klaxa> although... [00:42] <klaxa> i once encoded a 90 seconds clip 1080p with placebo preset [00:44] <Zeeflo> how long did it take? [00:49] <klaxa> Zeeflo: 3h 42m [00:49] <klaxa> cpu[2 x Intel(R) Core(TM)2 Duo CPU P7350 @ 2.00GHz (GenuieIIntel) @ 800MHz w/ 3072 KB L2 Cache] [00:50] <Zeeflo> ahh ok. A faster CPU would probably have helped hehe [00:50] <klaxa> well obviously :P [00:50] <Zeeflo> let me know if I can help with that ;) [00:51] <klaxa> oh you are interested in donating CPU time? :> [00:51] <Zeeflo> sure [00:51] <klaxa> i might come back to that :P [00:51] <Zeeflo> I actually have a server in the US you can rape for 5 more days if you want [00:51] <Zeeflo> dual xeon, 24 threads [00:51] <Zeeflo> 8 gb ram (only) [00:51] <Zeeflo> but ssd [00:51] <klaxa> ah too bad, might not be able to harvest that within those few days :P [00:51] <Zeeflo> and 1 tb [00:52] <Zeeflo> you can have ssh now! [00:52] <Zeeflo> its ready to go [00:52] <Zeeflo> i dont care about it [00:52] <klaxa> over the holidays i started writing some software to distribute encoding over multiple clients [00:52] <Zeeflo> its already cancelled [00:52] <klaxa> well yeah the software isn't failsafe and i have no data i need processed right now :S [00:52] <Zeeflo> ah ok.. Well. in the future as well. [00:52] <Zeeflo> I moved my data to some big ass servers in germany that does the same [00:53] <Zeeflo> which im keeping [00:53] <Zeeflo> let me know ;) [00:53] <klaxa> it's also not very... advanced, it's a naive high-level "hurr durr i call mkvextract and shit" approach [00:53] <klaxa> if you want to take a look and propose improvements, i'd love it https://github.com/klaxa/Distributed-encoding [00:54] <klaxa> i also thought about moving to ffmpeg with libx264 instead of using x264 itself because of libavfilter [00:55] <Zeeflo> well for clustering i use Beowuld [00:55] <Zeeflo> Beowulf [00:55] <Zeeflo> its already invented and works very well [00:57] <klaxa> i see... [00:57] <klaxa> how does it perform on end-user internet connections? [00:57] <klaxa> i assume not that performant? [00:58] <Zeeflo> Well, it performs as good as your internet connection is.. [00:58] <Zeeflo> But honestly, I would only cluster servers if theyre situated in a DC [00:59] <Zeeflo> or, if you have speeds like we do here in denmark/sweden [00:59] <klaxa> well there are a lot of people doing torrenting from home, i see it more like that ;) [00:59] <klaxa> you download a chunk provided by the server, encode it locally (i.e. actually disconnect from the server) then send it back [01:00] <klaxa> there are still problems regarding connections and software failing [01:00] <klaxa> i haven't found time and motivation to work on that [01:06] <Zeeflo> I dont know [01:06] <Zeeflo> i dont really mess with torrenting [01:07] <klaxa> it's more like a mentality comparision, although then again there is only one person that gains from it [01:08] <klaxa> i want private users to be able to contribute without having to buy expensive hardware [01:08] <klaxa> kinda crowd-sourcing, but less subtle [02:39] <llogan> sloof: just a bunch of boring old videos for some biologists [02:46] <llogan> klaxa: interesting license [02:46] <klaxa> wtfpl best license [02:51] <klaxa> it may be very liberal, but i really don't care what other people do with the code [02:52] <klaxa> i write it for myself and even if someone else just puts his/her name on it and claims it theirs, i don't really care, because for me it's enough to know what i have done [02:54] <llogan> klaxa: line 106 in client.py says "bitarte". [02:55] <klaxa> oh that... lol, thanks [02:55] <llogan> s/te/se [03:29] <Orphis> How do I find out how many bytes av_read_frame actually consumed in my buffer? [03:30] <Orphis> Hmmm [03:30] <Orphis> No, it's late, I'm confusing things [03:33] <jure> hi [03:33] <jure> is there another way to capture a windows screen with ffmpeg other than using UScreenCapture? [03:34] <klaxa> repeatedly slap the print-key? [03:43] <llogan> jure: http://ubuntuforums.org/showthread.php?p=8746719#post8746719 [03:44] <llogan> https://ffmpeg.org/trac/ffmpeg/wiki/How%20to%20grab%20the%20desktop%20%28sc… [03:44] Action: llogan would have preferred CamelCase wiki url [03:45] <llogan> oh..windows...see the second link [03:45] <llogan> UScreenCapture...failure. [04:11] <KeshlWare> Can I use -crf with two-pass encoding or otherwise increase -rc-lookahead over 250, oÉo? [08:38] <Orphis> Or do I convert planar audio to non planar audio ? [08:38] <Orphis> How * [08:46] <Orphis> A, found the issue, PEBKAC as usual [09:44] <Chel> im trying to recode my video file, from one format to another, and sound becomes very bad , i mean it like with a floating effect [09:44] <Chel> this is my command : ffmpeg -i /tmp/cut.mkv -acodec aac -ab 96k -vcodec libx264 -f mp4 -crf 22 -s 640x480 -strict experimental [09:45] <Mavrik> the fact that you had to add "experimental" parameter to use the internal aac encoder could tip you off to the fact it's not really finished yet ;) [09:45] <Mavrik> use libfdk_aac for AAC encode and it'll be ok [09:48] <Chel> -acodec libfdk_aac ? [11:15] <viric> Hello all [11:15] <viric> encoding with '-acodec libmp3lame', I end up with mplayer telling me this: [11:15] <viric> Requested audio codec family [mpg123] (afm=mpg123) not available. [11:15] <viric> Enable it at compilation. [11:15] <viric> Opening audio decoder: [ffmpeg] FFmpeg/libavcodec audio decoders [11:15] <viric> what is 'mpg123' doing there? :) [11:16] <durandal_1707> wrong channel [11:16] <durandal_1707> mpg123 is another decoder .... [11:17] <durandal_1707> which you have not installed so mplayer picks next available one [11:18] <viric> ah ok [11:18] <viric> I wondered that maybe ffmpeg+libmp3lame set some tags that make mplayer think about mpg123 [11:18] <viric> thank you [11:21] <viric> I made 4 videos with "-ac 1 -acodec libmp3lame -qscale:a 3", and a 'hardware tv player' shows one of them muted. In the computer, all work perfect. No idea what can be the difference of that one. [11:22] <viric> any suggestion? [11:24] <durandal_1707> hardware tv player may not support vbr [11:24] <durandal_1707> ask hardware tv player producers instead [11:25] <durandal_1707> or read libmp3lame manual/faq [11:57] <amaurea> Given two files with both video and audio, A and B, how do I create one with the video from A and the audio from B? [11:58] <amaurea> Seems like I have to use map [11:58] <Mavrik> amaurea: you pass both as input and use "-map" parameter to map video stream from file 1 and audio stream from file 2 [11:59] <amaurea> Mavrik: Thanks, I'll give it a shot [13:54] <gxk> which pixel format in ffmpeg equals to V4L2_PIX_FMT_YUV422UVP? [14:15] <cbsrobot> gxk: The format consists of two planes: one with the Y component and one with the CbCr components interleaved - not sure ffmpeg has it ... [14:23] <durandal_1707> it does not [14:25] <cbsrobot> ah durandal_1707 - you want do me a favor ? [14:25] <durandal_1707> it is similar to nv21/nv12 which is like yuv420 but with same nonsense as yuv422uvp... [14:26] <durandal_1707> cbsrobot: depends.... [14:26] <cbsrobot> maybe you could implement >8bit path in vf_drawtext - I tried lately but failed misarabely [14:27] <cbsrobot> atm all files > 8bit have to be converted to 8bit before the overlay & would be nice it would work without conversion [14:31] <cbsrobot> eh sorry it is in drawutils.c [14:39] <durandal_1707> cbsrobot: will explore how much work it will take and will send you invoice [14:39] <cbsrobot> ok - fine with me [14:41] <durandal_1707> ok, explore work cost 50$/h [14:43] <cbsrobot> sounds intriguing - but I will have to pass [14:43] <cbsrobot> unfortunately [14:52] <gxk> top [14:53] <gxk> sorry [16:46] <Samsagax> hi there! I have this problem: I'm trying to make a screencast and keep getting video and audio out of sync. When I try to play the recorder screencast I get the video going real fast and the audio going in normal speed. The line I'm using: [16:46] <Samsagax> ffmpeg -f alsa -ac 2 -i pulse -acodec pcm_s16le -f x11grab -s wxga -r 25 -i :0.0 -threads 0 -sameq -preset ultrafast -crf 0 output.mkv [17:12] <microchip_> how to pad with the pad filter? I need to pad 50 from both top and bottom of the image [17:14] <microchip_> it was so simple with -padtop and -padbottom [17:18] <cbsrobot> microchip_: what is complicated with -vf pad ? [17:19] <microchip_> cbsrobot: don't know how to use it [17:19] <cbsrobot> hehe [17:19] <microchip_> cbsrobot: how to add borders on top AND bottom of 72 pixels? [17:20] <cbsrobot> do you know the size of the video ? [17:20] <microchip_> cbsrobot: yes, 720x432 [17:20] <cbsrobot> (before pading) [17:21] <microchip_> yes [17:21] <microchip_> 720x576 [17:21] <cbsrobot> so you want a 16/9 video to fit in 720x576 ? [17:22] <cbsrobot> show me the command you have so far [17:22] <microchip_> cbsrobot: i don't have one yet [17:22] <cbsrobot> so just craft one [17:23] <microchip_> what i want to do is add borders on top & bottom of a 720x432 video to make it 720x576 [17:23] <microchip_> don't know how to do that unfortunately :( [17:23] <cbsrobot> hehe [17:23] <cbsrobot> we [17:23] <cbsrobot> ll [17:23] <cbsrobot> tha [17:23] <cbsrobot> ts [17:23] <cbsrobot> thats gonna be complicated then [17:23] <microchip_> hmm [17:23] <cbsrobot> what outputformat do you need ? [17:24] <cbsrobot> h264 + aac in an mp4 ? [17:24] <microchip_> cbsrobot: no, -target pal-dvd [17:24] <cbsrobot> aha mpeg for a dvd [17:24] <cbsrobot> that was easy [17:24] <microchip_> cbsrobot: i'm reading this, but -padtop & -padbottom are gone http://forum.doom9.org/archive/index.php/t-128609.html [17:25] <cbsrobot> you should use http://ffmpeg.org/ffmpeg-filters.html#pad instead [17:25] <microchip_> PhillipWyllie in that thread has the answer [17:25] <microchip_> cbsrobot: yes, but how? can you show me an example? [17:25] <cbsrobot> hehe [17:25] Action: microchip_ is dumb wrt ffmpeg filters [17:25] <cbsrobot> scroll a bit down and you see the examples [17:25] <cbsrobot> :-) [17:26] <microchip_> looking... [17:27] <cbsrobot> try: ffmpeg -i input.avi -vf pad=720:576:0:72,setdar=16/9 -target pal-dvd output.mpeg [17:27] <microchip_> ok [17:31] <microchip_> Input area 0:72:1280:608 not within the padded area 0:0:720:480 or zero-sized [17:32] <microchip_> why why why did somene needed to remove padtop & padbottom? [17:32] <Mavrik> why did you say your video is 720x432 if it isn't? O.o [17:32] <microchip_> it's scaled to that [17:32] <microchip_> input is varying [17:33] <cbsrobot> hehe [17:33] <microchip_> it could be 720p or 480p [17:33] <cbsrobot> that's why I asked [17:33] <microchip_> seriously, -padtop was so easy [17:33] <cbsrobot> ffmpeg -i input.avi -vf scale=720:432,pad=720:576:0:72,setdar=16/9 -target pal-dvd output.mpeg [17:33] <microchip_> ok i'll try [17:33] <cbsrobot> microchip_: you're just a bit lazy [17:34] <microchip_> yes i am! [17:34] <cbsrobot> hehe [17:56] <microchip_> cbsrobot: can you give me the pad params for a 720x380 video? thanks :) [17:57] <cbsrobot> microchip_: damn it - get you brain together and figure it out &. [17:57] <microchip_> ok [17:57] <cbsrobot> I can give you a hint [17:58] <cbsrobot> (576-432)/2 = .... [17:58] <microchip_> nice! [17:58] <cbsrobot> and then make it (576-380) / 2 [17:58] <microchip_> ok thanks [18:04] <KeshlWare> Ish there a maaaaagical way to increase --rc-lookahead over 250? Or do two-pass encoding with -crf or something oÉo? [18:09] <microchip_> cbsrobot: perfect! thanks a lot! :) [18:12] <JEEB> KeshlWare, by editing the x264 source code and editing the maximum. That said, bigger lookahead does not equal better quality/compression. Not only does it make stuff slower, but can also make movement seem faster than it is within the lookahead [18:15] <KeshlWare> oÉo What do you mean, oÉo? [18:15] <JEEB> exactly what I said [18:15] <KeshlWare> Can you reword it, oÉo? [18:16] <JEEB> if I recall correctly, x264 tries to see the "speed" of things going on in the lookahead, and it heightens the quants in "fast" scenes where the watcher wouldn't notice anyways [18:16] <JEEB> if your lookahead is too long, even stuff that isn't really fast might seem so within the longer lookahead :P [18:16] <KeshlWare> Si, oÉo, but it also tries to keep a sane bitrate, and one of the big problems with X264 is when you have an action movie. Near the end there's a lot of scenecuts that can throw it off unless it sees them /really/ far ahead. [18:17] <KeshlWare> Two-pass encoding solves this, and I can use -qr with that, but I don't care about keeping a static bitrate. [18:18] <JEEB> what? [18:19] <KeshlWare> There's three encoding methods for x264. [18:19] <JEEB> three rate control methods, yes I know [18:19] <KeshlWare> One of them is -crf, which tries to keep a certain quality. In fast scenes that change a lot it lowers the bitrate, thus lowering the quality of the frame, but it assumes that the viewer won't notice since so much is going on. [18:19] <KeshlWare> Usually this holds true (Unless you pause playback, but I figure it assumes they won't do that, either) [18:19] <KeshlWare> Thing is, you can't use two-pass encoding with it. [18:19] <JEEB> yes, I know [18:20] <KeshlWare> And one of its weaknesses is the fact it can't see infinitely far ahead. [18:20] <JEEB> because it has nothing to limit its rate control [18:20] <JEEB> as in, it doesn't have to limit its rate control, unlike abr [18:20] <JEEB> (where 2pass makes sense) [18:20] <KeshlWare> Since it needs to know ahead of time what's going to happen so it knows when and where it can lower and raise the quality of individual frames, this becomes problematic if part of the movie is mostly slow, then suddenly it speeds up near the end and has a lot of frame cuts. [18:20] <KeshlWare> And this is exactly the case in some clips I have. [18:21] <KeshlWare> So, I'd like to give it infinite lookahead or run two-pass encoding with -crf, because I don't need to limit the bitrate or the filesize. Smaller is ebtter in both cases, but there's no firm limit, and it's okay for it to be variable. [18:21] <KeshlWare> *better [18:22] <JEEB> ... [18:22] <KeshlWare> What's bothering me is how in the start of a clip, it's beautiful, then near the end the encoding just flops and a lot of stuff gets blocky. [18:22] <KeshlWare> Even with CRF at 18. <.< [18:22] <JEEB> abr is variable bit rate unless you set maxrate/bufsize to make it that [18:22] <JEEB> it's really hard to actually get constant bit rate with x264 :P [18:22] <KeshlWare> ... Oh. [18:22] <KeshlWare> Well that I didn't know. <.< [18:22] <KeshlWare> Thought ABR was constant. [18:23] <JEEB> ABR = Average Bit Rate [18:23] <KeshlWare> Yes, that name's highly misleading. x.x [18:23] <JEEB> as in, it tries to keep the average over the whole clip [18:23] <JEEB> why is it misleading? [18:23] <JEEB> it's like in all other encoders [18:23] <KeshlWare> ...Oh, it's not misleading then. Dang. [18:23] <KeshlWare> Does exactly what I don't want it to do. [18:23] <JEEB> I'm really not sure if hitting lookahead to the max is actually solving that to be honest, but if you want to set it longer you'll have to edit x264's source code :P [18:23] <JEEB> because the limit is set there [18:24] <KeshlWare> Yeah.. .... Sadly I'm on windows and building that ain't gunna happen. So dang. x.x [18:24] <JEEB> I'm pretty sure there are some negative concequences to setting it long [18:24] <JEEB> also you could try making qcomp a bit higher, 0.7 or something (default is 0.7) [18:24] <JEEB> ugh, 0.6 is default [18:25] <JEEB> when qcomp hits 1.0 it's pretty much constant quant :P [18:25] <KeshlWare> Whus quant, oÉo? [18:25] <JEEB> quantizer? [18:25] <KeshlWare> ... oÉo? [18:27] <KeshlWare> Ah, okay. Yay for Wikipedia. -É-. [18:46] <kode54> building should be fairly straightforward with either mingw or cygwin [18:46] <kode54> former if you're lucky, latter as a fallback [18:47] <kode54> a matter of tossing the right packages in there, make/make install, rinse repeat [18:47] <kode54> if I felt like throwing either into my Windows VM and building for you, I could probably change that [18:47] <kode54> but it's probably more trouble than it's worth [18:49] <kode54> probably less trouble than it would be for me to try to get cross compiling working [18:51] <KeshlWare> Oh, see, I'm cursed. [18:51] <KeshlWare> Literally. [18:51] <kode54> nah [18:52] <kode54> you could also try hex editing [18:52] <KeshlWare> My family was stupid and built our house on top of an anchient Indian burial ground. [18:52] <kode54> haha [18:52] <KeshlWare> So, seriously, crap happens. Weird crap that only happens to me. [18:52] <kode54> what are you trying to do here? [18:52] <KeshlWare> Generally, whenever I compile anything on windows, it goes horribly wrong. <É< [18:52] <KeshlWare> kode54: From what I can gather, change the number "250" to "250000000" or something else outlandishly big. [18:53] <kode54> and what would you hope to achieve with that? [18:53] <KeshlWare> The ability to toss a number bigger than 250 at -rc-lookahead, oÉo [18:53] <kode54> I gather that [18:53] <kode54> but what is the hope from encoding with higher parameters? [18:53] <kode54> is this a really high frame rate video? [18:54] <KeshlWare> I explained all this. .É. No, it just has a lot of scene cuts at the end that all occur /juuuuust/ over 250 frames apart so it doesn't see them ahead of time as it should (Maybe) and it ends up looking really crappy at the end. [18:54] <kode54> ah [18:54] <KeshlWare> Even the /credits/ look crappy, and those are rather low-motion/easy to predict. [18:54] <kode54> I can try to get that working [18:55] <kode54> http://doom10.org/index.php?topic=26.0 this alternative to doom9 seems to have a guide to compiling and has packages of the compilers and tools [18:55] <kode54> but I can try to get those working myself [18:55] <KeshlWare> Well, I already have MSYS and MinGW.. Lesse, what else is there that I need.. [18:56] <KeshlWare> ...There's Pthreads for windows? [18:56] <kode54> indeed there is [18:56] <KeshlWare> Nobody tells me these things! >É<' [18:57] <JEEB> you don't really need pthreads on win32/64 [18:57] <JEEB> because both ffmpeg and x264 support win32's threading [18:57] <KeshlWare> But I want Pthreads for myself. o_o [18:57] <JEEB> why? [18:57] <KeshlWare> Cuz that's nice and easy and win32's is evil o_o [18:58] <JEEB> ... [18:58] <KeshlWare> Same reason I use nCurses <É< [18:58] <JEEB> pthreads works on top of windows' threading [18:58] <JEEB> so it's the same damn thing [18:58] <KeshlWare> Different itnerface. [18:58] <KeshlWare> *interface [18:58] <JEEB> ... [18:58] <JEEB> jesus christ [18:58] <KeshlWare> To hide the blinking cursor in windows in a command prompt, using win32api, you need two structs and like six lines of code. [18:58] <KeshlWare> *Jesus Christ [18:59] <KeshlWare> To hide it with nCurses, you just type hide_cursor(); [18:59] <KeshlWare> So I use nCurses. oÉo. [18:59] <JEEB> I mean, you want to build x264 and ffmpeg specifically WITH WIN32 PTHREADS? [18:59] <JEEB> because you want to GRAB EXTRA POSSIBLE PROBLEM POINTS? [18:59] <JEEB> seriously? [18:59] <KeshlWare> I'm indifferent about it, but for /my own other unrelated projects/, I want pthreads. [18:59] <KeshlWare> And I just never knew it worked on windows, oÉo. [18:59] <JEEB> it doesn't [18:59] <JEEB> it's a layer on top of win32threads [18:59] <KeshlWare> Then why is it here? D: [18:59] <KeshlWare> To me, that means it works. D: [19:00] <JEEB> well yes, it enables you to kind of use pthreads on win32/64 [19:00] <KeshlWare> I'm just looking for a simpler interface so I can just type RunThisFunctionUntilItReturns(); instead of having to use two functions and needing to know the stacksize and stuff. [19:00] <KeshlWare> *stack size [19:00] <JEEB> you should look at the wrapper in x264 and ffmpeg [19:01] <JEEB> they pretty much share it [19:01] <JEEB> or use the win32 pthreads thingy, there's the old one + one from mingw-w64 projecgt [19:01] <JEEB> *project [19:02] <JEEB> also, fun fact. I'm pretty sure this is not intended, but x264's configure actually starts enabling pthreads in case it finds the pthreads headers, so you actually have to /tell/ it to use the native threading [19:02] <JEEB> (on !win32 it's a good thing, but on win32/64 I'm pretty sure it shouldn't override native threads with pthreads) [19:02] <JEEB> because on other architectures it generally means that you wouldn't have threading otherwise [19:03] <KeshlWare> Interesting, oÉo. [19:04] <KeshlWare> ...Any chance there's a port of Pthreads that'll run on Atmel devices? With really low overhead so it fits in ~1KB of RAM? <.< [19:04] <JEEB> no effing idea [19:04] <KeshlWare> D'aww .É. [19:06] <kode54> hexediting the binary may work as well, if you can find where it clamps the value [19:06] <KeshlWare> Yes, but I ain't so good at that, nor do I know how to add room (Cuz otherwise I might only be able to take it to 999, or 255, depending what exactly they did) [19:07] <KeshlWare> Blargh. Where do I need to put i686-pc-mingw32? Can't find that folder in my install anywhere,. [19:07] <KeshlWare> ...because I don't even have them in the first place. Okay. No wonder. ...Blah. x.x [19:08] <JEEB> just grab the newest mingw package from here http://files.1f0.de/mingw/ [19:08] <JEEB> and mount it as /mingw [19:08] <KeshlWare> JEEB: On windows, oÉo. [19:08] <JEEB> yes [19:08] <JEEB> withing msys [19:08] <KeshlWare> Ohhh.. [19:08] <JEEB> <msys dir>/etc/fstab controls the "mounts" [19:09] <KeshlWare> OH, so, they, just go in the root folder. [19:09] <JEEB> not necessarily [19:09] <JEEB> I have <msys root dir>/mingw as well, but I have left that be :P [19:09] <JEEB> I just mount another folder as /mingw [19:10] <KeshlWare> ... -Blank stare at the folder structors- .... Okay so what's the syntax for mount in mingw? <É< [19:10] <JEEB> mingw has no mount syntax [19:10] <KeshlWare> Then how... But... .É. [19:10] <JEEB> you just have the mounts in msys's etc/fstab [19:10] <JEEB> well, "mounts" [19:10] <KeshlWare> Oh right. [19:10] <KeshlWare> One moment, I will screw this up. oÉo. [19:11] <JEEB> so for example I have in <msys root dir>/etc/fstab [19:11] <JEEB> C:/JEEB/msys/mingw-w64-4.7.2 /mingw [19:11] <KeshlWare> Is there a way to comment out what's there? Just add a #? [19:11] <JEEB> yes [19:11] <KeshlWare> oÉo. [19:11] <JEEB> I have plenty of old /mingw lines there [19:11] <JEEB> commented out [19:11] <JEEB> lol [19:12] <JEEB> also it hates spaces in paths, so just don't use them. If you really need to, you'll have to use the short paths in fstab [19:12] <JEEB> like I have with the CUDA SDK [19:12] <JEEB> C:/PROGRA~1/NVIDIA~2/CUDA/v4.2 /cuda [19:12] <KeshlWare> Dun worry, I already have that common sense, oÉo. [19:13] <KeshlWare> I just never use spaces to begin with and obsessively only save stuff to folders that don't got them. [19:13] <KeshlWare> And for serious overkill, they usually have less than 9 characters in the name. [19:13] <kode54> building ffmpeg now [19:14] <kode54> so I can produce x264 binary [19:14] <JEEB> anyways, after you've set the mount point for /mingw , start msys from the msys.bat [19:14] <JEEB> and see if you get correct'ish response from gcc --version [19:14] <JEEB> i686-w64-mingw32-gcc --version and x86_64-w64-mingw32-gcc --version [19:15] <JEEB> just in case [19:15] <KeshlWare> well, it's longer than it wa s before, so seems right, oÉo. [19:15] <JEEB> if you grabbed one from the url I pointed at, it should say 4.7.2 for all [19:16] <KeshlWare> ...I cna't paste into it? .É. [19:16] Action: KeshlWare is just noticing this now. [19:16] <JEEB> depends on which shell you're using with msys [19:16] <KeshlWare> ...The normal one? .É. [19:16] <JEEB> or well, terminal emu [19:16] <JEEB> well, there is really no "normal" one :P [19:17] <JEEB> they used to use rxvt, they used to use standard windows cmd and they could be using mintty as default now [19:17] <KeshlWare> .É. Things are scary now cuz the guide's saying to poke the option with the scary red text and the red text says not to poke it unless I know what it does and Iunno what it does D: [19:17] <JEEB> all of those could copy and paste in various ways [19:17] <JEEB> KeshlWare, don't follow the guide [19:17] <JEEB> it's old [19:17] <KeshlWare> Then whus I do? D: [19:18] <JEEB> well I just asked if the output of those commands I listed looks more or less ok? [19:18] <KeshlWare> Si, oÉo, 4.7.2. [19:19] <JEEB> ok, grab http://www.tortall.net/projects/yasm/releases/yasm-1.2.0-win32.exe and save it as yasm.exe in <msys root dir>/bin [19:19] <JEEB> so you can call it out with just 'yasm' from the msys shell [19:20] <JEEB> (I generally put it into the msys /bin because I switch around /mingw toolchains rather often at times, so I don't have to care about copying it around or rebuilding it) [19:21] <KeshlWare> oÉo'd, oÉo. [19:23] <JEEB> tell me after you've done that and can launch yasm from the msys shell with just 'yasm' [19:23] <KeshlWare> I did, oÉo. [19:23] <KeshlWare> Sorry, sometimes I forget that english exists <É< [19:23] <JEEB> ok [19:23] <JEEB> did you have git around already? [19:24] <KeshlWare> Nu .É. [19:24] <JEEB> okies [19:24] <KeshlWare> On Linux yeah but, cross-compiling does /not/ happen in this ghouse. XD [19:24] <KeshlWare> *house [19:24] <JEEB> grab the newest 7z archive from http://code.google.com/p/msysgit/downloads/list [19:24] <JEEB> then extract it somewhere like the mingw [19:24] <JEEB> mount it to, say, /git [19:25] <kode54> I'm attempting to cross compile it for you now [19:25] <JEEB> (etc/fstab, similarily) [19:25] <kode54> I already had mingw installed for some wine related thing [19:25] <KeshlWare> kode54: 'Preciated, oÉo. (Watch it run fine on your system but not mine <.<) [19:25] <JEEB> then add "export PATH=${PATH}:/git/bin" to the end of the msys etc/profile file [19:25] <kode54> it's an i486 compiler, so it should be pretty safe [19:25] <kode54> (watch it slow down your encodes) [19:25] <KeshlWare> ...Yes. [19:25] <KeshlWare> x.x [19:25] <JEEB> *cough* [19:25] Action: KeshlWare is bothered. Has AVX D: [19:25] <kode54> it uses optimizations [19:26] <kode54> I mean [19:26] <kode54> i486 probably means it's built for i486 [19:26] <kode54> but can run AVX [19:26] <kode54> ffmpeg has avx I think [19:26] <KeshlWare> It does, oÉo. [19:26] <kode54> it seemed to assemble that stuff just fine [19:26] <JEEB> anyways, KeshlWare -- look at those lines of mine [19:26] <kode54> I had to apply a fix for compiling it with gcc 4.7 [19:27] <JEEB> extract the git 7z similar to the mingw package, mount it as /git in fstab and add it to PATH in etc/profile [19:27] <JEEB> after that you should be able to use 'git' in the msys shell [19:28] <JEEB> tell me when you can launch git from the terminal, can launch yasm from the terminal and all of the gcc commands output their versions properly. That's pretty much all of the prerequisites you'd need :) [19:28] <Nick-S> what image formats does ffmpeg image2 supports? [19:28] <KeshlWare> Git works, oÉo. [19:28] <KeshlWare> And other stuff /should/ work oÉo [19:29] <JEEB> so you mounted and added it to PATH like I noted? [19:29] <KeshlWare> Si, oÉo. [19:29] <JEEB> great [19:29] <JEEB> one little thing before we start grabbing code then [19:29] <Nick-S> anyone knows? i see gif and jpg in the docs but i'm wondering about tif [19:29] <JEEB> "git config --global core.autocrlf false" [19:29] <KeshlWare> I use Linux from time-to-time so I'm not completely lost here, but I'd never handle this without hand-holding, so danks, oÉo. [19:29] <JEEB> make sure you set this for git [19:29] <KeshlWare> Nick-S: PNG works, dunno about TIFF. [19:30] <JEEB> tell me after you've set that [19:30] <KeshlWare> JEEB: Poked oÉo [19:30] <JEEB> great [19:30] <JEEB> now let's grab x264 to your home folder [19:30] <JEEB> move to where you want to clone the source code, and then 'git clone git://git.videolan.org/x264.git x264' [19:31] <JEEB> and it should start cloning x264's source code into a folder called x264 [19:31] <KeshlWare> Doing, oÉo. [19:31] <KeshlWare> Is there verbose for this? I like seeing numbers do fancy stuff <É<. [19:31] <Nick-S> KeshlWare, hmm... [19:31] <JEEB> there's a verbose setting, but for whatever reason that seems not to be on by default on msys [19:31] <JEEB> for whatever reason [19:31] <KeshlWare> ...Okay is there verbose, cuz crap went wrong? [19:31] <JEEB> oh well [19:31] <JEEB> hm? [19:32] <KeshlWare> ...Oh. Typo. [19:32] <KeshlWare> Typed videolanD instead of videolan <É<. [19:32] <JEEB> lol [19:32] <KeshlWare> ... Scratch that, crap still goes wrong. [19:32] <KeshlWare> Sec, screenshotting cuz I can't copy/paste and I have a feeling you're gunna asked if I typed it right first. <É< [19:33] <KeshlWare> http://ompldr.org/vaDVqbA/crapwentwrong.png oÉo [19:33] <JEEB> o___O [19:33] <JEEB> also that seems like mintty [19:34] <JEEB> you should be able to use shift+insert to paste [19:34] <KeshlWare> Squee ^É^ [19:34] <JEEB> and ctrl+insert to copy [19:34] <JEEB> (also in the right-click menu) [19:34] <KeshlWare> There's no right0click meni. [19:34] <KeshlWare> And ctrl + Insert dun do anything oÉo [19:35] <JEEB> sure does have for me... [19:35] <KeshlWare> OH, you mean in the... No, see, this is the standard shell with colors. [19:35] <JEEB> oh [19:35] <KeshlWare> Just cmd.exe. [19:35] <Fjorgynn> nope [19:35] <JEEB> so if you right-click the window border on the top [19:35] <JEEB> you have no "options" there? [19:35] <KeshlWare> Yes, that works, oÉo, but I thought you meant in the backgroundish area. [19:36] <KeshlWare> Anyway, did I type stuff right? D: [19:36] <JEEB> that seems correct for me [19:37] <KeshlWare> So why error? D: [19:37] <JEEB> try without the tailing x264 once [19:37] <JEEB> (it will look for the repo name in the url) [19:37] <KeshlWare> Same output, oÉo. [19:37] <KeshlWare> Wait. [19:37] <KeshlWare> ... Is windows stupid enough that I need to escape froward slashes? [19:37] <KeshlWare> >w> [19:38] <KeshlWare> Dang. [19:38] <JEEB> no, and also you're in the msys shell [19:38] <KeshlWare> Si, oÉo. [19:38] <JEEB> not exactly windows's standard thing :P [19:38] <KeshlWare> But isn't that just cmd.exe with colors? .É. [19:38] <JEEB> no [19:38] <KeshlWare> oÉo.. [19:38] <JEEB> or well, it might be but you'd be in a msys shell in any case :P [19:39] <KeshlWare> oÉo. [19:39] <JEEB> if you right-click the window border on top, click options and have an about menu there, then you are in mintty as the terminal emulator :P [19:39] <KeshlWare> They dun has an about, oÉo. [19:40] <KeshlWare> Regardless. o_o Is this my juju acting up? D: [19:40] <JEEB> it sounds like something isn't letting git write [19:41] <KeshlWare> (Again, I am literally cursed. It is very possible that I am the only person with this problem right now and that it's completely unsolvable. I'll just add it to the stack of other unsolvable juju-cursed problems I have. <É<) [19:42] <kode54> I'm working on it [19:42] <kode54> have patience [19:42] <JEEB> you're far from the only one :P [19:42] <KeshlWare> kode54: But I wanna try at least D: [19:42] <JEEB> msysgit can be funky like that for some people [19:42] <kode54> KeshlWare: you're welcome to [19:42] <KeshlWare> JEEB: Should I mention that, on no system in my house, pywin works? [19:43] <KeshlWare> Talked in #python about it. It's a known but, nobody knows what causes it or how to fix it. They told me that I just have to accept the fact that the computer in question can't ever run it. All 10+ of my computers can't run it. /that/ is the extent of my juju. <.< [19:43] <JEEB> try to make sure there's nothing funky trying to block file system access? [19:43] <kode54> lol mlp encoder [19:43] <kode54> (yes I know that's a format) [19:43] <KeshlWare> JEEB: There's not, oÉo. [19:43] <JEEB> no antivirus or anything? [19:43] <KeshlWare> Unless ClamAV is stupider than I thought. [19:44] <KeshlWare> No, no on-demand scanning. Yeah, nothing blocking it. [19:44] <JEEB> :s [19:44] <KeshlWare> on-access* [19:44] <JEEB> ok, let's try via cygwin then I guess :D [19:44] <JEEB> that one's simpler to set up anyways [19:44] <kode54> I think this is called i486 simply because it's i486 binaries so that just about any system can run them [19:45] <JEEB> just grab the installer/package manager from http://www.cygwin.com/ [19:45] <KeshlWare> o_o Oh gosh I can go on for ages about the torture this things'b roughr upon me. [19:45] <KeshlWare> Before I go on, is there a way to prevent it from screwing with %PATH%? [19:45] <JEEB> yes, cygwin doesn't play with your general PATH at all [19:45] <KeshlWare> *thing's brought [19:45] <KeshlWare> ... I swear it used to. o_o [19:45] <KeshlWare> Had to reinstall an OS once after installing Cygwin o_o [19:45] <JEEB> I have cygwin installed and it hasn't modified my path at all [19:45] <Nick-S> can't i find in the code somewhere what formats are supported? [19:45] <JEEB> it's a completely separate ecosystem pretty much [19:46] <KeshlWare> Okay, wish me luck. [19:46] <JEEB> from the installer, when you get to the package selection [19:46] <JEEB> use the search to locate git [19:46] <JEEB> yasm [19:46] <JEEB> mingw-w64 [19:46] <JEEB> that should be enough [19:46] <KeshlWare> Just a note: Another of my unsvolable bugs is, on a recent system, the processor hangs unless I'm giving it constant input from a HID, OR unless Compiz is running and pouring rain down on the desktop. [19:46] <KeshlWare> And it only happens when Linux is booted. Windows is fine. [19:46] <JEEB> i686-w64-mingw32 and x86_64-w64-mingw32 [19:47] <JEEB> (for mingw-w64) [19:47] <KeshlWare> Do not underestimate this juju. <.< [19:47] <JEEB> that msysgit snafu doesn't sound like too weird [19:47] <Hans_Henrik> can i use ffmpeg to make a copy of all frames with a specific pixel color? something close to that? [19:47] <JEEB> either network or FS related, and I've seen others have similar things and google tells me others have had the same error pop up [19:48] <JEEB> not to mention that one of my friends had msysgit just not work for him for months [19:48] <JEEB> and then it suddenly worked later again [19:48] <Hans_Henrik> (i realize it would probably be annoyingly cpu-intensive, but i don't care about that currently) [19:48] <KeshlWare> JEEB: Reminds me of one of my laptops. [19:49] <KeshlWare> Bad update to a kernel module kicked out network access on it. No wifi, no ethernet. [19:49] <KeshlWare> Turned it off for a month, turned it back on.. Ethernet works fine now. Wifi's still dead. [19:49] <Hans_Henrik> ouch, [19:49] <KeshlWare> And nobody can fix it. It appears totally fine. [19:49] <Hans_Henrik> am i still in #ffmpeg ? [19:49] <KeshlWare> Correct module is loaded. Module's got the right checksum.. [19:49] <KeshlWare> Hans_Henrik: Yes oÉo [19:50] <Hans_Henrik> KeshlWare: sounds like a problem for ##linux :o [19:50] <Hans_Henrik> <KeshlWare> Turned it off for a month, turned it back on.. Ethernet works fine now. Wifi's still dead. [19:50] <Hans_Henrik> sounds more like a hardware issue, if "turning it off for a month" fixed anything.. [19:50] <KeshlWare> JEEB: The MinGW packages you mentioned aren't in Cygwin, oÉo. [19:51] <KeshlWare> Hans_Henrik: I doubt that. It works fine anywhere but Linux, as far as I'm able to test. [19:51] <JEEB> KeshlWare, there should be mingw-w64 stuff [19:51] <KeshlWare> JEEB: Nupe. oÉo. [19:52] <Hans_Henrik> KeshlWare: anything inportant on the linux installation? [19:52] <Hans_Henrik> important* [19:52] <JEEB> mingw64-i686-gcc-core [19:52] <JEEB> and mingw64-x86_64-gcc-core [19:52] <JEEB> it seems [19:52] <KeshlWare> Hans_Henrik: Yeah. Spent ages customizing it just right and, due to an intense reverse-juju incident, did something that should've been impossible. [19:52] <KeshlWare> Somehow I replaced the kernel's bootsplash early on with another image, without recompiling it. [19:53] <KeshlWare> I don't question it, and I'm afraid to get rid of it. <.< [19:53] <KeshlWare> And it just matches the laptop so nice .É.... [19:53] <Hans_Henrik> what GUI thing did you put on it? [19:53] <KeshlWare> JEEB: Dem's squeeful ^É^ [19:53] <Hans_Henrik> or, customize [19:53] <KeshlWare> Hans_Henrik: ...oÉo? [19:54] <KeshlWare> It's mainly the early boot splash, so, "none". [19:54] <Hans_Henrik> hmm [19:54] <JEEB> KeshlWare, speak English, please... [19:54] <KeshlWare> Screen that comes directly after Grub. [19:54] <Hans_Henrik> no idea what owo is :o [19:54] <KeshlWare> JEEB: Found them. [19:54] <KeshlWare> Hans_Henrik: :3 roated 90 degrees. [19:54] <JEEB> alrighty [19:54] <JEEB> also check what version yasm is in cygwin [19:54] <JEEB> the package selector should show the version [19:54] <Hans_Henrik> KeshlWare: seems correct [19:54] <KeshlWare> 1.2.0-1. [19:54] <JEEB> nice [19:54] <JEEB> it's new enough [19:55] <KeshlWare> Hans_Henrik: Well yeah. I know what oÉo is. XD [19:55] <JEEB> git, those mingw tools, yasm [19:55] <JEEB> should be enough [19:55] <KeshlWare> Kay, wish me luck. <.< [19:55] <KeshlWare> If I lag out, assume this set off a nuclear missile launch and it's aimed for my house and I had to evacuate. [19:55] <Hans_Henrik> ah wait a sec [19:55] <KeshlWare> That way you can't worry about something /worse/. [19:56] <Hans_Henrik> KeshlWare: just, where do you live? [19:56] <KeshlWare> Hans_Henrik: ... You're not near a silo are you? o_o [19:56] <Hans_Henrik> KeshlWare: no, not really, but long-range-missiles can go pretty far [19:56] <KeshlWare> .... The moon 8D [19:56] <KeshlWare> Pennsylvania [19:56] <KeshlWare> >W> [19:57] <Hans_Henrik> hmm ok, i should be fairly safe then; (way up in cold norway here) [19:57] <KeshlWare> Lucky. .É. [19:57] <Hans_Henrik> yup [19:57] <Hans_Henrik> why do you keep rotating :3 btw? [19:57] <KeshlWare> So I can do stuff like <É< [19:57] <Hans_Henrik> awesome! [19:57] <KeshlWare> I know right? OÉO [19:58] <KeshlWare> JEEB: Kay, it's in. [19:58] <JEEB> nice [19:58] <JEEB> start the cygwin shell [19:58] <KeshlWare> o-o How do I make the cursor bigger tellmehow o_o! [19:58] <JEEB> (now that should be mintty by default) [19:58] Action: KeshlWare panics. [19:58] <KeshlWare> Ogh, got it. [19:58] <KeshlWare> Kay. [19:58] <JEEB> options are under options when you right-click on the top bar [19:58] <KeshlWare> Carry on. [19:58] <JEEB> ok, so now you should be able to run git --version, yasm --version and i686-w64-mingw32-gcc --version [19:59] <JEEB> and all should actually return some info [19:59] <KeshlWare> 1.7.9, 1.2.0, 4.5.3 oÉo [20:00] <KeshlWare> Old D: [20:00] <JEEB> the gcc is a bit old, but that should be fine [20:00] <JEEB> now you can get back to cloning x264 [20:01] <KeshlWare> oÉo. [20:01] <KeshlWare> Whus address again, oÉo? [20:01] <JEEB> git clone git://git.videolan.org/x264.git [20:01] <KeshlWare> oÉo. [20:01] <JEEB> (it grabs the x264 from the x264.git to be as the directory name) [20:01] <KeshlWare> Squee numbers OÉO [20:01] <KeshlWare> Ish good, oÉo. [20:02] <JEEB> tell me when you've got x264 cloned [20:02] <KeshlWare> It be. [20:03] <JEEB> ? [20:03] <KeshlWare> It's cloned, oÉo. [20:03] <Fjorgynn> <3 [20:03] <JEEB> ok [20:03] <JEEB> move to the x264 directory and do ./configure --cross-prefix=i686-w64-mingw32- [20:04] <JEEB> see what that outputs [20:04] <KeshlWare> with the - at the end, oÉo? [20:04] <JEEB> yes [20:04] <KeshlWare> oÉo. [20:04] <KeshlWare> it says I can type make now, oÉo. [20:04] <JEEB> pastebin what you got :) [20:04] <KeshlWare> it also picks me up as X86. <É< [20:05] <KeshlWare> http://pastebin.com/B8wsgWBT oÉo [20:05] <JEEB> ahaha [20:05] <JEEB> yeah, you got the usual derp [20:05] <KeshlWare> oÉo..? [20:05] <JEEB> see thread: [20:05] <KeshlWare> Posix threads oÉo? [20:05] <JEEB> it picked up posix [20:05] <JEEB> add --enable-win32thread [20:05] <KeshlWare> Where, oÉo? [20:06] <JEEB> configure [20:06] <KeshlWare> oÉo. [20:06] <KeshlWare> Thread looks happy now, oÉo. [20:06] <JEEB> sweet [20:06] <JEEB> now to see what your prefix is [20:06] <kode54> how many do you want it to be? [20:06] <kode54> 250 is not enough [20:06] <JEEB> 'which i686-w64-mingw32-gcc' [20:07] <KeshlWare> kode54: Pretty much just "So high I'll never make a video long enough to hit it" [20:07] <kode54> I'll just make it 2000 for starters [20:07] <kode54> since it does add memory to several functions [20:07] <KeshlWare> kode54: I need more than that. XD [20:07] <kode54> because it scales the size of arrays [20:07] <KeshlWare> Video's almost 20,000 frames. [20:07] <KeshlWare> And I got 32 gigs of RAM, so, yeah. oÉo. [20:07] <kode54> this is going to be a 32 bit binary [20:07] <KeshlWare> ... Oh D: [20:08] <kode54> but yeah, I don't think it will be much of a problem [20:08] <kode54> I'll make it 50,000 then [20:08] <KeshlWare> JEEB: That reminds me, I'm probably gunna need more than 2 gigs of RAM for what I actually want to do. Can I do a 64-bit build too oÉo? [20:08] <KeshlWare> And the output's "/usr/bin/i686-w64-mingw32-gcc" [20:08] <JEEB> KeshlWare, switch to x86_64-w64-mingw32- as the prefix [20:08] <JEEB> that's the win64 one [20:08] <KeshlWare> oÉo. [20:08] <cbreak> ffmpeg works in 64 bit if that's what you mean [20:09] <kode54> sure it does [20:09] <KeshlWare> D: [20:09] <JEEB> also I have a feeling that just disabling mbtree or setting qcomp a bit higher than 0.6 (the default) [20:09] <KeshlWare> "No working C compiler found." [20:09] <JEEB> do you have a x86_64-w64-mingw32-gcc ? [20:09] <KeshlWare> ... Maybe D: [20:09] Action: KeshlWare gets. [20:09] <JEEB> try launching it from the terminal [20:09] <JEEB> if it's not there then it's not there [20:10] <JEEB> mingw64-x86_64-gcc-core should be the package name for the x86_64 stuff [20:10] <KeshlWare> It's there, oÉo. [20:10] <JEEB> ok, so you can launch x86_64-w64-mingw32 from terminal? [20:11] <JEEB> ugh [20:11] <JEEB> x86_64-w64-mingw32-gcc [20:11] <KeshlWare> Si, oÉo. [20:11] <KeshlWare> Says "no input files" so I figure it's there. [20:11] <JEEB> yeah [20:11] <JEEB> check config.log in x264's folder for what went wrong then [20:12] <JEEB> also I have a feeling that you might as well disable mbtree or heighten qcomp to fix your "problem" >_> [20:12] <JEEB> instead of making the lookahead lol long [20:12] <KeshlWare> Yes but I wanna know what happens D: [20:12] <KeshlWare> Curiousity sparks invention! Maybe I'm just a late bloomer! ...Really, really late. [20:13] <JEEB> well, in any case the config.log should be telling you what went wrong with x264's configuration [20:13] <KeshlWare> It's co ... ....Long. [20:13] Action: KeshlWare pastebins. [20:14] <KeshlWare> http://pastebin.com/fwXd4eqH oÉo [20:15] Action: JEEB double-blinks [20:15] <JEEB> oh right [20:15] <JEEB> it tries to use 32bit flags or something? [20:16] <KeshlWare> Iunno .É. [20:16] <KeshlWare> This is normally the part I seek help and professional medication advice D: [20:16] <JEEB> try adding --host=x86_64-w64-mingw32 [20:17] <JEEB> (this is without the tailing -) [20:17] <KeshlWare> Ish happy now ^É^ [20:17] <JEEB> nice [20:18] <JEEB> to change what you wanted [20:18] <JEEB> you'll have to open up common/common.h [20:18] <KeshlWare> Oh thank you o_o I just realized that I have no idea where it is.. XD [20:18] <JEEB> and find a line that says #define X264_LOOKAHEAD_MAX 250 [20:19] <JEEB> I didn't remember either, but since it's hardcoded to 250 I just did 'git grep "250"' in a x264 repository folder :P [20:19] <KeshlWare> oÉo. [20:19] <KeshlWare> ...I wouldn't have done that right. [20:19] <KeshlWare> >W> [20:19] <KeshlWare> First though, just, cuz, I have this hunch it'll go wrong, how do I build as-is, oÉo? [20:20] <JEEB> after you've run configure [20:20] <JEEB> hit make [20:20] <JEEB> and wait for a bit [20:20] <JEEB> you haven't set a prefix yet so don't do make install [20:20] <KeshlWare> D: [20:20] <KeshlWare> "make: command not found" [20:20] <JEEB> lawl [20:20] <KeshlWare> Juju! D: [20:21] <JEEB> time to close the cygwin shell and re-open the cygwin setup exe you installed some time ago [20:21] <JEEB> it's the package manager [20:21] <JEEB> so you just go through it until you get to the package selection and install make [20:21] <KeshlWare> oÉo. [20:21] <JEEB> gnu make should do [20:21] <KeshlWare> Ish makin' now, oÉo. Maybe. [20:22] <KeshlWare> yeah there it goes. [20:22] <KeshlWare> Oh good, -O3. I was gunna probably ask about that next <.< [20:22] <JEEB> also I think libtool and autoconf/automake should be useful if you want to compile in fdk-aac or something into ffmpeg [20:22] <JEEB> not like O3 makes it any faster really >_> [20:22] <JEEB> the yasm'ification does [20:22] <JEEB> (and threads) [20:22] <KeshlWare> Whus yasm oÉo? It sounds like yam. Yams are kewl -É- [20:22] <JEEB> yasm is the assembler that you installed :P [20:23] <JEEB> it compiles assembly code into binaries [20:23] <JEEB> object files [20:23] <KeshlWare> But dun GCC do that? oÉo [20:23] <KeshlWare> Also, appearntly, it got made. [20:23] <JEEB> GCC can only do inline asm, and that is evil and specific to GCC [20:23] <JEEB> then you can launch it with ./x264 --version [20:23] <JEEB> it's a normal windows binary [20:24] <KeshlWare> It works OÉO! [20:24] <JEEB> cool [20:24] <KeshlWare> make clean; editcrap; make, oÉo? [20:24] <JEEB> make clean, re-configure with a prefix set and --enable-static [20:24] <JEEB> check contents of usr [20:24] <JEEB> and see if you have the w64 directories there [20:24] <KeshlWare> I dun .É. [20:25] <KeshlWare> Well do and dun. [20:25] <KeshlWare> "i686-w64-mingw32 lib sbin src x86_64-w64-mingw32" (at the end) [20:25] <JEEB> yeah, you have both [20:25] <JEEB> although since cygwin has package management I guess we won't go using those as the prefix ^^; [20:26] <KeshlWare> I confus. .É. [20:26] <JEEB> we'll do it like it's done with *nix when you want to keep le usual folders pretty [20:26] <JEEB> in other words, ./configure --enable-static --cross-prefix=x86_64-w64-mingw32- --host=x86_64-w64-mingw32 --enable-win32thread --prefix=/home/yourusername/ownapps [20:27] <JEEB> this will then install the static library and the x264 app into /home/yourusername/ownapps as the prefix (it will make the bin, lib, include etc. folders there) [20:27] <KeshlWare> Ohhhh oÉo [20:28] <KeshlWare> Den make, oÉo? [20:28] <JEEB> after you've made your change? [20:28] <KeshlWare> ohright <É< [20:28] <JEEB> and then make install to install the stuff into the prefix [20:28] <JEEB> I hope you set the prefix with tab-autocompletion [20:29] <JEEB> lol [20:29] <KeshlWare> Wai? .É. [20:29] <JEEB> so you got it right? [20:29] <KeshlWare> I'm decently sure, oÉo. Not like it's long. [20:29] Action: KeshlWare is fixing stuff secretly. <.<' [20:29] <JEEB> sure as long as it's correct [20:30] <JEEB> and you know what you set the prefix as :) [20:30] <JEEB> oh, and since you don't really need the separate x264 app [20:30] <KeshlWare> Whus "make fprofiled" do oÉo? [20:30] <JEEB> you could add --disable-cli [20:30] <JEEB> make fprofiled runs the compiled stuff once through possible clips that you might or might not have around [20:30] <JEEB> you probably don't have any raw YCbCr clips around so just ignore it [20:31] <KeshlWare> oÉo.. Does it help make stuff sexier? [20:31] <JEEB> not really [20:31] <JEEB> placebo at best [20:31] <KeshlWare> I use -preset placebo D: [20:31] <JEEB> just ignore it since you don't have y4m files for it :P [20:31] <KeshlWare> Are you telling me it's all.. Lies? D: [20:31] <JEEB> you can play with it /later/ [20:32] <KeshlWare> Fiiiine <É< And I'll keep x264.exe, cuz documentation. [20:32] <JEEB> http://mewiki.project357.com/wiki/X264_Settings [20:32] <KeshlWare> So make && make install, oÉo? [20:32] <JEEB> yes [20:32] <KeshlWare> oÉo. [20:32] <JEEB> make, check that it compiled correctly, then make install to install it [20:33] <KeshlWare> Does Yasm work with all programs well, oÉo? Is it easy to make programs hug it? [20:33] <KeshlWare> And installed, oÉo. [20:33] <JEEB> it's one of the simpler things around, and x264 has a nice framework for using it with multiple architectures (32bit and 64bit, windows and linux etc.) [20:34] <JEEB> both ffmpeg and x264 use x264's yasm framework :3 [20:34] <KeshlWare> Dang, so I can't just hand a blob of Yasm and go "Here'ya go! Make it sexy!"? D: [20:34] <JEEB> uhh [20:34] <JEEB> it needs asm code [20:34] <KeshlWare> *a blob to Yasm [20:35] <KeshlWare> .... Ffmpeg's written in Assembly? [20:35] <KeshlWare> o_ [20:35] <KeshlWare> *o_o [20:35] <JEEB> parts of it, yes [20:35] <JEEB> just like parts of x264 as well [20:35] <KeshlWare> oÉo.. [20:35] <JEEB> it switches between asm or compiled C depending on your CPU's capabilities [20:35] <JEEB> function pointers, ho! [20:36] <JEEB> anyways, you can now move back to the directory that contains the x264 directory [20:36] <KeshlWare> So confused D: Task at paw, focusing.. [20:36] <JEEB> and clone ffmpeg [20:36] <KeshlWare> Mindjynx <É< [20:36] <JEEB> git clone git://git.videolan.org/ffmpeg.git [20:36] <KeshlWare> "fatal: repository 'ffmpegjustfindtheURLformeplease' does not exist" doesn't -- Oh thanks, oÉo [20:37] <kode54> my radar is getting lots of pings [20:38] <kode54> https://dl.dropbox.com/u/14849789/x264-Keshl.7z try this [20:38] <kode54> static zlib, ffmpeg (decoders only), and ffms2, and gpac [20:38] <kode54> all cross compiled from arch linux [20:39] <KeshlWare> kode54: Ish the syntax the same, oÉo? [20:39] <kode54> should be, it's just a straight up compile, with that range increased to 50000 [20:39] <kode54> who knows [20:39] <kode54> it may actually help [20:39] <kode54> for your weird video [20:39] <KeshlWare> JEEB: Clone'd, oÉo. Should I cryogenically freeze the clone now to preserve it in case of zombie invasion? [20:40] <JEEB> no real need [20:40] <kode54> the change in question is to X264_LOOKAHEAD_MAX in common/common.h [20:40] <kode54> which does increase the size of some stack based arrays [20:41] <kode54> who knows, maybe it will explode [20:41] <KeshlWare> JEEB: Okay, what next then, oÉo? [20:41] <KeshlWare> kode54: Perhaps. Or perhaps it'll hatch a raptor, oÉo. [20:41] <kode54> stop pinging my furdar [20:41] <KeshlWare> A raptor to fight the zombies. <É< [20:42] <KeshlWare> kode54: It's pinging for a reason. XD [20:42] <kode54> I already know [20:42] <kode54> but really [20:42] <kode54> try that binary [20:42] <kode54> it may work [20:42] <kode54> it at least runs [20:42] <kode54> I haven't tried to encode anything with it yet [20:42] <KeshlWare> Yes. The change was so crazy, that the results, miiiight just work. .... -Said with an Italian accent.- [20:43] <JEEB> KeshlWare, I'm writing down a base configure line, just make sure to switch all instances of /home/yourusername/ownapps with whatever prefix you used with x264 [20:43] <KeshlWare> ...Wait the freak was I running before? [20:43] <KeshlWare> JEEB: oÉo. [20:43] <JEEB> ugh.. did xchat just copy what I selected and what I had just written went out of the window... [20:43] <JEEB> meh [20:44] <KeshlWare> D: [20:44] <KeshlWare> kode54: Crap broke D: [20:44] <KeshlWare> But not in the way you thought D: [20:44] <KeshlWare> x264 -i IN.mkv -vcodec libx264 -preset medium -crf 18 -rc-lookahead 251 RC.mp4 [20:44] <KeshlWare> returns: x264: invalid option -- c [20:45] <JEEB> umm [20:45] <kode54> ahahaha [20:45] <JEEB> x264 is not ffmpeg [20:45] <JEEB> :P [20:45] <kode54> yeah [20:45] <KeshlWare> But he said the syntax is the same D: [20:45] <kode54> I built x264 [20:45] <kode54> oh [20:45] <kode54> no [20:45] <kode54> x264 has its own syntax [20:45] <KeshlWare> .É. [20:45] <kode54> I can try building an ffmpeg binary with encoders then [20:45] <JEEB> I think we'll get to that before that finishes :) [20:45] <KeshlWare> Aie's not the smartest thingy on the round shape .É. [20:46] <kode54> imaginative enough to come up with that on the spot [20:46] <KeshlWare> Imaginative yes, but smart questionable D: [20:47] <KeshlWare> JEEB: So what do I -- Oh wait right Xchat ate your clipboard. [20:47] <JEEB> anyways, ./configure --extra-cflags="-I/home/yourusername/ownapps/include -U__STRICT_ANSI__" --extra-ldflags="-L/home/yourusername/ownapps/lib" --prefix=/home/yourusername/ownapps --arch=x86_64 --target-os=mingw32 --cross-prefix=x86_64-w64-mingw32- --enable-gpl --enable-libx264 [20:47] <JEEB> there [20:47] <KeshlWare> Oh good you got it back ^É^! [20:47] <JEEB> I'm not sure if that -U__STRICT_ANSI__ if (still) needed, but I've needed when building for mingw at times [20:47] <JEEB> s/if/is/ [20:48] <JEEB> if configure's output only derps at your because of the url you used to clone ffmpeg not being source.ffmpeg.org, then you should be fine [20:48] <JEEB> s/your/you/ [20:48] <KeshlWare> It's taking a while. Either it's working or it's failing in slow motion, oÉo. [20:49] <JEEB> yes, it will take a while [20:49] <JEEB> shell on windows, fork() fork() fork() is not exactly fast [20:49] <KeshlWare> Done now oÉo [20:49] <KeshlWare> XDDD [20:49] <KeshlWare> Yes, cuz, see, windows only has spoons. >w> [20:49] <JEEB> you should be able to go through the configure's output [20:49] <JEEB> and it should seem fine? [20:49] <KeshlWare> meanwhile Linux has sprokives. ( http://savantemeritus.files.wordpress.com/2010/04/titaniumspork.jpg ) [20:50] <beastd> note that it is worth it to invoke configure with dash on Windows if you have it. it is noticably faster, at least factor 2... [20:50] <JEEB> if the configure's output seems fine [20:50] <JEEB> then you can try make :) [20:50] <JEEB> and wait for a... while :D [20:50] <KeshlWare> It says arch is x86, but C compiler is x86_64 D: [20:51] <JEEB> if it uses the x86_64-w64-mingw32 compiler, it's fine [20:51] <JEEB> it'll be x86_64 just fine [20:51] <KeshlWare> oÉo. [20:52] <KeshlWare> "WARNING: x86_64-w64-mingw32-pkg-config not found, library detection may fail." [20:52] <JEEB> yeah, that's fine [20:52] <JEEB> since you set the cflags and ldflags [20:52] <KeshlWare> And it complains about the thingy you said oÉo [20:52] <JEEB> yeah [20:52] <JEEB> that's usual [20:52] <KeshlWare> ...Uhoh .É. [20:52] <JEEB> the ffmpeg repository is actually hosted on videolan, so it's not really anything incorrect [20:52] <JEEB> if ffmpeg moves away from videolan, sure. But I don't see any reason for them to do it [20:53] <KeshlWare> http://pastebin.com/MLA7wW85 [20:53] <JEEB> yup, looks just fine [20:53] <KeshlWare> Scroll more D: [20:54] <JEEB> oh [20:54] <JEEB> lol [20:54] <JEEB> at least you haven't done anything wrong so just for now disable indevs I guess [20:55] <KeshlWare> How? D: [20:55] <JEEB> --disable-indevs [20:55] <JEEB> in the configure [20:55] Action: KeshlWare doo-doo-sing's the Jeopardy theme while waiting for ./configure. [20:56] <KeshlWare> Kay stuff's happening now, oÉo [20:59] <kode54> I'm building the reverse now [20:59] <kode54> ffmpeg that imports libx264 [21:00] <KeshlWare> Isn't that what this be do's, oÉo? [21:00] <kode54> yup [21:01] <KeshlWare> So how's you doing the reverse den? oÉo [21:03] <kode54> I already built a decoder-only ffmpeg that x264.exe would import [21:03] <kode54> but now I'm doing what you wanted in the first place [21:03] <KeshlWare> Ohhhh oÉo [21:03] <KeshlWare> JEEB: Uhoh .É. [21:04] <JEEB> hm? [21:04] <KeshlWare> http://pastebin.com/6HNBssUP [21:04] <JEEB> ahahaha [21:04] <JEEB> that's actually not a problem [21:04] <JEEB> it tries to compile documentation with the cygwin compiler [21:04] <JEEB> which you don't have installed yet [21:04] <KeshlWare> Ohhh oÉo [21:05] <JEEB> but it happens before linking ffmpeg itself [21:05] <JEEB> although wait... [21:05] <JEEB> yeah [21:06] <JEEB> that /should/ be just the documentation [21:06] <JEEB> now if I remembered the configure switch to disable that [21:07] <kode54> https://dl.dropbox.com/u/14849789/ffmpeg-Keshl.7z [21:07] <KeshlWare> Aw man, he beat us D: [21:07] Action: JEEB wonders if --enable-cross-compile is actually needed [21:08] <KeshlWare> kode54: Crap broke D: [21:08] <JEEB> and --disable-doc disables documentation [21:08] <kode54> how this time [21:08] <KeshlWare> "x264 [error]: malloc of size 6857216 failed" [21:08] <JEEB> yeah, it couldn't allocate memory [21:08] <kode54> yeah, I figured that 50000 limit was too big [21:08] <KeshlWare> So we needs x64 build? D: [21:08] <JEEB> KeshlWare, try configuring with --enable-cross-compile and --disable-doc added [21:09] <JEEB> then make clean [21:09] <JEEB> and make [21:09] <JEEB> or if you have more than one core, make -jX where X is the amount of cores you have [21:09] <kode54> let's rule out errors here [21:09] <KeshlWare> Do, do, do, doo doo do, do do.. [21:09] <KeshlWare> Si, but I've had bad experience with aything more than one job, oÉo. [21:09] <KeshlWare> *anything [21:10] <JEEB> sure [21:10] <JEEB> also I'm testing here if that dshow indev fails to compile here too [21:12] <kode54> redownload, to verify if I even made a working build (I dropped that limit back down to 250) [21:12] <kode54> wait [21:12] <kode54> it's not done uploading yet :[ [21:12] <kode54> there [21:13] <kode54> 50000 frames may not be possible with a 32 bit build [21:13] <kode54> since that's an awful lot of memory [21:13] <JEEB> KeshlWare, how's it doing? [21:14] <KeshlWare> .... Well. [21:14] <KeshlWare> You know the enter key? [21:14] <KeshlWare> It's important to hit dat. <É< Danks for reminding me. [21:14] <JEEB> :P [21:14] <KeshlWare> God gave us one mouth and two ears, because listening is twice as important as talking. [21:14] <kode54> so does the new set of binaries work at all? remember, limit of 250 [21:14] <KeshlWare> Likewise. [21:14] <KeshlWare> The keyboard has two enter keys. [21:15] <KeshlWare> kode54: Nu, tried with 251, oÉo [21:15] <kode54> I meant 250 period [21:15] <kode54> hmm weird [21:15] <kode54> maybe I fail at cross compiling [21:15] <kode54> oh, right [21:15] <kode54> maybe there's an issue [21:15] <kode54> I don't know [21:16] <kode54> I didn't want to install mingw in my vm because it doesn't have the free space [21:16] <kode54> good luck with your compilation efforts, though [21:16] <kode54> can't say I didn't try [21:16] Action: kode54 gives up [21:17] Action: kode54 boots VM to download more ponies with iTunes [21:18] <KeshlWare> oÉo; [21:18] <KeshlWare> .... Interesting. [21:19] <KeshlWare> kode54: For trying to help, has dis, oÉo. http://boingboing.net/wp-content/uploads/2013/01/312345_143684242453080_525… [21:19] <kode54> haha, thanks [21:22] <KeshlWare> Welcomes oÉo [21:25] <JEEB> KeshlWare, how's it going? [21:25] <KeshlWare> Crap worked OÉO [21:25] <KeshlWare> JEEB: Whut now oÉo? [21:26] <JEEB> then you should now have a ffmpeg.exe in there [21:26] <JEEB> ./ffmpeg in the folder [21:26] <JEEB> and check that it outputs something [21:26] <KeshlWare> It does. oÉo. [21:27] <JEEB> also if you check with ls -alh you'll probably find out that it's a bit bigger than you're used to [21:27] <JEEB> it should have debugging symbols :D [21:27] <KeshlWare> o_o Oh gosh do you expect me to commit code? [21:28] <JEEB> no [21:28] <KeshlWare> Good. It'd break. [21:28] <KeshlWare> <.< [21:28] <JEEB> you can now just use the ffmpeg binary [21:28] <KeshlWare> And nu, it's /smaller/ than I'm used to, ow. [21:28] <KeshlWare> But I has *_g.exe's that are bigger oow [21:28] <KeshlWare> *oÉo. [21:28] <JEEB> oh right [21:28] <kode54> _g are -g builds, I presume [21:29] <JEEB> yeah [21:29] <JEEB> that haven't been stripped [21:29] <kode54> aka debugging [21:29] <KeshlWare> oÉo. [21:29] <JEEB> the ffmpeg.exe is the one that's been stripped then [21:29] <JEEB> kind of forgot that [21:29] <KeshlWare> yay crap wo.. [21:29] <KeshlWare> D: [21:29] <KeshlWare> Windows says it crash'd D: [21:29] <JEEB> lul [21:29] <KeshlWare> Even setting rc-lookahead to 60 D: [21:30] <JEEB> you can test the _g one in a mingw gdb in a terminal [21:30] <JEEB> http://x264.fushizen.eu/builds/gdb/gdb-7.5.7z [21:30] <JEEB> extract that, open up cmd.exe or whatever [21:31] <JEEB> fire up 'gdb filename.exe' [21:31] <kode54> oops I just gave somebody a llama [21:31] <JEEB> see it load up the symbols (pick the _g.exe one) [21:31] <JEEB> then 'run -your -settings' [21:31] <JEEB> and it should tell you where it crashed [21:34] <KeshlWare> http://pastebin.com/SzmPzuBv oÉo [21:39] <kode54> bronies are the new furries. good thing I'm not a brony. [21:39] <KeshlWare> No. Bronies are mean D: [21:39] <KeshlWare> They pick on furries a lot >É< [21:39] <KeshlWare> In my experience anyway D: [21:39] <JEEB> KeshlWare, do 'bt' [21:40] <KeshlWare> http://pastebin.com/2PagfStv oÉo [21:40] <JEEB> bt full? [21:41] <KeshlWare> http://pastebin.com/Z9q4bX9V oÉo [21:42] <JEEB> try... running it all outside cygwin [21:42] <JEEB> although that might be a toolchain/compiler error :s [21:43] <KeshlWare> you mean gdb outside too oÉo? [21:43] <JEEB> yes [21:43] <JEEB> it's a mingw compile after all [21:43] <KeshlWare> oÉo. [21:45] <KeshlWare> http://pastebin.com/5uupKzGD oÉo [21:45] <JEEB> yeah... [21:45] <kode54> nope, I just can't do it [21:46] <JEEB> it seems like a problem with the toolchain or something [21:46] <kode54> I can't watch cartoons on speaker [21:46] <kode54> try another GCC package? [21:46] <kode54> was that TDM or something? [21:46] <JEEB> at least you can use those clones of x264 and ffmpeg to compile it with the mingw-w64 toolchain [21:46] <kode54> oh, right [21:46] <kode54> cygwin [21:46] <JEEB> that you got for msys/mingw [21:47] <JEEB> configure lines should be very similar, just remember to clean [21:47] <KeshlWare> oÉo. Danks. [21:47] <JEEB> (or you could build a new mingw toolchain for cygwin with zeranoe's scripts, lol) [21:47] <KeshlWare> For your help, has dis, oÉo. http://hyperboleandahalf.blogspot.com/2010/06/texas.html [21:52] <kode54> so randum [21:52] <KeshlWare> Si oÉo [21:53] <kode54> which reminds me of a story I've heard a number of times [21:53] <kode54> of one of my uncles [21:54] <kode54> when he and my mom would go on road trips through california decades ago [21:54] <kode54> they'd pass through the border checkpoints throughout the southern counties [21:54] <kode54> not actually crossing into mexico, but still [21:54] <kode54> my uncle [21:54] <kode54> they'd ask him if everyone in the vehicle is american [21:55] <kode54> he would respond "Si Señor" [21:55] <kode54> he was obviously caucasian [21:57] <KeshlWare_> .... HAH [21:57] <KeshlWare_> Wait. [21:57] <KeshlWare_> Lemme show you what I saw. [21:58] <KeshlWare_> kode54: http://pastebin.com/04Rh7Zbz [21:59] <kode54> what you missed [21:59] <KeshlWare> A lot .É. [21:59] <kode54> http://pastebin.ca/2304954 [22:00] <KeshlWare> .ca? [22:00] <KeshlWare> Canada? OÉO? [22:01] <kode54> I couldn't think of another handy pastebin at the moment [22:01] <kode54> considering I've heard bad things about specifically pastebin, I don't know if the .ca is related to the mess [22:01] <KeshlWare> But it's Canada right OÉO? [22:01] <kode54> no [22:02] <KeshlWare> D: [22:02] <kode54> pastebin.ca is another entity [22:02] <KeshlWare> But is .ca cCanada? D: [22:02] <kode54> yes [22:02] <kode54> dislike canada? [22:03] <KeshlWare> Nu, I like dem moosen .É. [22:04] <KeshlWare> What was the other idea you had, j0nr? oÉo? [22:04] <KeshlWare> jee* [22:04] <KeshlWare> ... [22:05] <KeshlWare> JEEB: * [22:06] <JEEB> <JEEB> at least you can use those clones of x264 and ffmpeg to compile it with the mingw-w64 toolchain <JEEB> that you got for msys/mingw <JEEB> configure lines should be very similar, just remember to clean [22:06] <KeshlWare> I meant with tthe qthingy oÉo [22:06] <JEEB> what [22:06] <KeshlWare> Doing it the sane way rather than using a massive lookahead, oÉo? [22:06] <JEEB> oh [22:07] <JEEB> first try raising qcomp [22:07] <JEEB> from 0.6 to 0.7 [22:07] <JEEB> and if that or 0.8 doesn't help, then just try disabling mbtree [22:07] <KeshlWare> How's I do the second oÉo? [22:08] <JEEB> same as the first one I'd think [22:08] <JEEB> set a x264opt [22:08] <JEEB> no-mbtree=1 it seems like? [22:09] <JEEB> (or mbtree=0) [22:11] <KeshlWare> ...Whus thingy? .É. [22:11] <KeshlWare> The. [22:11] <KeshlWare> Thingy SYNTAX [22:11] <KeshlWare> That word. [22:11] <KeshlWare> <.< [22:12] <JEEB> -x264opts setting=value [22:12] <JEEB> same for qcomp [22:12] <KeshlWare> oÉo. [22:12] <KeshlWare> ...Uups. [22:15] <KeshlWare> What's the magical paramater that scales down the output, oÉo? [22:16] <JEEB> in what sense? [22:16] <KeshlWare> Preferably by a factor of two or four? [22:16] <KeshlWare> Smaller output resolution, oÉo. [22:16] <JEEB> http://ffmpeg.org/ffmpeg-filters.html#scale [22:16] <JEEB> has examples too [22:16] <KeshlWare> oÉo. [22:18] <KeshlWare> ...I'm bad at this D: [22:19] <KeshlWare> "Unable to find a suitable output format for 'scale=iw/2:ih/2'" [22:25] <KeshlWare> Oh gotet oÉo [22:30] <kode54> would any of these videos be related to console gameplay recordings or emulator recordings? [22:31] <kode54> or is this one of your 3D rendering projects? [22:32] <KeshlWare> Not one of my projects, oÉo. The video is "Tears of Steel", made in Blender but I didn't have anything to do with it. When I play the .mkv the video is choppy. After converting it to h.264 it played fine at its full framerate, but then at the end weird crap happened. Trying to fix that now, oÉo. [22:32] <KeshlWare> http://mango.blender.org/download/ if wanted, oÉo [22:34] <kode54> did you encode the final video to a .mkv or .mp4? [22:34] <KeshlWare> .mp4, oÉo. [22:35] <kode54> and weird crap may depend on your video hardware, if you're using accelerated playback [22:35] <KeshlWare> FYI, due to weird audio hiccups, I passed "-acodec copy" [22:35] <KeshlWare> I am, but trust me it can handle it -- i7-3980qm and nVidia 680m. [22:36] <KeshlWare> What's weird is, although the bitrate never goes far above 12 meg/second with -preset medium, Placebo actually plays nicer. [22:37] <KeshlWare> The bitrate's about half, uses like 20 times the processing power to decode, but it's smoother somehow.. And the device it's playing off of can handle 80 meg bitrates fine. [22:43] <kode54> heh [22:43] <kode54> I only have i5 3570k and nvidia 670 [22:43] <KeshlWare> Not bad, oÉo. [22:44] <kode54> and a 120Hz monitor [22:44] <KeshlWare> Same here, oÉo. Shoulda mentioned that. [22:44] <KeshlWare> But this's at 30 FPS. (I usually do 60 even on this screen, though.) [22:44] <kode54> what are you using to play the video? [22:44] <KeshlWare> VLC, oÉo. [22:44] <kode54> VLC is kind of crap, in my opinion [22:44] <KeshlWare> 2.0.5 [22:44] <KeshlWare> ...oÉo Really? [22:44] <kode54> on Windows, you should probably be using MPC-HC [22:45] <KeshlWare> But it's made by the same guys who made x264 ain't it? D: [22:45] <kode54> I wasn't aware of that [22:45] <JEEB> under the same videolan umbrella that's all [22:45] <kode54> I know it uses ffmpeg for some things [22:45] <KeshlWare> Ah, oÉo. [22:45] <JEEB> but VLC 2.x is relatively OK [22:45] <KeshlWare> kode54: Take what I say with a grain of salt. Often I make mistakes. [22:45] <KeshlWare> Take that with salt too, oÉo. [22:45] <JEEB> compared to previous one [22:45] <JEEB> *ones [22:46] <JEEB> and with 2.1 they seem to be fixing stuff [22:46] <kode54> I think everyone should form their own opinions [22:46] <kode54> and accept the word of others merely as suggestions [22:46] <kode54> depending on the context [22:46] <KeshlWare> oÉo.. Shiny way to live. [22:47] <JEEB> it's just that directshow is a mess, I maintain a directshow playback pack too, so... :V [22:47] <kode54> MPC-HC has built-in decoders for most formats anyway [22:47] <JEEB> which suck [22:47] <kode54> alternative being the decently produced LAV filters [22:47] <JEEB> the splitters and decoders are much more hacked-up than f.ex. LAV's [22:47] <KeshlWare> kode54: VLC has them bui;lt-in too, original reason I got it. [22:48] <KeshlWare> But agreed, VLC's directshow is.. Blah. [22:48] <KeshlWare> Especially on windows. [22:48] <JEEB> the reason to use VLC [22:48] <JEEB> is that you don't want to touch DShow [22:48] <KeshlWare> oÉo [22:48] <JEEB> MPC-HC on the other hand, even with internal filters is a DShow player [22:48] Action: KeshlWare ish confus .É. [22:48] <kode54> yeah, I know dshow is bad [22:49] <kode54> if it didn't have awful frontends in Windows, I would recommend mplayer [22:49] <kode54> all frontends for similar code [22:49] <JEEB> mplayer2/mpv are nice, but yeah -- not exactly user-friendly. Even less because lachs0r is being derp with compiling windows binaries [22:50] <KeshlWare> What do you recommend for general use, oÉo? See, I like sharing my videos when I make stuff, but a lot of my friends aren't.. Shall we say, "computer-competent" [22:50] <KeshlWare> Some of them can only just barely use VLC, and I often have to remind them to use it to play my files. [22:50] <JEEB> for least computer competent people you want to keep away from dshow and stuff [22:50] <JEEB> so VLC is my general recommendation [22:50] <KeshlWare> oÉo. [22:50] <KeshlWare> Good, I didn't get it wrong then <É< [22:51] <JEEB> with win7+ you at least have H.264 and AAC decoders, so in theory WMP should work too [22:51] <JEEB> unless they have borked their MF/DShow [22:51] <KeshlWare> It does, but WMP's decoder sucks. [22:51] <JEEB> it sucks only if you try to feed 10bit H.264 at it or it tries to go and use DXVA too much [22:51] <JEEB> that can derp [22:51] <JEEB> and is hard to disable :V [22:52] <KeshlWare> ....Yes exactly <É< [22:52] <KeshlWare> And no, 8-bit, VLC can't do 10-bit either. [22:52] <KeshlWare> But I usually give people 60 FPS with -crf 18, oÉo. [22:52] <JEEB> uhh [22:52] <JEEB> VLC should be able to do 10bit H.264 just fine [22:52] <JEEB> and... 60fps content [22:52] <JEEB> yeah [22:52] <JEEB> you're just pressing for problems [22:52] <KeshlWare> And then people cna't play it at speed unless I do Ultrafast, but then some won't download it cuz they're on DSL, or Placebo, but then most can't play it. So yeah. <.< [22:55] <KeshlWare> 60 FPS just makes such a difference though .É. [22:56] <KeshlWare> So does 120 but I'm the only one of my usual friends who has a 120 hz screen so <É< [22:57] <kode54> no [22:57] <kode54> not ultrafast [22:58] <kode54> most people won't be able to play 60fps content at all without a beefy machine [22:58] <kode54> or hardware assisted decoding on a modern video card [22:58] <KeshlWare> What defines "modern" here, oÉo? [22:58] <kode54> for instance, a 1080p60 video I encoded a few years ago [22:58] <kode54> won't play back on a Core 2 Duo E8500 without multithreaded software decoding [22:58] <KeshlWare> ...That's usually the systems these people have. [22:58] <KeshlWare> .w> [22:58] <KeshlWare> *>w> [22:59] <kode54> with the paired GeForce 9800 Pro, hardware decoding doesn't work properly [22:59] <gwenael> hello [22:59] <kode54> it caps out at 45fps [22:59] <KeshlWare> Ohai oÉo [22:59] <kode54> yet [22:59] <KeshlWare> kode54: Really? 9800 won't do it? [22:59] <kode54> nope, not in hardware [22:59] <KeshlWare> o_o Whus I do then? [22:59] <kode54> but a Core i3 350M from 2010 can do it in hardware [23:00] <kode54> GTS 250 and newer should be able to do it in hardware as well [23:00] <kode54> for anyone whose system can't do it in hardware, they just need to be walked through disabling hardware decoding in their player [23:00] <KeshlWare> ...These are usually played on systems with 8000-9000 nVidia cards. <É< Is there like, a way to have the vide build a massive pre-rendered buffer? [23:00] <kode54> and unfortunately, if that player doesn't do multithreaded h.264 decoding, it will be even slower than their video card [23:01] <kode54> btw [23:01] <kode54> what I'm talking about [23:01] <gwenael> I can not make a cutting video yet simple, can anyone give me some help? [23:01] <kode54> is not the video card not handling 1080p60 raw frames [23:01] <kode54> I'm talking about not being able to handle hardware accelerated decoding, DXVA style [23:02] <kode54> a C2D E8500 or so should be able to handle full on software decoding, as long as you've got both threads working at the video at once [23:02] <kode54> I don't know if that's an issue [23:02] <KeshlWare> Could try it, oÉo. [23:02] <kode54> back when I was on that system, popular players didn't quite have ffmpeg-mt merged into ffmpeg mainline [23:02] <kode54> is ffmpeg-mt merged in now? [23:02] <JEEB> yes [23:02] <kode54> ah, good [23:03] <kode54> yeah, VLC should handle that fine now [23:03] <kode54> as long as it isn't trying to use DXVA [23:03] <kode54> I don't know if it's capable of that [23:03] <JEEB> I'll just have to note that 720p60 and 1080p60 rendering can be hard at times even with software decoding :P [23:03] <KeshlWare> How's I do, oÉo? Maybe it's easy enough to explain to at least some of my friends. [23:03] <JEEB> uhh [23:03] <kode54> yes [23:03] <JEEB> if you want something to play easily with non-knowledge'able people [23:03] <JEEB> don't use 60fps [23:03] <JEEB> seriously [23:03] <gwenael> how I spend a 60fps video to 1fps with start at ... end N second ? [23:03] <kode54> use 1080p24 [23:04] <JEEB> nah, 30fps is fine [23:04] <kode54> or 30 [23:04] <JEEB> at least much, much less prone to fail [23:04] <JEEB> since june or so of 2011, with first bits going in around march methinks @ ffmpeg-mt [23:04] <KeshlWare> 30 seriously looks bad. o_o [23:04] <KeshlWare> Here, lemme explain the siatuation. [23:04] <KeshlWare> The videos in question that I run at 60 FPS are typically CG slow-motion shots of water falling. [23:04] <JEEB> do some fancy thing with avisynth or ffmpeg with blurring or whatever [23:04] <JEEB> motion compensation, herp derp [23:05] <JEEB> I'm just going to say that if you're going to be using 60fps, you're not gonna get people to watch that shit easily [23:05] <KeshlWare> I render the frames with Luxrender. The software produces extremely life-like images. [23:05] <KeshlWare> Takes ages but it does. [23:05] <KeshlWare> Thing is, the way it renders, it tends to produce a lot of noise in animations (Lest you feel like getting 1 frame per day done) [23:05] <KeshlWare> Doubling the frame rate makes the noise almost impossible to notice. [23:05] <KeshlWare> And blurring it absoutely kills it, since then you can't see the really tiny droplets of mist. [23:06] <KeshlWare> So yeah. oÉo. [23:06] <JEEB> well you can't just fight the fact that 720p60 or 1080p60 is just not gonna be an option for quite some time now :V [23:06] <JEEB> for non-technical people to watch [23:06] <KeshlWare> But at least now I know that you do actually need a decent system to use it. <É< [23:06] <KeshlWare> So danks, oÉo [23:06] <KeshlWare> *to watch it [23:07] <KeshlWare> ....Semi-related question. [23:07] <mudlord> What you trying to do may I ask? [23:07] <mudlord> I have been trying to learn blender [23:07] <mudlord> for some animation related stuff [23:07] <KeshlWare> How bad do you think it'd play on a single core system, sub 2 ghz, with a 6800 GTX in it? XD [23:07] <kode54> KeshlWare: 60fps would be impossible there [23:07] <mudlord> well I thought those cards support HW video decoding [23:07] <kode54> at least at 720 or 1080p [23:07] <KeshlWare> kode54: Would it just chug along? [23:07] <KeshlWare> Or would it just like, poop out? [23:07] <kode54> indeedily doo [23:08] <kode54> chug along [23:08] <KeshlWare> mudlord: Lots of stuff, oÉo. [23:08] <kode54> chop chop chop [23:08] <KeshlWare> kode54: Okay, that explains some stuff then. I thought that'd just poop. [23:08] <mudlord> KeshlWare: know much about blender? [23:08] <mudlord> I want to learn [23:08] <mudlord> but it does seem like ew [23:08] <mudlord> the last time I checked it [23:08] <KeshlWare> mudlord: Not a whole lot, but enough to get people started sometimes oÉo [23:08] <KeshlWare> When is "last time"? [23:08] <KeshlWare> What version? 2.49 or below? [23:08] <mudlord> a few months ago [23:08] <mudlord> yeh [23:08] <KeshlWare> Okay, udnerstandable. [23:08] <KeshlWare> 2.49 and below was evil. [23:08] <mudlord> I just can't seem to render a single deformed cube [23:08] <KeshlWare> Totally different now. Now regular people can use it. <.< [23:09] <mudlord> Yay! [23:09] <KeshlWare> Kay, so. [23:09] Action: KeshlWare finds. [23:09] <kode54> now maybe mudlord can generate some content other than that fractal shader [23:09] <KeshlWare> http://en.wikibooks.org/wiki/Blender_3D:_Noob_to_Pro/Unit_1:_Knowing_Before… Read. [23:09] <kode54> which, while impressive, could use some variance [23:09] <KeshlWare> Don't skip arouund until you finish Unit 2 C. [23:09] <mudlord> :D [23:09] <KeshlWare> Just hit "next module" at the top right to turn pages, oÉo [23:09] <mudlord> kode54: I swear, those fractals are the last time I am going to use raytracing in distance fields [23:09] <mudlord> >_> [23:09] <KeshlWare> mudlord: Shhhh. [23:10] <KeshlWare> <.< [23:10] <KeshlWare> >.> [23:10] Action: KeshlWare finds. [23:10] <KeshlWare> http://fav.me/d5lxxad [23:10] <kode54> btw mudlord [23:10] <KeshlWare> Caustics made from fractles on glass. [23:10] <KeshlWare> <.< [23:10] <kode54> I haven't slept [23:10] <mudlord> Meds screwing up again? [23:10] <kode54> I took my medication at dinner last night like usual [23:10] <mudlord> oh [23:10] <mudlord> weird [23:10] <kode54> then I drank four shots of half caff espresso around 8:30pm [23:11] <kode54> mixed with about 8 ounces of brown water variety of the same ground blend [23:11] <mudlord> Nice refraction effects [23:11] <mudlord> oh [23:11] <KeshlWare> Si, oÉo. [23:11] <mudlord> maybe I should post a screenshot of what I am working on [23:11] <KeshlWare> It "only" took three weeks straight to render on a system with 12 cores at 4.4 ghz. [23:11] <KeshlWare> =D [23:11] <kode54> hey [23:12] <mudlord> mine works on the gpu, only, in REALTIME [23:12] <kode54> I did a fun render once [23:12] <KeshlWare> ...But I was rendering at 4320x7680. So there. [23:12] <kode54> check this [23:12] <kode54> the CNCD/Fairlight demo, Agenda Circling Forth [23:12] <kode54> which renders in realtime on video hardware [23:12] <kode54> I was running it at 1920x1080 [23:12] <kode54> with .kkapture forcing it to render 60fps [23:12] <mudlord> :D [23:13] <kode54> it took roughly two hours to render a 4m55s video [23:13] <kode54> full hardware [23:13] <kode54> I know your raytracer is obscene [23:13] <kode54> but this is something that was designed for real time presentation [23:13] <KeshlWare> yes, it's.. Massive overkill. XD [23:13] <kode54> oh [23:13] <KeshlWare> Assuming you know what you're doing, you *cannot* tell the results apart from photographs. [23:13] <kode54> and I sat there the whole time [23:13] <kode54> moving my mouse so the screen saver wouldn't kick in [23:14] <KeshlWare> kode54: ...Why not just turn the screen saver off? [23:14] <KeshlWare> XD [23:14] <mudlord> that is obscene [23:14] <kode54> I couldn't at that point [23:14] <kode54> because I already started it [23:14] <mudlord> http://i.imgur.com/7GdvoTf.jpg [23:14] <mudlord> my rendering [23:14] <KeshlWare> Oh, you'r elike me? [23:14] <mudlord> full raytracing in distance fields [23:14] <KeshlWare> Start: ..."OH CRAP SCREESANVER. Whelp 7 seconds in, can't stop now!" [23:14] <kode54> a bit OCD? maybe [23:14] <KeshlWare> mudlord: I see dead people O.O [23:14] <kode54> I also write C/C++ software [23:14] <mudlord> http://imgur.com/HG52BPx [23:15] <kode54> and I added multi-instance safety to a 68000 CPU emulator [23:15] <mudlord> :D [23:15] <KeshlWare> kode54: Hehe. <.< you wanna have an OCD war? XD I write in C and absutely refuse toi use pointers and structs. [23:15] <kode54> which involved pasting somestate * m68k, into about 10000 functions [23:15] <KeshlWare> Why? Because typically I compile for AtTiny chips with 256 *bytes* of RAM, and they simply take too much RAM. [23:15] <kode54> ah [23:15] <mudlord> And I almost finished a 52kb graphics demo which involves raytracing on the GPU *only* using shaders [23:15] <kode54> I stick to my own PC [23:15] <kode54> and maybe my router [23:16] <kode54> I refuse to deal with any library that isn't reentrancy save [23:16] <kode54> *safe [23:16] <mudlord> I will copy and paste the actual render code, its braindead simple [23:16] <KeshlWare> o_o Know what the three of us should do? [23:16] <kode54> make a game or something? [23:16] <KeshlWare> Yes, but make it the biggest AMOR possible. [23:16] <KeshlWare> With the smallest executable size. [23:16] <KeshlWare> ... Wait let me try to explain that better. [23:16] <mudlord> http://pastebin.com/GJzKvggY [23:17] <mudlord> thats literally the only C code in there [23:17] <kode54> https://github.com/kode54 I made this stuff [23:17] <mudlord> 09:09 <kode54> now maybe mudlord can generate some content other than that fractal shader [23:17] <mudlord> bored? [23:18] <KeshlWare> http://ideone.com/Y5hDv -- Average code I produce. I write directly in an obfusicated state from time to time <.< [23:18] <kode54> not bored enough to come up with artistic ideas [23:18] <kode54> when I knocked your shader [23:18] <kode54> I remembered [23:18] <KeshlWare> kode54: I can halps OÉO [23:18] <kode54> I don't have the artistic inspiration to make anything like that [23:18] <kode54> I make algorithmic code [23:18] <kode54> a means to an end [23:18] <KeshlWare> I'm not terribly artsy either oÉo [23:18] <KeshlWare> I just do simple geometric art and focus on fractles. [23:19] <mudlord> So you like the mandelbox fractal? [23:19] <KeshlWare> That produces stuff like http://fav.me/d5mg6kc and http://fav.me/d5pdrdo oÉo [23:19] <mudlord> and julia fractal? [23:19] <KeshlWare> Somehow people like them. So there. oÉo. [23:19] <kode54> KeshlWare: I did send you a llama already [23:19] <KeshlWare> Omg llama OÉO! [23:20] Action: KeshlWare didn't notice, doens't use DA often. [23:20] <mudlord> neigh? [23:20] <kode54> I only use it when I find something through google [23:20] Action: kode54 cyberstalker extraordinaire [23:20] <mudlord> oh yeh [23:20] <mudlord> http://www.pouet.net/prod.php?which=60888 [23:20] <KeshlWare> oÉo? What do you mean? [23:20] <kode54> I have an account [23:20] <mudlord> Image rendered in 4kb [23:20] <KeshlWare> mudlord: wat. [23:20] <mudlord> it pisses on your modeller made art [23:21] <mudlord> yes [23:21] <mudlord> 4kb size coding [23:21] <kode54> yes [23:21] <kode54> but [23:21] <KeshlWare> ...Wanna generate that at 16 mhz? [23:21] <KeshlWare> =D [23:21] <kode54> these people do the impossible [23:21] <kode54> because they're mensa or something [23:21] <kode54> bisqwit is mensa [23:21] <mudlord> I just make shitty fractals [23:21] <mudlord> D: [23:21] <mudlord> with shitty motion blur [23:21] <kode54> hey [23:22] <kode54> whatever passes your time [23:22] <kode54> makes you feel good [23:22] <kode54> feel productive [23:22] <mudlord> good point [23:22] <kode54> my software makes me feel productive [23:22] <kode54> I do things for other people a lot [23:22] <kode54> it gives me things to pass the time [23:22] <mudlord> actually, fizzer is not in mensa: he just is a graphics coder who has art skills [23:22] <mudlord> IF I had art skills, I will do 4k stuff [23:23] <kode54> mudlord: may as well be in mensa with those skills [23:23] <kode54> that's crazy awesome [23:23] <KeshlWare> kode54: Any chance you know C# and/or .net 2.0, oÉo? [23:23] <mudlord> but sadly I dont, so I focused on fractals in this prod [23:23] <kode54> I've done a bit of that [23:23] <mudlord> at least I understand fractals [23:23] <kode54> I mostly do pure C/C++, and I've worked with WTL and Qt [23:23] <kode54> but I've done a bit of C# [23:24] <kode54> I'm the Code Doctor [23:24] <KeshlWare> First of all, I love you for doing C <.< [23:24] <kode54> I did a PSF loader in pure C [23:24] <mudlord> KeshlWare: what models does blender now support? [23:24] <kode54> heard of the console music format PSF? [23:24] <mudlord> can it export to OBJ? [23:24] <KeshlWare> Second.. Decrypt this for me: Don't read the output please. <.< >.> [23:24] <KeshlWare> http://ideone.com/8WYOT [23:24] <KeshlWare> mudlord: Si, oÉo. And import. [23:24] <mudlord> Yay! [23:24] <KeshlWare> It's been used to.. Here, sec. [23:24] Action: KeshlWare finds. [23:25] <KeshlWare> Kay. [23:25] <KeshlWare> So. [23:25] <kode54> oh hell [23:25] <KeshlWare> The avreage movie in the US costs between $50,000,000 and $150,000,000 to produce, right? [23:25] <mudlord> yes [23:25] <mudlord> I make my demos on a $0 budget [23:25] <KeshlWare> Kay. [23:25] <kode54> from your embarrassment I was expecting erotic fiction or something [23:25] <KeshlWare> And you know the industry standard for CG stuff is Maya, right? [23:26] <mudlord> http://www.pouet.net/groups.php?which=11668 [23:26] <kode54> I remember bubba [23:26] <KeshlWare> kode54: .... Errrrr... So, uh, anyway, decode that for me ^É^' [23:26] <mudlord> thats what I done so far [23:26] <kode54> do you remember buba? [23:26] <kode54> you do know you can just cpp it, right? [23:27] <KeshlWare> kode54: I mean, make the code plaintext. [23:27] <kode54> I know [23:27] <kode54> C preprocessor can decode the macros to plaintext [23:27] <mudlord> eh, no: I am not decoding that [23:27] <KeshlWare> Yes, but not far enough. xwx [23:28] <mudlord> Oh [23:28] <mudlord> do you know anyone that does fonts [23:28] <mudlord> I need a small font for my prod [23:28] <mudlord> preferably in TGA [23:28] <mudlord> which I can then compress [23:28] <KeshlWare> Nu, oÉo. Wish I did but I dun. [23:28] <mudlord> darn [23:28] <mudlord> :< [23:28] <KeshlWare> If you need massive processing time though for some reason, ask me. Sometimes I loan my desktop out to people over SSH. [23:29] <mudlord> http://www.pouet.net/prod.php?which=60756 [23:29] <mudlord> cracktro style [23:29] <KeshlWare> Right now it's out of comission though. Bad RAM, waiting on an RMA> [23:29] <mudlord> yay [23:29] <mudlord> oh [23:29] <mudlord> that prod was simple technically too [23:29] <mudlord> around 3-4 FBOs [23:29] <mudlord> rendered in layers [23:29] <mudlord> then GLSL multitexturing to do lights [23:30] <mudlord> the latest one is even simpler [23:30] <mudlord> just fractals on GPU + extra framebuffer textures for motion blur [23:31] <KeshlWare> I'll still never get how people make stuff like this. o_o Which is weird considering my line of hobby. [23:31] <mudlord> just learn to code [23:31] <mudlord> that simple [23:31] <kode54> here it is expanded a bit [23:31] <kode54> http://ideone.com/dSEIme [23:31] <KeshlWare> Oh, I know how, sorta. [23:32] <KeshlWare> I just mean like... Gah I cna't explain it. [23:32] <mudlord> so you know about plotting pixels and stuff? [23:32] <KeshlWare> Nu, oÉo. I cna't make GUI's. That's probablt why I can't grasp this. XD [23:32] <kode54> bbl lunch [23:32] <kode54> have fun you two [23:32] <mudlord> yes, graphics API are a bit complicated [23:32] Action: KeshlWare huggles kode54 -É- [23:32] <mudlord> APIs [23:32] <mudlord> Oh I like hugz [23:32] <kode54> maybe hook up in another channel if this one isn't welcome to the off-topic discussion [23:32] <mudlord> :3 [23:33] Action: kode54 huggles back [23:33] <KeshlWare> mudlord: I mean, I know how computers work, and stuff, but for the life of me it's just.. Blargh./ [23:33] <mudlord> you mean demomaking? [23:33] <KeshlWare> See, I've been descriubed as "Programming Assembly with the C syntax" numerous times. [23:33] <KeshlWare> So appearntly I think too low-level for this stuff. [23:33] <mudlord> yeah, its a art [23:33] <mudlord> well [23:33] <mudlord> C is not too low level [23:33] <KeshlWare> *an o_o -tailtwitch- [23:33] <mudlord> I just abuse the compiler [23:33] <KeshlWare> The way I use it, it is. XD [23:33] <mudlord> and have some custom tools to spit out small code [23:33] <kode54> I code higher level C++ [23:33] <kode54> but yeah [23:33] <kode54> later [23:33] <KeshlWare> Anyway., demomaking I sorta get. [23:34] <mudlord> trick is to not use C++ STL, etc [23:34] <KeshlWare> I get how the code's compacted to crap and back. [23:34] <mudlord> all those sorta tricks to get small exes [23:34] <KeshlWare> I've done sort of OCDish compression too. [23:34] <mudlord> oh [23:34] <mudlord> so you dont get procedural content generation? [23:34] <KeshlWare> The fierst thing I learnt to program on, and my first language, was C on an old PsP. [23:34] <KeshlWare> 32 megs of RAM, 333 mhz clock. [23:34] <mudlord> psp :3 [23:34] <KeshlWare> And with OS overhead, it's really 4 megs of RAM and 80 mhz. [23:34] <KeshlWare> So I was already really compressed from the get-go. [23:35] <mudlord> I heard indeed the PSP was quite pathetic [23:35] <KeshlWare> And now I use Arduinos and Atmel chips. [23:35] <mudlord> now that is awesome [23:35] <KeshlWare> And I feel weird programming for a desktop/laptop <.< Cuz the RAM is just so freaking huge o_o (100 megs to me, is massive) [23:35] <mudlord> I need to show you something which is done on those sorta stuff [23:35] <KeshlWare> ((That coming from a guy with a laptop with 32 gigs of RAM <.<)) [23:35] <KeshlWare> So, anyway, my idea. [23:36] <mudlord> http://www.youtube.com/watch?v=L5W3b1gsnVI&noredirect=1 [23:36] <KeshlWare> Smallest demo program ever with the simplest, most useless thing like.. Gah sec. [23:36] <KeshlWare> Crap like http://www.youtube.com/watch?v=Z86V_ICUCD4 [23:36] <KeshlWare> And http://www.youtube.com/watch?NR=1&v=aWRdSKIpkfM&feature=fvwp [23:36] <mudlord> hahahah nice :3 [23:37] <KeshlWare> I love AMOR's. <.< [23:37] <mudlord> then you will love the video i linked, its also on embedded hw [23:38] <KeshlWare> Lately I've been doing.. Like. [23:38] <KeshlWare> HERE WE GO [23:38] <KeshlWare> kay [23:38] <KeshlWare> best example [23:38] <KeshlWare> totally got it worded right now. [23:38] <KeshlWare> <.< [23:38] <KeshlWare> Kay, so. My last project was http://ompldr.org/vZ2VvdQ/VertG1.mp4 [23:38] <KeshlWare> Looks nice, but it's bloated to heck and back, IMO. [23:38] <mudlord> aw [23:39] <KeshlWare> Rather than just running code that tells it to turn LED's on and off, the Arduino's listening via USB for char's being sent to it to tell it what to do. [23:39] <KeshlWare> And the program dirivng it is like 10 megs. [23:39] <mudlord> YAY [23:39] <KeshlWare> *driving [23:39] <mudlord> YAY! [23:39] <KeshlWare> XDDD [23:39] <KeshlWare> I wanna do this, but, all on one single chip. XD [23:39] <KeshlWare> Music too. [23:39] <KeshlWare> And bigger. [23:39] <mudlord> looks alright so far [23:39] <mudlord> needs a screen [23:40] <mudlord> so you can program graphics [23:40] <mudlord> and do 3D [23:40] <KeshlWare> Heck. Got something better. [23:40] <KeshlWare> http://www.youtube.com/watch?v=6mXM-oGggrM [23:40] <KeshlWare> (Not mine. <.<) [23:41] <KeshlWare> So basically, something like that, synced with music, and small enough that it can be a standalone device with only 1-3 chips (Cuz I might go nuts and decide to use 12 volt LED's, thus need tlc5940's) [23:41] <mudlord> I honestly seen better [23:41] <mudlord> you need a screen [23:41] <mudlord> :3 [23:41] <mudlord> do democoding on your own platform [23:41] <mudlord> :3 [23:41] <mudlord> plz [23:42] <KeshlWare> That's like a screen D: [23:42] <mudlord> OH [23:42] <mudlord> xF [23:42] <mudlord> *XD [23:42] <mudlord> silly me kind sir [23:43] Action: KeshlWare ponders. [23:43] <mudlord> hm [23:43] <mudlord> thats what we need [23:43] <mudlord> is there a offtopic channel to ffmpeg? [23:44] <KeshlWare> Dunno. Wanna go toooooo.... [23:44] <KeshlWare> ..... [23:44] <mudlord> :O [23:44] <KeshlWare> Iunno. #squee on furnet XD? [23:44] <mudlord> whats the full url of the network [23:44] <KeshlWare> Good question! [23:44] <KeshlWare> Sec [23:44] <KeshlWare> .w> [23:44] <mudlord> :O [23:45] <KeshlWare> irc://irc.furnet.org/#squee oÉo [23:45] <KeshlWare> Assuming your IRC client is smart, you should just be able to click that. [23:45] <KeshlWare> Dunno why I'm dragging you to Furnet actually, but, yeah. [23:45] <KeshlWare> I'm spontanious. Why not? [23:46] <mudlord> I like spontaneous people [23:46] <mudlord> they are cool [23:58] <mudlord> hm [23:58] <mudlord> kode54 left [00:00] --- Mon Jan 21 2013

1 0

ffmpeg-devel.log.20130119
by burek 20 Jan '13

20 Jan '13

[00:06] <pross-au> Compnn: thanks. i will pass on to michael m. [00:26] <cone-746> ffmpeg.git 03Stefano Sabatini 07master:176b72a28c7d: lavf/utils: provide feedback in case of invalid probe size [00:26] <cone-746> ffmpeg.git 03Stefano Sabatini 07master:6b7d9d515309: lavf/segment: add SegmentListEntry and use it [00:26] <cone-746> ffmpeg.git 03Stefano Sabatini 07master:d4890c10680e: lavf/segment: add support to segment list file entries sliding window listing [00:26] <cone-746> ffmpeg.git 03Stefano Sabatini 07master:1de369a632c4: lavf/segment: remove now unused field list_max_segment_time [00:39] <cone-746> ffmpeg.git 03Stefano Sabatini 07master:bbce913edfd8: lavf/segment: move reference stream selection code to a dedicated function [00:39] <cone-746> ffmpeg.git 03Stefano Sabatini 07master:d00df260ca33: lavf/segment: do not auto-select a video stream with attached_pic disposition [00:52] <cone-746> ffmpeg.git 03Stefano Sabatini 07master:040c461ccfea: doc/muxers: apply minor cosmetics to segment option descriptions [01:01] <llogan> saste: "duration specification" might be unclear for some users. [01:02] <saste> llogan, what do you suggest? [01:06] <llogan> saste: a variation of "must be a number in seconds, or sexagesimal format hh:mm:ss[.milliseconds]" perhaps. it's a lot more verbose though. tradeoffs. [01:07] <saste> llogan, no, because i don't want to update the whole manual if the syntax changes [01:09] <llogan> good point [01:10] <saste> originally it was taking an int (number of seconds, unspecified in the docs) [01:10] <saste> then i extended to take a duration specification, which is more flexible [01:10] <llogan> a good improvement [01:10] <saste> i could point to the ffmpeg-utils manuals, but i should do that in at least three different points [01:11] <llogan> nah. it was just thinking out loud more than anything. [01:11] <saste> so i'm rather supposing that the user knows what is a duration specification (which is not a valid assumption, but good for me) [01:13] <llogan> s/it/I [01:36] Action: ubitux is still alive [01:37] <ubitux> gtg, cya [01:37] <beastd> Hey ubitux [01:40] <ubitux> actually i have 5 more minutes [01:40] <ubitux> hey beastd [01:40] <ubitux> how is the subtitles encoding going on? [01:42] <beastd> Nicolas never replied on my 2nd mail AFAICT. So I will try to code a POC now and hope Nicolas will be happy with it :) [01:43] <beastd> ubitux: I try to get it done before you come back. So you can work on the next task once you returned ;) [01:44] <llogan> ubitux: are you traveling? [01:44] <ubitux> beastd: harass him! [01:44] <ubitux> llogan: yup [01:45] <ubitux> llogan: around here: http://lucy.pkh.me/checkpoint.jpg [01:46] <ubitux> and now gtg for good, see you [01:46] <beastd> ubitux: have fun! [02:25] <hyc> 'morning [02:27] <hyc> anyone up for an algorithm puzzle? I've got some C code that was reverse-engineered from an obfuscated binary. [02:27] <hyc> I believe it's an obfuscated implementation of something like HMAC MD5 [02:27] <hyc> but haven't been able to identify for certain [02:29] <hyc> would like to have a clean generic version of it, since this code does a lot of misaligned accesses. it would never run on sparc, only x86. [02:40] <cone-746> ffmpeg.git 03Michael Niedermayer 07master:d7a46ca179e4: sws: add AVOption flag for error diffusion dither [02:40] <cone-746> ffmpeg.git 03Michael Niedermayer 07master:ca447592b18a: sws: rgb/bgr 4/8 error diffusion dither [02:40] <cone-746> ffmpeg.git 03Michael Niedermayer 07master:8045c0167dbe: sws: Only reset dither state for bitexact mode [02:40] <cone-746> ffmpeg.git 03Michael Niedermayer 07master:2dc2e4f5af7a: sws: Enable full chroma resolution when needed for ED dither [02:56] <burek> if anyone is idle enough to check the new fate interface, you can see it here: http://ffmpeg.gusari.org/fate2/ [02:56] <burek> though, there's only 1 machine in the test folder, because folders were too big to just copy/paste to the test machine [03:02] <michaelni> burek, 0 Tests ? [03:03] <burek> 0 errors in tests [03:03] <burek> need to correct the title [03:07] <burek> fixed [03:08] <burek> when you click on the machine name, you get the history of its uploads [03:08] <burek> where you can check diff between each 2 subsequent uploads [03:08] <burek> among other logs [03:09] <cbsrobot> burek: not sure if I like the overlays, but they should at last have a onmouseout timout [03:10] <burek> I agree [03:10] <burek> I'll add that, though you can ctrl+click anyway, to get the content in a new page [03:11] <cbsrobot> and show the +- only if there is a difference ... [03:11] <burek> that way I would have to do diff for all the items [03:12] <burek> this way i do it only on mouse over [03:12] <burek> it saves cpu [03:12] <cbsrobot> it's just an idea &. [03:12] <cbsrobot> i know nothig about the internals &. [03:16] <burek> :beer: :) [03:16] <burek> thanks :) [03:17] <cbsrobot> and once looking at one machine, the "machine" row is not needed anymore [03:18] <burek> I agree, but then i have to write another css file just for that page :) i'll see if i can just hide that column through css or something :) [03:18] <cbsrobot> I just give some ideas - take what you want ... [03:19] <burek> ok [03:36] <michaelni> burek, 0 failed tests but how many tests did run ? [03:36] <michaelni> this matters too [03:36] <michaelni> if 0 tests are run for example :) [03:37] <hyc> lol [03:47] <cone-746> ffmpeg.git 03sedacca(a)comcast.net 07master:24c47f00fead: lavfi: EBU R.128 irregular time [04:08] <cone-746> ffmpeg.git 03Marton Balint 07master:1427c6842d21: ffplay: use videoq.serial to detect obsolete frames [04:08] <cone-746> ffmpeg.git 03Marton Balint 07master:f65e57124f8a: ffplay: flush video filters on seek [04:08] <cone-746> ffmpeg.git 03Michael Niedermayer 07master:7ba9f40ac602: Merge remote-tracking branch 'cus/stable' [04:29] <cone-746> ffmpeg.git 03Michael Niedermayer 07master:5c42078b9901: sws: fix ED mono dither black level [06:24] <cone-746> ffmpeg.git 03Michael Niedermayer 07master:96d1b7ffcab4: h264: unmark frames at the end [10:46] <cone-175> ffmpeg.git 03Carl Eugen Hoyos 07master:371cf6ee9a4e: Add APIchanges entry for av_codec_get_tag2(). [10:46] <cone-175> ffmpeg.git 03Carl Eugen Hoyos 07master:a5faae46a393: APIchanges: Add git version and fix version for avfilter_ref_get_channels(). [10:46] <cone-175> ffmpeg.git 03Jonas Bechtel 07release/1.1:397fafad230a: Fix opencv detection. [12:01] <durandal_1707> nevcairiel: why you keep your changes to lavf and do not report it to upstream? [12:01] <durandal_1707> s/lavf/lavc [13:07] <cone-175> ffmpeg.git 03Sean McGovern 07master:6d1114a806af: configure: Run SHFLAGS through ldflags_filter() [13:07] <cone-175> ffmpeg.git 03Diego Biurrun 07master:ef5d41a5534b: x86inc: Rename "program_name" to "private_prefix" [13:07] <cone-175> ffmpeg.git 03Michael Niedermayer 07master:1b03e09198e5: Merge commit 'ef5d41a5534b65f03d02f2e11a503ab8416bfc3b' [13:17] <cone-175> ffmpeg.git 03Diego Biurrun 07master:d633d12b2cc9: x86inc: Add cvisible macro for C functions with public prefix [13:17] <cone-175> ffmpeg.git 03Michael Niedermayer 07master:b45e0c2573fa: Merge commit 'd633d12b2cc999cee3ac25bf9a810fe7ff03726d' [13:27] <cone-175> ffmpeg.git 03Martin Storsjö 07master:f4facd2ce70b: x86: Add a Yasm-based emms() replacement [13:27] <cone-175> ffmpeg.git 03Diego Biurrun 07master:822b0728f0e3: x86: dsputil: Drop some unused macro definitions [13:27] <cone-175> ffmpeg.git 03Michael Niedermayer 07master:ed8ff70d9ee0: Merge remote-tracking branch 'qatar/master' [13:38] <cone-175> ffmpeg.git 03Michael Niedermayer 07release/0.8:56cc629a645e: Revert "h264: allow cropping to AVCodecContext.width/height" [14:49] <cone-175> ffmpeg.git 03James Almer 07master:4f8b73129bb3: doc/Makefile: Fix make docclean [14:49] <cone-175> ffmpeg.git 03James Almer 07master:b6576dca9976: configure: Fix Barcelona and Bulldozer compiler options with suncc [14:49] <cone-175> ffmpeg.git 03Matthieu Bouron 07master:61625565e461: lavf/id3v2: cosmetics: reindent [15:17] <cone-175> ffmpeg.git 03Michael Niedermayer 07master:938152196871: cmdline: fix -- to escape files starting with - [15:51] <cone-175> ffmpeg.git 03Nicolas George 07master:beb5d8f07d2a: ffmpeg: accept "chapters" as forced key frames. [16:16] <cone-175> ffmpeg.git 03Ronald S. Bultje 07release/1.1:c3b67720f913: h264: don't clobber mmco opcode tables for non-first slice headers. [16:16] <cone-175> ffmpeg.git 03Anton Khirnov 07release/1.1:6a4803a6a9ab: h264: fix ff_generate_sliding_window_mmcos() prototype. [16:16] <cone-175> ffmpeg.git 03Ronald S. Bultje 07release/1.1:9d60f608af34: h264: add 3 pixels below for subpixel filter wait position [16:16] <cone-175> ffmpeg.git 03Martin Storsjö 07release/1.1:c749bec8c3ef: theora: Skip zero-sized headers [16:16] <cone-175> ffmpeg.git 03Luca Barbato 07release/1.1:21ca4ab9449f: libcdio: support recent cdio-paranoia [16:16] <cone-175> ffmpeg.git 03Xi Wang 07release/1.1:c3c1db7c5637: rv30: fix masking in rv30_loop_filter() [16:16] <cone-175> ffmpeg.git 03Luca Barbato 07release/1.1:0a837b631757: fate: update ref after rv30_loop_filter fix [16:16] <cone-175> ffmpeg.git 03Michael Niedermayer 07release/1.1:002ad7cd3901: Merge remote-tracking branch 'qatar/release/9' into release/1.1 [16:40] <cone-175> ffmpeg.git 03Paul B Mahol 07release/1.1:ccf0cd967d32: 012v: remove double ; and return correct error code if ff_get_buffer() fails [16:40] <cone-175> ffmpeg.git 03Michael Niedermayer 07release/1.1:68a0477bc0af: error_concealment: Check that the picture is not in a half setup state. [16:40] <cone-175> ffmpeg.git 03Michael Niedermayer 07release/1.1:43c6b45a53a1: avcodec_decode_audio: do not trust the channel layout, use the channel count. [16:40] <cone-175> ffmpeg.git 03Michael Niedermayer 07release/1.1:3049d5b9b328: doc/RELEASE_NOTES [17:28] <durandal_1707> michaelni: you know some nice cuve(s) [0,1] <-> [0,1] ? [17:57] <michaelni> durandal_1707, whats nice ? x=y ? [19:57] <durandal_1707> michaelni: nice is something more spiral like sin/cos [20:00] <durandal_1707> ie something like half sin() but more like rotated S [20:00] <durandal_1707> otherwise sin looks more like x=y [20:48] <michaelni> durandal_1707, no clue what you are after but a parametric function within 0,1 that looks nice and has sin in it is sin(t*77)/2+.5,cos(t*101)/2+.5 [20:49] <michaelni> to see it in gnuplot "set parametric ; set samples 10000 ; plot sin(t*77)/2+.5,cos(t*101)/2+.5" [20:50] <Daemon404> michaelni, btw, i purposely havent applied those tablegen fixes, since a certain someone is opting to write proper host tools dep checks / headers [20:50] <Daemon404> just so you dont ask [21:04] <kierank> durandal_1707: are you going to fosdem [21:06] <durandal_1707> kierank: if you payme personal aircraft and body-guards + .... [21:07] <kierank> why [21:07] <durandal_1707> i think i cant go anywhere without visa, say until 2015 or so? [21:11] <Daemon404> some sort of country-specific thing? [21:13] <durandal_1707> kierank: i see nothing interesting of fosdem site... [21:13] <durandal_1707> s/of/on [21:13] <durandal_1707> Daemon404: i dont have passport... [21:14] <Daemon404> you cant get one? [21:15] <durandal_1707> :effort [21:23] <kierank> anyway i know people here don't believe me but meeting in person is not as bad as you imagine [21:24] <wm4> even if it's a libav dev? (or ffmpeg dev vice versa) [21:26] <kierank> mostly yes [21:28] <durandal_1707> kierank: i see nothing related to multimedia on fosdem [21:29] <kierank> it is more of a social thing [21:29] <Daemon404> you can go and ahrras poettering [21:29] <Daemon404> harass* [21:29] <llogan> what kind of beer will be on tap? [21:30] <wm4> why harass the guy who makes the linux desktop a reality [21:30] <Daemon404> is that a joke? [21:30] <Daemon404> :| [21:30] <durandal_1707> s/reality/nightmare [21:30] <Daemon404> ^ [21:30] <wm4> it's both a joke and the truth [21:32] <kierank> llogan: crazy belgian beer [21:32] <kierank> 10% stuff [21:34] <llogan> sounds good. [21:38] <durandal_1707> http://lists.libav.org/pipermail/libav-devel/2013-January/042394.html [21:53] <michaelni> durandal_1707, fixed long ago in ffmpeg [23:44] <cone-175> ffmpeg.git 03Michael Niedermayer 07master:f94220ac7ad0: sws: Fix error diffusion flag [23:44] <cone-175> ffmpeg.git 03Michael Niedermayer 07master:aef5c0b68f31: sws: Check for unsupported combination of ED/FULL_CHR flags [23:44] <cone-175> ffmpeg.git 03Michael Niedermayer 07master:89afa635bf10: sws: check that dither array is not null before memset [00:00] --- Sun Jan 20 2013

1 0

ffmpeg.log.20130119
by burek 20 Jan '13

20 Jan '13

[00:07] <derrink> Hi guys, i want to get FFmpeg latest version (1.1) on my CentOS 6.3 Server, i dont want to compile from source, any one keep any repo with precompiled latest ffmpeg for centos 6 ? [00:08] <klaxa> static builds [00:09] <derrink> the version in repoforge repo is very old, 0.5.2 [00:12] <viric> anyone running linux + firefox watching html5 videos? [00:12] <viric> Somehow I think videos play much worse if html5, than in flash. [00:12] <derrink> what version ffmpeg is the dropbox link? [00:15] <derrink> and how to install it ? [00:16] <klaxa> you extract the archive [00:16] <klaxa> and... well that's it [00:17] <klaxa> you can move the binary to a directory in your PATH [00:17] <derrink> oh [00:17] <viric> W 12 [00:22] <llogan> derrink: or just execute it ./ffmpeg -i ... [00:23] <derrink> yeah llogan i just added the binary to my PATH [00:23] <derrink> i didnt realize it was a binary in the package [00:23] <llogan> just double check that the static version is being used instead of your distro's version [00:39] <elkng> cant ffmpeg make 3D video into regular one ? I want to convert that video http://thepiratebay.se/torrent/6898884 [00:41] <llogan> lol [00:41] Action: llogan "deletes this page" from Firefox history [00:45] <elkng> llogan: so to hide it from your mom ? [00:45] <elkng> I first thought it was directs cut from original avatar [01:00] <klaxa> elkng: this video is in red-blue-3D you won't be able to really extract one of the two images [01:01] <klaxa> if it was saved in seperate video tracks it would have been a different story [01:05] <elkng> klaxa: yes, there are red and blue colors, and also white at the end, if you know what I mean [01:05] <klaxa> :D [01:05] <elkng> I mean credits [01:05] <klaxa> yeah i saw those [01:43] <KeshlWare> Is there a magical way to signify in an .mp4 file encoded with h.264 for it /not/ to be played in-frame in a browser? For some reason it just never works and I'd like to force people when I link to an .mp4 to download it and play it locally rather than in the window. [01:44] <klaxa> i highly doubt it [01:46] <KeshlWare> Dang. D: [01:46] <KeshlWare> Why doesn't it work in-frame, then? I've tried it on multiple systems, multiple OS's and browsers, but it just /never/ works. Instant blackscreen. [01:47] <KeshlWare> Video controls pop up, they sometimes move, sometimes don't, but are always responsive, and the screen's always black. [01:48] <klaxa> is it html5? [01:48] <KeshlWare> I compress using a line like: ffmpeg -r 30 -i pngstream.png -vcodec libx264 -preset medium -crf 18 OUT.mp4 [01:48] <KeshlWare> Dunno. I link directly to the .mp4 file, no HTML inbetween. if therre is any, the browser's adding it. [01:48] <KeshlWare> *If there [01:48] <klaxa> maybe, just maybe webm might work better [01:48] <klaxa> no idea though [01:48] <KeshlWare> Happens on Chrome, afari and Firefox. [01:48] <KeshlWare> *Safari [01:48] <KeshlWare> and Opera. [01:48] <klaxa> i don't like in-line video stuff [01:49] <KeshlWare> Neither do I. x.x [01:49] <drv> you probably need to run qt-faststart or a similar tool on the output file [01:49] <drv> that moves the index to the front of the file so it's streamable [01:50] <KeshlWare> Ah. [01:50] <drv> (qt-faststart is part of ffmpeg) [01:50] <KeshlWare> ...That'd make a lot of sense. Is this part of ffmpeg or -- Kay good. How's I do, oÉo? [01:51] <drv> pretty much just 'qt-faststart file.mp4 rewritten-file.mp4' [01:51] <drv> and then you can delete the original [01:51] <KeshlWare> "ql-faststart is not recognized as an internal or external command", oÉo [01:53] <drv> depending on what distro you are using, it might not be installed as part of the package [01:53] <KeshlWare> I'm on windows. D: [01:54] <drv> this looks promising http://ffmpeg.zeranoe.com/blog/?p=59 [01:54] <KeshlWare> Squee ^É^ [01:55] <KeshlWare> Dankies, oÉo. [01:58] <KeshlWare> drv: Nupe, oÉo. Didn't do it. [01:58] <KeshlWare> http://ompldr.org/vaDRxZA/OUT2.mp4 [02:00] <cbsrobot> KeshlWare: try: ffmpeg -r 30 -i pngstream.png -vcodec libx264 -preset medium -crf 18 -movflags +faststart OUT.mp4 [02:02] <KeshlWare> cbsrobot: Nupe, oÉo. http://ompldr.org/vaDRxZA/OUT2.mp4 [02:02] <KeshlWare> Er [02:02] <KeshlWare> ... [02:02] Action: KeshlWare facedesks and uploads the right file. [02:04] <KeshlWare> Appearntly the site thinks it matches the previous upload. [02:04] <KeshlWare> Checksum must be identical or something. [02:04] <cbsrobot> nah [02:04] <cbsrobot> clean your cache [02:04] <KeshlWare> Yeah, it's the same. [02:04] <KeshlWare> Just did md5sum on each. [02:05] <cbsrobot> hmm what version are you using ? [02:05] <cbsrobot> ffmpeg version [02:05] <KeshlWare> N-47062-g26c531c oÉo [02:06] <KeshlWare> November 25th 2012, some nightly build. [02:06] <cbsrobot> try to update [02:06] <drv> hmm, i'm not sure what chrome doesn't like about that file [02:07] <KeshlWare> Chroime, opera, Firefox and Safari, oÉo. [02:08] <llogan> drv: there is also "-movflags faststart" so you don't have to use qt-faststart [02:08] <KeshlWare> Updated to 1.1, different sum.. [02:08] <drv> interesting, that must be fairly recent... guess i missed it [02:08] <KeshlWare> And smaller output, oÉo. [02:08] <KeshlWare> Barely but yeah. [02:09] <KeshlWare> http://ompldr.org/vaDRxaw/OUT4.mp4 Nupe. [02:09] <llogan> drv: yeah, it is fairly recent. just means you need to get back to some ffmpeg code contributin' to refresh! [02:09] <drv> if only time grew on trees... ;) [02:09] <KeshlWare> Trees grow on time, oÉo. That close enough for'ya? [02:10] <drv> i can't find any obvious way to tell what's going on with the chrome video playback stuff, no obvious output in the dev console or anything [02:10] <llogan> ah, didn't see that cbsrobot already mentioned movflags [02:10] <drv> this may be related http://stackoverflow.com/questions/13513621/creating-chrome-compatible-mp4-… [02:11] <KeshlWare> Ah, baseline might be it. [02:11] <KeshlWare> Too bad I can't use that. [02:12] <cbsrobot> ffmpeg -i OUT2.mov segfaults here! [02:12] <KeshlWare> .mov? These are .mp4's, oÉo. [02:13] <cbsrobot> also OUT4.mp4 segfaults [02:13] <cbsrobot> mov, mp4 its almost the same [02:13] <KeshlWare> Doesn't segfault here, oÉo. [02:13] <KeshlWare> Almost, yes, but what if you happen to have OUT2.mov on your hard drive, in the same folder, from someone else? .É. [02:14] <cbsrobot> KeshlWare: can you pastebin the output of ffmpeg -i OUT4.mp4 [02:14] <KeshlWare> Not easily. Lemme see if I can redirect it, one momeeeent.. [02:14] <KeshlWare> Nope. Not easily. Gimme like ten minutes. [02:14] <cbsrobot> you can use: ffmpeg -report -i OUT4.mp4 [02:15] <cbsrobot> it will write a log file [02:15] <KeshlWare> Says "at least one output file must be spe" ...Oh. [02:15] <KeshlWare> .... Nope that don't work either. [02:15] <cbsrobot> doesnt matter [02:15] <cbsrobot> ah [02:15] <KeshlWare> Just gimme ten minutes. [02:15] <cbsrobot> strange [02:15] <drv> this is my (somewhat outdated) system ffmpeg: http://pastebin.com/SeevWMsS [02:16] <drv> i presume the issue is the color space based on that stack overflow post [02:16] <cbsrobot> can you pastebin the followin report: ffmpeg -report -i pngstream.png [02:16] <KeshlWare> Oh, easuier than I thought: http://pastebin.com/c6Uf4m1F [02:16] <cbsrobot> what is pngstream btw ? [02:16] <KeshlWare> Frame_%04d.png [02:17] <KeshlWare> Nope. Says I gotta supply an output file. [02:17] <KeshlWare> When I do, it complains it cna't use .txt [02:17] <cbsrobot> ok [02:17] <KeshlWare> I figure it wants a movie, which, don't make sense.. <.< [02:17] <cbsrobot> add -pix_fmt yuv420p [02:17] <KeshlWare> To where oÉo? [02:18] <cbsrobot> ffmpeg -r 30 -i Frame_%04d.png -vcodec libx264 -preset medium -crf 18 -movflags faststart -pix_fmt yuv420p OUT9.mp4 [02:19] <KeshlWare> ...oh, the, logs are made. [02:19] <KeshlWare> Weirdo thingy. [02:20] <drv> hmm, my just-built git ffmpeg does not crash with ffmpeg -i OUT4.mp4... [02:20] <KeshlWare> Squee OÉO http://ompldr.org/vaDRxbw/OUT9.mp4 [02:20] <KeshlWare> Choppy cuz buffer but sqyee OÉO [02:20] <KeshlWare> *squee OÉO [02:21] <cbsrobot> wut ? [02:21] <cbsrobot> crf 18 is quite high quality [02:21] <KeshlWare> Ish happynoise, oÉo. [02:21] <cbsrobot> use crf 27 [02:22] <KeshlWare> And yes, it is. I'm a bit obsessive. [02:22] <KeshlWare> I really can see the difference between 18 and 25. It bothers me. >w> [02:22] <cbsrobot> ok - leave it then [02:24] <KeshlWare> Now, I'm crazy and have more RAM than any sane person needs. Is there a magical way to preload all the uncompressed .png's into RAM and then try each preset, oÉo? Sometimes medium compresses better than placebo, for example. (I also have more time than any sane person, and more processing power. <.< Plus the videos are short.) [02:24] <llogan> KeshlWare: generally you want to use the highest crf value that still gives and acceptable quality. [02:24] <llogan> *an acceptable... [02:25] <llogan> placebo is a joke and a waste of time [02:25] <KeshlWare> I know it only gives like a 2% return over veryslow, but 2% is 2%. [02:25] <llogan> less than that [02:25] <KeshlWare> I use Luxrender for serious stuff. Believe me, I obsess over anything greater than .0001%. [02:27] <KeshlWare> And when I use pngcrush, I use -brute. <.< [02:34] <cbsrobot> drv: mine still segfaults [02:34] <drv> hmm, got a backtrace? maybe i am just lucky [02:36] <cbsrobot> http://pastebin.com/07WvJytT [02:36] <cbsrobot> drv: ^ [02:37] <drv> ah, i am building on linux with gcc, maybe it is some corner case that only gets tickled with llvm [02:37] <drv> (or llvm optimization bug) [02:37] <cbsrobot> maybe - I'll file a bug [02:41] <KeshlWare> Just outta curiousity, what are the different settings the presets adjust? [02:42] <cbsrobot> how hard the encoder should try to encode [02:43] <KeshlWare> Well, I figured that. XD But I mean, what exactly do they pass? I know they've just got a list of different paramaters, and I wanna know what those are. [02:43] <cbsrobot> !x264 [02:43] <KeshlWare> I confus. .É. [02:44] <cbsrobot> KeshlWare: see http://mewiki.project357.com/wiki/X264_Settings [02:44] <KeshlWare> oÉo. [02:44] <KeshlWare> Dankies ^É^ [02:47] <llogan> KeshlWare: also see "x264 --fullhelp" [02:47] <llogan> and the console output will show you what settings are being used [02:47] <llogan> or: strings input.mp4 | grep x264 [02:49] <cbsrobot> llogan: can you please test http://ompldr.org/vaDRxaw/OUT4.mp4 ? [02:49] <cbsrobot> just ffmpeg -i OUT4.mp4 [02:49] <cbsrobot> thanks burek [02:49] <burek> :beer: [02:50] <KeshlWare> llogan: I don't have x264.exe, oÉo. [02:54] <burek> http://pastebin.com/RDKYQUq3 [02:54] <drv> x264 --fullhelp has all the details http://pastebin.com/YcbviuEc [02:55] <burek> i think he's on win :) [03:01] <KeshlWare> I said I'm on windows1 D: Anyway danks, was about to say the wikipage doens't got what I need. <É< [03:15] <llogan> cbsrobot: sorry, endless stream of people in my office. works for me. [03:16] <cbsrobot> np [03:16] <cbsrobot> I filed a bug already [03:16] <cbsrobot> thanks anyhow [08:32] <sloof> I'm trying to apply a URL watermark to a video. I'm using a transparent PNG. Right now it's just size 14 bold Arial text. It looks awful when encoded into an otherwise nice looking 1080p video. [08:33] <sloof> Is there a guideline to watermark text font type/size etc so the watermark is legible. It looks like terrible aliasing around the font. [09:35] <p1r473> hey guys, im having some trouble decoding an h264, the output is just gray video [09:35] <p1r473> my decoder is set to libva and it plays fine in vlc [09:44] <relaxed> p1r473: You're using ffmpeg? [09:47] <p1r473> oh I think it uses hcenc or something [09:47] <p1r473> but maybe you still know the answer even if this is a ffmpeg room [09:47] <relaxed> I have no idea what you're doing. [09:48] <p1r473> im burning a mkv to a dvd using avstodvd but the video encoding of the h264 is coming out as gray [09:48] <p1r473> even though i can watch it in vlc, and my ffdshow is set to use libvacodec [09:50] <relaxed> Sorry, no idea. [09:51] <xiphmont> aaaand looks like mpegts is completely borked. Damn. [09:53] <Mavrik> it isn't? [09:55] <xiphmont> what isn't? [10:39] <xiphmont> hm. seeking in ffplay works. older code doesn't. Something must have shifted. OK, n/m, more testing... [10:40] <durandal_1707> seeking on what? [10:52] <brx_> how do i create a 5 second mp4 video from a single jpg (no audio required) [10:56] <relaxed> brx_: -loop 1 -i blah.jpg -t 5 [11:00] <brx_> thanks relaxed [15:07] <Hfuy> Hello. [15:07] <Hfuy> The ProRes encoder seems to give me files which are many tens of percent bigger than the Quicktime encoder on the same material with the same requested profile; is this normal? [15:37] <brontosaurusrex> getting "buffer queue owerflow" with certain input videos (trying to overlay a logo), full -report is here http://pastebin.com/7UhVuHAA , clues? [15:45] <cbsrobot> Hfuy: at least there is no chromashift ! [15:50] <brontosaurusrex> i mean, does this look like : borken filter or wrong usage on my part, or ...? (Could a fresh git compile solve this?) [15:57] <brontosaurusrex> can i increase the internal buffer somehow? should i? how? [16:00] <brontosaurusrex> hrm, if i do an prores intermediate, everything is fine [16:04] <zmode> hi. is it possible to use `-vf "ass=file.ass` with a stream instead of a file? (something like `-vf "ass=#0:2"`) i'd like to not have to extract the subtitles before streaming my video [16:28] <cbsrobot> zmode: mkv ? afaik atm not. [16:29] <cbsrobot> but extracting subtitles is cheap [16:30] <zmode> cbsrobot: yup, mkv. thank you. [16:31] <brontosaurusrex> cbsrobot: any clues about that buffer owerflow? [16:35] <cbsrobot> brontosaurusrex: no - but try to find the smallest command line that triggers the buffer overflow [16:36] <brontosaurusrex> ok [16:39] <brontosaurusrex> seems to be anything with -filter_complex [16:48] <brontosaurusrex> ill try to do a fresh git compile of ffmpeg .... [17:32] <brontosaurusrex> well, no, fresh git did not help [17:36] <brontosaurusrex> ehm, what would be a good intermediate format? (prores is slow) [17:36] <brontosaurusrex> something lossless perhaps? [17:36] <cbsrobot> utvideo ? [17:37] <brontosaurusrex> is it fast? [17:37] <cbsrobot> try it [17:37] <brontosaurusrex> ok [17:37] <cbsrobot> or you could try to pipe it [17:37] <cbsrobot> with yuv4mpeg [17:37] <brontosaurusrex> an example? [17:38] <brontosaurusrex> pipe should be must faster right? [17:38] <cbsrobot> hmm I never remember the bot commands [17:38] <cbsrobot> ah so [17:39] <cbsrobot> hehe [17:39] <cbsrobot> so [17:39] <cbsrobot> no [17:39] <cbsrobot> damn it [17:40] <durandal_1707> lol [17:40] Action: durandal_1707 kicks ffloger [17:40] <cbsrobot> burek - I give you a 6pack beer if you fix that ... [17:41] <cbsrobot> he should also add a command how the bot can send him an email when he's not online .... [17:42] <cbsrobot> btw. durandal_1707 you still have something to do ? [17:43] <cbsrobot> I had another idea what you could if you have time ;-) [17:48] <brontosaurusrex> so the 2nd ffmpeg would just be: ffmpeg -i - or ... ? [17:49] <brontosaurusrex> well, seems to be working [17:50] <cbsrobot> yes [17:50] <brontosaurusrex> damn cool [17:50] <cbsrobot> brontosaurusrex: please fiel a bugreport [17:51] <cbsrobot> btw if you use -strict -1 or similar you can also use -pix_fmt yuv444p10 and several others for the yuv4mpeg format [17:51] <brontosaurusrex> does it matter if input is 420p8 ? [17:52] <cbsrobot> no - just saying [17:52] <brontosaurusrex> ok [17:57] <AL13N> i cannot find any documentation regarding the possible expressions used by ffmpeg in filters. is there a doc regarding this? furthermore, it seems a simple "if(1\,640\,-1)" doesn't work, do i need to set some option to enable expressions in ./configure? [17:58] <cbsrobot> AL13N: http://ffmpeg.org/ffmpeg-utils.html#Expression-Evaluation [18:07] <Orphis_> Is there any up to date tutorial I could use to decode some audio data using ffmpeg APIs in C ? [18:08] <Orphis_> There's of course http://dranger.com/ffmpeg/tutorial01.html but it's a bit old [18:09] <Mavrik> Orphis_: "doc/examples" subfolder in ffmpeg source has up-to-date examples [18:09] <Mavrik> for demuxing and decoding [18:09] <Orphis_> Oh, great [18:09] <Orphis_> That's too obvious to be found :P [18:10] <cbsrobot> let's hide it then [18:23] <Orphis_> What would be the best strategy with the current protocols in ffmpeg to read a non seekable buffer that is managed externally to decode some mp3 data ? [18:24] <Orphis_> The data is fed to my application by some external code and it fills a buffer that I'd like ffmpeg to decode [18:24] <Orphis_> A pipe would be nice, but it's supposed to work on some non Unix platforms too [18:25] <Orphis_> Then, I could use tcp, but isn't that a bit overkill? [18:25] <Mavrik> Orphis_: check the AVIO functions [18:25] <Mavrik> create your own AVIO context and pass pointers to functions in it (check doxygen for what do you need) [18:25] <Mavrik> ffmpeg will call you when it requires data [18:26] <Orphis_> Alright [18:26] <Mavrik> and you'll feed it to its buffer [18:26] <Mavrik> Orphis_: http://ffmpeg.org/doxygen/trunk/avio_8h.html#a853f5149136a27ffba3207d852017… [18:26] <Mavrik> define your own "read_packet" function [18:26] <Mavrik> fill the passed buffer with data when it gets called and return number of bytes filled [18:27] <Mavrik> when just do what the examples do for decode :) [18:27] <Orphis_> That's easy enough [18:50] <raket> is it possible to generate a ffserver.conf based on ffmpeg -i file.avi? [19:09] <maxlapshin> Hi. Can anybody help me with libavcodec/libavformat usage. I transcode mpeg2 to H.264 and get h264 frames delayed. Thus I need to sync audio and video. I try to write them with av_interleaved_write_frame and I thought that it has internal sync queue, but it refuses to accept non-monotonic timestamps [19:09] <maxlapshin> what is the right way to sync audio and video dts (I need to add constant delay to audio)? Do I need to write my own queue? [19:15] <Mavrik> um [19:15] <Mavrik> av_interleaved_write_frame DOES have a queue [19:15] <Mavrik> however no muxer will accept non-monotonic timestamps [19:15] <Mavrik> you just need to make sure your PTS/DTS values always increase [19:17] <maxlapshin> Mavrik: video PTS cannot always increase because of b-frames [19:17] <Mavrik> point there [19:17] <maxlapshin> yet audio is increasing and video is increasing [19:17] <maxlapshin> I'll pastebin now, one moment [19:19] <maxlapshin> http://pastebin.com/DGcNav28 [19:20] <Mavrik> ah I see [19:20] <maxlapshin> https://github.com/erlyvideo/flussonic/blob/0b2751626e61b800bfe3eff0d3eeba4… [19:20] <maxlapshin> and this is source code [19:21] <Mavrik> that's a rather huge difference in PTS [19:21] <maxlapshin> this PTS is in 90000 base [19:21] <Mavrik> I guess you'll have to manually reorder them before muxing :\ [19:21] <maxlapshin> it is not milliseconds [19:21] <Mavrik> I know, but that's still a second of difference [19:21] <maxlapshin> Maybe I'm doing it wrong? [19:21] <raket> where can i get all the ffpreset files for baseline etc? [19:21] <maxlapshin> more than second, about 1.5 seconds [19:22] <maxlapshin> but how can it be a problem if libx264 can give up to 5 seconds delay? [19:22] <Mavrik> raket: uhm... those are obsolete [19:22] <maxlapshin> how ffmpeg is working? [19:22] <Mavrik> raket: what are you trying to achieve [19:22] <raket> Mavrik: well. i need the baseline preset so i can stream rtsp to my old phone :-) [19:23] <Mavrik> raket: baseline is a profile, presets (if you're using libx264) are "slow, medium, fast" etc. and set encoding speed [19:23] <Mavrik> raket: so just do "-profile baseline" and you'll get proper output :) [19:23] <Mavrik> maxlapshin: 5 seconds delay? huh? [19:24] <Mavrik> maxlapshin: MPEG-TS container requires increasing DTS it seems [19:24] <Mavrik> so you'll have to reorder packets when muxing by DTS [19:24] <raket> Mavrik: what should it be in ffserver.conf? AVProfile baseline? [19:25] <maxlapshin> Heh. It's a pity =) So I need separate queue for each output stream [19:25] <maxlapshin> I'm writing multibitrate live transcoder [19:26] <Mavrik> raket: try "AVOptionVideo profile baseline" [19:44] <raket> Mavrik: haha ultra error there :( [20:23] <durandal_1707> cbsrobot: ? [21:31] <brontosaurusrex> raket,bitrate based would look like: -vcodec libx264 -b:v number -preset slow -profile baseline -threads 0 [21:31] <brontosaurusrex> and profiles are bitrate and video-size/framerate related as well [21:33] <brontosaurusrex> or maybe not and i mixed that with levels, in any case its cool [21:34] <llogan> brontosaurusrex: -profile:v [21:34] <brontosaurusrex> right [21:35] <brontosaurusrex> how would i artistically calculate logo size overlay, if video input can be of any size? [21:36] <brontosaurusrex> i need to slap a logo in a bottom corner [21:36] <sacarasc> Do it as a percent of the main video? [21:36] <brontosaurusrex> just taking some percentage of width seems pretty cool so far, but ... [21:37] <Mavrik> brontosaurusrex: why don't you just scale image first and then slap logo on it? :) [21:37] <brontosaurusrex> sacarasc: yeah, should i include height percentage as well? [21:38] <sacarasc> I probably wouldn't... [21:38] <brontosaurusrex> Mavrik: examples: a. input video is 400*300, b. input video is 1920x1080 and so on [21:38] <sacarasc> It would skew it. [21:38] <Mavrik> yeah, but is your output video also arbitrary? [21:38] <brontosaurusrex> yeah, output should be the same size as input [21:38] <brontosaurusrex> but this will only work until some certain size i guess [21:39] <brontosaurusrex> probably i should upscale everything smaller than 1280*720 to just that [21:40] <brontosaurusrex> well, its a kinda silly idea anyway > i mean to hardcode the video, no beauty can be expected [21:40] <brontosaurusrex> the logo even* [21:52] <llogan> just give it 15 pixels padding. done. lazy. [21:53] <brontosaurusrex> allready done in logo itself, even lazier [21:58] <brontosaurusrex> sacarasc> It would skew it. < nah, i mean to take cubic proportion of how much space it would take, scaling would still be proportional [21:58] <brontosaurusrex> anyway, have enough for today [22:50] <Orphis_> Just curious: is there any buildscript for ffmpeg using cmake? [23:55] Action: llogan hits 50% in 85 hour encode session [00:00] --- Sun Jan 20 2013

1 0

ffmpeg-devel.log.20130118
by burek 19 Jan '13

19 Jan '13

[01:38] <cone-420> ffmpeg.git 03Michael Niedermayer 07master:4aed4f58465f: mlpdec: dont leave a invalid huff_lsb in the context. [01:38] <cone-420> ffmpeg.git 03Michael Niedermayer 07master:d270c3202539: avcodec_decode_audio: do not trust the channel layout, use the channel count. [05:23] <cone-871> ffmpeg.git 03Michael Niedermayer 07release/1.1:9bfda9df71c4: mpeg12: Support decoding dimensions that are a multiple of 4096 [05:23] <cone-871> ffmpeg.git 03Michael Niedermayer 07release/1.1:d7cff9f8e8f1: mpeg12enc: Correctly mask dimensions [05:23] <cone-871> ffmpeg.git 03Michael Niedermayer 07release/1.1:cc8ab9865650: mpeg12enc: check dimension validity [05:23] <cone-871> ffmpeg.git 03Michael Niedermayer 07release/1.1:722bfe4e7c73: swr: fix handling of timestamps that cause multiple drops or silence injections [05:23] <cone-871> ffmpeg.git 03Michael Niedermayer 07release/1.1:e7475335b1e0: doc/examples: fix assignments in if() [05:23] <cone-871> ffmpeg.git 03Michael Niedermayer 07release/1.1:07174ed8417f: aacenc: Fix assignments in if() [05:23] <cone-871> ffmpeg.git 03Michael Niedermayer 07release/1.1:d39400fed753: libfaac: Fix assignments in if() [05:23] <cone-871> ffmpeg.git 03Michael Niedermayer 07release/1.1:dacac91973e9: libfdk-aacenc: Fix assignments in if() [05:23] <cone-871> ffmpeg.git 03Michael Niedermayer 07release/1.1:d4a08e560d11: libmp3lame: Fix assignments in if() [05:23] <cone-871> ffmpeg.git 03Michael Niedermayer 07release/1.1:59956a5957d6: libopencore-amr: Fix assignments in if() [05:23] <cone-871> ffmpeg.git 03Michael Niedermayer 07release/1.1:65a4b90840a6: libspeexenc: Fix assignments in if() [05:23] <cone-871> ffmpeg.git 03Michael Niedermayer 07release/1.1:b48e2513606f: libvo-aacenc: Fix assignments in if() [05:23] <cone-871> ffmpeg.git 03Michael Niedermayer 07release/1.1:3be8aeb14e48: libvorbisenc: Fix assignments in if() [05:23] <cone-871> ffmpeg.git 03Michael Niedermayer 07release/1.1:ab471e17e444: nellymoserenc: Fix assignments in if() [05:23] <cone-871> ffmpeg.git 03Michael Niedermayer 07release/1.1:16dc41de279a: ra144enc: Fix assignments in if() [05:23] <cone-871> ffmpeg.git 03Michael Niedermayer 07release/1.1:6a0633e961f4: svq1enc: Fix assignments in if() [05:23] <cone-871> ffmpeg.git 03Michael Niedermayer 07release/1.1:2338eda8d8fd: tiff: Fix assignments in if() [05:23] <cone-871> ffmpeg.git 03Michael Niedermayer 07release/1.1:17704500fb0c: vsrc_testsrc: Fix assignments in if() [05:23] <cone-871> ffmpeg.git 03Michael Niedermayer 07release/1.1:9348514a6713: lavf/mux: Fix assignments in if() [05:23] <cone-871> ffmpeg.git 03Michael Niedermayer 07release/1.1:4d6d8d9ae943: rtmpproto: Fix assignments in if() [05:23] <cone-871> ffmpeg.git 03Michael Niedermayer 07release/1.1:30f0cd2f1e76: h264: fix () placement [06:39] <Darkarnium> .win 4 [06:39] <Darkarnium> >_> [10:31] <cone-3> ffmpeg.git 03Jonas Bechtel 07master:a003c5bd4ff1: Fix opencv detection. [11:34] <durandal_1707> so nobody work on video cmp filter :( [11:57] <durandal_1707> nevcairiel_: ffmpeg does have j2k native decoder but it is experimental .... [11:57] <durandal_1707> it also have wrapper to libopenjpeg [13:42] <cone-810> ffmpeg.git 03Martin Storsjö 07master:e33db35b4a91: theora: Skip zero-sized headers [13:42] <cone-810> ffmpeg.git 03Martin Storsjö 07master:973b4d44f1e8: float_dsp: Add #ifdef HAVE_INLINE_ASM around vector_fmul_window [13:42] <cone-810> ffmpeg.git 03Luca Barbato 07master:43e0e0c4e56d: libcdio: support recent cdio-paranoia [13:42] <cone-810> ffmpeg.git 03Luca Barbato 07master:80ac87c13dc8: lavc: support ZenoXVID custom tag [13:42] <cone-810> ffmpeg.git 03Michael Niedermayer 07master:17596198cacb: Merge commit '80ac87c13dc8c6c063e26a464c5c542357c0583f' [14:00] <cone-810> ffmpeg.git 03Xi Wang 07master:783e37f7ef3b: rv30: fix masking in rv30_loop_filter() [14:00] <cone-810> ffmpeg.git 03Luca Barbato 07master:56ef1ef1f758: fate: update ref after rv30_loop_filter fix [14:00] <cone-810> ffmpeg.git 03Justin Ruggles 07master:f07ef2d9c9e9: lavr: call mix_function_init() in ff_audio_mix_set_matrix() [14:00] <cone-810> ffmpeg.git 03Justin Ruggles 07master:b90632d5964d: lavr: always reset mix function names and pointers in mix_function_init() [14:00] <cone-810> ffmpeg.git 03Anton Khirnov 07master:940b8b586124: h264: avoid pointless copying of ref lists [14:00] <cone-810> ffmpeg.git 03Michael Niedermayer 07master:8ac0d95bd371: Merge remote-tracking branch 'qatar/master' [14:05] <Compn> pross-au : how many gpl cases has xvid taken against companies ? [14:05] <Compn> good success rate ? [14:06] <Compn> pross-au : i ask because i see a decoder has the xvid copyright line : https://ffmpeg.org/trac/ffmpeg/ticket/317 [14:08] <Compn> also wondering if it should be reported somewhere or if i can just pass xvid gpl to you :) [15:16] <cone-810> ffmpeg.git 03Michael Niedermayer 07master:13178d002bed: h264/get_lowest_part_list_y: remove unneeded operations [15:39] <cone-810> ffmpeg.git 03Paul B Mahol 07master:d084c5c742b6: vf_fade: support YUVA422P & YUVA444P [15:52] <cone-810> ffmpeg.git 03Michael Niedermayer 07master:fd165ace7b27: golomb: check log validity before shifting [17:47] <durandal_1707> saste: named options for fade do not work at all [17:50] <saste> durandal_1707, command? [17:53] <durandal_1707> ./ffplay -f lavfi -i testsrc -vf fade=type=in [18:15] <saste> durandal_1707, It accepts the parameters: type:start_frame:nb_frames[:options] [18:15] <saste> it could be extended to support shorthands though [18:26] <durandal_1707> saste: well i expected to output of -h full have sense for fade filter [18:27] <saste> durandal_1707, WIP [18:27] <saste> that was pre-av_opt_set_from_string()/shorthands [19:10] <cone-810> ffmpeg.git 03Stefano Sabatini 07master:44d5a28b7d9e: doc/filters: avoid @example for showing the drawbox shorthand syntax [20:04] <saste> cehoyos, missing APIchange entry for av_codec_get_tag2() [20:08] <saste> wbs, we have an HLS patch on list, maybe you want to review/port it [20:09] <saste> [PATCH] HTTP options in HLS (again) [20:09] <saste> depends on the http cookie patches already applied [21:59] <beastd> llogan: did you change preset fast to medium on purpose here: https://ffmpeg.org/trac/ffmpeg/wiki/x264EncodingGuide?action=diff&version=3… [22:00] <beastd> i think it's fine. just was a bit confused because the comment only told about option reordering [22:32] <llogan> beastd: yes, since most people will probably want to use a slower setting than fast (even if they don't know it) [23:22] <cone-746> ffmpeg.git 03Stefano Sabatini 07master:2852bd704a34: lavfi/fade: accept shorthand syntax [23:22] <cone-746> ffmpeg.git 03Stefano Sabatini 07master:24bbc416a4d8: doc/filters: create a dedicated "Examples" subsection for fade, and use @itemize [23:23] <durandal11707> saste: working on any new filter? [23:23] <saste> durandal11707, not at the moment [23:23] <saste> but i have this lua thing in mind, which is going to be terrific :) [23:24] <durandal11707> i'm doing afade [23:24] <saste> durandal11707, what about histogram? [23:24] <saste> also this reminds that i have the sox and ladspa patch rotting in my hd... [23:25] <durandal11707> sox patch? [23:25] <saste> sox wrapper [23:25] <durandal11707> for what? [23:25] <saste> sox filters [23:26] <durandal11707> about histogram: i was wondering to make it more generice... [23:26] <durandal11707> i dislike wrappers [23:27] <cone-746> ffmpeg.git 03Michael Niedermayer 07master:646ade76793c: sws: Support error diffusion dither for mono output [23:27] <saste> durandal11707, at least they help native ports [23:27] <saste> see mp [23:28] <durandal11707> michaelni: why you are working on mono? [23:29] <durandal11707> michaelni: i guess your could code gbrp/rgba64 output in 4 minutes [23:30] <saste> do we know when ubitux is going to be available again? [23:30] <saste> ebur patches... [23:30] <durandal11707> in next ~30 days [23:30] <michaelni> durandal11707, iam working on mono just for fun :) [23:31] <durandal11707> michaelni: than do pal8. thats fun too [23:33] <durandal11707> michaelni: could those tables be generated at runtime? [23:34] <michaelni> the dither tables, yes but thats not fun [23:35] <durandal11707> i just want to make stuff extensible and easy to maintain/improve because it is not fun to scroll 3gb of tables [23:36] <saste> durandal11707, no fade FATE test? [23:37] <durandal11707> offcourse if code that generates table should be generic enough to generate more useful tables of size bigger that size of code that generates tables ...... [23:37] <durandal11707> saste: there is fade test.... [23:38] <saste> ok because I see no ref changes (so I suppose the test is not affected by your changes) [23:38] <durandal11707> what my changes? [23:38] <saste> added pixel formats [23:39] <durandal11707> the test tests single format [23:39] <durandal11707> look at coverage [23:39] <durandal11707> if it can be improved add more pixfmts otherwise do not care [23:39] <saste> indeed [23:40] <saste> btw i never liked that fade with the frame count [23:40] <durandal11707> but alpha case is not covered ... [23:40] <durandal11707> can be easiliy extended to seconds [23:41] <saste> yes i suppose [23:41] <durandal11707> also fading to only black is little disapointing [23:42] <saste> durandal11707, i have an old patchset implementing colorfading [23:42] <durandal11707> you mean new filter? [23:42] <saste> durandal11707, basically [23:43] <saste> but for black fading fade is more efficient [23:43] <llogan> i think there is a feature request for color fading [23:43] <durandal11707> i dislike such approach it is not rocket science to fade something to certain color [23:43] <saste> durandal11707, not rocket science but someone needs to do the work [23:44] <saste> also there was some syntax problem, now it may be solvable [23:44] <durandal11707> the fading is also linear ..... [23:44] <llogan> #1822 [23:45] <saste> durandal11707, alternatively, there is hue which can be used for black fading [23:45] <durandal11707> but not for non-black? [23:46] <saste> only brightness fade-in/out [23:46] <durandal11707> those are just functions of time on image pixels [23:47] <durandal11707> so some generic filter that use eval should be fine [23:48] <saste> durandal11707, sure you can use geq [23:48] <saste> but it only works with YUV [23:49] <durandal11707> geq accepts time as argument? [23:49] <saste> also a dedicated filter is more efficient [23:49] <saste> sure [23:49] <durandal11707> so what geq that fade to pink would look like? [23:51] <saste> durandal11707, you have to compute pink in YUV coordinates, which is not very handy [23:51] <saste> ping on lavf/utils: provide feedback in case of invalid probe size [23:52] <durandal11707> saste: whatever, random color [23:53] <cone-746> ffmpeg.git 03Paul B Mahol 07master:c79be51dfc80: vf_fade: use AVPixFmtDescriptor to find out if pixel format have alpha [00:00] --- Sat Jan 19 2013

1 0

ffmpeg.log.20130118
by burek 19 Jan '13

19 Jan '13

[01:05] <amaurea> Can it get the number of frames in a video stream without decoding all of it? [01:06] <sacarasc> I believe that some containers put that info in the headers. But not all do. [01:06] <pzich> if you pass just the input file with nothing else it should tell you what it can about the streams [01:06] <pzich> I think [01:07] <amaurea> where is that information available in the programming api, if the container provides it? [01:07] <amaurea> somewhere in libavformat, I guess? [01:08] <amaurea> (not storing the number of frames is very strange, I think. Like a picture format that does not store the number of pixels, only the intended display dimensions in cm or something) [01:09] <sacarasc> MPEG2, for example, you can cut and paste bits together as if they were letters of a string... [01:11] <amaurea> sacarasc: so the total duration is not available in mpeg2? [01:12] <pzich> the weirdest thing is videos without a fixed pixel size, which digital broadcast TV uses [01:12] <pzich> that or they're just abusing the MPEG2TS format [01:12] <amaurea> i.e. videos vith varying resolution? [01:12] <amaurea> I've seen that in the output from some emulators too [01:12] <pzich> yeah, like when they splice an SD ad into an HD TV show [01:16] <amaurea> Anyway, I'm writing a program that plays a video backwards by seeking to near the end, decoding some frames into a buffer, playing the buffer backwards, then seeking a bit earlier, and repeating. But I haven't found out how to get the list of the frames I can jump directly to yet. [01:23] <beware> can i find out if an existing xvid avi file uses features like qpel, GMC, etc? [03:40] <p4plus2> Is there any way to pause an ffmpeg desktop recording or would I have to stop it and splice together the videos at a later point? I attempted "killall -STOP ffmpeg" abd "killall -CONT ffmpeg" which kinda works but that doesn't provide a good pause creating lots of sounds artifacts and cutoff sound. [04:40] <burek> p4plus2 there is no pause/resume feature in ffmpeg [05:15] <p4plus2> (19:42:42) burek: p4plus2 there is no pause/resume feature in ffmpeg [05:15] <p4plus2> ah well, shouldn't be hard to make a script to fake it by setting up "sessions" which splice together all parts on the final stop [08:53] <Zeeflo> i dont quite get the documentation for setting the PlayResX and PlayResY [08:53] <Zeeflo> can someone explain it to me? [08:54] <Zeeflo> Also, would I need a -preset when I define -b:v 1200k -maxrate 1200k -bufsize 2400k ? [09:04] <viric> Zeeflo: preset only determines how much cpu will you dedicate to the encoding [09:04] <viric> no? [09:04] <viric> (at least in x264) [09:17] <Zeeflo> oko [10:30] <studybot_> hi [10:30] <studybot_> how can I encode a pcm data to for example mp3? [10:47] <zap0> just do it. [11:16] <retard> studybot_: PRES BUTAN [11:22] <studybot_> well [11:22] <studybot_> i found the command to do it [12:37] <juha__> what is the correct way to convert h264 video from 59 fps to half/ 25 xvid to make it compatible with mpeg4 dvd player?? I made a conversion and it seems that the player can't handle 59 fps [12:37] <juha__> as it plays too slowly [12:38] <durandal_1707> -framerate ? [12:39] <juha__> that does nothing, it fails [12:39] <juha__> even setpts filter fails [12:53] <klaxa> -r ? [12:58] <juha__> i'll try it now if it works.. [12:59] <juha__> yes it works now [13:02] <juha__> i had tried to convert it from a faulty file.. it wasnt a good idea to convert from mkv container to avi [13:02] <juha__> it lost some frame information [13:03] <Marcin_PL> Hello& I can't find, with what was the option -cpq replaced. [13:03] <Marcin_PL> -cqp ofcourse [13:06] <Marcin_PL> I'm using that with x64 codec [13:07] <Marcin_PL> omg. x264. [13:10] <JEEB> cqp? [13:10] <JEEB> I don't think there ever was such a setting with x264 [13:10] <JEEB> -crf for constant rate factor (closest thing we have to 'constant quality') [13:12] <Marcin_PL> Well, OK, let me paste my old line [13:12] <Marcin_PL> from a script [13:13] <Marcin_PL> ffmpeg -i src_$1.avi -an -vcodec libx264 -cqp 27 -preset veryfast tmp_$1.avi [13:14] <Marcin_PL> That worked well& Now I got errror unrecognized option. [13:14] <JEEB> I have an idea what that might be, but I'm not really sure why you would use that [13:14] <JEEB> as the only reason constant quantizer should be used is for a) lossless encoding (qp 0) or b) for development purposes (I don't think you are a developer for x264) [13:15] <Marcin_PL> Naah, lol [13:15] <JEEB> end users should use either bit rate based, or crf. qp is a very dumb rate control system [13:15] <JEEB> in your case I would guess crf should be good [13:16] <JEEB> also avi output with H.264 makes me cringe [13:16] <JEEB> because I'm pretty sure that veryfast still has b-frames [13:16] <Marcin_PL> I'm using that on poor cpu (intel nano), and I got encode fast at possibly good quality and litthe size, to up via 3G modem to youtube [13:16] <Marcin_PL> outdoor :) [13:17] <JEEB> anyways, you probably want to use mkv or mp4 container for the encode, and crf is what you want. You'll probably have to test different numbers with it though [13:17] <Marcin_PL> Maybe it's a way [13:17] <Marcin_PL> But I used avi for now [13:18] <JEEB> well, avi just doesn't handle b-frames correctly [13:18] <JEEB> that's all [13:18] <divVerent> basically [13:18] <divVerent> to explain it somewhat differently [13:18] <divVerent> CRF is closer to "constant quality" than constant QP is [13:18] <JEEB> it's the closest we have to that, as far as I know [13:19] <divVerent> QP merely stands for quantization parameter, which could as well be "anything" [13:19] <Marcin_PL> OK, so maybe You can propose to me some other codecs? At src i got MJPEG+PCMs16le [13:19] <divVerent> it rather defines "roughly" how much data is lost at an intermediate step of encoding [13:19] <JEEB> Marcin_PL, just use -crf instead of -cqp [13:19] <JEEB> and find a nice number [13:19] <divVerent> but there are scenes where losing a lot data is ok, and there are scenes where not much can be freely lost - and deciding this is the task of a good codec [13:19] <JEEB> and try switching to f.ex. mkv [13:19] <divVerent> and basically CRF is one way to try doing this [13:20] <divVerent> basically, it's unfortunate that "quality" starts with Q too... so people assumed a constant QP means constant quality :P [13:22] <divVerent> the rough idea of CRF's scale was "CRF value N means roughly the same output file size as fixed QP N, but with quantization parameters redistributed among frames in a way yielding higher quality" [13:22] <divVerent> but it's really just rough and depends a lot on the input [13:23] <Marcin_PL> Yes, I'd rather want constant quality than large pixels on quick moves or huge file at static scene [13:23] <divVerent> then CRF is what you want [13:23] <Marcin_PL> co cbr/cfr i guess is wrong for me. [13:23] <Marcin_PL> Oh? [13:23] <divVerent> constant bitrate yields those large pixels [13:23] <divVerent> that's what you do NOT want [13:23] <Marcin_PL> Well then, OK& What should I use for ancoding? x264 is good? [13:24] <divVerent> in fact, CRF quite likely can even use a LOWER QP (thus higher accuracy) in your fast moving frames, depending on the input [13:24] <divVerent> x264 is the best we have [13:24] <divVerent> so yes [13:25] <Marcin_PL> And the container? [13:25] <divVerent> what do you want to play it on? [13:26] <divVerent> on PCs or other devices too? [13:27] <divVerent> for PC only, use MKV, it's the probably most flexible container, also supports subtitles and such stuff properly [13:27] <Marcin_PL> PC. And upload to youtube [13:27] <divVerent> mkv it is, then [13:27] <divVerent> for reference, if you want to play it on e.g. iPhone/iPad, MP4 would have been your only choice [13:28] <Marcin_PL> I understand [13:29] <Marcin_PL> But mp4 as a container may be OK for all use, right? [13:29] <divVerent> if you don't need to e.g. embed text subtitles, or freeform metadata, or attachments, or such stuff - sure [13:29] <divVerent> but beware, MP4 is basically Quicktime, and Quicktime stinks ;) [13:30] <divVerent> also, mp4's main catch is that you have only a small choice of audio codecs to pick from [13:30] <divVerent> while mkv supports basically everything, even flac if you really want it [13:32] <Marcin_PL> ok, I'm switching to x264+mp3 at mkv& [13:32] <divVerent> that'll work, yes [13:32] <Marcin_PL> What about that video constant quality then? [13:32] <divVerent> crf will do what you want [13:34] <Marcin_PL> OK, so let's summary: ffmpeg -i src_$1.avi -an -vcodec libx264 -crf 27 -preset veryfast tmp_$1.mkv [13:35] <Marcin_PL> I'll test with different crf [13:35] <divVerent> sure, that looks good... except that you turned off audio for now [13:35] <Marcin_PL> And I prefer use acodec copy later to mux with lame-based mp3 [13:35] <divVerent> why? ffmpeg can use lame too [13:35] <divVerent> the libmp3lame codec in ffmpeg is basically identical to LAME's output [13:36] <divVerent> if you compiled it in, that is [13:36] <Marcin_PL> In theory& [13:36] <divVerent> in practice too [13:36] <divVerent> they use the same code [13:36] <divVerent> it just may sometimes be tricky to map the options from LAME to ffmpeg [13:36] <Marcin_PL> I'm using lame --preset medium $1.wav $1.mp3 [13:37] <Marcin_PL> And that's the point I think, and that's a half minute more to encode [13:37] <divVerent> haha, exactly --preset is currently not supported by ffmpeg for LAME :P [13:37] <Marcin_PL> :) [13:37] <divVerent> so you'd have to figure out which options --preset medium implies [13:37] <Marcin_PL> So I'm doing that in a script :) [13:37] <divVerent> sure, you already have the script, so it's good :P [13:37] <Marcin_PL> Easier is do like that, and easier it will be to switch to aac e.g. [14:30] <Marcin_PL> I tested it out, at crf 27 I got 10x compession and quite goot quality, i'll test that parameter out more. Big thanks for input. [14:47] <StaRetji> Howdy folks. I've a problem which I think someone here can help me solve. I have a folder with video files and I've made a script to load one after another and encode to flash rtmp server (make live stream from multiple video files). However, I have problem, ffmpeg encoding stops on each video file change. [14:49] <StaRetji> I understand why, it stops, then after it loads video2.mp4 it resumes encoding to rtmp server, but this is enough to break playing on xbmc :/ [14:49] <StaRetji> so, is there a way to force ffmpeg to encode to rtmp and do not stop, no matter what? [14:49] <sanaris> You want concatenate files in stream? There is whole man in docs, with variants [14:50] <StaRetji> yes! [14:50] <StaRetji> I read ffmpeg concate, but not sure if I understand it [14:51] <StaRetji> ffmpeg -i opening.mkv -i episode.mkv -i ending.mkv -filter_complex '[0:0] [0:1] [0:2] [1:0] [1:1] [1:2] [2:0] [2:1] [2:2] concat=n=3:v=1:a=2 [v] [a1] [a2]' -map '[v]' -map '[a1]' -map '[a2]' output.mkv [14:55] <StaRetji> I dont understand what is this lol, I was hoping to work like: for i in videos/*/*.avi; do concat -f flv rtmp://10.0.1.1/live/video [14:58] <sanaris> solution to interrupts lies somewhere in streaming software [14:59] <StaRetji> ehm [14:59] <StaRetji> okay :D [15:01] <sanaris> 'pause' things are provided by many streaming utils, just save time 'n' quality dont recode where is no much need)) [15:01] <StaRetji> well, here problem is, that on input when video changes, encoding stops [15:02] <StaRetji> so, streaming stops, and receiving stops also [15:02] <StaRetji> I will play with concat, thx sanaris [15:05] <sanaris> then You will need some 'new funny opening' or voice from another ip.. [15:05] <sanaris> broadcasting client + server [15:09] <StaRetji> btw, how to prevent audio/video out of sync? I've tried -async, but nothing changes. Input stream goes out of sync after several hours. Any suggestions? [15:44] <stulluk> Hi, how to transcode an MKV file ( h264 video, ac3 audio) into TS file ( h264 video, AC3 audio) ? I mean, just changing container, no transcoding, by using ffmpeg ? [15:45] <juanmabc> -vcodec copy -acodec copy [15:45] <stulluk> Is that possible on 216Mhz Arm Cpu? Is this enough speed? [15:46] <juanmabc> copy would be the less intensive of operations [15:46] <juanmabc> i.e. try it yourself, i guess [16:03] <juha__> im trying to encode h264 mkv to xvid avi and ffmpeg fails at same position everytime with error [matroska,webm @ 0x1ff8720] Read error at pos. [16:03] <sacarasc> Is the file complete? [16:03] <juha__> no [16:03] <juha__> yes yes it is [16:03] <juha__> it plays well in vlc [16:04] <juha__> but ffmpeg fails always at that same position [16:05] <juha__> is it possible to skip the errors and just continue? [16:29] <afm> morning all&. compiling the latest git ffmpeg& ran into an issue with librtmp& http://pastebin.com/Tddacwam [16:36] <durandal_1707> afm: and where is config.log? [16:36] <afm> will add& stby [16:40] <afm> updated http://pastebin.com/ds13xrAb [16:43] <grepper> afm: do you have the development package of librtemp installed ? librtmp-dev or somesuch ? [16:43] <afm> neg just the git rtmpdump [16:44] <durandal_1707> afm: you have librtmp.pc? [16:45] <durandal_1707> what "pkg-config --libs --cflags librtmp" gives you? [16:46] <afm> Perhaps you should add the directory containing `librtmp.pc'&.. found in /usr/local/lib/pkgconfig/librtmp.pc [16:47] <afm> looking for how to add& thanks [16:53] <afm> score! thank you for the help [17:06] <simonpatapon> hi guys [17:06] <simonpatapon> i'm trying to capture my webcam with avconf and would like to know hot to set autofocus and if I can zoom [17:11] <Mavrik> that's not something you set in ffmpeg :) [17:14] <simonpatapon> ok i'll look in the v4l chan [17:14] <simonpatapon> thx!!! [17:41] <tekzilla> any documentation on the pattern matching for the input filename(s) ? [17:42] <tekzilla> is the %..d notation printf-like ? [17:55] <cbsrobot> tekzilla: yes [17:55] <durandal_1707> tekzilla: AFAIK pattern matching is for image2 demuxer only any other just use your shell [18:02] <beastd> tekzilla: as others said -- docs for image2 are here http://ffmpeg.org/ffmpeg-formats.html#image2-1 [18:05] <tekzilla> ah ok great, thanks everyone [18:07] <beastd> tekzilla: you are just curious? or you have a use case? [18:11] <tekzilla> beastd: i'm using it to compile a bunch of frame images from cctv cameras into a video stream [18:12] <beastd> tekzilla: ah nice. then image2 to should be good for you, right? [18:12] <beastd> tekzilla: do you use glob pattern or printf-like pattern now? [18:14] <tekzilla> beastd: yes image2 seems to be what i need [18:15] <tekzilla> using a printf-like pattern, when using a glob like *-capture.jpg i got a segfault! [18:16] <tekzilla> or even with *.jpg.. [18:16] <beastd> tekzilla: huh? which ffmpeg version? also did you escape your glob pattern properly like \*-capture.jpg? [18:16] <tekzilla> well no i didnt, guess there was my problem then [18:17] <tekzilla> 0.8.4-6:0.8.4-1 [18:17] <beastd> depends can you try if it works with proper escaping? (you need the glob chars to get through to ffmpeg, if you don't escape them they will be expanded by your shell) [18:20] <tekzilla> mh when i escape the asterisk, it doesnt seem to match [18:22] <beastd> tekzilla: what is ffmpeg saying as error. [18:22] <cbsrobot> tekzilla: are you on ubuntu ? [18:22] <beastd> tekzilla: OMG. i think there is trouble waiting for us [18:22] <tekzilla> it just returns to the shell immediately because it doesnt seem to match any files to process [18:22] <tekzilla> cbsrobot: debian [18:22] <beastd> i guess you don't have ffmpeg installed at all :( [18:23] <tekzilla> ah ok its some bastardized version ? [18:23] <beastd> tekzilla: can you post the output of your command on pastebin? [18:23] <cbsrobot> tekzilla: you can read about it ^ [18:25] <beastd> tekzilla: also look if you can get FFmpeg: http://ffmpeg.org/download.html#LinuxBuilds [18:27] <tekzilla> ok thanks for clearing that up! if i invole it without args i also get the notice about the switch to avconv [18:27] <elbeardmorez> hi can ffmpeg process mpeg-ps / vcds ? i'm just looking for stream info. mplayer copes (though only with a track number explicitly specified). do i need to rip the track from the disc? [18:28] <beastd> tekzilla: that is a shameless scam by libav people. [18:28] <tekzilla> i`ll have to read some more about to find out if this is actually a bad thing, or if there are any deb repos with the actual ffmpeg so i dont have to maintain it by hand [18:28] <beastd> FFmpeg is actively developed [18:29] <beastd> tekzilla: We have problems getting near Debian because a developer of the forked ffmpeg version is also debian developer responsible for many debian multimedia packages including ffmpeg previously [18:31] <tekzilla> yeah it sounds like a big mess and the fork seems rather pointless :/ [18:33] <beastd> tekzilla: it is especially messy because the fork did not rename their libs (instead they named even named the fork after the common stem of the libs). so now it is even technically difficult to provide e.g. debian ffmpeg packages let alone political issues. [18:35] <LukeL> gd forkers [18:43] <mgeary> hi folks [18:44] <mgeary> i've captured some .dv files from a MiniDV camera with iMovie, and i am hoping to convert them to mp4 with ffmpeg. FFmpg gave me this message when i started: "Incompatible pixel format 'yuv411p' for codec 'libx264', auto-selecting format 'yuv422p'" [18:44] <mgeary> and my resulting file had audio, but the video was completely black [18:45] <mgeary> any suggestions? [18:45] <durandal_1707> mgeary: you play file in what player? [18:45] <mgeary> QT pro [18:46] <durandal_1707> than use yuv420p [18:49] <mgeary> durandal_1707: that's not in my -formats list [18:49] <mgeary> i have "yuv4mpegpipe"...? [18:50] <mgeary> are you saying use it like this: ffmpeg -i in.dv -vcodec yuv420p out.mp4 ? [18:52] <durandal_1707> yuv420p is pixel format [18:52] <mgeary> so&? can you give me a sample ffmpeg command? [18:53] <durandal_1707> add -pix_fmt yuv420p [18:53] <durandal_1707> because qt pro plays only h264 with that pix fmt [18:53] <mgeary> ah [18:53] <mgeary> ok [18:53] <mgeary> trying that now. thank you [18:54] <durandal_1707> IIRC this is in FAQ [18:54] <tekzilla> beastd: hmpf, sounds like theres no easy way to a clean ffmpeg install then.. i dont fancy compiling all the dependencies, not using ffmpeg that much [18:55] <tekzilla> but thanks for pointing me to this whole issue [18:55] <tekzilla> as lazy as i am i might just wait it out.. we'll see after some reading :) [19:00] <beastd> tekzilla: waiting it out can go a long way. the fork started in early 2011... [19:03] <beastd> tekzilla: on which linux distro are you? also ffmpeg does not have many deps so building should be really straight forward! anyway you can have a look at our wiki, where you will find compilation guides amongst many more things. [19:24] <viric> Some JVC camcorders say to have "DSC Function" [19:24] <viric> Can anyone guess what can be that about? [19:31] <teratorn> viric: I think that might be the feature that lets it act as a digital camera [19:31] <teratorn> for still shots? [19:36] <viric> Ahh [19:36] <viric> umh may be [21:36] <amaurea> Can I use avformat_seek_file to do frame-accurate seeks by setting min_ts and max_ts to the same number? [21:40] <amaurea> also, is there a way to get a full list of the frames one can seek efficiently to? [00:00] --- Sat Jan 19 2013

1 0

ffmpeg-devel.log.20130117
by burek 18 Jan '13

18 Jan '13

[00:01] <burek> btw, im going to finish this fate thing [00:02] <burek> and then we might think about setting up a machine (maybe even virtual machines) which could produce different builds of ffmpeg (rpm, deb, static, etc) [00:08] <burek> does ffmpeg support this thing: http://en.wikipedia.org/wiki/Active_Format_Description [00:14] <kierank> burek: not well [00:17] <cehoyos> burek, kierank: Do you have samples? [00:18] <burek> I've asked because of the forum post http://ffmpeg.gusari.org/viewtopic.php?f=11&t=787 [00:18] <kierank> it is not about samples [00:18] <kierank> it is about the way afd is implemented [00:18] <kierank> it is a frame accurate piece of metadata [00:18] <kierank> but it's implemented in avcodeccontext [00:19] <kierank> anyway it'll change when the avbuffer stuff gets in [00:20] <Daemon404> kierank, you must have missed the drama^Wdiscussion on ffmpeg-devel about avbuffer [00:20] <Daemon404> /if/ it gets in [00:20] <Daemon404> lololol [00:21] <kierank> yeah whatever [00:21] <kierank> if libav has a feature ffmpeg will merge it [00:22] <saste> resistance is useless, you will be merged [00:23] <cone-942> ffmpeg.git 03Michael Niedermayer 07master:f674cc776f20: h264: always copy block_offset in thread update [00:23] <cone-942> ffmpeg.git 03Michael Niedermayer 07master:97d190283ee2: h264: always copy linesizes in thread update [00:23] <cone-942> ffmpeg.git 03Michael Niedermayer 07master:765a3ba0187b: h264: add assert to check that block_offset is consistent. [00:25] <burek> does -fflags +genpts work on muxer level? i.e. can it be used with -vcodec copy ? [00:29] <saste> burek, feel free to file a feature request, those flags need to be documented [00:29] <saste> anyway gn [00:31] <burek> gn :beer: :) [01:39] <cone-942> ffmpeg.git 03Diego Biurrun 07release/0.10:e287201c77dc: x86: Require an assembler able to cope with AVX instructions [01:39] <cone-942> ffmpeg.git 03Reinhard Tartler 07release/0.10:398d4e866975: Prepare for 0.8.5 Release [01:39] <cone-942> ffmpeg.git 03Anton Khirnov 07release/0.10:3fca5799c686: ivi_common: check that scan pattern is set before using it. [01:39] <cone-942> ffmpeg.git 03Luca Barbato 07release/0.10:145317d22073: vp56: make parse_header return standard error codes [01:39] <cone-942> ffmpeg.git 03Luca Barbato 07release/0.10:211badf0689d: vp56: release frames on error [01:39] <cone-942> ffmpeg.git 03Sami Pietila 07release/0.10:9837f1969384: vp8: reset loopfilter delta values at keyframes. [01:39] <cone-942> ffmpeg.git 03Luca Barbato 07release/0.10:abe345251a1c: yuv4mpeg: reject unsupported codecs [01:39] <cone-942> ffmpeg.git 03Justin Ruggles 07release/0.10:0ba0e3195517: flacenc: ensure the order is within the min/max range in LPC order search [01:39] <cone-942> ffmpeg.git 03Anton Khirnov 07release/0.10:50b8e4c8fd25: avconv: only apply presets when we have an encoder. [01:39] <cone-942> ffmpeg.git 03Diego Biurrun 07release/0.10:576834b08e2e: Ignore generated aviocat tool. [01:39] <cone-942> ffmpeg.git 03Diego Biurrun 07release/0.10:dcf8f259d107: build: Add 'check' target to run all compile and test targets. [01:39] <cone-942> ffmpeg.git 03Michael Niedermayer 07release/0.10:ea3b68c99d24: Merge commit 'dcf8f259d107838ff3778343dcb762398130a1a3' into release/0.10 [01:57] <cone-942> ffmpeg.git 03Carl Eugen Hoyos 07release/0.10:4fc22e85c011: Revert "x86: Require an assembler able to cope with AVX instructions" [02:04] <cone-942> ffmpeg.git 03Diego Biurrun 07release/0.10:26b177b8f7ab: build: fix 'clean' target [02:04] <cone-942> ffmpeg.git 03Janne Grunau 07release/0.10:52adbc0e1793: h264: Fix parameters to ff_er_add_slice() call [02:04] <cone-942> ffmpeg.git 03Anton Khirnov 07release/0.10:3750104e9d80: id3v2: fix reading unsynchronized frames. [02:04] <cone-942> ffmpeg.git 03Anton Khirnov 07release/0.10:f65ec488a94d: avconv: fix copying per-stream metadata. [02:04] <cone-942> ffmpeg.git 03Kostya Shishkov 07release/0.10:e5ea6539d484: indeo3: ensure that decoded cell data is in 7-bit range as presumed by decoder [02:04] <cone-942> ffmpeg.git 03Kostya Shishkov 07release/0.10:c55ca9876975: indeo3: when freeing buffers, set pointers referencing them to NULL as well [02:04] <cone-942> ffmpeg.git 03Anton Khirnov 07release/0.10:56c1e18a5225: mpeg12: do not decode extradata more than once. [02:04] <cone-942> ffmpeg.git 03Luca Barbato 07release/0.10:a5290800f571: mp3: properly forward mp_decode_frame errors [02:04] <cone-942> ffmpeg.git 03Luca Barbato 07release/0.10:3e700cc66be2: vp6: properly fail on unsupported feature [02:04] <cone-942> ffmpeg.git 03Diego Biurrun 07release/0.10:01a4e7f623a2: lavf: Bump minor version to distinguish branch and master version numbers [02:04] <cone-942> ffmpeg.git 03Michael Niedermayer 07release/0.10:52cea9ce9209: Merge commit '01a4e7f623a2e6dc95862f9a56c777f058d7bfaf' into release/0.10 [02:12] <cone-942> ffmpeg.git 03Carl Eugen Hoyos 07master:c345100efc9b: Fix detection of struct v4l2_frmsize_discrete. [02:18] <cone-942> ffmpeg.git 03Carl Eugen Hoyos 07release/1.0:aa1af0d6cace: Fix detection of struct v4l2_frmsize_discrete. [02:18] <cone-942> ffmpeg.git 03Carl Eugen Hoyos 07release/1.1:ccc421955858: Fix detection of struct v4l2_frmsize_discrete. [02:31] <cone-942> ffmpeg.git 03Diego Biurrun 07release/0.10:bb35a42e93c1: APIchanges: Fill in missing commit hashes [02:31] <cone-942> ffmpeg.git 03Alex Converse 07release/0.10:a4a63bf5b55f: aacdec: Fix an off-by-one overwrite when switching to LTP profile from MAIN. [02:31] <cone-942> ffmpeg.git 03Janne Grunau 07release/0.10:d282e5ce7286: lavf: avoid integer overflow in ff_compute_frame_duration() [02:31] <cone-942> ffmpeg.git 03Janne Grunau 07release/0.10:522e97bd9e91: flashsv: check for keyframe before using differential coding [02:31] <cone-942> ffmpeg.git 03Janne Grunau 07release/0.10:6cd92c388095: h264: enable low delay only if no delayed frames were seen [02:31] <cone-942> ffmpeg.git 03Janne Grunau 07release/0.10:b6592b402cd2: flashsv: make sure data for zlib priming is available [02:31] <cone-942> ffmpeg.git 03Janne Grunau 07release/0.10:f1b3cc02ec5e: h264: error out on unset current_picture_ptr for h->current_slice > 0 [02:31] <cone-942> ffmpeg.git 03Luca Barbato 07release/0.10:6b70965f398e: ppc: always use pic for shared libraries [02:31] <cone-942> ffmpeg.git 03Janne Grunau 07release/0.10:1d98811b957d: h264: slice-mt: get last_pic_dropable from master context [02:31] <cone-942> ffmpeg.git 03Janne Grunau 07release/0.10:f620c12067a2: h264: check sps.log2_max_frame_num for validity [02:31] <cone-942> ffmpeg.git 03Victor Lopez 07release/0.10:a335ffd7f4cd: h264: fix sps parsing for SVC and CAVLC 4:4:4 Intra profiles [02:31] <cone-942> ffmpeg.git 03Michael Niedermayer 07release/0.10:c961ce969a61: Merge commit 'a335ffd7f4cdaaa6a8fe4187f6f06b0418eea19a' into release/0.10 [02:41] <cone-942> ffmpeg.git 03Justin Ruggles 07release/0.10:77e6676d3eb3: alacdec: do not be too strict about the extradata size [02:41] <cone-942> ffmpeg.git 03Justin Ruggles 07release/0.10:dfb7a638e6b9: opt: avoid segfault in av_opt_next() if the class does not have an option list [02:41] <cone-942> ffmpeg.git 03Martin Storsjö 07release/0.10:d1d329932fd4: rtsp: Recheck the reordering queue if getting a new packet [02:41] <cone-942> ffmpeg.git 03Dale Curtis 07release/0.10:9ded14fcb8a5: Fix uninitialized reads on malformed ogg files. [02:41] <cone-942> ffmpeg.git 03Luca Barbato 07release/0.10:6eebba08e188: oggdec: check memory allocation [02:41] <cone-942> ffmpeg.git 03Luca Barbato 07release/0.10:03fec31cd76f: oggdec: free the ogg streams on read_header failure [02:41] <cone-942> ffmpeg.git 03Janne Grunau 07release/0.10:06312bbb1018: h264: check context state before decoding slice data partitions [02:41] <cone-942> ffmpeg.git 03Janne Grunau 07release/0.10:adef01c370c9: h264: check ref_count validity for num_ref_idx_active_override_flag [02:41] <cone-942> ffmpeg.git 03Reinhard Tartler 07release/0.10:cf2cab5b2ab8: Update Changelog [02:41] <cone-942> ffmpeg.git 03Michael Niedermayer 07release/0.10:492eb0aa14b6: Merge remote-tracking branch 'qatar/release/0.8' into release/0.10 [02:45] <cone-942> ffmpeg.git 03Carl Eugen Hoyos 07release/0.10:74b04e6a359f: Fix detection of struct v4l2_frmsize_discrete. [02:45] <cone-942> ffmpeg.git 03Carl Eugen Hoyos 07release/0.11:91e016865ccc: Fix detection of struct v4l2_frmsize_discrete. [03:00] <cone-942> ffmpeg.git 03Diego Biurrun 07release/0.8:7b91e52eb9fd: x86: Require an assembler able to cope with AVX instructions [03:00] <cone-942> ffmpeg.git 03Mans Rullgard 07release/0.8:a2ae183a382f: h264: allow cropping to AVCodecContext.width/height [03:00] <cone-942> ffmpeg.git 03Kostya Shishkov 07release/0.8:e39fc137aeac: vc1dec: add flush function for WMV9 and VC-1 decoders [03:00] <cone-942> ffmpeg.git 03Anton Khirnov 07release/0.8:d3e2f35f7add: bmpdec: only initialize palette for pal8. [03:00] <cone-942> ffmpeg.git 03Max Lazarov 07release/0.8:0892a6340f86: eval: fix swapping of lt() and lte() [03:00] <cone-942> ffmpeg.git 03Martin Storsjö 07release/0.8:a81c1ea2eb66: h263: Add ff_ prefix to nonstatic symbols [03:00] <cone-942> ffmpeg.git 03Janne Grunau 07release/0.8:99008ba3667a: rv34: use AVERROR return values in ff_rv34_decode_frame() [03:00] <cone-942> ffmpeg.git 03Mina Nagy Zaki 07release/0.8:10c244cc89e8: lavfi: avfilter_merge_formats: handle case where inputs are same [03:00] <cone-942> ffmpeg.git 03Michael Niedermayer 07release/0.8:7a0ff7566b66: avsdec: Set dimensions instead of relying on the demuxer. [03:00] <cone-942> ffmpeg.git 03Aneesh Dogra 07release/0.8:42c3a3719b05: bytestream: add a new set of bytestream functions with overread checking [03:00] <cone-942> ffmpeg.git 03Anton Khirnov 07release/0.8:4c849c69910c: dfa: check that the caller set width/height properly. [03:00] <cone-942> ffmpeg.git 03Kostya Shishkov 07release/0.8:2d63f9b4effc: dfa: add some checks to ensure that decoder won't write past frame end [03:01] <cone-942> ffmpeg.git 03Kostya Shishkov 07release/0.8:c0df6a24ce9e: indeo: check custom Huffman tables for errors [03:01] <cone-942> ffmpeg.git 03Kostya Shishkov 07release/0.8:601fa565823b: indeo: clear allocated band buffers [03:01] <cone-942> ffmpeg.git 03Kostya Shishkov 07release/0.8:3c0f84402b2b: indeo: check for invalid motion vectors [03:01] <cone-942> ffmpeg.git 03Janne Grunau 07release/0.8:8148833193c6: indeo5: prevent null pointer dereference on broken files [03:01] <cone-942> ffmpeg.git 03Michael Niedermayer 07release/0.8:aa097b4d5fd4: indeo5: check tile size in decode_mb_info(). [03:01] <cone-942> ffmpeg.git 03Michael Niedermayer 07release/0.8:44da556815fc: lagarith: check count before writing zeros. [03:01] <cone-942> ffmpeg.git 03Michael Niedermayer 07release/0.8:4a636a5e4368: wmaprodec: check num_vec_coeffs for validity [03:01] <cone-942> ffmpeg.git 03Anton Khirnov 07release/0.8:05f5a2eb62db: avidec: use actually read size instead of requested size [03:01] <cone-942> ffmpeg.git 03Michael Niedermayer 07release/0.8:6996a2f79689: cavsdec: check for changing w/h. [03:01] <cone-942> ffmpeg.git 03Michael Niedermayer 07release/0.8:aa45b90804ab: alsdec: Check k used for rice decoder. [03:01] <cone-942> ffmpeg.git 03Michael Niedermayer 07release/0.8:597d709eb42a: Merge commit 'aa45b90804ab21175b8c116bd8e5eb4b4e85fbcb' into release/0.8 [03:09] <cone-942> ffmpeg.git 03Michael Niedermayer 07release/0.8:e3e369f6962c: alsdec: Check that quantized parcor coeffs are within range. [03:09] <cone-942> ffmpeg.git 03Thilo Borgmann 07release/0.8:1b48a426a96f: alsdec: Fix out of ltp_gain_values read. [03:09] <cone-942> ffmpeg.git 03Mans Rullgard 07release/0.8:7e070cf2025f: alsdec: remove dead assignments [03:09] <cone-942> ffmpeg.git 03Thilo Borgmann 07release/0.8:9474c9302844: alsdec: fix number of decoded samples in first sub-block in BGMC mode. [03:09] <cone-942> ffmpeg.git 03Sean McGovern 07release/0.8:dd14723602f2: wmapro: prevent division by zero when sample rate is unspecified [03:09] <cone-942> ffmpeg.git 03Anton Khirnov 07release/0.8:562d6fd5b577: avidec: return 0, not packet size from read_packet(). [03:09] <cone-942> ffmpeg.git 03Anton Khirnov 07release/0.8:8168a7cec951: vf_pad: don't give up its own reference to the output buffer. [03:09] <cone-942> ffmpeg.git 03Franz Brauße 07release/0.8:fb3189ce8b8a: smacker audio: sign-extend the initial 16-bit predicted value [03:09] <cone-942> ffmpeg.git 03Anton Khirnov 07release/0.8:5754176b5bfb: yuv4mpeg: return proper error codes. [03:09] <cone-942> ffmpeg.git 03JindYich Makovi ka 07release/0.8:11ecd8574a2e: h264: avoid stuck buffer pointer in decode_nal_units [03:09] <cone-942> ffmpeg.git 03Luca Barbato 07release/0.8:ec6719f6551d: mpegaudiodec: fix short_start calculation [03:09] <cone-942> ffmpeg.git 03Alex Converse 07release/0.8:549b8083d6e0: tiffenc: Check av_malloc() results. [03:09] <cone-942> ffmpeg.git 03Reinhard Tartler 07release/0.8:642d758a2d82: Update RELEASE file for 0.7.7 [03:09] <cone-942> ffmpeg.git 03Anton Khirnov 07release/0.8:9f8071245491: ivi_common: check that scan pattern is set before using it. [03:09] <cone-942> ffmpeg.git 03Luca Barbato 07release/0.8:700fb8c8dd06: vp56: make parse_header return standard error codes [03:09] <cone-942> ffmpeg.git 03Luca Barbato 07release/0.8:7fd7950174f9: vp56: release frames on error [03:09] <cone-942> ffmpeg.git 03Sami Pietila 07release/0.8:bfbff1c7483c: vp8: reset loopfilter delta values at keyframes. [03:09] <cone-942> ffmpeg.git 03Luca Barbato 07release/0.8:f3f22f183fa2: yuv4mpeg: reject unsupported codecs [03:09] <cone-942> ffmpeg.git 03Justin Ruggles 07release/0.8:3d0c9c9af687: flacenc: ensure the order is within the min/max range in LPC order search [03:09] <cone-942> ffmpeg.git 03Reinhard Tartler 07release/0.8:ce8910d861c1: h264: Fix parameters to ff_er_add_slice() call [03:09] <cone-942> ffmpeg.git 03Luca Barbato 07release/0.8:4ede95e69cf9: vp6: properly fail on unsupported feature [03:09] <cone-942> ffmpeg.git 03Alex Converse 07release/0.8:b143844ea0f6: aacdec: Fix an off-by-one overwrite when switching to LTP profile from MAIN. [03:09] <cone-942> ffmpeg.git 03Michael Niedermayer 07release/0.8:3f1a58db6f7a: Merge commit 'b143844ea0f6246e0d5a938d743e2e8a98453bec' into release/0.8 [03:21] <cone-942> ffmpeg.git 03Janne Grunau 07release/0.8:10ff052c6013: lavf: avoid integer overflow in ff_compute_frame_duration() [03:21] <cone-942> ffmpeg.git 03Janne Grunau 07release/0.8:5fa739e685bc: h264: enable low delay only if no delayed frames were seen [03:21] <cone-942> ffmpeg.git 03Luca Barbato 07release/0.8:08d9fd611eac: ppc: always use pic for shared libraries [03:21] <cone-942> ffmpeg.git 03Janne Grunau 07release/0.8:4457e6137d83: h264: check sps.log2_max_frame_num for validity [03:21] <cone-942> ffmpeg.git 03Victor Lopez 07release/0.8:884a9b0d298a: h264: fix sps parsing for SVC and CAVLC 4:4:4 Intra profiles [03:21] <cone-942> ffmpeg.git 03Justin Ruggles 07release/0.8:a39c6bf1b878: alacdec: do not be too strict about the extradata size [03:21] <cone-942> ffmpeg.git 03Martin Storsjö 07release/0.8:808187965570: rtsp: Recheck the reordering queue if getting a new packet [03:21] <cone-942> ffmpeg.git 03Dale Curtis 07release/0.8:55065315caf1: Fix uninitialized reads on malformed ogg files. [03:21] <cone-942> ffmpeg.git 03Luca Barbato 07release/0.8:910c1f235283: oggdec: check memory allocation [03:21] <cone-942> ffmpeg.git 03Reinhard Tartler 07release/0.8:3bc9cfe66e1a: oggdec: free the ogg streams on read_header failure [03:21] <cone-942> ffmpeg.git 03Janne Grunau 07release/0.8:ad025377462c: h264: check context state before decoding slice data partitions [03:21] <cone-942> ffmpeg.git 03Janne Grunau 07release/0.8:dd0c5e0fa909: h264: check ref_count validity for num_ref_idx_active_override_flag [03:21] <cone-942> ffmpeg.git 03Michael Niedermayer 07release/0.8:685321e4bd84: Merge remote-tracking branch 'qatar/release/0.7' into release/0.8 [04:44] <cone-942> ffmpeg.git 03Reinhard Tartler 07release/1.1:8069b44ebfa2: Prepare 9.1 Release [04:44] <cone-942> ffmpeg.git 03Justin Ruggles 07release/1.1:a55700541756: x86: lavr: use the x86inc.asm automatic stack alignment in mixing functions [04:44] <cone-942> ffmpeg.git 03Reinhard Tartler 07release/1.1:c1555ae4b62a: update Changelog [04:44] <cone-942> ffmpeg.git 03Luca Barbato 07release/1.1:42bd6d9cf681: oggdec: free the ogg streams on read_header failure [04:44] <cone-942> ffmpeg.git 03Luca Barbato 07release/1.1:c01be297ce5a: oggdec: make sure the private parse data is cleaned up [04:44] <cone-942> ffmpeg.git 03Reinhard Tartler 07release/1.1:0135dd73bb23: update Changelog [04:44] <cone-942> ffmpeg.git 03Ronald Bultje 07release/1.1:ebd3aa429c38: vp3: Fix double free in vp3_decode_end() [04:44] <cone-942> ffmpeg.git 03Dale Curtis 07release/1.1:ca2e3f113188: matroska: Fix use after free [04:44] <cone-942> ffmpeg.git 03Vladimir Pantelic 07release/1.1:58baa367d696: vc1dec: prevent a crash due missing pred_flag parameter [04:44] <cone-942> ffmpeg.git 03Luca Barbato 07release/1.1:fbde7b2d0aeb: zmbv: Reset the decoder on keyframe errors [04:44] <cone-942> ffmpeg.git 03Marcin Juszkiewicz 07release/1.1:bc182a6acaf8: configure: enable pic for shared libs on AArch64 [04:44] <cone-942> ffmpeg.git 03Diego Biurrun 07release/1.1:fadebd256e91: APIchanges: Fill in missing hashes and dates; fix a version number typo. [04:44] <cone-942> ffmpeg.git 03Luca Barbato 07release/1.1:7a2ee770f520: lavc: check for overflow in init_get_bits [04:44] <cone-942> ffmpeg.git 03Luca Barbato 07release/1.1:71e00caeab89: lavc: introduce the convenience function init_get_bits8 [04:44] <cone-942> ffmpeg.git 03Luca Barbato 07release/1.1:e44d56b18d46: lavc: set the default rc_initial_buffer_occupancy [04:44] <cone-942> ffmpeg.git 03Luca Barbato 07release/1.1:72eca26bf95d: libx264: use the library specific default rc_initial_buffer_occupancy [04:44] <cone-942> ffmpeg.git 03Michael Niedermayer 07release/1.1:76477c3843c2: Merge remote-tracking branch 'qatar/release/9' into release/1.1 [04:44] <cone-942> ffmpeg.git 03Michael Niedermayer 07release/1.1:05ed9b700599: oggparsevorbis: fix vorbis_cleanup return type [05:56] <cone-942> ffmpeg.git 03James Almer 07master:b454c64e0311: latmenc: Check for LOAS sync word [12:26] <powerkill> Hi [12:29] <durandal_1707> hi [13:50] <cone-420> ffmpeg.git 03Justin Ruggles 07master:b805c725a31f: idcin: fix memleaks in idcin_read_packet() [13:50] <cone-420> ffmpeg.git 03Michael Niedermayer 07master:2dc9bcad7d57: Merge remote-tracking branch 'qatar/master' [15:08] Action: durandal_1707 request for >8 chan support [15:14] <michaelni> nicolas patches should make >8 chans closer to working [15:17] <cone-420> ffmpeg.git 03Michael Niedermayer 07master:c57c1329a995: mux: simplify chunking interleaver code [15:17] <cone-420> ffmpeg.git 03Michael Niedermayer 07master:69a96f9d4cf6: mux: de-obfuscate chunked interleaver checks [15:17] <cone-420> ffmpeg.git 03Michael Niedermayer 07master:46730e7a0c1a: mux/chunked interleaver: better align duration chunks. [15:30] <Compn> durandal_1707 : i didnt find any multi channel samples [15:55] <durandal_1707> michaelni, ubitux: anything done for channel reargament in lswr like in lavr? [16:01] <wm4> swr doesn't have that? [16:07] <durandal_1707> wm4: no it only have one using channel layouts which is limited [16:08] <wm4> durandal_1707: so this would be needed for arbitrary channel reordering for interleaved formats? [16:09] <durandal_1707> for any formats, it currently have only swr_set_channel_mapping for input [16:10] <durandal_1707> thought that could be used for outout too [16:14] <cone-420> ffmpeg.git 03Carl Eugen Hoyos 07master:12cb21efa611: Make the latm decoder less verbose on initialisation. [16:14] <durandal_1707> hmm, i think i could already use this to make lavd/oss usable [16:16] <ubitux> hey, 'got a few minutes to spare for you guys :) [16:17] <ubitux> durandal_1707, wm4: i won't work on channel mapping capabilities in swr again, but afaik the channel mapping is working just fine [16:17] <ubitux> what can't you do? [16:18] <ubitux> you can reorder/dup/silence channels as you wish [16:18] <durandal_1707> nothing i was unaware that swr had it all the time unlike avr [16:18] <ubitux> check how it's done with pan filter [16:18] <ubitux> try stuff like pan=5.1:c0=c3:c2=c1 etc [16:19] <ubitux> i think there are a few examples [16:19] <wm4> why are there multiple filters using swr, instead of making one filter that exposes all of swr? [16:19] <cehoyos> ubitux: http://thread.gmane.org/gmane.comp.video.ffmpeg.devel/157646 I am slightly against fixing three issues in one patch, but if the OP will not separate it and you don't have time... [16:19] <ubitux> 20:27:04 < Compn> have to check if we have vobsub muxer [16:20] <ubitux> 20:27:06 < Compn> i dont think so [16:20] <ubitux> 20:27:11 < Compn> demuxer only ? [16:20] <ubitux> yes, demuxer only [16:20] <ubitux> but you can mux "vobsub" into mkv for instance [16:20] <ubitux> not as idx/sub standalone though [16:20] <ubitux> cehoyos: ohai! perfect timing :) [16:21] <cehoyos> Salut [16:21] <Compn> yay its carl :) [16:21] <ubitux> cehoyos: can you guys wait until i come back? [16:21] <cehoyos> Sure, I just wanted to mention it since you wrote "got a few minutes" [16:21] <ubitux> asking for a split was a good idea, thanks [16:22] <ubitux> yeah, but not sure i can really review this in a few minutes [16:22] <cehoyos> np [16:24] <ubitux> wm4: because some filters do more advanced stuff than just options mapping with swr [16:24] <ubitux> ? [16:25] <wm4> ubitux: such as? [16:26] <ubitux> cat libavfilter/af_{amerge,aresample.pan}.c|less [16:26] <ubitux> cat libavfilter/af_{amerge,aresample,pan}.c|less [16:26] <mateo`> ubitux: \o/ [16:26] <ubitux> the syntax of the filter is a nice way to express complex usages [16:26] <ubitux> mateo` e [16:27] <wm4> most I see is option parsing [16:27] <ubitux> exactly [16:28] <wm4> in #libav-devel I claimed that the libavresample API design was backwards, because it puts so much different functionality behind a big API [16:28] <wm4> but that got me eventually kicked [16:28] <durandal_1707> what? can i have log? [16:28] <wm4> so I guess they know why it has to be that way [16:29] <wm4> though I can't see anything (except maybe that lavr has to copy the input for each filter invocation) [16:34] <wm4> I didn't know that libpostproc and libswscale were once the same thing (postproc/ sub directory in arch-mplayer) [16:34] <wm4> that's... scary [16:34] <cehoyos> [23:36] <cehoyos> durandal_1707 : You asked January 5th about commit 697b476c075a - it was about ticket #1747 [16:35] <durandal_1707> i dont remember what was yesterday .... [16:36] <j-b> wow, the amount of clulessness on this ticket is... [16:36] <cehoyos> The commit message says "ticket 1474" and you asked. [16:36] <cehoyos> j-b: Which ticket? [16:37] <durandal_1707> what is wrong with people that put wavpack/flac/dts in wav masked as it is pcm [16:37] <j-b> the Widevine [16:37] <cehoyos> Of course, but when closed the last drm ticket, I was flamed... [16:37] <durandal_1707> not just wav but aiff, au and others like shorten in nist [16:38] <cehoyos> (And apart from the fact that I had no clue myself, the xbmc devs sent him to trac) [16:38] <cehoyos> durandal_1707 : Samples? [16:39] <durandal_1707> cehoyos: hell no, i would rm that stuff if i found it immediatelly - was refering to dts in wav [16:39] <cehoyos> dts-in-wav is easy to find;-) [16:39] <cehoyos> But I don't remember flac-in-wav [16:42] <durandal_1707> wm4: if you have time look at ffmpeg git log from start..... [16:45] <durandal_1707> j-b: are you aware that ffmpeg support certain encrypted adx files [16:50] <j-b> durandal_1707: yes, and? [16:51] <j-b> durandal_1707: also OMA [17:01] <durandal_1707> omg what a brain damage, oss lavd have code to flip stereo [17:02] <durandal_1707> it is crappy hack for broken oss implementation [17:23] <wm4> omg bugs in lavd [17:24] <durandal_1707> wm4: what bug? [17:25] <wm4> commenting on what you just said (ok not a bug just a crackpot hack) [17:27] <shahriman> found a chrome bug [17:27] <shahriman> Daemon404: you here? [17:28] <durandal_1707> huh, did BBB forgot to reconfigure email client? [17:30] <durandal_1707> wm4: another nonsense, lavd oss code does pointless memcpy [17:33] <cone-420> ffmpeg.git 03Carl Eugen Hoyos 07master:ad406f7e4029: Only skip MLP header in mpeg files if the codec actually is MLP. [17:44] <durandal_1707> i will just reply with lgtm [17:46] <cone-420> ffmpeg.git 03Carl Eugen Hoyos 07release/1.0:c287efcc4479: Only skip MLP header in mpeg files if the codec actually is MLP. [17:46] <cone-420> ffmpeg.git 03Carl Eugen Hoyos 07release/1.1:ed12d1ecadb6: Fix compilation with --disable-everything. (cherry picked from commit f023003ce610a8fd6377cf4a8e98002ac3117ef4) [17:46] <cone-420> ffmpeg.git 03Carl Eugen Hoyos 07release/1.1:a95306e2d752: Only skip MLP header in mpeg files if the codec actually is MLP. [18:34] <durandal_1707> michaelni: i call swr_convert once before loop with output samples 0 [18:34] <durandal_1707> and in loop with output samples to some fixed number [18:34] <durandal_1707> and output i get is messed up [18:39] <durandal_1707> ignore, it was error in code [18:42] <durandal_1707> why is asyncts disabled? [18:42] <durandal_1707> ah, i have libavr disabled [18:45] <durandal_1707> so what it can do that others cant? [19:00] <durandal_1707> cehoyos: please do not add me to Reviewed-by for such trivial changes, thanks [19:02] <fbernon116> hi group [19:02] <cone-420> ffmpeg.git 03Paul B Mahol 07master:8fdd24455e97: .gitignore: add /doc/examples/pc-uninstalled [19:02] <fbernon116> I got a question about libavcodec. Can I post on this forum, or there is another one? [19:21] <durandal_1707> anyone need audio fade filter? [19:36] <cehoyos> durandal_1707: So why did you review it? [19:37] <``durandal_1707> cehoyos: to make you happy [19:40] <cehoyos> durandal_1707 : Ticket #1315 [19:49] <``durandal_1707> lol, i could add different kind of curves [19:51] <``durandal_1707> i have sample that reports referenced qt chapter not found [19:51] <cehoyos> use_absolute_path ? (just guessing) [19:52] <``durandal_1707> some folk reported it on #ffmpeg and did not created bug report when i tell him to do it [19:53] <``durandal_1707> cehoyos: you mean it is in another file? [19:54] <cehoyos> That was what I thought when I read "referenced qt chapter not found" but I may be completely wrong... [19:54] <``durandal_1707> it was created by Apple Keynote [19:58] <``durandal_1707> lol this dcc thing is broken [19:59] <cehoyos> It has never worked here... [20:00] <``durandal_1707> ok, opening bug... [20:00] <cehoyos> You could upload to incoming or datafilehost... [20:06] <``durandal_1707> why trac(a)avcodec.org? [20:06] <``durandal_1707> avcodec.org redirects to ffmpeg :) [20:11] <cehoyos> fcolor-diagnostics ? [20:17] <``durandal_1707> cehoyos: your patch returns 0 on error and 1 on non-error [20:17] <cehoyos> I don't think there is an error-case [20:17] <cehoyos> (But that may only be imo) [20:18] <``durandal_1707> yes, ignore me, i'm lost now [20:32] <``durandal_1707> can i use eval to calculate sin/cos/log and such ? [20:33] <``durandal_1707> eg, user gives it as argument to filter [20:33] <``durandal_1707> aeval already do this.... [20:34] <``durandal_1707> aeval_src_ [20:34] <``durandal_1707> which is far less usefull [20:35] <``durandal_1707> now all i need is to get started.... [20:40] <burek> is there a filter like -vf scale=MaxX:MaxY:someflag [20:40] <burek> which will make scale work with given X and Y as maxX and MaxY, keeping aspect ratio [20:40] <burek> just like 640:-1 [20:40] <burek> or -1:480 [20:40] <burek> depending on the aspect ratio of the input video [20:41] <cehoyos> just like 640:-1 [20:42] <cehoyos> or -1:480 [20:42] <burek> yes, but we don't know the aspect ratio of the input, so we can't just write 640:-1 or -1:480, but need to tell scale filter that 640:480 is max allowed for both X and Y [20:43] <burek> and it should decide if it will use 640:-1 or -1:480 [20:43] <cehoyos> Why not? [20:43] <burek> otherwise one should have to use ffprobe/ffmpeg to first determine which is greater (X or Y) [20:43] <cehoyos> Sorry, now I understand: If you search the -user mailing list, you will find a mail from Stefano (?) containing a sample command line [20:44] <burek> ok, i'll try, thanks :) [20:49] <cone-420> ffmpeg.git 03Carl Eugen Hoyos 07master:a1a707f7280a: Add a new function av_codec_get_tag2(). [20:49] <cone-420> ffmpeg.git 03Carl Eugen Hoyos 07master:9a038a95d276: Use new function av_codec_get_tag2() in ffmpeg.c. [20:52] <``durandal_1707> cehoyos: "new function" is redundant in commit log [20:52] <cehoyos> burek: I was thinking of this mail: http://thread.gmane.org/gmane.comp.video.ffmpeg.user/41141/focus=41199 but it seems the answer is incomplete [20:52] <cehoyos> I suspect such a formula could work though [20:57] <burek> cehoyos, i guess.. but just in case, i created a feature request [20:57] <burek> because in the source code it boils down to one simple if(width>height) [20:58] <cehoyos> burek: I fear your feature request is missing a command line etc... [21:10] <burek> btw, where are the docs for all the expressions used in filters? [21:10] <burek> like sin, min, max, etc? [21:10] Action: burek sets mode: +beer llogan [21:18] <llogan> man...i already have two growlers in the fridge, 5 gallons of mead in the fermenter, and another gallon of wine going too... [21:21] <burek> :) [21:24] <llogan> burek: http://ffmpeg.org/eval.html [21:25] <burek> oh great :) thanks :) [21:25] <llogan> i didn't actually look at it. maybe it's not what you were asking for. </lazy> [21:26] <burek> i was hoping there was an IF(expression,result_true,result_false) just like in mysql :) [21:26] <burek> but ok [21:27] <burek> i wanted to see if i could do: -vf scale=IF(iw>ih,'640:-1','-1:480') [21:27] <burek> :) [21:49] <beastd> burek: there is an if expression but it is harder to use because it returns 0 if not true. so last time i used addition to emulate if-else . [21:50] <beastd> burek: like this: -vf "geq=p(X\,Y):if(gt(Y\,H/2)\,128)+ifnot(gt(Y\,H/2)\,cr(X\,Y)):if(gt(Y\,H/2)\,128)+ifnot(gt(Y\,H/2)\,cb(X\,Y))" [21:50] Action: Daemon404 wipes blood from his eyes [21:50] <beastd> but maybe others have better ideas [21:50] <burek> ooh i see if(expr, if()) [21:51] <burek> wait, it's not that [21:51] <beastd> at least i do not know how to do it otherwise with current ffmpeg [21:51] <burek> let me first debug what it says [21:53] <burek> what does cr() do? [21:53] <burek> red chroma? [22:04] <beastd> burek: it switches chroma planes in one half of the picture and sets both to 128 in the other. luma plane is preserved. but that is just a stupid example. i just meant you could maybe use the "if(...)+ifnot(...)" to emulate if-else. [22:05] <burek> i see [22:06] <burek> but maybe if(a1, if(a2, a3)) might be a better example [22:07] <burek> i don't quite see how if+ifnot can create else... :S [22:07] <burek> oh, you mean like XOR? [22:07] <beastd> burek: dunno, but yours will return 0 if not a1. so was that what you meant to do? [22:08] <burek> if not a1, mine should return the result of the 2nd if [22:08] <beastd> burek: you tested it. i can't see why it should ATM [23:11] <``durandal_1707> can filter write to requested frame directly / no need to copy? [23:35] <``durandal_1707> hmm only if right perms are set, but this gonna be removed .... [23:44] <cone-420> ffmpeg.git 03Michael Niedermayer 07master:e1d7d4bd13cd: mpegvideo: reset context state on failed thread update. [00:00] --- Fri Jan 18 2013

1 0

ffmpeg.log.20130117
by burek 18 Jan '13

18 Jan '13

[00:00] <burek> grab it to an image file [00:00] <burek> and use what-the-font [00:00] <burek> :) [00:00] <Zeeflo> yea [00:00] <Zeeflo> excactly [00:16] <Zeeflo> What the hell! [00:16] <Zeeflo> Its not creating a stream with the subtitle! [00:17] <Zeeflo> -vf "scale=1280:720,subtitles=wolfman.srt" [00:18] <Zeeflo> [Parsed_subtitles_1 @ 0x2750800] Neither PlayResX nor PlayResY defined. Assuming 384x288 [00:18] <Zeeflo> but its not in the output file! [00:19] <saste> Zeeflo, you on Mac? [00:19] <Zeeflo> no [00:19] <Zeeflo> its being encoded on a debian headless server [00:20] <Zeeflo> with bureks latest build [00:20] <Zeeflo> i think maybe, i need to install some fonts.. [00:20] <Zeeflo> i cant think of anything else [00:20] <saste> Zeeflo, dunno but i used the subtitles filter before and it works fine [00:20] <Zeeflo> perhaps the ms-fonts [00:20] <saste> may be a libass issue [00:21] <Zeeflo> could it not be because of missing fonts? [00:21] <Zeeflo> but then it should give me a warning shouldnt it? [00:21] <saste> Zeeflo, yes this is what i suppose [00:21] <Zeeflo> ill try and install ms fonts [00:21] <burek> [Parsed_subtitles_1 @ 0x2750800] Neither PlayResX nor PlayResY defined. Assuming 384x288 <- that's just a warning, nothing more [00:21] <saste> but with libass/fontconfig you never knows [00:21] <burek> did you use -vcodec copy [00:22] <Zeeflo> nope [00:22] <Zeeflo> ill pastebin the whole thing [00:22] <Zeeflo> http://pastebin.ca/2303615 [00:22] <Zeeflo> but I do suspect its because its missing system fonts [00:23] <burek> could you provide complete output [00:23] <Zeeflo> sure [00:24] <saste> Zeeflo, loglevel debug may help [00:25] <Zeeflo> yea [00:26] <Zeeflo> but i do have absolutely no fonts installed.. [00:26] <Zeeflo> the server is brand new [00:26] <Zeeflo> http://pastebin.ca/2303620 [00:26] <Zeeflo> ms fonts is the msttcorefonts right? [00:27] <saste> yes fontconfig issue [00:27] <Zeeflo> and i assume its gonna want the arial font [00:27] <saste> i find a bit weird that libass doesn't complain hard and exits though [00:27] <Zeeflo> most likely [00:27] <Zeeflo> Yes [00:28] <Zeeflo> I would have gotten a warning like i mentioned earlier [00:28] <Zeeflo> but it did not complain [00:29] <Zeeflo> im installing the corefonts and gonna try again [00:44] <Zeeflo> Works flawlessly [00:44] <Zeeflo> Yeaha! [00:44] <Zeeflo> no need to convert to ass anymore! [00:49] <Zeeflo> I will say Adieu and be on my merry way again! [00:49] <Zeeflo> have a splendid evening! [02:51] <PMacDiggity> Hello, I'm having some trouble with libfaac, using this command: "ffmpeg -i $1 -codec:1 libx264 -q 16 -s 1920x1080 -vf yadif=0:0 -pix_fmt yuv420p -acodec libfaac -ab 192k $1_x264.mp4" I get this output: http://pastie.org/5700798 which, best as I can figure says the audio should be 192k like I asked, but when I listen to it (VLC/FFplay?QT), or extract the stream with QuickTime it says it's 33kbps, and sounds like it, any ideas? [02:52] <klaxa> i think it's generally advised to use fdk-aac for aac encoding [02:53] <PMacDiggity> ok, thanks [03:00] <llogan> PMacDiggity: what is the -q 16 for? [03:00] <llogan> if ffmpeg is interpreting that as -q:a 16, then it may be ignoring your -ab [03:01] <llogan> because -q:a and -b:a (or -ab) are mutually exclusive [03:02] <PMacDiggity> thanks, I'll check that as well [03:04] <llogan> and libx264 ignores -qscale if -q is being applied to the video. that is why -crf 23 (the default) is being applied to your video stream [03:09] <PMacDiggity> yup, that was it, thanks [03:14] <llogan> PMacDiggity: consider using -q:a (range is 0-250) instead of -b:a with libfaac. [03:14] <llogan> https://ffmpeg.org/trac/ffmpeg/wiki/AACEncodingGuide [03:15] <llogan> https://ffmpeg.org/trac/ffmpeg/wiki/x264EncodingGuide [03:51] <Fedcomp> hi. Can i ask? [03:51] <klaxa> always [03:52] <Fedcomp> how to connect two ffmpeg? [03:52] <Fedcomp> i mean i want to send some file from pc 1 over network to pc 2 [03:52] <klaxa> piping over ssh maybe? [03:52] <Fedcomp> pc 2 will receive it and encode. It looks stupid but it's for livestreaming [03:52] <klaxa> or using udp directly [03:52] <Fedcomp> not fits for livestream [03:53] <Fedcomp> tell me please how to do it over udp [03:53] <Fedcomp> i tryed some ways but i dunno how exactly to do it. [03:53] <Fedcomp> ffmpeg doesn't looks like want to listen port [03:54] <Fedcomp> ffmpeg -re -i "output.flv" udp://192.168.0.103:1924 [03:54] <Fedcomp> pc1 [03:54] <Fedcomp> ffmpeg -i udp://127.0.0.1:1234 test.flv [03:54] <Fedcomp> not working [03:54] <Fedcomp> oops, udp://192.168.0.103:1924 both [03:54] <Fedcomp> not working. [03:55] <klaxa> 192.168.0.103 is the first computer? [03:55] <Fedcomp> second [03:55] <Fedcomp> first one is 192.168.0.100 [03:55] <klaxa> then try that ip instead [03:55] <klaxa> you stream to your local network point, then kinda grab it from there with the second ffmpeg instance [03:56] <klaxa> it's like when you access a website, google doesn't push the data to you, you pull it from them [03:56] <Fedcomp> okay i'll try now. [04:06] <Fedcomp> on local machine sending and receiving working just fine. On linux machine old ffmpeg 2010 year [04:06] <Fedcomp> is there any possible problems with ffmpeg 2010? [04:06] <Fedcomp> built on Jan 29 2012 23:56:18 with gcc 4.1.2 20080704 (Red Hat 4.1.2-51) [04:07] <klaxa> can you get a newer one working? compiling from source would be the best option for most optimized results [04:08] <Fedcomp> i tryed to compile ffmpeg once, there's a lot of dependecies. [04:08] <Fedcomp> i'll try probably [07:30] <zabidin> hi [07:30] <zabidin> can ask about ffmpeg here? [07:31] <zap0> yes [07:35] <zabidin> i'm compiling from source for sles11 sp1. using this command: ./configure --enable-gpl --enable-libmp3lame --enable-libvorbis --disable-yasm [07:36] <zabidin> i get this error: http://pastebin.com/WpEfQsYZ [07:36] <zabidin> i have install libvorbis but still get those error. [07:53] <beware> can i use ffmpeg to remux an avi without anu junk chunks etc to make it more compact? [07:53] <beware> any* [09:42] <_jm> hello, ffprobe shows an error on some media file (contains h264 stream), can someone help with the meaning of the error ? [09:44] <_jm> the exact message is: missing picture in access unit with size XX [09:46] <_jm> this message appears several times and the sizes is equal to my SPS and PPS NALU [09:49] <_jm> full output is here: http://pastebin.com/sQ7xiG4U [09:50] <_jm> the stream begins with SPS, PPS and an IDR but i still see some garbage when i play the file [10:27] <TeknoJuce> is there a way to get ffserver to capture and stream with less than 1ms of delay on a smaller clocked device like a 500Mhz processor, I can get this from MPEG_Streamer but not from ffmpeg is MPEG_Streamer based on anything to do with ffmpeg? [10:44] <b_jonas> I would like to encode only a particular set of frames from a video, given by a list of timestampts (or list of frame numbers). How do I do that? [11:18] <klaxa> b_jonas: look at -ss and -t [11:23] <b_jonas> klaxa: the problem is, I don't want an interval of consecutive frames [11:23] <klaxa> ah um... hmmm... [11:24] <klaxa> i assume those frames you want are not in equal time-dinstances? if they are, see http://ffmpeg.org/trac/ffmpeg/wiki/Create%20a%20thumbnail%20image%20every%2… [11:24] <b_jonas> indeed [11:25] <b_jonas> if they're equal time distances, I could just use a framerate [11:25] <klaxa> true [11:25] <klaxa> the inefficient and naive process would be to go through the video n times for n frames and always extract the desired frame [11:25] <klaxa> but that would be... well inefficient and naive [11:26] <b_jonas> that would be slow, yes [11:26] <b_jonas> another inefficient process would be to decode all frames then delete the ones I don't need [11:27] <Mavrik> mornin [11:31] <b_jonas> hmm, there's a "select" filter that can keep and drop frames, though I'd have to pass a complicated expression to it to compare with a list [11:37] <powerkill> hello [11:39] <powerkill> is there a ffmpeg guru present ? [11:42] <simonec77> Hi , I'm triyng to compile ffmpeg libraries (I have download c99wrappr.exe ecc.. from the network, I do not build from myself) with msvc toolchain, whit enable shared and DEBUG, but I got an error, thi is my "config.log" file is:http://pastebin.com/8NV2U8EM [11:43] <simonec77> any idea? [11:57] <viric> simonec77: '-m' are not parameters for the msvc toolchain [11:57] <viric> ah [11:57] <viric> well, I don't know c99wrappr at all. [11:58] <powerkil_> viric do you code on ffmpeg ? [11:58] <viric> not at all. [11:59] <viric> Take me as a person you don't know that you have simply met through irc [11:59] <powerkil_> :) [12:01] <powerkil_> anyone coding in ffmpeg ? can help me ? I have some depreacted function in my patch that need update [12:16] <cbsrobot> powerkil_: join #ffmpeg-devel for developpement [12:16] Last message repeated 1 time(s). [12:16] <cbsrobot> or at least give a hint what goes wrong [12:17] <cbsrobot> powerkill ^ [12:26] <powerkill> j #ffmpeg-devel [12:28] <powerkill> url_get_filename is deprecated but I don't know what to use instead [12:28] <powerkill> I'm trying to port a patch to the last ffmpeg version [12:34] <durandal_1707> powerkill: use URLContext.filename ? [12:37] <powerkill> so instead of url_get_filename(post_pb, mov->filename, sizeof(mov->filename)); ? [12:38] <durandal_1707> url_get_filename is gone in latest version [12:39] <powerkill> Yes I see I have log of deprecated http://pastebin.com/eskr629y [12:39] <powerkill> *lot [12:39] <durandal_1707> what version you are using? [12:39] <powerkill> 1.1 [12:39] <powerkill> I was using 0.7 and want to use 1.1 now [12:40] <durandal_1707> the 1.1 do not have it at all [12:40] <powerkill> So I have to update all the function and it will work right ? [12:41] <powerkill> I already update url_is_streamed to seekable [12:42] <powerkill> but url_get_filename url_fileno url_connect put_flush_packet av_metadata_get ... are function that are not available anymore [12:43] <durandal_1707> can we know for what program you are working on? [12:43] <powerkill> for outputing smoothstreaming format [12:45] <powerkill> urlcontext.filename is not a pointer anymore [12:46] <dott_infernus> hi guys, is there any way to use ffmpeg function in c++ [12:46] <dott_infernus> ?? [12:47] <durandal_1707> powerkil_: there is char * filename in URLContext [12:48] <gxk> #videolan [12:48] <gxk> sorry [12:49] <powerkil_> durandal_1707: can you help me with the error if I send you the code ? [12:51] <durandal_1707> powerkil_: what error? [12:52] <powerkil_> http://pastebin.com/H6vRY95W here are all the error [12:54] <durandal_1707> from where you got ism.c code? [12:54] <powerkil_> I convert url_is_streamed(s->pb) with !(s->pb->seekable) [12:56] <powerkil_> from http://smoothstreaming.code-shop.com/ [12:58] <durandal_1707> so you paid for it? [12:59] <powerkil_> there is a commercial and a open licence [12:59] <durandal_1707> code can be downloaded? [12:59] <durandal_1707> for free? [13:00] <powerkil_> yes http://www.unified-streaming.com/support/documentation/ffmpeg-addon/ [13:00] <powerkil_> just have to register on their site [13:03] <gxk> need recommendation for arm able to encode 1080p30 with ffmpeg? [13:13] <simonec77> sorry: what is the correct command line to complile ffmpeg with toolchain msvc and debug support for msvc? [13:22] <durandal_1707> simonec77: it is in documentation available on ffmpeg.org [13:44] <b_jonas> simonec77: http://ffmpeg.org/platform.html#Microsoft-Visual-C_002b_002b [14:43] <Eric___> I have one question on HLS standard using ffmpeg. Can we force client to always look for the highest bit rate defined in the manifest file? instead of looking for the lowest bit rate in the manifest file? [14:53] Last message repeated 1 time(s). [14:58] <simonec77> durandal_1707: I have followed the documentation at http://ffmpeg.org/platform.html#Microsoft-Visual-C_002b_002b but in this nothing is related to building ffmpeg in toolchian MSVC with the MSVC debug support... to be alble to debug whit MSVC needs: *.lib, *.dll, headers, *.pdb and *.ilk , I'm alble to compile obtaining *.lib, *.dll, headers bu NOT the *.pdb and *.ilk necessary to be able to debuf FFMPEG libraries in [15:03] <JEEB> pdb is only needed when the symbols are wanted to be outside of the binary itself AFAIK. I would guess an --enable-debug library of ffmpeg would work. Not sure how many symbols the standard configured library has [15:11] <simonec77> this is my command: ./configure --toolchain=msvc --enable-shared --extra-cflags="-Z7" --extra-ldflags="-DEBUG" [15:11] <simonec77> it run fine, then I do make [15:11] <simonec77> and run fine [15:12] <simonec77> but no *.pdb and no *.ilk in the libraries foledr... [15:12] <simonec77> these are necessary to be able to debug FFMPEG libraries in MSCV envoronment... [15:18] <JEEB> I disagree with the ilk file, and pdbs just contain the debug symbols in a separate file instead of the binary itself [15:18] <JEEB> also ilk files are only done during debug compilation [15:18] <JEEB> --enable-debug [15:18] <JEEB> try with that [16:55] <Pinhole> I have an ts stream with 6 mpeg4 programs in it. I want to convert it to mpeg2 ts with all 6 programs. I'm not finding any way to do that. Any hints? [18:59] <Pinhole> how do you transcode a mpts with mp4 programs into a mpts with mpeg2 programs? [19:40] <octavio_> Hi i have a quiestion, when i run ./configure i got a error message and the last line of the config.log is "./configure: /tmp/ffconf.VhP28398.sh: /bin/sh: bad interpreter: Permission denied" please help me [19:41] <octavio_> i have run .configure as root, and the /bin/sh exists [19:46] <beastd> octavio_: i can't think of any cases where you should run ./configure as root. so please don't do it. [19:47] <octavio_> beastd: thank you i will not [19:48] <beastd> octavio_: i never had this error myself but it sounds like configure has probelems executing the e.g. generated shell script [19:48] <beastd> maybe it is a problem with your /tmp. you could create a directory in your home and try to use that [19:49] <octavio_> yes, i have set the enviroment variable TMPDIR in my home dir and this fix the problem, thank you! [19:49] <octavio_> :) [19:49] <beastd> octavio_: good! you solved your problem already. [19:50] <beastd> octavio_: but i guess it is better you use some dedicated temp dir like ~/fftemp or so. it is easier to cleanup if temporary files stay for some reason [19:52] <octavio_> yeah i see that, ok i will try, thank you so much [19:56] <AL13N> burek: llogan: the thing is, that with -s 640x480, the resulting video file is 640x422 which is correct. but, the image preview is 640x480, breaking the aspect ratio [19:56] <AL13N> burek: llogan: and, it's only webm with this issue, mp4 and ogg doesn't have this [19:57] <AL13N> if i do 640:-1 it will not be the expected result if the source file was rotated and thus had 1080x1920 [20:17] <burek> AL13N, what is your goal? [20:17] <burek> to scale down the video [20:17] <burek> or to stretch it to fit 640x480 [20:18] <AL13N> burek: my goal was to convert any video to within 640x480 [20:18] <AL13N> keeping aspect ratio [20:19] <AL13N> -s did that perfectl [20:19] <burek> you cannot do that [20:19] <AL13N> i did mp4, ogg and webm [20:19] <AL13N> burek: ? what do you mean, i cannot do that? [20:19] <burek> you can either keep the aspect ratio or to stretch it to 640x480 [20:19] <burek> you can't do both [20:19] <AL13N> ah [20:19] <AL13N> not stretch [20:19] <AL13N> to within 640x480 [20:19] <burek> that's called stretching [20:19] <AL13N> ie: those are the maximum resolutions [20:20] <burek> oh [20:20] <burek> then use 640:-1 [20:20] <burek> or -1:480 [20:20] <burek> depending on which ordinate you want to fix [20:20] <AL13N> 640x200 is ok and 200x480 is also ok [20:20] <AL13N> burek: any [20:21] <burek> well choose one [20:21] <burek> and then choose the right syntax according to that [20:21] <AL13N> burek: anyway, i used ffmpeg -i file.MOV -s 640x480 -r 12 file.{MP4|OGG|WEBM} [20:21] <AL13N> it worked [20:21] <AL13N> the video keeps aspect ratio and it respects the max resolutions [20:22] <burek> AL13N, it doesn't work that way [20:22] <burek> i don't know what's the meaning of your "it worked" [20:22] <burek> but it's not implemented to work that way [20:22] <AL13N> burek: the documentation said so [20:22] <burek> so the correct way to say it is: it shouldn't have worked the way you used it [20:22] <burek> can you give me a link [20:22] <AL13N> i'll relook [20:23] <burek> if you specify both X and Y in -s [20:23] <burek> it stretches video [20:23] <burek> to satisfy both of your ordinates [20:23] <burek> if you put -1 for any of those, it will auto-calculate it, to keep the aspect ratio [20:23] <AL13N> in any case, it did work, but when i open up dolphin and see the 3 videos, 2 preview images are correct, and the 3rd one (webm) displayed the preview in 640x480 aspect ratio, even though the movie was 640x422 [20:23] <burek> as i said [20:24] <burek> it didnt work [20:24] <AL13N> burek: but the problem is that i don't know [20:24] <burek> period. [20:24] <burek> what worked is probably a bug [20:24] <burek> and you shouldn't rely on it [20:24] <AL13N> i don't know if it's 1920x1080 or 1080x1920 [20:24] <AL13N> so i can't just put -1 on height [20:24] <burek> can you give me a link to the docs that you've read [20:24] <burek> so i can check what is says [20:25] <burek> it* [20:32] <AL13N> burek: i have a hard time locating it [20:32] <AL13N> burek: ok, so maybe it's a bug, so what do you propose i use instead? [20:34] <burek> as suggested, use -1 [20:35] <AL13N> burek: but as i said, if i use 640:-1 and it's 1080x1920, i'll have a height that exceeds 480 by far [20:35] <burek> then use -1:480 [20:36] <AL13N> burek: but, i don't know in advance the resolution of the source [20:38] <AL13N> at least the -s work; perhaps one could just transform the bug into a feature and document it... [20:40] <AL13N> burek: anyway, can you tell me if webm has an included preview image (or any other container) and how to find out if it is? [20:41] <burek> what is a preview image? [20:43] <AL13N> well, at least some image formats have an included small image (sort of a thumbnail) so that the image doesn't need to get rendered to get a thumnail. i wonder if any video formats have this feature too [20:44] <AL13N> i donno if i'm explaining myself well here [20:45] <AL13N> maybe this more related to the libraries who do the actual video codecs... [20:45] <klaxa> like metadata thumbnails? [20:45] <burek> i think i understand, but i dont think any of those do support such thing [20:45] <AL13N> klaxa: yes, that sounds like what i mean [20:45] <burek> it is the same anyway, you will have to extract either that thumb or 1 frame from video stream [20:45] <burek> so basically in both cases you'll have to extract it [20:46] <AL13N> oh yes, i'm not really asking about that [20:47] <AL13N> this is more for non-video related programs, that they don't need a video decoder to get a preview/thumbnail [20:47] <AL13N> i suspected some containers (or video codecs themselves) would be able to do this [20:47] <klaxa> well but they will need a demuxer [20:47] <burek> i see.. well try google.. i didnt hear of any so far [20:47] <AL13N> klaxa: true, maybe [20:48] <klaxa> you might be able to add it to matroska as an attachment [20:48] <AL13N> well, i was thinking that maybe HTML5 video tags (from browsers) would be able to show the image while the file is caching/loading [20:48] <AL13N> if such a thing exists [20:49] <AL13N> but, if it isn't much heared of here, it'll not really be true [20:49] <klaxa> doesn't look like attachments are supported in webm [20:49] <AL13N> anyway: burek, klaxa about that earlier issue, do you know of a solution how i can do what i want for scaling? [20:50] <klaxa> script it :V [20:50] <klaxa> and do calculation automatically [20:50] <AL13N> klaxa: how? i'd need to know the resolutions beforehand? [20:50] <AL13N> or not? [20:51] <klaxa> well yeah, if you don't know the orientation beforehand, you can't really predict the resolution you want [20:51] <AL13N> klaxa: yes... [20:51] <AL13N> klaxa: so, how would you go about it? [20:52] <klaxa> maybe ffprobe and parse with grep? actually i think you can use variables within one ffmpeg command [20:52] <klaxa> referencing input resolution [20:52] <AL13N> klaxa: it's intended for webview and i have a place of maximum 640x480 to put it [20:52] <AL13N> klaxa: that sounds interesting [20:52] <AL13N> klaxa: got any search terms or links for that kind of thing? [20:52] <klaxa> currently looking at the documentation for that [20:53] <klaxa> more like looking for it [20:53] <AL13N> awesome [20:54] <klaxa> maybe this? http://ffmpeg.org/ffmpeg-scaler.html [20:54] <klaxa> no idea how to use that though :S [20:56] <klaxa> uh... probably not... [20:58] <AL13N> unless i could script with if() and max() and stuff, it doesn't seem likely [20:59] <burek> AL13N, if you compiled your ffmpeg [20:59] <burek> you can patch it a little bit [20:59] <burek> add a single if(width>height) in scale filter [20:59] <burek> which will determine if -1:480 or 640:-1 should be used [21:00] <burek> :) [21:00] <burek> or use ffprobe first to determine which dimension is greater [21:00] <burek> and based on that use one of those 2 for scale [21:01] <burek> I've created a feature request, but I can't promise it will be implemented any time soon [21:01] <burek> https://ffmpeg.org/trac/ffmpeg/ticket/2150 [21:01] <AL13N> actually, i just read that complex filters can use some functions [21:01] <burek> so, i dunno [21:01] <burek> if ffmpeg supports some kind of conditional functions/expressions, like IF() [21:02] <burek> you might do something like: -vf scale=IF(width>height,640:-1,-1:480) [21:02] <burek> :) [21:02] <burek> -might +could [21:02] <burek> but i think it doesnt [21:03] <burek> hmh [21:03] <burek> try this [21:03] <burek> scale=640:-1,scale=-1:480 [21:05] <AL13N> after each other? [21:05] <burek> yes [21:05] <AL13N> won't work [21:06] <AL13N> the second one will upscale again if the video would've been 640x422 [21:06] <burek> try it [21:07] <AL13N> as i thought: 853x480 [21:08] <burek> ok, so you need OR and not AND [21:08] <burek> well, it is a problem :) [22:06] <viric> hm it looks like 'ffmpeg' completely decodes the video stream before the start point, if given "-ss". [22:08] <viric> Can anything make ffmpeg seek? The input is mkv [22:09] <klaxa> i'm not sure, i know i knew something at some point, i'll go through my logs [22:09] <viric> thank you! [22:09] <llogan> placement of -ss matters. [22:09] <viric> ahh [22:09] <klaxa> ah yeah that was it [22:09] <klaxa> no need to go through my logs :) [22:09] <viric> before -i! [22:09] <burek> http://ffmpeg.org/trac/ffmpeg/wiki/Seeking%20with%20FFmpeg [22:11] <viric> amazing [22:11] <viric> thank you! [22:13] <burek> :beer: :) [22:13] <viric> that wiki helps a lot [22:13] <viric> (not only for that page. For many others too :) [22:17] <llogan> a VP8/9/Vorbis/webm guide is probably next in line...not that i use those formats (excluding vorbis) [22:18] <pzich> would fast seek be a reason for audio seeming to be out of sync (and faster than the video), or should I be looking for another issue? [22:21] <burek> viric thanks, we do are best, with the free time that we all have [22:21] <burek> -are +our [22:21] <viric> llogan: yes, that'd be great [22:25] <viric> What can this be? [swscaler @ 0x28088e0] Warning: data is not aligned! This can lead to a speedloss [22:26] <juanmabc> viric: old ffmpeg [22:26] <viric> mh it's 1.0 [22:26] <juanmabc> mmm, then shouldn't [22:26] <viric> I'll get 1.1 [22:28] <viric> I'm using yadif,crop as filters [22:29] <viric> ffmpeg1.1 says the same, but the [swscaler..] appears in blue instead of white :D [22:30] <juanmabc> haha [22:30] <beastd> viric: would be a good idea to file a bug report (including full uncut output of a sample commandline triggering the message) [22:31] <viric> Ok [22:31] <viric> a trac ticket? [22:32] <beastd> viric: yes [22:34] <viric> http://ffmpeg.org/trac/ffmpeg/ticket/2151 [22:36] <beastd> viric: normally you submit the ffmpeg output with higher log level. look here for instructions: http://ffmpeg.org/bugreports.html [22:36] <viric> oh [22:37] <beastd> anyway thanks for the report [22:37] <viric> I can do it [22:38] <beastd> That would be nice. [22:39] <viric> done [22:40] <viric> any idea? it seems it's the crop that does that [22:40] <``durandal_1707> what is problem? [22:41] <viric> it tells me about "data is not aligned! This can lead to a speedloss" [22:41] <``durandal_1707> and ? [22:41] <viric> juanmabc said it's not normal to have that on a recent ffmpeg [22:41] <viric> and beastd asked for a ticket about it [22:46] <``durandal_1707> well you are croping to weird dimension, so it is not bug [22:47] <viric> ok [22:47] <viric> I thought os. [22:47] <viric> so [22:47] <viric> I tried to use multiples of 16, but it didn't change anything [22:47] <viric> what is the alignment about? [22:48] <``durandal_1707> multiplies of 16 for what? [22:48] <viric> the crop numbers [22:49] <viric> (question apart... anyone knows how can I delay audio vs video?) [23:31] <llogan> viric: is it synched correctly if you leave out the yadif filter? [23:32] <viric> audio-video? [23:32] <llogan> yes. that is what you were asking about I assume. [23:32] <viric> well, it's not that ffmpeg desyncs anything [23:32] <viric> I've an original desynched [23:32] <llogan> oh [23:32] <viric> and I wanted to more or less correct it [00:00] --- Fri Jan 18 2013

1 0

ffmpeg-devel.log.20130116
by burek 17 Jan '13

17 Jan '13

[02:01] <cone-169> ffmpeg.git 03Ronald S. Bultje 07master:2c85d7c01548: h264: add 3 pixels below for subpixel filter wait position. [02:04] <Daemon404> michaelni, there look to be some small differences like [02:04] <Daemon404> - 9.463782453291253760e+17, 1.052965658439974912e+18, 1.162323703413866496e+18, 1.274318858307502080e+18 [02:04] <Daemon404> + 9.463782453291253760e+17, 1.052965727159451648e+18, 1.162323840852819968e+18, 1.274318858307502080e+18 [02:04] <Daemon404> "small" [02:04] <Daemon404> mind you, these are present whenever we use libm.h impls at runtime as well [02:05] <Daemon404> only for the cbrtf replacements [02:09] Action: Daemon404 sends a proper reply [02:16] <kierank> i wonder if the intel aac encoder is any good [02:16] <Daemon404> is it sw or hw? [02:16] <kierank> dunno, installing now [02:16] <Daemon404> lul [02:17] <Daemon404> for now, i fairly satisifed with fdk [02:17] <Daemon404> though i wish it had a proper vbr mode [02:17] <Daemon404> well, 'quality' mode. [02:33] <llogan> kierank: so...how was it? [04:30] <joshack> ffmpeg noob here. Looks like I got most things installed well excpet mencoder/mplayer [04:30] <joshack> and apparently I need that working for stitching teh avi files ffmpeg whipped up for me [04:31] <joshack> after ./configure and make && make install .. some of the output tells me the place it's going to be installed [04:31] <joshack> Install prefix: /usr/local [04:31] <joshack> Data directory: /usr/local/share/mplayer [04:31] <joshack> Config direct.: /usr/local/etc/mplayer [04:31] <joshack> yeah, there's nothing there and then v [04:31] <joshack> -bash: mencoder: command not found [04:32] <joshack> anyone want to point a loser in the right directoin [04:33] <highgod> Hi I want to ask a question,does mpeg2video has 444 or 422 format?thanks. [04:58] <Compn> highgod : i could not get anyone to review your patch :( [05:06] <highgod> Compn,I'm fixing the patch now,hehe. [05:16] <highgod> Compn:when I finish the the patch,can I submit again? [05:16] <michaelni> highgod, mpeg2 allows 422 and 444 [05:18] <highgod> @michaelni,I want to check whether the mpeg2 is 420 at the codec init stage.I how can I get the format? I use av_parser_parse2, but I didn't get it [05:20] <cone-169> ffmpeg.git 03Michael Niedermayer 07master:31c4a1b7d0a0: h264: do not mess up cur_chroma_format_idc during thread update [05:24] <michaelni> highgod, the info is stored in the ext sequence header or whatever its called [05:25] <michaelni> might be tricky to access this reliably from codec init [05:26] <michaelni> with parser split it might end in extradata but not sure how reliable this is [05:28] <Compn> michaelni : i think highgod needs to know the 444/422 for passing to hwaccel ... [05:28] <Compn> michaelni : btw highgod is author of the dxva2 patch , need someone to review it :) [05:28] <Daemon404> does 4:4:4 even exist in practice? [05:28] <Compn> highgod : yes, submit new patch when you are finished [05:30] <Compn> Daemon404 : dunno [05:30] <Compn> its in the specs and there seems to be a few encoders for it [05:57] <bcoudurier> why do I know iso number of mpeg-2 by heart ? :) [05:58] <bcoudurier> actually you cannot encode 444 because it's not allowed in any profile [05:59] <bcoudurier> even high is 420 or 422 [06:07] <cone-169> ffmpeg.git 03Michael Niedermayer 07master:2d372d3a3ffb: h264: document h264_set_parameter_from_sps() re-calling behavior [06:07] <cone-169> ffmpeg.git 03Michael Niedermayer 07master:5c6283e5c3bf: mpegvideo: Increase MAX_MV for HD video [06:07] <cone-169> ffmpeg.git 03Michael Niedermayer 07master:bbe56bcd6bef: motion_est: Limit motion vector search range to MAX_MV [06:08] <michaelni> bcoudurier, btw, ill commit the chunk patches tomorrow or so with "[PATCH 2/2] mux/chunked interleaver: better align duration chunks." [06:09] <michaelni> that should mostly solve the issues [06:25] <highgod> I want to check whether the mpeg2video is 420,if it is not,I will stop the hardware decoder process. [08:16] <highgod> I find a function try_decode_frame, can I write a function like it to get the mpeg2 chroma info 444 or 422 or 420 [08:52] <highgod> Can I add a function like mpeg_decode_sequence_extension to get the mpeg2 chroma format at init stage in my added file?thanks [09:25] <j-b> 'morning [11:28] <cone-942> ffmpeg.git 03Stephan Hilb 07master:36810215fa25: lavd/v4l2: improve debug message [11:28] <cone-942> ffmpeg.git 03Stephan Hilb 07master:f245a2086a10: lavd/v4l2: update broken link to v4l2 video capture example [11:49] <cone-942> ffmpeg.git 03Diego Biurrun 07master:0b22107d9504: rv34_parser: Adjust #if for disabling individual parsers [11:49] <cone-942> ffmpeg.git 03Justin Ruggles 07master:23098bbd5093: vf_fps: add final flushed frames to the dropped frame count [11:49] <cone-942> ffmpeg.git 03Diego Biurrun 07master:dae1d507af94: x86: Add PAVGB macro to abstract pavgb/pavgusb instruction via cpuflags [11:49] <cone-942> ffmpeg.git 03Michael Niedermayer 07master:68f92a70f1f4: Merge commit 'dae1d507af94261bafd3b11549884e5d1eca590e' [11:55] <durandal_1707> what is CONFIG_RESAMPLE_AUDIOPHILE_KIDDY_MODE ? [11:56] <durandal_1707> it appears it was removed [12:03] <michaelni> its a mode for thouse who can hear a needle drop on the moon from earth [12:05] <cone-942> ffmpeg.git 03Joakim Plate 07master:f924d52975ec: dvdsubdec: Support palette in mkv [12:05] <cone-942> ffmpeg.git 03Ronald S. Bultje 07master:fb845ffdd335: h264: add 3 pixels below for subpixel filter wait position [12:05] <cone-942> ffmpeg.git 03Michael Niedermayer 07master:06af724c56c9: Merge commit 'fb845ffdd335a1efd6dfd43e8adeb530397b348e' [12:10] <mateo`> durandal_1707: does the data size of an aiff chunk should be always even ? (even if the declared chunk size is odd) [12:12] <cone-942> ffmpeg.git 03Martin Storsjö 07master:3130fa51a5d6: lavu: Add a fate test for the HMAC API [12:12] <cone-942> ffmpeg.git 03Martin Storsjö 07master:c2603aa25b75: lavf: Add a fate test for the SRTP functions [12:12] <cone-942> ffmpeg.git 03Martin Storsjö 07master:0eecafc948b7: configure: Make the new srtp protocol depend on the rtp protocol [12:12] <cone-942> ffmpeg.git 03Michael Niedermayer 07master:9ea65c65f700: Merge commit '0eecafc948b74c247ebbc59f18f508db5d590d0b' [12:12] <mateo`> durandal_1707: in that case if (size & 1) file_size--; could be ok [12:16] <durandal_1707> michaelni: it is not exported from configure [12:16] <divVerent> durandal_1707: don't you see it? it's the placebo flag [12:16] <durandal_1707> mateo`: i do not remmember, find spec [12:17] <divVerent> sort of like using gold plugs for digital audio for better audio quality [12:17] <durandal_1707> than name it PLACEBO and not AUDOPHILE_KIDDIE [12:17] <divVerent> but then the kiddies would know [12:17] <cone-942> ffmpeg.git 03Martin Storsjö 07master:42364fcbcac9: srtp: Mark a few variables as uninitialized [12:17] <cone-942> ffmpeg.git 03Martin Storsjö 07master:977d4a3b8a2d: rtpdec_mpeg4: Check the return value from malloc [12:17] <cone-942> ffmpeg.git 03Michael Niedermayer 07master:1459f34251c3: Merge commit '977d4a3b8a2dbc2fb5e747c7072485016c9cdfaa' [12:17] <mateo`> durandal_1707: i didn't find any specification for the ID3 chunk :( [12:17] <divVerent> but seriously, it does something, but there is no way to set it on configure [12:18] <divVerent> it mainly does all resampling decisions in foating point instead of integer [12:18] <michaelni> divVerent, double precission floating point! [12:18] <divVerent> yes [12:18] <divVerent> change it to long double NOW!!!1 [12:18] <durandal_1707> mateo`: aiffenc put_meta use FFALIGN(size, 2) [12:18] <divVerent> ;) [12:19] <michaelni> divVerent, sent patch and lets make a news entry on the front page [12:19] <durandal_1707> how many bits it takes? [12:19] <divVerent> michaelni: haha [12:19] <divVerent> michaelni: extra points if for some reason, it doesn't even DO the caluclations as long double? ;) [12:19] <durandal_1707> at least ha folks will be happy [12:20] <divVerent> wonder if that can be snuck in somehow [12:20] <divVerent> i.e. an accidental conversion to double [12:21] <divVerent> haha [12:21] <divVerent> this "extra point" is already given [12:21] <divVerent> dst[dst_index] = av_clip_int16(lrintf(val)); [12:21] <durandal_1707> mateo`: lol, i wrote that aiff code [12:21] <divVerent> so we convert to float first [12:21] <divVerent> then to int16 [12:21] <divVerent> luckily float DOES have enough precision ;) [12:21] <durandal_1707> mateo`: i used sox/libsndfile as reference [12:22] <divVerent> michaelni: haha, that should do ;) [12:22] <divVerent> if we change the type to long double [12:22] <divVerent> the filter array is STILL built using double math [12:22] <divVerent> so chanigng this from double to long double seems to be a pure placebo [12:23] <divVerent> at least the resampling itself is done using long double math then [12:24] <cone-942> ffmpeg.git 03Nicolas George 07master:0e79fe37e5c5: lavd/v4l2: init return value. [12:25] <divVerent> michaelni: http://dpaste.com/878947/ don't commit this ;) [12:26] <divVerent> it's 100% placebo as the filter is still done using double math [12:26] <michaelni> haha :) [12:26] <michaelni> anyway resample now resides in libswresample [12:26] <divVerent> that too [12:26] <divVerent> makes it EVEN more pointless [12:27] <michaelni> and it has a double precission mode without ifdefs ;) [12:27] Action: divVerent just imagines coat hangers with golden 3.5mm audio plugs attached [12:27] <divVerent> you do remember this test where some guys ABX'd coat hangers vs expensive cables [12:27] <divVerent> and the coat hangers won [12:27] <michaelni> lol [12:28] <divVerent> http://gizmodo.com/363154/audiophile-deathmatch-monster-cables-vs-a-coat-ha… I think this is it [12:28] <michaelni> are we missing a AV_SAMPLE_FMT_LDBL :/ [12:28] <divVerent> no, it was a draw [12:28] <divVerent> well, the coat hangers DO have an advantage [12:28] <divVerent> they're quite thick, thus very low resistance [12:29] <divVerent> a long double sample format... unlikely we ever need this [12:29] <durandal_1707> michaelni: yes we are missing it because you never know how it can be used (not just for pure listening) [12:29] <pross-au> michaelni: no i want AV_SAMPLE_FMT_EVAL. [12:29] <divVerent> long double is nasty unaligned stuff [12:29] <divVerent> pross-au: you mean AV_SAMPLE_FMT_SYMBOLIC? [12:30] <divVerent> which uses symbolic math and a computer algebra system to calculate stuff? [12:30] <durandal_1707> and how would you encode/convert to id? [12:30] <durandal_1707> *it [12:30] <pross-au> that might please 'em [12:30] <divVerent> convert TO it? easy [12:30] <divVerent> convert FROM it? easy, once the values of all the variables are known ;) [12:30] <wm4> btw. does anyone know how important 24 bit formats are for audio APIs [12:30] <pross-au> .m matlab file [12:31] <divVerent> matlab is not symbolic math [12:31] <divVerent> so we still need a CAS [12:31] <wm4> because ffmpeg doesn't support that format [12:31] <mateo`> durandal_1707: for the metadata chunk, quoting the aiff spec "ckSize is the size of the data portion of the chunk, in this case the text. [12:31] <divVerent> a CAS that produces matlab formulas though [12:31] <divVerent> so once you ran ffmpeg, you can then use matlab to do the real processing [12:31] <divVerent> ;) [12:31] <divVerent> wm4: IIRC it's somewhat important, can't tell for which APIs [12:31] <divVerent> the main advantage is that it's "close" to 32bit float [12:32] <mateo`> durandal_1707: + http://muratnkonar.com/aiff/index.html mention that "chunkSize is the number of bytes in the chunk, not counting the 8 bytes used by ID and Size fields nor any possible pad byte needed to make the chunk an even size." [12:32] <durandal_1707> mateo`: yes, but next chunk may still be aligned [12:32] <divVerent> but sound APIs tend to rather use 32bit float [12:32] <mateo`> durandal_1707: so i guess FFALIGN should not be used to write the size of the chunk there [12:32] <wm4> that would require extra hacks to do the conversion past lavr's/swr's ass [12:32] <durandal_1707> mateo`: it is for compability [12:32] <mateo`> durandal_1707: howerver if (size & 1) avio_w8(pb 0) is still relevant [12:32] <divVerent> was mainly saying, if your input is 24bpp, this can be losslessly converted from/to float32 [12:32] <mateo`> durandal_1707: ok ... [12:33] <wm4> input is never 24 for the user, because lavc doesn't have that [12:34] <wm4> not sure what it does with 24 bit raw input [12:34] <divVerent> S32, I suppose [12:37] <saste> av_probe_input_buffer() always fails if max_probe_size < PROBE_BUF_MIN [12:38] <saste> so my question, is, does it make sense to be able to specify a probesize lesser than PROBE_BUF_MIN? [12:39] <divVerent> "yes", because PROBE_BUF_MIN is not defined/documented in the header [12:40] <divVerent> in other words, I wouldn't blame the caller for not knowing PROBE_BUF_MIN is 2048 [12:43] <saste> right now ffmpeg fails with invalid argument (correct), but with no hint to the user of what's going wrong [12:43] <saste> in each it may be senseful to raise the minimum value accepted by probesize [12:43] <saste> *in each case [12:44] <divVerent> maybe just document the fact that EINVAL may be returned when max_probe_size is "too small" to get a clear result? [12:45] <saste> divVerent, it is an application level problem, i'm going to add a message in the lib [12:46] <divVerent> you mean in the header file? [12:46] <divVerent> or an av_log one [12:51] <saste> divVerent, av_log [12:51] <cone-942> ffmpeg.git 03Martin Storsjö 07master:a7ba3244131d: rtpdec_mpeg4: Check the remaining amount of data before reading [12:51] <cone-942> ffmpeg.git 03Justin Ruggles 07master:e034cc6c60c7: lavc: Move vector_fmul_window to AVFloatDSPContext [12:51] <cone-942> ffmpeg.git 03Michael Niedermayer 07master:5c7e9e16c961: Merge remote-tracking branch 'qatar/master' [14:09] <cone-942> ffmpeg.git 03Michael Niedermayer 07master:cae11e40315a: mux: fix chunked_duration rounding anomaly [14:09] <cone-942> ffmpeg.git 03Michael Niedermayer 07master:f0cf017dee3c: configure: suppress "enumerated type mixed with another type" for icc [15:11] <kierank> durandal_1707: the error message is not strange at all [15:28] <durandal_1707> mateo`: what is unclear to you about id3 tag? [15:28] <durandal_1707> in aiff it is always 2byte aligned [15:29] <durandal_1707> so skiping 1 byte if not aligned is enough [15:29] <durandal_1707> and that is already done for metadata [15:31] <durandal_1707> kierank: messages that telly you that arbitary bug is not going to happen are weird [15:31] <mateo`> durandal_1707: this is clear now. however in the case i explained, do you agree that the declared chunk size is invalid ? [15:34] <durandal_1707> the one in file? [15:36] <mateo`> durandal_1707: the id3 chunk size is one byte less than the actual read data [15:46] <durandal_1707> mateo`: yes, if all read data is used [15:50] <mateo`> durandal_1707: at least they are consumed by the ff_id3v2_read function [16:08] <cone-942> ffmpeg.git 03Nicolas George 07master:0942aa463736: lafv/matroska: add A_OPUS/EXPERIMENTAL codec name. [16:37] <mateo`> do we have g726 samples ? [17:45] <saste> anyone experience with HLS streaming? how much can latency be reduced? [17:47] <kierank> you are limited by the chunk size afaik [17:47] <saste> kierank, also i noticed that i can hardly control what the player displays [17:48] <kierank> yes it's rather black magic [17:48] <kierank> because of the wide range of implementations [17:48] <kierank> this is why i don't work on HLS :) [17:48] <saste> the problem is that it is advertised like "live" streaming when in the truth you can hardly control the latency [17:49] <saste> for an interactive application 10 seconds is not "live" anymore [17:49] <kierank> this is what i say to all the people who think adaptive streaming is as good as broadcast tv [17:49] <kierank> 10 seconds is ridiculous to say the least [17:49] <kierank> not including cdn caching etc [17:52] <saste> kierank, 10s delay for VOD is not that bad [17:52] <kierank> yes that's true [17:52] <saste> it is a *big* problem if you have an interactive application (for example a guy talking on the other hand) [17:52] <kierank> but channel surfing is a problem [17:53] <kierank> (zapping) [17:53] <saste> flash/rtmp is suited for that, but now people is trying to use HLS for doing the same stuff [17:54] <kierank> flash/rtmp allows you to set a buffer time [17:54] <kierank> all adaptive streaming etc is implementation defined [17:54] <saste> so it seems [19:21] <burek> hi guys :) [19:21] <burek> happy ny and all the trailing holidays :) [19:46] <burek> http://www.ffmpeg.org/ffmpeg-formats.html#segment_002c-stream_005fsegment_0… [19:46] <burek> segment_time time [19:46] <burek> Set segment duration to time. Default value is "2". [19:46] <burek> is this in seconds, milliseconds, .. ? [19:47] <burek> also segment_times times [19:50] <saste> burek: they are duration specifications [19:50] <burek> also, is it correct to write "-segment_list_flags +live" or "-segment_list_flags live" (documentation is not precise about the usage of +) [19:50] <burek> saste [hh:mm:ss.mmm] ? [19:50] <saste> +foo => add foo to the current value [19:51] <saste> burek: see ffmpeg-utils(1) [19:51] <burek> hm, it should be documented in some more clear way [19:51] <burek> imho [19:53] <burek> btw, if -segment_time 2 means 2 seconds, then there might be a bug in segment muxer http://ffmpeg.gusari.org/viewtopic.php?f=12&t=772 (provided that the guy is using the latest ffmpeg) [19:55] <saste> burek: no, all are complaining about the same thing, noone reads (carefully) the docs [19:55] <saste> it depends on I-frames on the input [19:56] <burek> well if many are complaining, it might be a good idea to put a note there or something, to lower the number of complaints [19:56] <burek> btw, this guy is re-encoding, he doesn't use -codec copy [19:57] <saste> burek, it is already in the docs, i'll make it more explicit [19:58] <burek> thank you :) [20:05] <Compn> is there any warning message for user [20:05] <Compn> when he selects -codec copy but uses encoder options ? [20:05] <Compn> 'warning, you are copying video, any encoding options will be ignored...' [20:06] <Compn> or e.g. ffmpeg -vcodec copy -vf yadif ... the vf will be ignored [20:06] <Compn> mencoder also lacks this warning message [20:09] <burek> yes, that would be highly useful if possible to implement [20:10] <saste> we lack developers... [20:24] <burek> is it possible to create vobsub (.idx/.sub) form .bmp files? [20:25] <burek> using ffmpeg [20:25] <burek> from* [20:27] <Compn> have to check if we have vobsub muxer [20:27] <Compn> i dont think so [20:27] <Compn> demuxer only ? [20:28] <burek> well, i think demuxer only would help too, if we can save the output into something useful [20:28] <burek> that ffmpeg supports [20:30] <burek> a guy has got a lot of bmp images, that he ripped using subrip from some video and he edited those bmp images and now would like to put all that into some subs format or so [20:30] <burek> http://ffmpeg.gusari.org/viewtopic.php?f=12&t=788 [20:31] <Compn> well there are bmp2vobsub tools ... [20:33] <shahriman> guys, who looks after the mpeg-ps demuxer? [20:35] <burek> Your search - bmp2vobsub - did not match any documents. :) [20:37] <Compn> but uh [20:37] <Compn> then i'm wrong [20:37] <Compn> but i think ubitux was talking about bmp and vobsub [20:38] <durandal_1707> shahriman: there is bug ? [20:38] <shahriman> durandal_1707: yep. http://git.videolan.org/?p=ffmpeg.git;a=commit;h=759901f817cb481c989af7bec4… [20:38] <shahriman> this is the culprit commit [20:38] <shahriman> problem is [20:38] <shahriman> i can't share sample [20:39] <Compn> why not shahriman ? [20:39] <shahriman> i don't own it [20:39] <Compn> its covered under fair use [20:39] <shahriman> no it's not [20:39] <Compn> if you mark it private, we will delete it after bug fixed [20:40] <shahriman> IANL [20:40] <shahriman> IANAL* [20:40] <Compn> i've followed copyright law for a long time. its covered under fair use [20:40] <Compn> but i'm no lawyer either :) [20:41] <shahriman> sometimes users upload stuffs under CC, sometimes they allow anyone to download the original [20:41] <shahriman> but in this case, the user didn't do any of those [20:42] <Compn> it doesnt matter [20:42] <Compn> for educational or research purposes , especially a small clip , runs under the fair use in us copyright law [20:44] <durandal_1707> what is bug about? [20:46] <durandal_1707> perhaps only first say 1023 bytes is needed to reproduce it? [20:59] <Compn> since when does michael work on ogg code? :) [20:59] <Compn> chained ogg [21:09] <durandal_1707> if i type git clean -f i get : Not removing doc/examples/pc-uninstalled/ [21:10] <durandal_1707> perhapts it should be added to .gitignore? [21:13] <shahriman> durandal_1707: http://git.videolan.org/?p=ffmpeg.git;a=blobdiff;f=libavformat/mpeg.c;h=f01… [21:13] <shahriman> it's the fourth hunk [21:13] <shahriman> that's responsible [21:14] <shahriman> I don't know where the number 6 came from [21:14] <shahriman> i don't have the spec, so can't verify [21:15] <durandal_1707> that still does not explain bug/crash [21:15] <durandal_1707> where it crashes? [21:17] <durandal_1707> michaelni: you have sample for above mentioned commit? [21:22] <shahriman> durandal_1707: i never said it crashes [21:22] <durandal_1707> so what is problem? [21:22] <shahriman> it just breaks the audio decoding [21:23] <shahriman> i get corrupted audio output [21:23] <durandal_1707> it is detected as wrong codec? [21:23] <cone-942> ffmpeg.git 03Michael Niedermayer 07master:b7bc49a957a1: mips: move vector_fmul_window_mips to libavutil [21:23] <cone-942> ffmpeg.git 03Michael Niedermayer 07master:1191db31c1ed: mux: fix chunked interleaver [21:24] <michaelni> durandal_1707, i might have one but i dont know how to find it ... [21:24] <michaelni> maybe searching trac would produce a sample [21:24] <michaelni> or ask carl [21:25] <shahriman> it's not really enough to have a particular sample [21:25] <shahriman> you need the spec to be sure that the decoder is doing the right thing [21:25] <shahriman> i doubt if the user submitted sample is broken [21:26] <shahriman> i.e. broken muxer [21:26] <shahriman> or encoder [21:27] <shahriman> durandal_1707: it's not detected as a wrong codec, because I can still hear the original audio. only with a lot of clicking artifacts [21:28] <shahriman> and this error: [21:28] <shahriman> [pcm_s16be @ 0x7f2830000e00] Invalid PCM packet, data has size 2 but at least a size of 4 was expected [21:28] <shahriman> Frame changed from size:0x0 to size:720x576 [21:28] <shahriman> [pcm_s16be @ 0x7f2830000e00] Invalid PCM packet, data has size 1 but at least a size of 4 was expected [21:28] <shahriman> ofc windows media player plays it fine [21:32] <michaelni> if someone provides me with a sample that doesnt work then i can try to fix it ... [21:33] <michaelni> otherwise without spec and without sample i have no idea ... [21:34] <shahriman> michaelni: well you accepted a patch without any reference in the first place which caused the breakage [21:35] <durandal_1707> shahriman: looks like extra skip is done [21:37] <durandal_1707> so if you undo 4th and 5th hunk it plays fine? [21:39] <nevcairiel> this is the original issue that the patch fixed, http://roundup.libav.org/issue1731 .. there was a sample once, no idea if its still in incoming [21:39] <shahriman> durandal_1707: undoing 4th does the trick [21:39] <nevcairiel> DVD Audio isn't exactly a format that has documentation available everywhere on the web =p [21:40] <shahriman> understandable [21:43] <durandal_1707> i cant understand how by undoing 4th hunk issue is fixed [21:46] <shahriman> durandal_1707: yeah sorry 4 and 5 [21:50] <durandal_1707> shahriman: same :( [21:59] <cone-942> ffmpeg.git 03Paul B Mahol 07master:641bbd967119: vima: switch to init_get_bits8() [23:36] <saste> wtf is this the segmenting year? [23:36] <cehoyos> durandal_1707 : You asked January 5th about commit 697b476c075a - it was about ticket #1747 [23:37] <kierank> saste: yes [23:41] <Daemon404> carl on irc? wtf [23:41] <Daemon404> my world is shattered [23:41] <saste> well rtsp issues also are going strong [23:42] <saste> who want to debug a mov issue for me ? [23:43] <Daemon404> lavf's mov/mp4 demuxer is the bane of my existence [23:43] Action: Daemon404 is seriously considering writing liblsmash.c [23:44] <Daemon404> (read: no wayyyyy) [23:45] <saste> Daemon404, i got a nice crash coming from the hls demuxer [23:45] <saste> chained crash [23:45] <Daemon404> not touching hls [23:45] <llogan> what if i said..."c'mon!"? [23:59] <burek> "Please setup a official RPM Repository, where latest releases are updated." [23:59] <burek> I think we spoiled our users with those static builds :) [00:00] --- Thu Jan 17 2013

1 0

ffmpeg.log.20130116
by burek 17 Jan '13

17 Jan '13

[00:00] <llogan> also the FFmpeg wiki is coming along nicely so the arch wiki could link to specific ffmpeg wiki pages to decrease duplication [00:01] <damaru> can ffmpeg target a certain window, like the gnash player for instance ? [00:01] <stqn> ah thats interesting& I had problems finding good up to date information on webm encoding [00:02] <llogan> the ffmepg wiki lacks such an article, AFAIK. [00:02] <llogan> probably because vp8/webm is usually a waste of time [00:02] <stqn> eh& why? [00:02] <damaru> well webm has been quite usefull for javascript video player [00:02] <llogan> compared to H.264 produced by x264 [00:02] <damaru> (or I am doing something really wrong) [00:03] <llogan> but i would like to see a VP8/9 in webm article anway for those who want to use it anyway [00:04] <llogan> damaru: like this: "-i :0.0+200,100" x and y offsets [00:05] <llogan> or run xwininfo and click on your desired window. it will tell you the offsets and frame size [00:05] <damaru> kool, I'll play around to get the right offset [00:05] <damaru> kool [00:13] <damaru> sweet!! all working, thanks a lot, that was a lot of knowledge in a short ammount ;) [00:13] <damaru> one last question while i am here, is there some place I can find receipe for ffmpeg exporting for html5 video ? [00:22] <damaru> thanks :D [00:32] <tyebud> Is it possible to extract a sprite track from a quicktime movie using ffmpeg? [00:33] <cbsrobot> tyebud: I was looking at your file [00:33] <cbsrobot> I don't see any file referneced to it [00:34] <tyebud> I opened it up in quicktime pro 7 and I'm seeing a sprite track embedded in the file/ It shows as 200~ kb [00:34] <cbsrobot> basicly you have 3 tracks: video, data and subtitle [00:34] <cbsrobot> what do you want to do with it ? [00:37] <tyebud> Well, I've got a web based video player that handles time based events [00:37] <tyebud> I'm trying to essentially extract the stop points in the .mov file so I can use them in my web player [00:40] <cbsrobot> tyebud: so you want to extract a list of timecode and chapters ? [00:41] <tyebud> Correct [00:56] <cbsrobot> hmmm not sure [00:56] <cbsrobot> I have to sleep first [00:57] <tyebud> It's a tough one! [00:57] <tyebud> I've been researching all day. not many people have shown interest in doing something like this. [01:03] <cbsrobot> tyebud: still alive ? [01:03] <tyebud> Yep [01:03] <cbsrobot> try: [01:03] <cbsrobot> ffmpeg -i keynote_animation_test.mov test.srt [01:04] <cbsrobot> so you get the subtitles out [01:04] <cbsrobot> I guess that is enough to create the chapters aswell [01:04] <cbsrobot> as I see it you file does not have any chapters [01:06] <cbsrobot> ah wait - maybe it has [01:06] <tyebud> The resulting srt file doesn't have any contents [01:06] <cbsrobot> tyebud: update your ffmpeg [01:08] <cbsrobot> it seems in the sprite track, the tref > chap atom references to track 3 [01:08] <tyebud> How are you viewing the atom info? [01:08] <cbsrobot> but it does not seem to have a track 3 (counting from 0) [01:09] <cbsrobot> but then again I don't know what this sprite track is ... [01:09] <tyebud> From my reading, the sprite track contols the starting and stopping [01:10] <tyebud> From what I understand it can be used for dvd menu type stuff [01:11] <tyebud> So, in QT Pro, I'm seeing this in the tracks view - http://cl.ly/MBwA [01:11] <tyebud> there's a disabled chapter track [01:12] <cbsrobot> I use atom inspector to see the qt atoms [01:12] <cbsrobot> you can find it somewhere [01:12] <tyebud> win or mac? [01:12] <cbsrobot> mac [01:13] <cbsrobot> try: http://connect.apple.com/cgi-bin/WebObjects/MemberSite.woa/wo/5.1.17.2.1.3.… [01:14] <cbsrobot> you need to login [01:16] <tyebud> Which one am I looking at? [01:17] <cbsrobot> which what [01:17] <tyebud> I see a list of several different sdks [01:17] <cbsrobot> ha [01:17] <cbsrobot> no idea [01:18] <cbsrobot> you have to dig a bit [01:18] <tyebud> Oh I found the atom inspector [01:22] <cbsrobot> don't get radioactive with all these atoms [01:23] <tyebud> haha [01:23] <tyebud> I'm going to get radiation poisoning.. [01:24] <tyebud> The Sample to Time atom looks interesting [01:29] <tyebud> Alright, I need to get out of here for a bit. Thanks for all your help, cbsrobot [01:29] <cbsrobot> np [02:59] <stqn> bye [06:20] <Abhijit> hi [06:20] <Abhijit> definition of avcodeccontext is inside avcode.h but i found out that in libavg programming they have used avcodeccontext wihtout including acvodec.h [06:21] <Abhijit> how they have done it? any idea? [06:21] <Abhijit> help please [06:40] <Abhijit> hello [06:40] <Abhijit> any help please [07:06] <praveenmarkandu> hi trying to run this command: ffmpeg -i test.mov -c:v libx264 -c:a libfaac -vf "scale=720:-1" output.mp4 [07:06] <praveenmarkandu> it tells me something is wrong with my bitrate, hieght or width [07:07] <praveenmarkandu> but when i do a -s 720x404 it works fine [07:07] <praveenmarkandu> can i not use -vf "scale=720:-1" [08:01] <AL13N> praveenmarkandu: why do you want to use -vf scale? doesn't -s 720x404 works as well? [08:02] <AL13N> i'm using -s 640x480 and the videos come out as 640x422 [09:56] <praveenmarkandu> AL13N: i want to keep the aspect ratio the same and only adjust the width [09:56] <praveenmarkandu> is that possible? [10:26] <cbsrobot> praveenmarkandu: sure [10:33] <praveenmarkandu> cbsrobot: what command do i use to do that? [10:34] <praveenmarkandu> -vf "scale:720:-1" doesnt work [10:34] <cbsrobot> pastebin the command you used [10:34] <cbsrobot> and the full output [10:34] <praveenmarkandu> i found that somewhere online [10:34] <cbsrobot> in irc speak: [10:36] <praveenmarkandu> cbsrobot: http://pastebin.com/RxaRN1aN [10:36] <praveenmarkandu> it says the height isnt divisible by 2 [10:37] <praveenmarkandu> .... common, there must be a way to get around that [10:38] <cbsrobot> yeah but it depends on the pix fmt used [10:38] <cbsrobot> yuv420p cannot be odd [10:38] <cbsrobot> so use the nearest approximation: 306 [10:38] <cbsrobot> or 308 [10:39] <cbsrobot> i guess its 306 [10:39] <praveenmarkandu> found this bug: https://ffmpeg.org/trac/ffmpeg/ticket/309 [10:39] <praveenmarkandu> i guess it hasnt been addressed [10:39] <praveenmarkandu> i just want it to be automatic. i have to batch a lot of videos. all with different source resolutions [10:41] <cbsrobot> so tuse the workaround [10:41] <cbsrobot> d [10:42] <cbsrobot> *use [10:42] <praveenmarkandu> oh -2? [10:43] <cbsrobot> try it [10:45] <praveenmarkandu> cbsrobot: -2 doesnt work, 854:trunc(ow/a/2)*2 seems to start. will look at the output once it is done [10:50] <brontosaurusrex> is it usual that US broadcasts are 60p? [11:24] <brontosaurusrex> will ( ffmpeg -i video.mov -i logo.png -filter_complex "scale=1230:800,overlay=0:0" ) scale logo as well, or only video? [11:31] <Mavrik> brontosaurusrex: if you switch order it will [11:31] <Mavrik> so -vf overlay=0:0,scale=1230:800 [11:31] <brontosaurusrex> so logo is intact in this case right? [11:32] <Mavrik> yes [11:32] <Mavrik> because you scale the iimage first [11:32] <brontosaurusrex> good [11:32] <Mavrik> and then apply the logo [11:32] <brontosaurusrex> another question: proper framerate conversion possible in ffmpeg? and proper ivtc? [11:33] <Mavrik> I don't know what "proper" is [11:33] <Mavrik> ffmpeg can convert framerate [11:33] <Mavrik> it can't do telecine as far as I know [11:34] <Mavrik> it basically just either duplicates or drops frames as needed [11:34] <Mavrik> so for proper PAL/NTSC conversion you'll need something else :\ [11:34] <brontosaurusrex> ok [11:47] <brontosaurusrex> but fractional framerates cant be used? like 29.97 ? [11:48] <Mavrik> mhm [13:52] <_julian> hi [13:52] <_julian> which protocols do I have to enable on build to get rtsp support? [13:59] <Mavrik> rtsp? :) [14:20] <_julian> Mavrik: was some more, udp and rtp for example :) [14:26] <shevy> hey guys [14:26] <shevy> I have a .mp3 file which I want to cut with ffmpeg [14:26] <TeknoJuce> is there a way to get ffserver to capture and stream with less than 1ms of delay on a smaller clocked device like a 500Mhz processor, I can get this from MPEG_Streamer but not from ffmpeg is MPEG_Streamer based on anything to do with ffmpeg? [14:26] <shevy> this is possible via -ss [14:27] <shevy> my question is - can I combine several -ss together in the same ffmpeg commandline? [14:29] <TeknoJuce> shevy ffmpeg -i son.wav -i video_origine.avi video_finale.mpg [14:31] <gxk> how to create avi file from video and 2 separate audio files in the way that at playtime all audio sources should be heard at the sane time? [14:32] <Mavrik> shevy: not really, no [14:33] <shevy> ok [14:40] <TeknoJuce> https://chrome.google.com/webstore/detail/chrome-remote-desktop/gbchcmhmhah… [14:41] <TeknoJuce> oops [14:41] <TeknoJuce> wrong window [14:59] <gxk> can to play video with 2 audio files at the same time? [15:32] <Olive6767> csrobot: Hello :) Got your messages, -metada:s works great, thanks ! :) [15:34] <cbsrobot> Olive6767: spell my nick correctly or use the tab key to autocomplete ! [15:34] <cbsrobot> but nice to hear [15:34] <Olive6767> oops! sorry :-) [15:46] <brontosaurusrex> gxk: ffmpeg -i video -i left.mono.wav -i right.mono.wav -map 0:v -map 1:a -map 2:a , ect < example [15:47] <brontosaurusrex> shevy: a combo of ss and duration [16:00] <_julian> how can I check if a frame which I read with av_read_frame() is an I frame? [16:02] <mpfundstein> +Namegduf: Oh, i thought unsafe was depricated in 1.0 [16:04] <Mavrik> _julian: ehm [16:04] <Mavrik> AVFrame has a "key_frame" field? [16:13] <_julian> Mavrik: but AVFrame stores the decoded frames doesn't it? I want to know if I am at a keyframe before starting to decode [16:29] <shevy> brontosaurusrex a combo? ffmpeg -ss 22 -ss 33 ? [17:15] <DonGnom> how do i encode a side-by-side video to get a 2D version from (dropping one side)? [17:17] <Mavrik> DonGnom, use the crop filter (http://ffmpeg.org/ffmpeg-filters.html#crop) to crop away a side [17:20] <DonGnom> Mavrik: hm but isnt it then little malformed? (the video is 1080p but only one half used for each eye but full height for both of the videos so isnt the result then 960x1080 instead of 1920x1080? [17:20] <Mavrik> DonGnom, yeah, it's anamorphic [17:20] <DonGnom> that means? [17:21] <Mavrik> that means it's stored as a 960x1080 image which gets stretched on TV [17:21] <Mavrik> (the "SAR" flag tells it to do that) [17:21] <Mavrik> you can of course stretch image with "scale" filter to 1920x1080 if you want [17:22] <Mavrik> just make sure you scale AFTER cropping [17:22] <Mavrik> :) [17:22] <DonGnom> Mavrik: ok so i can do something like -codec:v copy -<crop stuff> -codec:a copy and for the chapter metadata too and i get a anamorphic video or i can add the scale stuff and get a "normal" video? [17:22] <DonGnom> and with "after cropping" you mean in an extra step? [17:22] <Mavrik> DonGnom, first of all, you can't drop with "codec:v copy", you have to reencode [17:23] <Mavrik> and I mean to make sure your filters are in the right order [17:23] <Mavrik> so [17:23] <DonGnom> ok [17:23] <Mavrik> -vf crop=<crop parameters>,scale=1920:1080 [17:23] <DonGnom> thanks! [17:23] <Mavrik> or better yet [17:23] <Mavrik> -vf crop=<crop parameters>,scale=1920:1080,setsar=1:1 [17:24] <DonGnom> Mavrik: ok thank you very much ill try that [17:42] <DX099> hello all, what is the relationship between Totem and ffmpeg ? can Totem can play all videos format that ffmpeg can once the later is installed ? [17:43] <Mavrik> um [17:43] <Mavrik> ffmpeg is actually a program to transcode videos from one format to another [17:43] <saste> DX099, totem makes use of the gstreamer framework [17:43] <Mavrik> and yeah, Totem uses gstreamer to decode videos [17:43] <DX099> ok [17:43] <saste> which in turn makes use of the ffmpeg libraries (libavcodec/libavformat in particular) [17:43] <Mavrik> yeah [17:43] <Mavrik> complicated enough for ya? ;) [17:43] <saste> so in theory it should support as many formats as supported by the linked libraries [17:44] <DX099> even though the gstreamer-ffmpeg package still asks for libavcodec even when ffmpeg is installed ? [17:45] <durandal_1707> because it is build for some other ffmpeg lib [17:46] <DX099> ok [17:46] <DX099> but when it comes to mplayer and vlc, they're working directly with ffmpeg, right ? [17:47] <Mavrik> actually neither uses ffmpeg libraries currently [17:47] <Mavrik> they both have their own forks [17:47] <DX099> ok [17:49] <DX099> so on linux, apart from CLI ffplay and gstreamer far related cousin, no player actually uses ffmpeg to decode videos the way it can be done on Windows ? [17:49] <Mavrik> DX099, most distro-installed players use gstreamer (and by extension ffmpeg libs) with exception of mplayer and vlc [17:49] <DX099> ok [17:56] <saste> Mavrik, Mplayer still uses standard FFmpeg libs, VLC can be compiled with FFmpeg or Libav libs [17:58] <videoman> Using version git-2013-01-13-d676598 for HLS, segment_time parameter seems to be ignored [18:16] <videoman> thanks, ffmpeg ignoring segment_time : http://pastebin.com/h0DyABwR [18:34] <saste> videoman, read the segment manual about key frames [18:35] <videoman> Will do, thanks saste [18:35] <saste> you can use -force_key_frames, a better solution is WIP, check for example http://gitorious.org/~saste/ffmpeg/sastes-ffmpeg/commits/extend-segment-201… [20:11] <AL13N> i have an odd thing: i use -s 640x480 -r 12 and i convert to mp4/ogg/webm; however, the image preview for the webm is stretched into 640x480, while the others have 640x422, subsequently taking an image with -vf frames=1 also stretches the image [20:11] <AL13N> why would the image preview of mp4 and ogg not be stretched and webm would be stretched? [20:29] <llogan> AL13N: what is generating the previews? [20:30] <llogan> burek: haven't seen you here for a while [20:30] <burek> hi llogan :) holidays :D [22:35] <AL13N> burek: llogan: https://gist.github.com/4551150 [22:35] <AL13N> from "ffmpeg -i testfotos/DSC_0222.MOV -s 640x480 -frames:v 1 file3.jpg 2>&1 | wgetpaste" [22:42] <AL13N> also : "ffmpeg -ss 00:00:10 -i testfotos/DSC_0222.MOV -s 640x480 -r 12 file.webm 2>&1 | wgetpaste" has this output: https://gist.github.com/4551173 [22:42] <AL13N> the preview i see in a file browser is 640x480 exactly, while the video has the aspect ratio from before [22:43] <AL13N> the weird thing is that if i do the same but with .ogg or .mp4, the preview i see in a browser has the same aspect ratio as the video [22:43] <AL13N> so, this is only webm [22:43] <AL13N> burek: llogan: any idea? [23:43] <llogan> AL13N: not really, unfortunately. what file browser? how do yo uknow it isn't the issue? [23:46] <burek> AL13N 1920/1080 = 1.77 and 640/480=1.33 [23:46] <burek> that's expected result that you are getting [23:46] <burek> so, i guess everything is working fine [23:49] <llogan> AL13N: so you can do this instead: "-vf scale=640:-1" [23:50] <Zeeflo> burek: u here? [23:50] <Zeeflo> What was the command to include .srt subs? [23:50] <Zeeflo> like. "ass=blabla.ass" [23:51] <klaxa> -vf ass=somefile.ass [23:51] <klaxa> iirc [23:51] <Zeeflo> dude, i asked for srt :D [23:51] <klaxa> ah uh... [23:53] <burek> Zeeflo http://www.ffmpeg.org/ffmpeg-filters.html#subtitles-1 [23:53] <Zeeflo> Burek: Thank you [23:53] <burek> -vf subtitles=sub.srt [23:53] <burek> :beer: [23:54] <Zeeflo> :D [23:54] <Zeeflo> when will you release an updated build? [23:54] <burek> of what [23:54] <Zeeflo> ffmpeg [23:54] <Zeeflo> standaloneš [23:54] <klaxa> i couldn't find that page via google, what the hell [23:55] <burek> Zeeflo ffmpeg.static.64bit.2013-01-16.tar.gz [23:55] <burek> it's up-to-date i guess [23:55] <burek> klaxa [23:55] <Zeeflo> Have you changed anything in it since december? [23:55] <burek> try from there [23:55] <burek> "FFmpeg filters" [23:56] <Zeeflo> alright! Getting a new copy then [23:56] <burek> Zeeflo, I've done a lot of stuff since december [23:56] <burek> I really don't remember exactly what was it about :S [23:56] <llogan> klaxa: the man page was segmented recently [23:56] <klaxa> ah yeah, well when i punched "ffmpeg-filters" into google that never showed up, then i looked for "libavfilter" but that documentation wasn't what i was looking for :V [23:56] <Zeeflo> i just noticed that the filesize has been the same for like 2 months or something [23:56] <klaxa> good to know thanks :) [23:56] <burek> Zeeflo, I've added ffprobe too [23:56] <Zeeflo> yea [23:56] <Zeeflo> I see! [23:56] <burek> a lot of people complained for it not being there [23:56] <Zeeflo> I just got it [23:56] <burek> that's why the binary got bigger [23:57] <burek> -binary +archive [23:57] <Zeeflo> i hardly ever use probe, but its good to have [23:58] <Zeeflo> burek: if I use srt subs what will the standard font and size be? [23:58] <Zeeflo> do you know? [23:58] <burek> i didn't test it [23:58] <burek> try asking ubitux i think [23:58] <burek> i believe he worked on that filter [23:59] <Zeeflo> ubitux: what is the standard font and size if I use the srt filter? [23:59] <Zeeflo> ill make a test and see how it looks! [00:00] --- Thu Jan 17 2013

1 0