Ffmpeg-devel-irc

Download

ffmpeg-devel-irc@ffmpeg.org

May 2013

1 participants
62 discussions

ffmpeg-devel.log.20130505
by burek 06 May '13

06 May '13

[00:34] <cone-149> ffmpeg.git 03Michael Niedermayer 07master:369a12082635: xbmdec: fix off by one error in scanf() [01:10] <cone-149> ffmpeg.git 03Michael Niedermayer 07master:adaa7743f5fd: src_movie: fix scanf string [02:01] <cone-149> ffmpeg.git 03Claudio Freire 07master:c545876d1bcf: AAC encoder: Fixed a rounding bug in psy's channel bitrate computation. [04:24] <slavik262> Stupid question - I'm trying to upload sample files to the FTP server to go along with a bug report, but I keep getting "553 Can't open that file: Permission denied" [04:24] <slavik262> on a related note, I can't cd into "incoming" as the guide describes [04:25] <slavik262> are the instructions out of date? [04:34] <cone-149> ffmpeg.git 03Claudio Freire 07master:1ef82cc60d3b: AAC encoder: Fix rate control on twoloop. [05:36] <Compn> slavik262 : what ftp addy you using ? [05:37] <Compn> oh something did change [05:37] <Compn> it goes directly to incoming now [05:38] <Compn> j-b has been running our incoming server for a while [05:42] <slavik262> Compn: so, potential fix might be? [05:43] <slavik262> like I said, it refuses to accept any files [05:43] <slavik262> login is anonymous, no password, correct? [05:52] <Compn> talk to j-b , seems broken [05:52] <Compn> yes [05:52] <Compn> j-b : ping [05:52] Action: Compn afk [10:01] <cone-652> ffmpeg.git 03Reimar Döffinger 07master:3bb1c58e9532: Avoid incorrect deprecated pixel format warning. [10:01] <cone-652> ffmpeg.git 03Reimar Döffinger 07master:d82729e5d929: Detect and skip mj2-style jp2c header. [10:24] <cone-652> ffmpeg.git 03Reinhard Tartler 07master:768e44d044bc: build: tune down the output of lcov/gcov [10:24] <cone-652> ffmpeg.git 03Michael Niedermayer 07master:fbede75a2580: Merge commit '768e44d044bc32de1e58c198a3708d186af7e966' [10:41] <cone-652> ffmpeg.git 03Reinhard Tartler 07master:95e8ac60f8b7: build: normalize coverage.info [10:41] <cone-652> ffmpeg.git 03Michael Niedermayer 07master:336ffb837705: Merge commit '95e8ac60f8b7317ce23c648250fdc8fb68d65a07' [11:00] <cone-652> ffmpeg.git 03Reinhard Tartler 07master:b1ee8eec849f: configure: Add basic valgrind-memcheck support [11:00] <cone-652> ffmpeg.git 03Michael Niedermayer 07master:4451be5ce1fc: Merge commit 'b1ee8eec849f17fce2f411a297de9cbf2edaeb3a' [11:00] <cone-652> ffmpeg.git 03Michael Niedermayer 07master:86c9cf643b0b: configure: fix valgrind-memcheck arguments [11:24] <cone-652> ffmpeg.git 03Luca Barbato 07master:72e228b274a9: nut: K&R formatting cosmetics [11:24] <cone-652> ffmpeg.git 03Michael Niedermayer 07master:057c5d2e1646: Merge commit '72e228b274a98af1aba9588415d0cd87fde44663' [11:39] <cone-652> ffmpeg.git 03Diego Biurrun 07master:6fee1b90ce3b: avcodec: Add av_cold attributes to init functions missing them [11:39] <cone-652> ffmpeg.git 03Michael Niedermayer 07master:0aa095483d19: Merge commit '6fee1b90ce3bf4fbdfde7016e0890057c9000487' [11:50] <cone-652> ffmpeg.git 03Diego Biurrun 07master:093804a93cc5: avfilter: Add av_cold attributes to init/uninit functions [11:50] <cone-652> ffmpeg.git 03Michael Niedermayer 07master:a8ff830b79d0: Merge commit '093804a93cc5da3f95f98265a5df116912443cec' [11:56] <cone-652> ffmpeg.git 03Diego Biurrun 07master:e445647b4fdf: avresample: Add av_cold attributes to init functions missing them [11:56] <cone-652> ffmpeg.git 03Michael Niedermayer 07master:b507c00dc03c: Merge commit 'e445647b4fdf481b13b2743b303d84de4f43bedd' [12:00] <cone-652> ffmpeg.git 03Diego Biurrun 07master:4d7ed3e73541: build: output-example: Add avutil to ELIBS in link command [12:00] <cone-652> ffmpeg.git 03Michael Niedermayer 07master:77570facdb41: Merge commit '4d7ed3e73541983459a7913c2034eba69ebc113d' [12:33] <megaTherion> to convert an picture into different format, swscale functions are the correct things to use? [12:33] <megaTherion> I've read an older tutorial which said to use img_convert, but that is deprecated obviously [12:40] <cone-652> ffmpeg.git 03Diego Biurrun 07master:e0f8be6413b6: avformat: Add AVPROBE_SCORE_EXTENSION define and use where appropriate [12:41] <cone-652> ffmpeg.git 03Michael Niedermayer 07master:f083b4c33883: Merge commit 'e0f8be6413b6a8d334d6052e610af32935c310af' [12:45] <cone-652> ffmpeg.git 03Diego Biurrun 07master:a8a45dfe25ec: avs: Increase magic value score above file extension score [12:45] <cone-652> ffmpeg.git 03Diego Biurrun 07master:c0c2b96c136e: sctp: Add a note on why sctp_recvmsg and sctp_sendmsg have ff_-prefixes [12:45] <cone-652> ffmpeg.git 03Michael Niedermayer 07master:fdccb84e3b94: Merge commit 'c0c2b96c136ec2cbfd808eefc26de399f447d62f' [12:53] <cone-652> ffmpeg.git 03Hendrik Leppkes 07master:20c86571ccc7: lavfi: let gcc realign the stack on public graph driving functions [12:53] <cone-652> ffmpeg.git 03Michael Niedermayer 07master:e7f8493fd8c9: Merge commit '20c86571ccc71412781d4a4813e4693e0c42aec6' [12:58] <cone-652> ffmpeg.git 03Diego Biurrun 07master:1fda184a8517: avutil: Add av_cold attributes to init functions missing them [12:58] <cone-652> ffmpeg.git 03Michael Niedermayer 07master:3c200aa6930d: Merge commit '1fda184a85178cfd7b98d9e308d18e1ded76a511' [13:05] <cone-652> ffmpeg.git 03Diego Biurrun 07master:2b677ffca54a: swscale: Add av_cold attributes to init functions missing them [13:05] <cone-652> ffmpeg.git 03Michael Niedermayer 07master:d5dbd84c9ab6: Merge commit '2b677ffca54a5fbef9c8860841c32f28ecd68f70' [13:10] <cone-652> ffmpeg.git 03Martin Storsjö 07master:2cca96e670e7: configure: Check for getenv [13:10] <cone-652> ffmpeg.git 03Michael Niedermayer 07master:35623d833eec: Merge remote-tracking branch 'qatar/master' [13:41] <cone-652> ffmpeg.git 03Michael Niedermayer 07master:a1f234be11ea: make the ordering of the return type and av_cold consistent [13:41] <cone-652> ffmpeg.git 03Michael Niedermayer 07master:097f668047c8: avformat: replace "AVPROBE_SCORE_MAX / X" by AVPROBE_SCORE_EXTENSION / Y [13:51] <cone-652> ffmpeg.git 03highgod0401 07master:548101b553cb: lavu/opencl: add opencl public error API [14:29] <cone-652> ffmpeg.git 03Michael Niedermayer 07master:cdfa1c89b715: avcodec: Set frame.width/height on all exit pathes of ff_get_buffer() [15:55] <cone-652> ffmpeg.git 03Rainer Hochecker 07master:b37cc5995b88: fix vdpau vc1 interlace modes [16:27] <cone-652> ffmpeg.git 03Michael Niedermayer 07master:54021aea5704: configure: set valgrind-memcheck toolchain when --valgrind is specified [16:27] <cone-652> ffmpeg.git 03Michael Niedermayer 07master:fb90e7b86d02: configure: Use the tool specified with --valgrind instead of a hardcoded "valgrind" [16:49] <cone-652> ffmpeg.git 03Rainer Hochecker 07master:fe1f36547d0b: register AV_PIX_FMT_VDPAU for hwaccel [16:49] <cone-652> ffmpeg.git 03Michael Niedermayer 07master:2d8dddd92659: Merge branch 'vdpau' of https://github.com/FernetMenta/FFmpeg [19:07] <kierank> Daemon404: holy shit [19:07] <kierank> an aac patch [19:13] <Compn> you saying people dont like working on aac ? [19:14] <kierank> yes [19:23] <Compn> kierank : is it because aac is a terrible codec or because ffmpeg code isnt easy to work on ? [19:23] <Compn> or both [19:23] <kierank> because of historical "problem" of lame being good [19:25] <Compn> lame as in the lame project or lame as in bad [19:25] <Compn> lol [19:25] Action: Compn goes back to whatever it is [19:26] <kierank> lame [19:26] <kierank> project [19:29] <Compn> so people preferred mp3 back in the day [19:29] <Compn> but now aac is de-facto standard in mkv and mp4 [19:30] <kierank> no it's not [19:30] <kierank> people preferred mp3 and so no decent aac encoder was developed [19:31] <Compn> isnt this a decoder patch ? [19:31] <kierank> no [19:31] <Compn> oh was looking at the trac not ml [19:31] <Compn> nevermind :) [19:32] <Compn> yes i remember aac encoder fracass [19:40] <cehoyos> Is anybody able to encode lagarith? (Compn?) [19:42] <Compn> uhh [19:42] <Compn> the old version LAGS ? [19:42] <Compn> no i guess not [19:42] <Compn> but i have windows if you want me to install a codec [19:45] <cehoyos> I don't know where to find the codec (and I don't know how to use it) but I I think I am searching for newer lagarith encoder (if there are old and new versions) [19:47] <Compn> is there a bug or something i should read ? [19:48] <Compn> http://lags.leetcode.net/codec.html [19:49] <Compn> should be the only version [19:50] <cehoyos> I would need the following video as lagarith: ffmpeg -f lavfi -i color=red,format=bgr24 -t 1 -vcodec rawvideo out.avi [19:52] <Compn> color @ 025755e0] Missing key or no key/value separator found after key 'red' [19:52] <Compn> lavfi @ 02577700] Error initializing filter 'color' with args 'red' [19:52] <Compn> olor=red,format=bgr24: Invalid argument [19:52] <Compn> ffmpeg version N-51433-g551f683 Copyright (c) 2000-2013 the FFmpeg developers [19:52] <Compn> built on Mar 28 2013 21:52:57 with gcc 4.8.0 (GCC) [19:53] <Compn> maybe my ffmpeg needs an update [19:53] <Compn> a month old [19:55] <cehoyos> Try ffmpeg -f lavfi -i color=c=red,format=bgr24 -t 1 -vcodec rawvideo out.avi [19:56] <Compn> that worked [19:56] <Compn> what mode lagarith ? [19:56] <Compn> RGB (default) [19:57] <Compn> RGBA [19:57] <Compn> YUY2 or YV12 [19:57] <cehoyos> RGB please [19:57] <cehoyos> Actually both RGB and RGBA please [19:58] <Compn> [lagarith @ 0126de60]Unsupported Lagarith frame type: 0x6 [19:58] <Compn> ehe [19:59] <cehoyos> ;-) [19:59] <Compn> should i dcc this to you or ? [19:59] <cehoyos> Or upload it to datafilehost.com [19:59] <cehoyos> (if dcc succeeds, I am happy.) [20:00] <Compn> smaller than i thought :) [20:01] <Compn> try this > http://174.84.194.75:10000 [20:01] <Compn> http://174.84.194.75:10000/lagarith-red-rgb.avi [20:02] <Compn> http://174.84.194.75:10000/lagarith-red-rgba.avi [20:02] <Compn> oops dcc might have worked if i didnt move the files for that [20:03] Action: Compn afk [20:03] <cehoyos> Is the file blue or red for you? [20:05] <Compn> red [20:05] <cehoyos> ;-( [20:05] <cehoyos> (I will have to fix my patch) Thank you! [20:05] <Compn> np [20:05] <Compn> swapuv :P [20:05] <cehoyos> It's rgb [20:05] <Compn> oh [20:05] <Compn> :P [20:05] <cehoyos> Probably swaprb [20:05] <Compn> that too [20:06] <Compn> did you expect lagarith to change red to blue ? [20:06] <Compn> red input to blue output that is [20:06] <cehoyos> Could you try if the "RGBA" output file changes if you use following input: ffmpeg -f lavfi -i color=c=red,format=bgr32 -t 1 -vcodec rawvideo out.avi [20:06] <cehoyos> So the output file is blue for you [20:06] <cehoyos> ? [20:07] <cehoyos> With the original lagarith decoder? [20:07] <Compn> with the original decoder its red [20:07] <cehoyos> So what is blue? [20:07] <Compn> i was asking why you got blue result [20:07] <cehoyos> Sorry, I skipped one line (the one with the question)! [20:07] <cehoyos> No, the output file is blue here indicating a bug in my code. [20:08] <Compn> this outputs a blue file for me >> ffmpeg -f lavfi -i color=c=red,format=bgr32 -t 1 -vcodec rawvideo out.avi [20:08] <Compn> i mean [20:08] <Compn> out.avi is blue [20:09] <Compn> in both ffplay and vdub raw decoder [20:09] <Compn> is this bug or intended ? [20:10] <cehoyos> But did the byte on position 0xba change from 0x18 to 0x24 ? [20:10] <cehoyos> In the encoded file. [20:11] <Compn> i didnt test encoding it [20:11] <cehoyos> (format=rgb32 to fix red vs blue) [20:11] <Compn> i wanted to know if you wanted me to have a blue input [20:11] <Compn> before i encode it [20:11] <cehoyos> I don't care;-) [20:11] <Compn> killing me... [20:11] <cehoyos> But if you change from bgr32 to rgb32 it will be red (I expect) [20:11] <cehoyos> Which is probably better given that my output here is wrong. [20:12] <Compn> so you want me to test with red input or blue input ? [20:13] <Compn> http://174.84.194.75:10000/lagarith-blueinput-rgba.avi [20:13] <Compn> changing bgr32 to rgb32 does indeed make a red input [20:14] <Compn> http://174.84.194.75:10000/lagarith-redinput-rgba.avi [20:14] <Compn> opening both with lagarith original decoder produces the correct colors [20:14] <Compn> blueinput is blue output, redinput is red output [20:15] <cehoyos> Perfect, thank you! [20:15] <Compn> np [20:16] Action: Compn afk [20:31] <cone-652> ffmpeg.git 03Carl Eugen Hoyos 07master:a07ac1f7888f: Fix type of shared flac table ff_flac_blocksize_table[]. [21:00] <cone-652> ffmpeg.git 03Carl Eugen Hoyos 07master:4ec509ce3d1f: Support demuxing of audio-only vmd files. [21:00] <cone-652> ffmpeg.git 03Carl Eugen Hoyos 07master:7c23922c2be9: Cosmetics: Reindent Sierra vmd demuxer after last commit. [21:01] <cone-652> ffmpeg.git 03Carl Eugen Hoyos 07release/0.10:48f0a2c2394c: Fix type of shared flac table ff_flac_blocksize_table[]. [21:01] <cone-652> ffmpeg.git 03Carl Eugen Hoyos 07release/0.11:7dd042e657d9: Fix type of shared flac table ff_flac_blocksize_table[]. [21:01] <cone-652> ffmpeg.git 03Carl Eugen Hoyos 07release/0.7:11dd8eb48ba4: Fix type of shared flac table ff_flac_blocksize_table[]. [21:01] <cone-652> ffmpeg.git 03Carl Eugen Hoyos 07release/0.8:e2d529424f3f: Fix type of shared flac table ff_flac_blocksize_table[]. [21:01] <cone-652> ffmpeg.git 03Carl Eugen Hoyos 07release/0.9:6642804c22fd: Fix type of shared flac table ff_flac_blocksize_table[]. [21:01] <cone-652> ffmpeg.git 03Carl Eugen Hoyos 07release/1.0:6e2fb5cd73ed: Fix type of shared flac table ff_flac_blocksize_table[]. [21:01] <cone-652> ffmpeg.git 03Carl Eugen Hoyos 07release/1.1:46e1d05991bc: Fix type of shared flac table ff_flac_blocksize_table[]. [21:01] <cone-652> ffmpeg.git 03Carl Eugen Hoyos 07release/1.2:0a53c7016f77: Fix type of shared flac table ff_flac_blocksize_table[]. [22:31] <durandal_1707> saste: your comment on volume example in doc with 6db? [22:32] <saste> durandal_1707, didn't read it yet [22:38] <ubitux> durandal_1707: what does "not work"? [22:39] <durandal_1707> example [22:39] <ubitux> it seems to work fine here [22:39] <ubitux> yeah but how does it "not work"? [22:40] <durandal_1707> ubitux: what you actualy tried? [22:42] <ubitux> i copy pasted the example and it worked (volume boost by 6dB) [22:42] <ubitux> ./ffplay in.mp3 -af volume=volume=-6dB:precision=fixed [22:42] <ubitux> (well i tried -6dB to ear the diff too) [22:44] <durandal_1707> ubitux: what version you used? [22:45] <ubitux> HEAD? [22:45] <durandal_1707> well i really do not know but looks like you use different ffmpeg version, is there another fork? [22:46] <ubitux> not that i know of [22:46] <ubitux> maybe your strtod function doesn't behave the same on your system [22:46] <durandal_1707> i tried with ffmpeg and ffplay and it errors out: [volume @ 0x29d00a60] [Eval @ 0xbf8fc698] Invalid chars 'b' at the end of expression '6db' [22:46] <durandal_1707> ubitux: bingo!! [22:46] <ubitux> 'B' [22:46] <ubitux> not 'b' [22:50] <durandal_1707> ubitux: where it that offending strtod? [22:50] <ubitux> did you try "6dB" instead of "6db" anyway? [22:51] <ubitux> git grep av_strtod i guess [22:51] <durandal_1707> but there is used strtod alone [22:51] <durandal_1707> dB works [22:51] <ubitux> then it's ok? [22:52] <ubitux> "db" is not supposed to be supported [22:52] <durandal_1707> its does not red out [22:52] <saste> db is "deci-byte", iirc [22:53] Action: durandal_1707 hides [22:55] Action: durandal_1707 case sensitive hurts [23:00] <cehoyos> durandal_1707: See the log here for two samples [23:03] <durandal_1707> cehoyos: of what? (what log?) [23:04] <durandal_1707> cehoyos: what about patches you attach to bug reports and forget about it? [23:04] <cehoyos> See irc log of today (now) for the sample you requested five minutes ago on ffmpeg-devel [23:04] <cehoyos> I don't think I ever forgot a patch, which one do you mean? [23:04] <durandal_1707> aha, logs are updated in real time? [23:05] <durandal_1707> cehoyos: one of unitialized stuff g guy reported [23:05] <cehoyos> As exaplained, I don't know if there is a bug, if you believe there is one, please push. [23:07] <durandal_1707> cehoyos: i dont have link for softpulldown samples, because i dont have log of that, because i joined channel too late, so i will get it next day hopefully when new log arrives in archive [23:09] <cehoyos> Softpulldown?? [23:09] <durandal_1707> forget it, whatever log you are talking about, i do not have access to it [23:11] <cehoyos> durandal_1707: http://174.84.194.75:10000/lagarith-blueinput-rgba.avi http://174.84.194.75:10000/lagarith-redinput-rgba.avi [23:11] <cehoyos> I did not see at what time you arrived. [23:12] <durandal_1707> aha.... [23:15] <durandal_1707> you could upload that server, but anyway if decoded output matches ref decoder it should be ok [23:15] <durandal_1707> *to server [23:16] <durandal_1707> cehoyos: you have any samples that works with mplayer -vf softpulldown ? [23:17] <durandal_1707> the one that does not use lavf demuxer, afaik that one does not set all flags [23:18] <durandal_1707> eg samples that make use of soft telecine [23:23] <durandal_1707> ubitux: how timeline works with volume filter (or any audio filter) [23:23] <durandal_1707> ? [23:24] <Daemon404> i still find it amusing that rff flags have to be hadled by an external filter. [23:24] <ubitux> durandal_1707: just like the video filters [23:25] <ubitux> durandal_1707: -af "volume=volume=-6dB:enable='between(t,2,5)'" [23:25] <ubitux> 'works for me [23:25] <durandal_1707> ubitux: its sample accurate? [23:26] <ubitux> frame accurate [23:26] <durandal_1707> that is not that useful.... [23:27] <durandal_1707> Daemon404: for people that still watch stuff on analog TV (in 2013) [23:27] <Daemon404> rff flags are no for that [23:27] <Daemon404> they need them to accurately reproduce, e.g. hybrid dvds [23:27] <Daemon404> VFR dvds, etc [23:28] <saste> should i still work on my per-frame evaluation volume thing? [23:29] <ubitux> saste: would be nice [23:29] <ubitux> av_eval_expr_is_const! [23:30] <durandal_1707> //FIXME replace the crap with get_bits(): in mpegvideo_parser.c [23:30] <saste> ubitux, i noticed that when people ask me something, it usually takes more time for me to do it [23:30] <saste> it looks like *work* ;-) [23:30] <ubitux> :( [23:30] <ubitux> you asked for it here [23:31] <saste> uhm no av_eval_expr_is_const is a different problem [23:31] <ubitux> it's somehow related [23:31] <saste> (and it implies some boring work, testing, benchmarking, etc.) [23:31] <ubitux> since it can allow to disable it automatically [23:32] <saste> i would disable per-frame evaluation by default [23:32] <saste> similar to what i did in overlay [23:32] <saste> indeed it can be done in half an hour [00:00] --- Mon May 6 2013

1 0

ffmpeg.log.20130505
by burek 06 May '13

06 May '13

[04:19] <slavik262> Stupid question - I'm trying to upload sample files to the FTP server to go along with a bug report, but I keep getting "553 Can't open that file: Permission denied" [04:20] <slavik262> am I going about it the wrong way? [08:22] <DMJC> is there a really good tutorial for implementing a video player using libavformat etc? [08:23] <DMJC> I'm trying to fix the homeworld 1 sourcecode. Someone's implemented an avformat based video player for BINK animations but it's crashing when trying to run. [08:24] <DMJC> I think it's got something to do with the code that determines the format of the file. [08:57] <DMJC> ok I've now got video playback working [08:57] <DMJC> but when I try to play a second video it fails [08:57] <DMJC> debug output it looks like the second run through isn't working [09:22] <DMJC> completely stumped [09:37] <DMJC> can I put a log file up for someone to look at? [09:42] <DMJC> http://filebin.ca/g1uUgyVaMSl/homeworld.log [09:42] <DMJC> it's a 1.6mb xml format log from valgrind [10:55] <DMJC> Fixed it :D [10:55] <DMJC> someone forgot to initialise to zero... bad programmer.. [11:01] <zap0> someone still writes C code like its 80's ? [14:25] <alex88_> hi guys, someone ever tried to open a vsd file? [14:25] <alex88_> It's from a security camera [15:55] <yns> I'm wondering where to find documentation on ffmpeg's -bufsize flag? [15:56] <JEEB> it sets the vbv bufsize (that is then used in conjunction with the vbv maxrate) [15:57] <JEEB> it is the amount of buffer over which the maxrate is calculated, and thus has to preferably match between the encoder and the decoder [15:57] <yns> thanks for your answer but iew, newbie here [15:57] <yns> vbv? [15:58] <JEEB> vbv is a model basically, techy stuff that you really don't have to know that much about in most cases [15:58] <yns> okay [15:58] <yns> basically I'm trying to change the datarate of a video [15:58] <yns> so -b:v [15:58] <yns> but quicktime doesn't play the video anymore afterwards [15:59] <yns> tought bufsize could help maybe [15:59] <JEEB> then it's most probably up to something else. maxrate and bufsize are mostly used when you are making something for consumption over a limited amount of bandwidth (such as networks or removal discs) [16:00] <JEEB> post a full pastebin with the actual command you used, and the terminal output [16:00] <JEEB> and link it here [16:00] <yns> okay [16:01] <yns> http://pastebin.com/ZvjCgtgJ [16:02] <JEEB> I have this lingering thought that QuickTime just doesn't support MJPEG like that :P [16:03] <yns> No it does :) [16:03] <yns> Or doesn't it support *this kind* of MJPEG? [16:03] <yns> Other MJPEG's played back, without fiddling with the -b:v tag? [16:04] <yns> -? [16:04] <JEEB> so you have created MJPEG videos with this exact setup and played it with QT? [16:05] <yns> Without the -b:v part yes [16:05] <JEEB> it's also rather short of a test, but I am pretty sure libavcodec's MJPEG encoder isn't exactly applying your rate control at all :P [16:05] <JEEB> yns, same container output? [16:05] <yns> ah no sorry [16:05] <JEEB> yeah [16:05] <yns> that was different yess [16:05] <yns> it was .mov [16:05] <JEEB> yeah, I'm pretty sure that's it :P [16:05] <yns> don;t need the mp4 anyway [16:05] <yns> oh ok :) [16:05] <yns> So basically: [16:06] <yns> I'm trying to use the video as a source for some creative programming, and I think my HDD speed is too slow (the video playback lags). So I thoght to reduce the datasize (which I assumed must be bitrate) [16:06] <yns> -b:v is the way to go you think? [16:06] <JEEB> umm [16:06] <JEEB> while that is a 20mbps source, I'm pretty sure it's not your HDD that is being the bottleneck [16:07] <JEEB> also looking at those 31 frames' log the -b:v isn't doing /anything/ [16:07] <yns> Oh :( [16:07] <yns> Actually I'm playing more video's at once [16:07] <yns> So therefore I thought it could be my HDD [16:08] <JEEB> also if you want to make the thing less HDD-intensive I have NFI why you're re-encoding it into a format that is less CPU-intensive while can be huge as hell for certain rather obvious reasons [16:09] <JEEB> anyways, I'd recommend you check playback with a few apps and see what exactly is your bottleneck :P [16:09] <yns> hm, well with apps it all works fine [16:09] <yns> and using h.264 I had the idea CPU was the problem [16:10] <JEEB> "with a few apps" I meant that if you're having trouble with playing something back, try that scenario with various other playback apps (mpv/VLC f.ex.) and then monitor your system to see what exactly is your bottleneck :P [16:11] <yns> hm ok [16:13] <yns> Well yeah turns out to be like I thoght [16:14] <yns> h.264 takes a little more that 50% CPU usage, so playing 2 could be problematic, HDD usage is really low [16:14] <yns> And MJPEG is liek 10% CPU but 60% HDD usage [16:17] <yns> So I guess trying to lower the HDD usage is what I'd like to try [16:18] <yns> But since -b:v doesn't work, what can I try to lower datarate? [16:25] <JEEB> yns, setting the compression "quality" lower, I guess? [16:26] <yns> -qscale:v ? [16:27] <JEEB> or just -q:v , the values go from 1-31 it seems [16:27] <JEEB> higher is higher quantizer [16:27] <JEEB> and thus more compression, less quality [16:28] <yns> thanks, where did you find that info? documentation says "The meaning of q is codec-dependent" is there a list of codecs with meanings of q somewhere? [16:29] <JEEB> from a random post on the internet, you would probably have to dive into the source code or the specification of a format to find the actual limits of quantizer values :) [16:30] <yns> hehe okay [16:30] <yns> I'll just see what 1-32 gives me:) [16:32] <yns> Data rate stays pretty high (11Mb/s) with -q:v 8, the video starts to look ugly already. Other ways to go? [16:34] <JEEB> use formats that compress better, like H.264? there you would use the -crf option instead of -q btw. possibly use -tune fastdecode with the libx264 encoder [16:35] <yns> hm okay [16:39] <yns> any documentation on the -crf flag? [16:40] <JEEB> there was a way to pull out the libx264-specific help, but I don't remember it :P [16:40] <JEEB> it's a straight mapping to the libx264's crf parameter [16:40] <JEEB> which is the closest we have to "constant quality" right now with video encoders afaik [16:41] <JEEB> values go from 1 to 51 (actually higher than that, but the last ones are 51 + filtering for exceptional cases) [16:41] <JEEB> you can start from 23 or so [16:42] <yns> the lower the better quality? [16:42] <JEEB> yes [16:42] <yns> and the libx264 encoder is automatically being used when converting to h264? [16:42] <JEEB> there is no other encoder for H.264 in libavcodec atm [16:43] <yns> okay just being a newbie here :) [16:45] <yns> I get really ugly results when reducing the dimensions with h264 (with -s), that something normal with h264? [16:48] <yns> Doing this ffmpeg -i /Users/yns/Image/Samples/SpoutingHorn-720.mp4 -s 854x480 -vcodec h264 -g 1 /Users/yns/Image/Samples/SpoutingHorn-480_h264_kfrm.mp4 [16:48] <yns> Lowering quality after this should be disastrous [16:51] <JEEB> ugh @ -g 1 [16:51] <JEEB> why are you using that? [16:51] <JEEB> also why aren't you setting a -crf value (23 should be default)? [16:51] <JEEB> (and then you should be able to turn it up and down) [16:51] <JEEB> I mean, if you want to compress I'm pretty sure intra-only compression isn't exactly the best thing to do :P [16:53] <yns> hm, I'm using -g 1 so I can acces random points in the video quickly, for the creative coding stuff [16:53] <yns> intra-only compression? [16:55] <yns> Ah yeah, even not resizing the dimensions makes it ugly as hell. Is that the -g 1? [16:55] <JEEB> show me the full log :P [16:55] <yns> :P [16:56] <JEEB> I just want to see what the hell you're doing badly [16:56] <JEEB> so another pastebin pl0x [16:57] <JEEB> also make sure you're not only testing with quicktime, as many versions of it have bugs with decoding completely valid H.264, although it shouldn't really be "ugly as hell" [16:57] <JEEB> also, no. If you were doing it right -g 1 would only bloat your bit rate [16:57] <JEEB> crf would more or less control the quality [16:57] <yns> http://pastebin.com/TUhUcrns [16:59] <JEEB> generally seems fine, crf gets used and set to the default of 23... [16:59] <JEEB> if it really just looks bad, then lower the crf value [16:59] <yns> hm [17:00] <yns> Well yeah but I wanted to compress :( [17:00] <JEEB> then drop the fucking -g 1 or set a slower -preset [17:00] <JEEB> http://mewiki.project357.com/wiki/X264_Settings#preset list of presets [17:01] <JEEB> medium is the default [17:01] <yns> Hm okay, Maybe it doesn't look that "bad as hell" it just looks grainly, like there are small dots all over the screen [17:02] <JEEB> also I'm pretty damn sure you don't want to re-encode the audio in that file [17:02] <JEEB> might want to set -c:a copy [17:03] <yns> http://postimg.org/image/u11q0ecgv/ [17:03] <yns> no wait [17:04] <yns> no the screenshot is accurate [17:04] <yns> it's just less anoying when the image aint moving [17:05] <yns> but -c:a is not going to do anything for video q right? [17:05] <JEEB> no, it will not [17:05] <JEEB> it will just tell ffmpeg to copy over the audio, and not re-encode [17:05] <yns> hm [17:06] <yns> Well I guess I should conclude h264 is not that much better in compression :) [17:06] <JEEB> ... [17:06] <JEEB> I don't know where to start to comment on that sentence [17:06] <yns> haha [17:06] <yns> sorry [17:07] <JEEB> ok, you have a baseline 1280x720 file that most probably is not all-intra and depending on how sane the encoder was that did it, either looks somewhat better or somewhat worse [17:07] <JEEB> you hit it with crf 23 and a one-frame GOP [17:07] <JEEB> with preset medium [17:08] <JEEB> and haven't even encoded with it to the end to know the full result [17:08] <JEEB> (to know the full end average rate you got with that crf value) [17:09] <yns> okay that seems logical indeed. what do you mean by that last sentence? that i havent tried everything yet? [17:10] <yns> -crf looks fine indeed [17:10] <JEEB> you are basically cutting the hands and legs off the format (making it intra-only), and not even checking the slower presets, and then crying that it doesn't compress automagically better than a relatively compressed file that by all means that I know of isn't all-intra [17:10] <JEEB> and then you fucking cry out that "H.264 is not that much better in compression" [17:11] <JEEB> until we get good HEVC encoders, H.264 (esp. with libx264) is going to be the best compression you're going to get from anything :P [17:11] Action: sacarasc gives JEEB a hug. [17:12] <yns> oh sorry man didn't mean to offend you. I'm just a huge newbie when it comes to video [17:12] <JEEB> you didn't offend me per se, I'm just facedesking with all of my might at some of this stuff [17:13] <yns> well anyway I'm sorry. [17:13] <yns> what's all-intra, and intra-only [17:14] <JEEB> same thing, if all is intra then it is also intra-only, and it means that you are not using inter prediction at all (predicting between frames) [17:14] <JEEB> ffmpeg -i welp.input -c:v libx264 -crf 23 -preset veryslow -g 30 -c:a copy out.file , try this at least :P [17:14] <yns> okay :) [17:16] <yns> trying [17:17] <yns> the all-intra is a result from adding keyframes to every frame with -g 1? [17:17] <JEEB> -g 1 means that your GOP is of length of one frame, and then after that one frame at max you have to put out another random access point (usually in case of libx264, an IDR frame) [17:18] <yns> which results in all-intra? [17:18] <JEEB> well, since your GOP is of length one frame then you can't really have any inter frames if the next frame is going to be of another GOP already :P [17:19] <yns> hehe ok :) [17:19] <JEEB> which leads to all frames being intra-only because there's zero frames that then predict from them [17:19] <yns> ah ok [17:19] <yns> i see the veryslow preset is indeed very slow :) [17:20] <yns> wat else is it trading of for the higher quality? [17:20] <JEEB> anyways, I'm not sure what you need is attainable within your needs, partially because I'm tired as hell to try and guide you through it. If you need something that's quickly decode'able, then use something like that (ffvhuff, mjpeg, ut video, whatever). But do expect that if you already have somewhat compressed stuff, you're not going to *compress* them as well [17:20] <yns> or is it just compressing slower? [17:21] <JEEB> or heck, x264 as well, but in any case you're not going to get any magical pixie dust compression [17:22] <yns> okay well thanks a lot for your help, i've learned a lot! basically setting the preset to veryslow makes it encode slower? (which is actaually no problem for me) [17:22] <JEEB> yes, it will use slower algorithms and more data to compress [17:22] <JEEB> in other words, better decisions are made [17:23] <yns> okay [17:23] <JEEB> and also of course it enables the use of certain features in certain cases in case they were decided to be disabled with faster presets [17:23] <yns> the file is really small indeed (just saw the result of your suggested command) but it lags a bit even when playing it back in quicktime? [17:24] <yns> (the quality of the file looks marvelous as well by the way) [17:25] <JEEB> dunno why it'd lag unless quicktime sucks a dique, or your CPU sucks *shrug* [17:25] <yns> hehe [17:25] <yns> well it's not that bad, 2.4Ghz Core2Duo [17:26] <yns> cpu usage is also only about 20% [17:26] <yns> well anyway [17:26] <yns> thanks a lot for your help [17:26] <yns> maybe one day I find the compression settings that do it for me :) [17:26] <JEEB> anyways, QuickTime *does* indeed have random bugs and other crap [17:27] <JEEB> so if you're limited to it I wish thee good luck [17:27] <JEEB> you'll most probably have to cut off a lot of alternatives just because "QuickTime doesn't like it" [17:28] <yns> Well I'm using openFrameworks which uses QuickTime API to playback videos, like most creative coding environments on OSX [17:28] <yns> I don;t have the luxury of "just the desire to watch a nice movie" [17:29] <JEEB> I wasn't even noting that, I'm just noting that you'll have "plenty of fun" with the QT itself and its framework :P [17:29] <yns> hehe :P [17:29] <JEEB> and any good result you might find might not be optimal purely by the reason of "QuickTime decided to pee on your legs" [17:29] <JEEB> just keep that in mind [17:30] <yns> it is what it is. actually I'm at the point of giving up video, concludng my computer is just too slow [17:31] <sacarasc> What specs is your computer/ [17:31] <JEEB> 20% CPU usage does sound like you should be able to do more, though. But suit yourself :) [17:31] <yns> 2.4Ghz Core2Duo, 5400RPM HDD, 4GB RAM [17:31] <yns> 256MB VRAM [17:33] <JEEB> I'd probably use something like ffvhuff or Ut Video or something if I had to make an app that shows multiple clips at the same time, but neither are supported by QT so (Ž [17:34] <yns> Hm [17:35] <yns> I was told MJPEG is the way to go for fast scrubbing and stuff [17:35] <yns> The thing is that playing back the video is just a little beginning in what I want to do [17:36] <JEEB> anything that is intra-only or has short GOPs is fast to seek in [17:36] <yns> I want to acces the pixels change their rgb values, look at what coordinates there are all-white pixels and do somehting with that. that kind of stuff [17:36] <JEEB> not in any way limited to MJPEG [17:37] <yns> No but since I wan't to that crazy stuff I really need something that is super easy on CPU for just playing back the file [17:37] <JEEB> then all of those I mentioned go as well [17:37] <yns> those things I describe eat CPU like hell, and HDD is not much of a problem in that case [17:37] <JEEB> that said, HDD and "what is supported by lolQT" become the problem [17:37] <yns> yes but they are not supported by QT :) [17:37] <JEEB> Ut Video actually is [17:38] <yns> which == "no option" [17:38] <JEEB> it has a decoding component for the QT framework [17:38] <yns> oh ok [17:38] <yns> ok than I'll have to look into that [17:38] <yns> But I have spent DAYS already converting stuff [17:38] <yns> I'm tired of it [17:38] <JEEB> it's a lossless format that's all-intra and consists of simple huffman coding [17:39] <JEEB> will eat space (much less than raw video naturally), but should otherwise be within your area :P [17:39] <yns> okay :) [20:24] <Astin> Hello [20:24] <Astin> I installing ffmpeg on ubuntu [20:24] <Astin> why ERROR: libass not found [20:25] <Astin> I installed libass-dev [20:25] <Astin> T.T [20:25] <JEEB> see the config.log [20:25] <Astin> Thank you [20:26] <Astin> I saw it [20:26] <Astin> but [20:26] <Astin> just ./configure ~~~~ _decoder [20:28] <Astin> last [20:28] <Astin> gcc -Wl,--as-needed -o /tmp/ffconf.Ke9NC7MQ /tmp/ffconf.zVyCDmHn.o -lm -pthread -lz -lrt [20:28] <Astin> check_pkg_config libass ass/ass.h ass_library_init [20:28] <Astin> ERROR: libass not found [20:28] <JEEB> pastebin the whole config.log [20:28] <JEEB> and link here [20:29] <Astin> yes [20:29] <Astin> I delete and [20:29] <Astin> I reconfigure [20:29] <Astin> I installed libass-dev on ubuntu T.T [20:31] <Astin> I'm trying http://forum.serviio.org/viewtopic.php?f=7&t=9511 [20:32] <JEEB> why can't you just post on the pastebin the config.log when you configure with --enable-libass :P [20:32] <JEEB> and then link it here [20:33] <Astin> okay [20:34] <Astin> Thank you [20:34] <Astin> I don't know about pastebin [20:34] <Astin> but i got it now T.T sorry [20:34] <JEEB> pastebin is a type of site nowadays, the most well-known is pastebin.com [20:34] <JEEB> you can post pieces of text that aren't just a single line in there :P [20:35] <Astin> http://pastebin.com/EkBUuGEe [20:35] <Astin> here T.T [20:35] <Astin> and Thank you [20:36] <JEEB> :s [20:36] <JEEB> oh, it just checks with pkg-config? [20:36] <Astin> pkg-config? [20:37] <JEEB> if you try to run pkg-config --version in a terminal, does it say that you don't have such a thing, or does it properly show a version? [20:37] <Astin> 0.025 [20:38] <Astin> 0.25 [20:38] <JEEB> ok [20:38] <JEEB> well, all I can say is that the libass-dev package should have the .pc file :s [20:39] <JEEB> what Ubuntu are you on? [20:39] <Astin> 12.04 [20:39] <Astin> Thanks for your answer . [20:40] <Astin> Um .. When I installing libass manually [20:40] <Astin> configure: error: Package requirements (freetype2 >= 9.10.3) were not met: [20:40] <Astin> No package 'freetype2' found [20:40] <Astin> Consider adjusting the PKG_CONFIG_PATH environment variable if you [20:40] <Astin> installed software in a non-standard prefix. [20:40] <Astin> Alternatively, you may set the environment variables FREETYPE_CFLAGS [20:40] <Astin> and FREETYPE_LIBS to avoid the need to call pkg-config. [20:40] <Astin> See the pkg-config man page for more details. [20:41] <JEEB> means that you don't have freetype's development stuff :P [20:41] <Astin> How can I .. [20:42] <Astin> googling T.T [20:48] <Astin> I don't have freetype2.pc [20:48] <Astin> on /usr/lib/pkgconfig [20:50] <Astin> Thanks JEEB [20:50] <Astin> with your help [20:50] <Astin> I got it [22:40] <mapreduce> I've compiled ffmpeg from the latest on github, and calling avcodec_find_decoder_by_name("mjpeg") gives me NULL. How can I see what codecs ffmpeg was compiled with? [22:46] <c_14> ffmpeg -codecs [00:00] --- Mon May 6 2013

1 0

ffmpeg-devel.log.20130504
by burek 05 May '13

05 May '13

[01:21] <cone-356> ffmpeg.git 03Michael Niedermayer 07master:64819bfc7a1f: snow: Fix off by 1 error in reference picture management [03:14] <cone-356> ffmpeg.git 03Michael Niedermayer 07master:f18d2dff1194: evrcdec: fix sign error [10:12] <cone-39> ffmpeg.git 03Christophe Gisquet 07master:5a97469a4fd1: x86: sbrdsp: Implement SSE2 qmf_deint_bfly [10:12] <cone-39> ffmpeg.git 03Christophe Gisquet 07master:8394d9a676db: sbrdsp: Unroll sbr_autocorrelate_c [10:12] <cone-39> ffmpeg.git 03Christophe Gisquet 07master:4a7af92cc80c: sbrdsp: Unroll and use integer operations [10:12] <cone-39> ffmpeg.git 03Michael Niedermayer 07master:711c8ee71d71: Merge commit '4a7af92cc80ced8498626401ed21f25ffe6740c8' [10:30] <cone-39> ffmpeg.git 03Diego Biurrun 07master:a5f8873620ce: silly typo fixes [10:30] <cone-39> ffmpeg.git 03Michael Niedermayer 07master:0104570fb65b: Merge commit 'a5f8873620ce502d37d0cc3ef93ada2ea8fb8de7' [10:38] <cone-39> ffmpeg.git 03Luca Barbato 07master:69467fb64ba4: wav: return meaningful errors [10:38] <cone-39> ffmpeg.git 03Michael Niedermayer 07master:23a7c32a7ade: Merge remote-tracking branch 'qatar/master' [13:27] <cone-39> ffmpeg.git 03Hendrik Leppkes 07master:d18341fb1121: aacdec: free frame buffer if no audio was decoded [15:05] <cone-39> ffmpeg.git 03Michael Niedermayer 07master:e9e207ece7a2: dfa: Put our pointer check back. [16:58] <cone-39> ffmpeg.git 03Michael Niedermayer 07master:c7a760565663: wmaprodec: Fix null pointer dereference in decode_frame() [17:11] <cone-39> ffmpeg.git 03Nicolas George 07master:cc24afb45cfd: lavf: call the new seek API from the old. [17:11] <cone-39> ffmpeg.git 03Nicolas George 07master:3fc66dfef22c: lavf: filter out AVSEEK_FLAG_BACKWARD in new API. [17:11] <cone-39> ffmpeg.git 03Nicolas George 07master:d608a27d9e28: lavf/concatdec: remove invalid check for AVSEEK_FLAG_BACKWARD. [17:11] <cone-39> ffmpeg.git 03Nicolas George 07master:fa245e432bdc: lavf/matroskaenc: return an error for unsupported types. [17:11] <cone-39> ffmpeg.git 03Michael Niedermayer 07master:30db867cb7be: Merge remote-tracking branch 'cigaes/master' [19:06] <megaTherion> Hi [19:23] <megaTherion> how would I decode an AVFrame*? I've read the example in decoding_encoding.c however they use a complete different approach (by reading the file in via fstream rather than ffmpeg functions) [20:11] <michaelni> megaTherion, a AVFrame holds decompressed data, a AVPacket holds the compressed frame [20:32] <cone-578> ffmpeg.git 03Michael Niedermayer 07master:c40f51e15b91: h264: Always decode MMCOs into temprary array [21:12] <j-b> good afternoon [21:13] <j-b> michaelni: ping [21:16] <michaelni> j-b, pong [21:16] <j-b> michaelni: will you do us an invoice for the VC1i ? [21:16] <megaTherion> michaelni: oh I see, so the AVFrame is decompressed already - by accessing data I can store it right into a picture? [21:16] <michaelni> j-b, carl found another bug [21:17] <michaelni> in vc1 [21:17] <j-b> so? [21:17] <j-b> michaelni: http://wiki.videolan.org/Bounties#VC-1_interlaced [21:19] <michaelni> hmm [21:20] <michaelni> j-b, there was also at least 2 or 3 patches related to interlaced vc1 on ffmpeg-dev that where from someone else [21:20] <megaTherion> michaelni: by using av_read_frame I get the AVPacket's correct? then I use avcodec_decode_video2 to decompress them? [21:20] <j-b> michaelni: so? [21:20] <j-b> michaelni: B-frame is now in? [21:20] <j-b> Claudio: Y U NO split patches? [21:21] <michaelni> j-b, B frame and interlacing should be supported but iam not the only one who made that happen [21:22] <michaelni> so iam not sure about "just" taking the money for that, but the technical side of writing an invoice would be no problem [21:22] <michaelni> and theres at least one bug left in there [21:22] <j-b> if you don't want to claim the money, don't [21:22] <j-b> but it will not be said, that we don't hold to our words [22:02] <michaelni> j-b, is it ok if i send you an invoice over 500euro, and the other 500 would then be for everyone else who worked on vc1i (Sebastian Sandberg, nevcairiel, possibly others)? And if no one claims the other 500 euro then you could consider that part a donation to videolan [22:03] <j-b> michaelni: https://github.com/mstorsjo/libav/commits/vc1i seems that you did most of it [22:03] <j-b> michaelni: but yes, if you can do a proper invoice, please do [22:08] <michaelni> j-b, what name/company should i put on the invoice ? "videolan videolan street 1" ? and do you need it signed & scanned or is gpg signed email of some openoffice generated pdf ok ? [22:08] <j-b> michaelni: http://www.videolan.org/contact.html [22:09] <j-b> a pdf is enough [22:40] <ubitux> the aac improvements look interesting :) [22:42] <michaelni> j-b, sent invoice for 500 euro [23:10] <megaTherion> If I get multiple "header damaged" on avcodec_decode_video2 (for mpeg4) what would that mean? Am I doing something wrong? I dont get them if I try converting the video via ffmpeg cli. [23:18] <michaelni> megaTherion, do you pass the same stuff into avcodec_decode_video2() that ffmpeg does ? [23:19] <megaTherion> michaelni: same video if you mean that - I do, fixed my previous issue and now decoding works - I counted the frames decompressed are 2872, 13 are lacking [23:23] <ubitux> btw, can we consider the aac improvements a good reason to remove the experimental bit from the encoder? [23:23] <ubitux> or there is some other blocking stuff? [23:24] <nevcairiel> if the quality is in an actual usable state now, i'm all for it [23:25] <michaelni> megaTherion, i meant is the same data/packets/bytes passed in, you said it behaves different to ffmpeg CLI [23:26] <megaTherion> michaelni: I cant tell you, Im bloody beginner if it comes to ffmpeg devel - I just did a fast ffmpeg -i source.mp4 -vcodec mpeg4 dest.avi and there were no such errors... [23:32] <michaelni> megaTherion, you could just dump the bytes with av_log() to stderr and compare if ffmpeg and your code feed the same to avcodec_decode_video2() [23:41] <megaTherion> michaelni: it seems, probing for the stream index while(av_read_frame(...)) skips these messages [23:41] <megaTherion> so I guess I tried decoding frames which didnt belong to the video stream.... [23:41] Action: megaTherion has much to learn yet [23:54] <megaTherion> michaelni: what I dont understand, the nb_frames within AVStream* have a higher number - but if I probe the packets then and they are skipped (due not belonging to the video stream) - why are they within the same data in the first way? [23:54] <megaTherion> Shouldnt I only get packets belonging to my chosen (video) stream anyways... [00:00] --- Sun May 5 2013

1 0

ffmpeg.log.20130504
by burek 05 May '13

05 May '13

[00:00] <relaxed> How is it not specific? [00:00] <Jorky> for instance album mettallica-ride the lightning album (mp3) to metallica-ride the lightnig album (wma) but 128k wma [00:00] <sacarasc> The thing about that script, Jorky, is that it will do exactly what you asked on as many different albums as you want. [00:01] <Jorky> but in what quality? [00:01] <relaxed> Jorky: for i in *mp3; do ffmpeg -i "$i" -c:a wmav2 -b:a 128k "${i%.*}".wma; done [00:01] <Jorky> ok [00:01] <Jorky> now I get it [00:01] <Jorky> thnx [00:01] <Jorky> my native language isn't english so I need more time and different words to understand [00:23] <satonio> hi [01:28] <brontosaurusrex> how would i scale to say 16px X 1px and export to some sort of filmstrip vertical image? [01:28] <brontosaurusrex> input is video [02:10] <cypod> Hello [02:10] <cypod> I had a question about compiling ffmpeg [02:11] <cypod> I have build libvpx to a custom location and want to using that version instead of the system default [02:18] <llogan> cypod: see --extra-cflags and --extra-ldflags [02:18] <llogan> for ffmpeg [02:19] <cypod> so would that be --extra-ldflags=-L/path/libvpx/lib ? [02:22] <llogan> cypod: looks good to me [02:26] <cypod> llogon : and the plug-in I'm trying to load is libvpx-v1.1.0/libvpx.so which points to libvpx.so.1.1.0 [02:42] <brontosaurusrex> how would i pipe from ffmpeg to imagemagick convert? [02:44] <brontosaurusrex> i'd like this: [02:44] <brontosaurusrex> ffmpeg -i in.mp4 -sws_flags neighbor -vf scale=2:2 -f image2 out-%05d.png && convert +append out-*.png -resize 1x1\! -resize 1x64\! out.png [02:44] <brontosaurusrex> as a pipe [02:45] <cypod> is there anyway to load the libvpx plugin without recompiling ffmpeg? [02:45] <brontosaurusrex> cypod, is that vp8 encoder? [02:46] <cypod> I think so, hoping to create some webm [02:46] <brontosaurusrex> yeah, i played with that [02:46] <brontosaurusrex> It ended up with piping from ffmpeg to vpxenc [02:46] <brontosaurusrex> or whatever its called [02:47] <brontosaurusrex> because vpx cli was like 123412536 chars long and i just couln't convert that to ffmpeg mappings [02:47] <brontosaurusrex> i can find an example if you wish [02:50] <llogan> brontosaurusrex: maybe "ffmpeg -i input -f image2pipe - | convert [input options] - [output options] ..." [02:50] <brontosaurusrex> llogan, let me try, i think that was the first i tryed [02:51] <cypod> where to get vpxenc ? [02:52] <cypod> think I found it there in the build dir, yes examples please [02:53] <brontosaurusrex> cypod, hold on [02:56] <brontosaurusrex> cypod, a two-pass example http://paste.debian.net/2000/ [02:58] <brontosaurusrex> actually thats a sort of quality-based encode [02:58] <brontosaurusrex> irc [02:58] <brontosaurusrex> anyway, i remember it was a pain to figure it out [02:59] <cypod> in that examle is vp8 the encoder or is it vpxenc? [03:01] <cypod> was thinking it would be : vpxenc --best -o source.mp4 -w 1920 -h 1080 newfile.webm [03:02] <brontosaurusrex> its vpxenc [03:03] <brontosaurusrex> thats from a script [03:03] <brontosaurusrex> ~ [03:04] <cypod> maybe get better result with -p 2 [04:38] <voidah> m [04:38] <voidah> oops [04:39] <voidah> is there a way to take an audio file as input (.mp3) and "shink" it so that it become of an exact specified length? [04:39] <voidah> I have a .mp3 file that is 1m23 long, and would like to shrink it to 1:21 [04:57] <voidah> said otherwise: I would like to play my sound faster [04:59] <voidah> :w [04:59] <voidah> oops [05:12] <voidah> ah https://lists.ffmpeg.org/pipermail/ffmpeg-devel/2013-April/142777.html [05:12] <voidah> seems to be the same issue i'm having [05:12] <voidah> aptempo might be ok finally [05:12] <voidah> recompiling [09:24] <samek_> Hi, I've got two video live streams put on static image and i want to sync audio with only one video (second video is without audio), how would i sync audio with correct video, since it seams it's syncing it to the second one. [09:29] <samek_> right now i have /usr/local/bin/ffmpeg -y -f mpegts -i /tmp/SPOT.pipe -f mpegts -i /tmp/PIPE.pipe -f image2 -loop 1 -i /tmp/SIT.png -map 0:v -map 0:a -async 2 -map 1:v -map 2:v but's it's not syncing ok. [09:41] <smj> is FAAC better than LAME, as it should? [09:42] <Keshl> Depends what you mean by "better: [09:42] <Keshl> *" [09:43] <smj> better quality per bitrate [09:43] <Keshl> FAAC outputs .faac files, which are lossless. Every detail you put into them is preserved and compressed, so it sounds exactly the same on the way out. [09:43] <Keshl> However, files and bitrates are substantally larger than LAME. [09:43] <Keshl> LAME is lossy, but it most humans will not hear a difference. [09:43] <Keshl> Outputs .mp3 files. [09:43] <smj> are you talking about FLAC? [09:43] <Keshl> .. [09:43] <Keshl> Yes, yes I am. [09:43] <Keshl> Ignore me. [09:43] <Keshl> >w> [09:43] <smj> I will [09:44] <Keshl> Danks. <É< [09:44] <Keshl> For the record though, the standard says that h.264 uses .mp4 which uses FAAC. [09:44] <Keshl> Just saying. [09:44] Action: Keshl might be wrong. [09:45] <smj> how does a file format use an encoder? [09:45] <Keshl> ...Iunno D: -Hides in a corner- D: [09:45] Action: Keshl isn't the smartest tatertot here. [09:45] <TheSchaf> so i have to convert alot of mp3s from 441000 to 480000 and 128 kbit/s to 192 kbit/s and i did that using ffmpeg. is there any tips on how i can keep "quality"? i heard transcoding is bad :p [09:46] <smj> At least you bother to type well. [09:46] <smj> unlike me :I [09:46] <Keshl> People who dun need to be slap'd with fish, oÉo. [09:46] <Keshl> 'cept not mIRC fish. Them's bad. <É< [09:48] <Mavrik> smj, most tests show that FAAC (being a reference encoder) doesn't produce significantly better output than libmp3lame [09:48] <Mavrik> (at the same bitrate and settings) [09:48] <smj> oh, so it's a waste of CPU [09:48] <smj> ? [09:48] <Keshl> Depends. [09:48] <Mavrik> TheSchaf, why wuld you do a silly thing like that? :P [09:48] <Mavrik> smj, depends. Rather tell us what do you want to achieve. [09:48] <Keshl> Devices that follow standards are going to support .mp4's with FAAC in them always. [09:48] <Keshl> it's a guarentee. [09:49] <Keshl> They may or may not support mp3. [09:49] <TheSchaf> Mavrik, i'm creating content for a game [09:49] <TheSchaf> it only supports that one format :P [09:49] <Keshl> TheSchaf: Fix the game. D: [09:49] <TheSchaf> i dont have its source [09:49] <Keshl> Get it D: [09:49] <Mavrik> TheSchaf, it only supports one bitrate? [09:50] <TheSchaf> yes [09:50] <Mavrik> TheSchaf, that seems weard. I get supporting 48000 Hz sample rate though [09:50] <Mavrik> *wierd [09:50] <TheSchaf> its a music dance game [09:50] <TheSchaf> DDR like [09:50] <TheSchaf> so it has alot of dirty hacks to sync to audio [09:50] <Mavrik> TheSchaf, anyway, simce you're upsampling to a certain output bitrate and samplerate there's nothiing you can do [09:50] Action: Keshl twitches. ( http://hyperboleandahalf.blogspot.com/2010/04/alot-is-better-than-you-at-ev… ) [09:50] <Mavrik> you'll lose a little quality but there's no magic. [09:51] <TheSchaf> mhh, ok [09:51] <Keshl> basically yeah. Somehow data has to be created, that's going to be carried over to what humans hear, it's gunna make a difference. [09:51] <Mavrik> getting original music source in CD / losless format and transcoding that would be magic though ;) [09:51] <Keshl> There's no lossless way to do it, contrary to what common sense says. [09:51] <Keshl> Well, from a computing perspective it is lossless, but what you hear will be different nonetheless. [09:52] <Keshl> HopefullythatmakessenseImmagobacktoalesssmartchannelnowforaseckay<w< [09:52] <TheSchaf> no way for me to get the music in original format [09:53] <Mavrik> well you have no choice then [09:55] <smj> Mavrik, I'm trying to get the best possible stream quality out of my hardware. Until today, I thought twitch.tv supported only MP3 audio, but one clip I downloaded from there had AAC audio [09:57] <Mavrik> smj, yeah, they probably use AAC [09:57] <Mavrik> but that has little to do with your hardware. [09:58] <Mavrik> smj, encode audio in high enough bitrate so it sounds good and thats it [09:58] <Mavrik> smj, you can use fdk_aac encoder (the Fraunhofer one) which is currently the best of the bunch [09:59] <Keshl> oÉo... Well now, that's weird. Does ffmpeg handle video that barely changes from frame to frame well? [09:59] <Mavrik> Keshl, nothing to do with ffmpeg, everything to do with the encoder and settings. [10:00] <Mavrik> smj, but in reality, your bitrate and samplerate settings will make far greater difference than choice of mp3 or aac. Especially if you're not using ridiculusly low bitrates. [10:01] <Mavrik> Keshl, x264 with large GOP can probably encode those quite efficiently, especially with stillimage tuning [10:01] <Keshl> Mavrik: Kay, I'm doing "ffmpeg -r 60 -i Frame_%04d.png -i "pathto.wav" -vcodec libx264 -preset medium -crf 10 -movflags faststart -pix_fmt yuv420p OUT.mp4 -- This results in q=18 (my goal) on pretty much every frame. However, white specs randomly appear. [10:01] <Keshl> Turning the CRF up or down seems to fix them. [10:01] <Keshl> Just seems that, anywhere near 7-15 it adds white specs randomly. [10:01] <smj> my CPU isn't fast enough for 720p, so I'm encoding at 480p, and I want to keep audio bitrate around 100-120, is that low? [10:01] <Keshl> There's no white in my frames. [10:02] <Mavrik> smj, depends on kind of audio :) [10:02] <Mavrik> smj, we lived with 128kbps mp3s for a LONG time ;) [10:02] <Mavrik> does it sound right? if it does, yeah, that's ok :P [10:03] <Mavrik> Keshl, probably a question for x264 devs then. [10:03] <Keshl> Should also mention that, during encode, ffmpeg randomly sorta freezes. [10:03] <Mavrik> and be prepared to show a sample of input and output :) [10:03] <Mavrik> hm. [10:03] <Keshl> The time= suddenly climbs up /really/ fast, then holds still while frame= catches up. [10:03] <Keshl> But the video plays fine, audio's in sync and all, just the white specs randomly appear. [10:03] <Mavrik> Keshl, set fps on output as well. [10:03] <Keshl> Never had to before. oÉo [10:04] <Keshl> And what's weird is, it worked /just fine/ on similar video yesterday. [10:04] <Keshl> No updates. [10:05] <Keshl> Mavrik: Also, how do I do that other stuff you said, oÉo? [10:05] <Keshl> Maybe just by chance it'll fix it, might be useful to the x264 devs to know if it does. [10:05] <Mavrik> you say "-tune stillimage" [10:05] <Mavrik> if you have a still image [10:05] <Mavrik> and test that [10:05] <Keshl> It's not /still/, it's just that most of it stays static. [10:06] <Mavrik> anyway, sleep time. [10:06] <Keshl> Here, sec. It'll make more sense if I be shows, oÉo. [10:06] <Keshl> D'aww .É. [10:07] <smj> So how do I use this... fdk_aac? [10:08] <Fjorgynn> I just fdk_aac [10:08] <Fjorgynn> but you must compile it yourself [10:08] <smj> aha, --enable-libfdk-aac [10:08] <Fjorgynn> yeah [10:08] <Fjorgynn> and download it from git [10:09] <Fjorgynn> if you don't have it [10:10] <smj> https://github.com/mstorsjo/fdk-aac Is this the official one, if there is such? [10:15] <Keshl> Very interesting, -tune stillimage fixed it. [10:15] <Keshl> Practically doubled the fielsize, but it fixed it. [10:15] <Keshl> *filesize [10:17] <xxthink> how to enable aac decoding? [10:17] <xxthink> I use this configure script, but the result says the decoder does not support aac [10:19] <xxthink> http://pastebin.ca/2373091 [10:20] <xxthink> this is the output [10:20] <xxthink> http://pastebin.ca/2373093 [10:21] <xxthink> I debug for about 3 hours [10:21] <xxthink> and don't know the reason why aac decoder is not enabled [10:31] <xxthink> close --disable-fft --disable-lpc is ok [13:39] <xlinkz0> how can i see what PTS the first frame of the video has after i seek? [16:43] <nilsge> greetings [16:43] <nilsge> can I use ffmpeg to edit wav riff chunks? [16:43] <nilsge> fmt, fact, cue, smpl and others (see http://www.sonicspot.com/guide/wavefiles.html ) [18:59] <dax_roc> Evening all [19:00] <dax_roc> I'm trying to stream a mpegts to ffserver with ffmpeg -i /dev/video0 -re -vcodec copy -qp 0 -acodec copy -f mpegts tcp://$HOST:$PORT/feed1.ffm and it's failing with av_interleaved_write_frame(): Connection reset by peer [19:01] <dax_roc> The source is a happauge hdpvr (1st gen) [19:02] <dax_roc> If i stream via mpegts udp I can watch the stream no problem [19:07] <dax_roc> Anyone have any suggestions why http://bpaste.net/show/rhGR3LcN8yULX1Qdpl9x/ [19:59] <smj> has overlay filter been tested with stream inputs and inconsistent frame rates? it looks a lot like it goes by the lowest one [20:31] <dax_roc> Can I take a remote streem and feed it to a local dev and have another ffmpeg transcode it or can I do this directly ? [21:14] <thee_l> Hey [21:14] <thee_l> So I am wondering what signals can be used to close ffmpeg in a safe way. [21:15] <thee_l> I hope to script some streaming to justin.tv and it would really be great it I could use the kill command to shut down ffmpeg [21:15] <thee_l> *without worrying anything could go wrong [21:15] <thee_l> Basically a signal equivalent to pressing 'q' [21:16] <thee_l> anyone? [21:18] <thee_l> anyone know of any signals that are handled like a 'q' [21:25] <tonsofpcs> 14 maybe? [21:25] <tonsofpcs> (just guessing) [21:30] <thee_l> woops, sorry didn't realize you responded [21:31] <goo_> Hi. Is there a filter in ffmpeg corresponding to Levels in Photoshop? [21:31] <thee_l> @goo like: 'multiple' 'add' 'screen' [21:31] <thee_l> oh wait [21:31] <thee_l> nevermind [21:32] <thee_l> I think imagemagick might have what you need. [21:32] <thee_l> check it out http://www.imagemagick.org/script/index.php [21:32] <goo_> well, I was searching for that too but i found it's blend=all_mode=multiply [21:33] <goo_> Yes, but I was searching for something faster than imagemagick... [21:33] <thee_l> oh, I gotcha [21:34] <thee_l> I couldn't tell you, I don't know much about ffmpeg, but I'm pretty sure effects on single images isn't really it's specialty, I could be very wrong though [21:34] <tonsofpcs> imagemagick is pretty fast... [21:37] <thee_l> @tonsofpcs I guess I'm looking for a signal which is specifically used for the purpose of shutting down ffmpeg safely just like 'q' [21:37] <goo_> yes, it's fast but for video I thought there could be something more specific. [21:37] <thee_l> oh it's a video [21:37] <thee_l> I thought you were talking about individual images. [21:38] <thee_l> yeah so @tonsofpcs I am looking for something definite, thank you though [21:39] <goo_> yes, already tried to feed ffmpeg with Imagemagick filtered pictures but I think it's too slow [21:39] <tonsofpcs> goo_: what processing are you trying to do? [21:41] <goo_> I need to multiply a solid color, modify levels, brightness and contrast, overlay pictures [21:41] <goo_> i already find how to do most of this operations but i don't find levels, onli curves [21:47] <tonsofpcs> erm... brightness/contrast are just black/white curves... [21:47] <tonsofpcs> brightness is the black level - higher brightness and the lowest possible level is higher; contrast is the white level - lower contrast and the highest possible level is lower [21:51] <thee_l> I'm going... see ya, hope your processing works out goo_ and thanks anyway tonsofpcs [21:51] <thee_l> bye [21:52] <goo_> thank you thee_l [21:55] <goo_> Ok, @tonsofpcs I think I need to read some theory, thank you ;) [22:01] <tonsofpcs> goo_: that is the theory, there's nothing more to it. [00:00] --- Sun May 5 2013

1 0

ffmpeg-devel.log.20130503
by burek 04 May '13

04 May '13

[00:23] <ubitux> durandal11707: use comma for aliases [00:26] <durandal11707> ubitux: what? [00:27] <ubitux> for asetpts [00:27] <ubitux> @item S, NB_SAMPLES [00:27] <ubitux> otherwise it means 2 differents things with similar semantics [00:28] <ubitux> (such as x and y) [00:32] <cone-25> ffmpeg.git 03Michael Niedermayer 07master:2e2b6d7ab37d: fate: Adjust filter-trim-time parameters [03:24] <ubitux> michaelni: so you added a rootkit in sws with xyz support! you evil! [03:44] <michaelni> ubitux, for a moment you scared me ... ;) [03:56] <ubitux> yeah rest assured, it seems the rootkit wasn't really spotted yet [03:56] <ubitux> we just know there is one [04:11] <BBB> xyz is inherently evil, bc conversion requires linearization [04:11] <BBB> I wrote some code for that at some point, you guys just use luts I suppose? [04:13] <michaelni> yep, the code uses luts, and iam not too happy about that [04:15] <BBB> I tried other approaches and they were slower [04:15] <BBB> and yes they were simd [04:16] <BBB> with fairly limited precision [04:16] <BBB> but you can make the luts be generated through the simd, then it's pretty cool [04:16] <BBB> (luts in 19bit intermediates are kinda slow to generate) [04:20] <cone-25> ffmpeg.git 03highgod0401 07master:f685aee18f7a: lavu/opencl: fix the compile bug on linux [04:21] <BBB> if you want I'll see if I can release the code for that stuff [04:21] <BBB> shouldn't be a big deal [04:22] <michaelni> if its faster and or more precisse it would be interresting [04:23] <BBB> than luts? no it isn't :( [04:23] <BBB> assuming you make the luts 8bit through a general pow() floating point call [04:23] <BBB> maybe with haswell it gets faster [04:24] <michaelni> maybe we could put more than 1 sample in a LUT [04:25] <michaelni> adjacent pixels have similar colors so the table doesnt need to be x^2 sized [04:25] Action: michaelni tries to stay awake but fails [10:42] <j-b> good moroning [10:44] <j-b> lol @MSE bug [10:44] <michaelni> morning [10:45] <michaelni> btw, that git irker thing still has issues with commit messages containiung '|' or something [10:45] <j-b> o_O [10:45] <j-b> don't use | then :) [10:45] Action: j-b runs [10:46] <michaelni> libav uses it alot lately it seemss [10:46] <michaelni> so when merging we use it too ... [10:46] <j-b> ok [10:46] <j-b> I'll ask thresh [10:47] <michaelni> thanks [10:48] <cone-418> ffmpeg.git 03Diego Biurrun 07master:9c112a615892: x86: dsputil: Move avg_pixels8_mmx() out of rnd_template.c [10:48] <cone-418> ffmpeg.git 03Diego Biurrun 07master:932806232108: x86: dsputil: Move avg_pixels16_mmx() out of rnd_template.c [10:48] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:35ef98013dfc: Merge commit '932806232108872655556100011fe369125805d3' [10:54] <cone-418> ffmpeg.git 03Diego Biurrun 07master:bf7c3c6b157f: x86: dsputil: Move cavs and vc1-specific functions where they belong [10:54] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:05599308e988: Merge commit 'bf7c3c6b157f7938578f964b62cffd5e504940be' [11:00] <cone-418> ffmpeg.git 03Diego Biurrun 07master:153f31b1af5e: cavs: Remove cavsdata.h header, unused since 88386feefd [11:00] <cone-418> ffmpeg.git 03Diego Biurrun 07master:6b753015ca6f: aac: Remove aacpsy.h header, unused since 78e65cd7726 [11:00] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:1ba4cbf734b3: Merge commit '6b753015ca6f93d16a05e1a97da1fa42e4da60ec' [11:11] <cone-418> ffmpeg.git 03Luca Barbato 07master:13f57a4d881e: wav: K&R formatting cosmetics [11:11] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:482c1ac849fa: Merge commit '13f57a4d881ec669b87bf82f424bfff502433060' [11:16] <cone-418> ffmpeg.git 03Martin Storsjö 07master:4d810ad2e943: configure: Use a different variable name in push/popvar [11:16] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:085a4e6b16bf: Merge commit '4d810ad2e943a59658ddf00a397121c2b62f7157' [11:21] <cone-418> ffmpeg.git 03Martin Storsjö 07master:ffb068ce8e99: configure: Don't do enable_deep_weak on disabled variables [11:21] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:6190aae90dbe: Merge commit 'ffb068ce8e99df121d18a5b9b1d820f3ea8697fb' [11:38] <cehoyos> j-b: Hi, any news on c99 and ftp? [11:39] <cone-418> ffmpeg.git 03Anton Khirnov 07master:803d21f38bda: lavc: schedule AVCodecContext.lowres for removal on next bump. [11:39] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:88ef7f6d2ba7: Merge commit '803d21f38bdafe7c4e2571a9ae7052013a12923b' [11:39] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:2ee29b5fb936: avcodec: move lowres down with the next ABI bump and add accessors [11:45] <cone-418> ffmpeg.git 03Anton Khirnov 07master:b333f3a22a4d: lavf-regression: use -frames instead of -t for image tests [11:45] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:550defe16eea: Merge commit 'b333f3a22a4db4cf65d6a0457ac82ecbe7c7ac44' [12:03] <cone-418> ffmpeg.git 03Christophe Gisquet 07master:566b7a20fd0c: x86: float dsp: butterflies_float SSE [12:03] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:e91339cde25a: Merge commit '566b7a20fd0cab44d344329538d314454a0bcc2f' [12:21] <cone-418> ffmpeg.git 03Masaki Tanaka 07master:507b1e454cf9: avcodec: Add field order information to AVCodecParserContext. [12:21] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:6fc58a8cb670: Merge commit '507b1e454cf9953da3e18f33c9bd1fca78c97cb5' [13:07] <ubitux> durandal_1707: what's wrong with extractplanes? [13:07] <cone-418> ffmpeg.git 03Masaki Tanaka 07master:f43c14d0467f: mpegvideo_parser: Set field_order. [13:07] <cone-418> ffmpeg.git 03Masaki Tanaka 07master:accde1bd8756: vc1_parser: Set field_order. [13:07] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:9fd5e75bdf62: Merge commit 'accde1bd8756936e1757b42fc9bad2eb5d192f8a' [13:09] <durandal_1707> ubitux: too long [13:10] <ubitux> mapplanes? [13:10] <ubitux> xplane? airplane? [13:10] <durandal_1707> deplane [13:11] <durandal_1707> planefun [13:11] <ubitux> slctpln [13:11] <ubitux> xtrctpln [13:11] <durandal_1707> too cryptic [13:12] <ubitux> pshuffl [13:12] <ubitux> ask intel for more variants [13:12] <J_Darnley> durandal_1707: extract_video_sample_plane(s) [13:13] <cone-418> ffmpeg.git 03Anton Khirnov 07master:0574bc06d23a: matroskaenc: cosmetics, reorder the #includes [13:13] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:2e0cc34292d2: Merge commit '0574bc06d23a6690a16c5926dddc9be359a3d7bb' [13:13] <ubitux> _please [13:16] <ubitux> durandal_1707: pickplanes [13:17] <ubitux> pp already taken :( [13:17] <durandal_1707> but really. filters named with only two letters ....... [13:18] <J_Darnley> Why is the length of a filter name a problem? [13:18] <ubitux> durandal_1707: "plane" could be ok too btw [13:18] <ubitux> but extractplanes is fine, really :p [13:19] <iive> deplane [13:19] <ubitux> 13:10:34 <@durandal_1707> deplane [13:19] <iive> argh... [13:19] <iive> ok, deplane wins [13:20] <iive> is there plane_split? [13:20] <durandal_1707> channelsplit [13:20] <iive> consistency is good thing. [13:20] <iive> (if you don't go diego with it). [13:21] <ubitux> iive: your sentence is not ordered alphabetically [13:21] <iive> lol [13:39] <cone-418> ffmpeg.git 03Anton Khirnov 07master:e3b225a4fe0f: matroskaenc: add an option to put the index at the start of the file [13:39] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:653782718915: Merge commit 'e3b225a4fe0ff1e64a220b757c6f0a5cf9258521' [13:45] <ubitux> i wonder if we shouldn't do just like with mp4 for this ^ [14:15] <cone-418> ffmpeg.git 03Anton Khirnov 07master:b4a5a292274f: avconv: improve -re implementation [14:15] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:22dc25cb29ab: Merge commit 'b4a5a292274f904f404f40f826c51e6fc9cfb8fe' [14:18] <cehoyos> __gb__: Ping, are you online? [15:12] <cone-418> ffmpeg.git 03Anton Khirnov 07master:76d23f40314f: avconv: do not send non-monotonous DTS to the muxers. [15:12] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:2b7f595aca84: Merge remote-tracking branch 'qatar/master' [15:12] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:9fe2a62cc4bc: ffmpeg: move non monotone prevention code down [15:12] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:ee3824f6f696: ffmpeg: limit non monotone workaround to audio & video streams [15:12] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:6b2fbdb047da: ffmpeg: Check dts before use in non monotone dts workaround [15:12] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:26b89f62740b: ffmpeg: use a max variable to hold the maximum dts [15:12] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:007ee09d7ccf: ffmpeg: fix the used maximum dts in the monotone dts workaround [15:12] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:30c05f6b02a4: ffmpeg: fix pts handling in the non monotone dts workaround [15:12] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:b2a1d92b7b30: ffmpeg: port loglevel selection code to new dts workaround code [15:12] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:af2d96433624: ffmpeg: set correct log context for non monotone dts workaround warnings [15:12] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:0966623604a9: ffmpeg: drop old non monotone dts workaround. [15:21] <cehoyos> michaelni: Do you want to merge two more commits? "compat:msvc" fixes ticket #2049 [15:27] <michaelni> cehoyos, i always want to merge bugfixes [15:27] <cehoyos> ;-) [15:28] <michaelni> where do i find them ? [15:28] <cehoyos> Latest avconv commits (two more wrt your merge) [15:50] <cone-418> ffmpeg.git 03Luca Barbato 07master:af4cc2605c7a: id3v2: check for end of file while unescaping tags [15:50] <cone-418> ffmpeg.git 03Martin Storsjö 07master:769d921f3e4d: compat: msvc: Make sure the object files are included when linking statically [15:50] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:760f7d3a90bf: Merge remote-tracking branch 'qatar/master' [15:51] <cehoyos> Thank you! [17:09] <cone-418> ffmpeg.git 03Clément BSsch 07master:e85ea7d387a3: lavfi/frei0r: fix crash when parameters are not set. [18:08] <durandal_1707> michaelni: i need grp16 O [18:21] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:90c2e40ebb9c: vc1dec: vc1_mc_4mv_chroma4: match addressing between compensation and MC [18:21] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:7e30bfcb9176: vc1dec: Fix mixed field/frame intensity compensation [18:21] <cone-418> ffmpeg.git 03Michael Niedermayer 07master:d29f9f4a0e0a: vc1dec: fix lutuv use in vc1_mc_4mv_chroma4() [18:22] <michaelni> j-b, according to carl above should have been the last known/vissible interlaced vc1 bug [18:22] <JEEB> yay [18:23] <nevcairiel> oh right, i wanted to run my test files through it [18:23] <nevcairiel> is it bitexact to reference now? [18:23] <michaelni> no [18:23] <michaelni> i made no attempt to make it bitexact [18:23] <nevcairiel> is the loopfilter still missing for interlaced? [18:24] <michaelni> yes i think so but it seems not to lead to any vissible difference according to carl [18:24] <michaelni> at least for the files he looked at [18:25] <nevcairiel> i'll build git master and see if i can find a file with issues [18:25] <michaelni> ok [18:35] <durandal_1707> should i wait saste for asetps doc patch? [18:36] <cone-418> ffmpeg.git 03Paul B Mahol 07master:ce322f4c191f: extractplanes filter [18:41] <j-b> michaelni: no way? [18:47] <cone-418> ffmpeg.git 03Stefano Sabatini 07master:f51aa92b0f16: lavfi/testsrc: add support for color interactive command [19:17] <cone-418> ffmpeg.git 03Paul B Mahol 07master:e6d4e687558d: lavu: add planar 16 bit and 8 bit rgb with alpha [19:58] <michaelni> j-b, i didnt investigate what effects not doing loop filtering could have, it likely can lead to vissible effects [19:58] <michaelni> If someone reports a case where it does show artifacts then ill look into implementing it [20:01] <j-b> http://streams.videolan.org/issues/6365/0002_vc1_notworking.m2ts [20:01] <j-b> http://streams.videolan.org/issues/5887/ [20:01] <j-b> https://trac.videolan.org/vlc/attachment/ticket/7246/Cruise%202012_07_29_19… [20:05] <nevcairiel> those seem all to play fine [20:05] <tempusfol> Hello, I've encountering some problems compiling ffmpeg. I've tried #ffmpeg, I got no answer after pasting my ./configure and full logs. My I discuss about that here? [20:05] <tempusfol> s/My/May/ [20:18] <durandal_1707> tempusfol: i dont see logs on #ffmpeg [20:19] <tempusfol> durandal_1707: I've posted there 3 hours ago. [20:19] <durandal_1707> found it [20:23] <cone-418> ffmpeg.git 03Paul B Mahol 07master:b937d3694e54: doc/filters: document asetpts in single place [20:26] <durandal_1707> so nobody cares for broken ilpack and softpulldown? [20:28] <durandal_1707> the ilpack have some inline assembly that could be used in fieldmatch [20:31] <durandal_1707> i could teach trim filters to use AV_OPT_TYPE_DURATION [21:38] <durandal_1707> ubitux: how is 3d lut going? [21:52] <ubitux> durandal_1707: in progress [21:53] <ubitux> mostly done TBH [21:53] <ubitux> but i'm working on several other things, please be patient :) [23:11] <trose> looking for some info about pipe2image png support [23:11] <trose> anyone know how to make this work or is it not supported? [23:36] <cone-356> ffmpeg.git 03Michael Niedermayer 07master:e756635964ed: eamad: Check against minimum supported dimensions [00:00] --- Sat May 4 2013

1 0

ffmpeg.log.20130503
by burek 04 May '13

04 May '13

[00:20] <Zarx> im trying to encode only a certain frame range of a video, and also encode the audio from that same frame range. I had thought the command to trim it would look something like this: -vf select='gte(n\,100)*lte(n\,200)',aselect='gte(n\,100)*lte(n\,200)' [00:20] <Zarx> am I wrong? [00:21] <Zarx> the video part works, but when I add the part to trim the audio, that breaks it [00:23] <durandal11707> Zarx: because aselect is audio filter and thus cant work with -vf [00:23] <Zarx> haha [00:23] <durandal11707> use -filter_complex [00:23] <Zarx> ok, thanks [00:25] <durandal11707> also n for aselect means number of samples, not frames [00:26] <Zarx> oh... am I going to have to use timecodes then? [05:01] <smj> using a webcam overlay drops streaming fps to 3-5 from 60 [05:04] <smj> to me that sounds like it's waiting to get a fresh frame from the webcam [05:04] <smj> is there a way to do this asynchronoysly? [05:11] <smj> it might be this http://ffmpeg.org/pipermail/ffmpeg-user/2011-May/000934.html [05:26] <smj> in overlay filter documentation... "Be aware that frames are taken from each input video in timestamp order, hence, if their initial timestamps differ, it is a a good idea to pass the two inputs through a setpts=PTS-STARTPTS filter to have them begin in the same zero timestamp, as it does the example for the movie filter." [05:26] <smj> what's a PTS? some sort of time stamp? [05:31] <relaxed> smj: http://en.wikipedia.org/wiki/Presentation_time_stamp [05:31] <smj> yeah, found that [06:03] <smj> that setpts thing doesn't do much... [06:06] <smj> so there's https://lists.ffmpeg.org/pipermail/ffmpeg-user/2011-April/000346.html https://lists.ffmpeg.org/pipermail/ffmpeg-user/2011-April/000395.html and http://ffmpeg.org/pipermail/ffmpeg-user/2011-May/000939.html [06:16] <smj> and https://ffmpeg.org/trac/ffmpeg/ticket/1339?cversion=3&cnum_hist=2 [06:46] <cryptopsy> how to trim duration a-b out of a file that's a-c in duration? [06:47] <cryptopsy> nvm [07:16] <Mavrik> Gandalfar, wat [11:46] <EvilTengil> I'm trying to figure out how I can bundle some data with each frame. Is this possible with "data streams" (AVMEDIA_TYPE_DATA)? [12:07] <zap0> i too would like to know. [12:08] <durandal_1707> what data? [12:08] <EvilTengil> 6 doubles [12:08] <EvilTengil> the video is supposed to be sent from a phone with sensors [12:45] <rvnikauj> hi ffmpeg users and dev [12:46] <rvnikauj> i have an issue with cli usage under archlinux [12:46] <rvnikauj> i have some .VOB which i want to convert to theora/ogg [12:46] <rvnikauj> but only the 2.0 track, which is the third one [12:47] <rvnikauj> so i try this from the wiki [12:47] <rvnikauj> at trac [12:47] <rvnikauj> http://pastebin.com/UiXm4ser [14:49] <sacredchao> Hi, can I use ffmpeg to extract an AC-3 stream from a decrypted M2TS file that came from a BR? [14:50] <sacredchao> I want to leave all streams alone, and simply extract only one of the 2 AC-3 audio streams in the M2TS files [14:54] <Magicking> sacredchao: You might be looking for the -map option http://ffmpeg.org/trac/ffmpeg/wiki/How%20to%20use%20-map%20option [14:56] <dolevo> hi all [14:57] <dolevo> I have a question related to hardware accelerated decoding on AMD radeon. Is there anybody who can help me? [15:02] <durandal_1707> no [16:08] <theos> hey [16:09] <theos> is there a way to reduce the size of mp3 files? [16:10] <durandal_1707> no [16:11] <durandal_1707> unless you want to transcode them which is really bad idea [16:11] <theos> like reduce bitrate etc? [16:13] <theos> no? [16:16] <durandal_1707> yes you can - its called transcoding - and its bad idea [16:18] <theos> why is it bad idea? [16:27] <relaxed> theos: you lose quality [16:28] <sacarasc> theos: It's like taking photo of a photo. It won't be as good as the original photo and only a shade of the original. [16:29] <theos> got it [16:31] <sacarasc> If you have the original wave files, though, you could reencode that and it would sound better than reencoding the mp3 ¶ mp3. [16:58] <EvilTengil> Does anyone know what the side_data is on a AVFrame? Can I add my own data there? [17:07] <tempusfol> Hello, make fails with ffmpeg's last git with "/usr/bin/ld: libavutil/lls.o: relocation R_X86_64_PC32 against undefined symbol `memset@@GLIBC_2.2.5' can not be used when making a shared object; recompile with -fPIC". May I proceed pasting my ./configure here? [17:10] <tempusfol> well, I was asking if this was the correct channel to discuss about this, I'll assume "yes" [17:18] <atog> When I tell ffmpeg to record audio from an HTTP stream for a specific duration, I receive a warning that it's estimating duration from the bitrate. How do I get rid of this warning and make it record based upon the actual time it's been recording? [17:31] <tempusfol> Last full log (only last error showed) http://fpaste.org/10255/75950281/ btw, I've tried with --enable-static --disable-shared as well. It's part of a script that rebuilds for me mplayer,ffmpeg,x264,xvidcore,fr3ior and that has served me well for over two years. [18:00] <atog> anyone have any suggestions for my duration issue? [18:01] <Magicking> atog: You download the file first and don't use http [18:02] <atog> Magicking: it's an audio stream [18:03] <atog> Magicking: What I want to do is copy the a portion of a live audio stream for a specific duration [18:04] <atog> Magicking: I just don't see why ffmpeg is trying to estimate the duration. If I tell it to record for 10 minutes, I expect the output to be 10 minutes long [18:08] <klaxa> atog: have you tried doing that? [18:08] <klaxa> i mean just recorded and checked? [18:08] <klaxa> the "estimating duration from the bitrate" doesn't really apply on streams i think [18:11] <Fjorgynn> Where is my cake? [18:11] <Fjorgynn> what audio stream is that? [18:13] <Fjorgynn> happy friday [18:15] <Magicking> atog: You are just anoyed by the warning ? [18:16] <atog> klaxa: I've got several streams that I used cron to start at a specific time and told ffmpeg to record for a fixed amount of time. The cron job usually starts recording a few seconds after the minute.. and the file time indicates that ffmeg stopped recording 8 or more seconds before it should have [18:17] <klaxa> that's weird, indeed [18:19] <atog> klaxa: I can only assume the difference is due to the fact that ffmpeg is due to the initial duration estimate [18:19] <klaxa> how do you invoke the command? [18:20] <klaxa> and what kind of stream is it? [18:20] <atog> klaxa: it's an mp3 stream [18:20] <klaxa> fixed bitrate? [18:20] <atog> no idea [18:20] <atog> presumably [18:21] <klaxa> still, what command do you use to record the stream? [18:21] <atog> current command is: ffmpeg -i <http stream> -acodec copy -t 300 output.mp3 [18:22] <klaxa> what ffmpeg version are you using? [18:22] <atog> klaxa: 1.0.6 [18:23] <klaxa> humm.. [18:24] <klaxa> that's weird indeed [18:26] <atog> the stream is supposed to be a 64kb mp3 stream [18:41] <atog> klaxa: time ffmpeg -i <http stream> -acodec copy -t 60 test.mp3 ran for 50s [18:43] <sine`> audio vorbis is that oog ? [18:43] <sine`> or ogg [18:44] <klaxa> atog, pretty weird :S [18:44] <klaxa> sine`: it's called ogg [18:45] <sine`> ok so i have a video that has the vorbis 44100 hz, stereo ,s16 default [18:45] <sine`> how do i just rip or demux the audio without reencoding it [18:46] <klaxa> ffmpeg -i my_file.mkv -c:a copy audiostream.ogg [18:46] <klaxa> hmm... [18:46] <klaxa> maybe add -vn in there too [18:46] <klaxa> so: ffmpeg -i my_file.mkv -c:a copy -vn audiostream.ogg [18:46] <Fjorgynn> yeah -vn is awesome [18:47] <Fjorgynn> if it's ogg by default [18:47] <klaxa> if you want to make it super fancy use -map [18:47] <Fjorgynn> and if you want mp3 use lame to convert it from wave [18:48] <sine`> i got the ogg [18:48] <sine`> hopefully my phone will play that [18:48] <Fjorgynn> :) [18:48] <Fjorgynn> maybe [18:48] <sine`> dont want to reencode it again uggh [18:49] <Fjorgynn> what phone you have? [18:50] <sine`> note 2 [18:50] <sine`> should be ok [18:52] <sine`> yes [18:52] <sine`> fine [18:52] <sine`> schweet [18:52] <sine`> thanks [18:52] <sine`> time for a run bye [18:55] <Fjorgynn> :D [19:36] <voidah> hi guys [19:37] <voidah> Is there general tips concerning sound/video out-of-sync issue when using x11grab? [19:37] <voidah> or a place where I could find the answer [19:37] <voidah> I tried so many thing [19:38] <klaxa> editing the file afterwards would be easiest i think [19:39] <voidah> this is what I would like to avoid if possible [19:40] <voidah> in despair, I tried recordmydesktop [19:40] <voidah> there is not sync issue with it [19:40] <voidah> but I _need_ to use ffmpeg [19:40] <klaxa> doesn't recordmydesktop just call ffmpeg? [19:40] <voidah> I don't know, and their website doesn't specify that [19:41] <voidah> out-of-sync issue with ffmpeg only is visible on long video [19:41] <durandal_1707> what codec you use? [19:41] <voidah> libx264 [19:41] <voidah> acodec: mp3 [19:41] <voidah> (tried flac too for acodec) [19:42] <durandal_1707> tried ultrafast preset? [19:42] <voidah> yes [19:42] <klaxa> oh it does in fact not simply call ffmpeg [19:42] <voidah> I'm now trying veryfast [19:42] <voidah> to see if it makes a difference [19:42] <voidah> I record audio and video separately [19:43] <voidah> the duration is not even the same [19:43] <voidah> on a ~10 minutes record, the sound is ~10 seconds longer [19:43] <voidah> the sound is perfectly in sync in the first ~5 minutes of the video [19:43] <atog> klaxa: I'm going to switch to streamripper since it seems to record for the correct amount of time [19:43] <voidah> then the longer the video is, the bigger the gap is [19:43] <klaxa> atog kk [19:44] <voidah> I tried going though pulse, of directly with hw:0,0 [19:44] <voidah> trought [19:44] <voidah> same difference [19:45] <klaxa> hw:0,0 is alsa though [19:45] <voidah> s/of/or [19:49] <voidah> last test: [19:49] <voidah> video is exactly 10:00 long [19:49] <voidah> sound is 10:02 [19:49] <voidah> both capture were started at the same time [19:50] <voidah> and stopped at the same time [19:51] <voidah> the sound and video are perfectly in sync at the beginning [19:51] <voidah> but out of sync at the end [19:51] <voidah> I'm sure I made something wrong, and if someone have tips, they are welcomed [19:53] <voidah> is there a way to merge the video and audio files, shrinking a little bit the audio file so that it fits the length of the video file? [19:53] <voidah> with ffmpeg of course [19:56] <klaxa> did you look at -async? [19:56] <voidah> nope, I google it [19:58] <voidah> should I use it when I merge the 2 files? [19:59] <voidah> and does it works with any files types? [20:00] <klaxa> wait... are you recording audio and video with two ffmpeg instances? [20:00] <voidah> yes [20:00] <voidah> I tried doing it when the same instance too [20:01] <klaxa> but? [20:01] <voidah> I get the same problem [20:02] <voidah> what method would you recommend? [20:04] <klaxa> try using one instance and apply -async and see if that helps [20:04] <voidah> Ok I'm trying it now [20:05] <voidah> I get many of these message when using async [20:05] <voidah> [matroska @ 0x84ce60] st:0 PTS: 67854 DTS: 67854 < 67892 invalid, clipping [20:08] <voidah> it's mayby a good sign [20:08] <voidah> still 5 minutes to record [20:14] <voidah> klaxa: man.. perfectly in sync [20:14] <voidah> with async [20:14] <voidah> I'll test it again just to be 100% sure [20:21] <durandal_1707> tempusfol: you did make distclean? [20:27] <voidah> klaxa: it does'nt seems to work finally... [20:27] <voidah> here is the command used: ffmpeg -f alsa -ac 2 -i pulse -f x11grab -r 20 -s 1920x1080 -i :0.0 -acodec pcm_s16le -async 1000 -vcodec libx264 -preset ultrafast -threads 0 output.mkv [20:27] <tempusfol> durandal_1707: yes [20:32] <tempusfol> durandal_1707: now I always get "libavcodec/libavcodec.so: undefined reference to `x264_encoder_open_132'" [20:33] <durandal_1707> tempusfol: what libx264 version? [20:34] <tempusfol> durandal_1707: http://paste.fedoraproject.org/10306/67606024 [20:37] <durandal_1707> x264 is build shared? [20:38] <tempusfol> I've tried right now once again --enable-static --disable-shared, it ends with "libx264.c:(.text.unlikely+0xb08): undefined reference to "x264_encoder_open_132" [20:38] <tempusfol> x264 is build.... let me check [20:39] <voidah> tempusfol: remove everything that contains static of dynamic [20:39] <voidah> in your configure options [20:39] <voidah> it will work [20:39] <voidah> I once had a similar proble [20:39] <tempusfol> x264 is built with "./configure --enable-shared --enable-visualize --enable-strip --extra-cflags='-O2 -march=native -mtune=native' --prefix=/opt/bleedingvideo" [20:39] <voidah> m [20:40] <tempusfol> voidah: I'll try that as well [20:43] <tempusfol> voidah: ...no luck. It always ends with `undefined reference to `x264_encoder_open_132'.. I'll guess I'll grab a different version of x264 and try again? [20:44] <voidah> ok then I don't know [20:45] <voidah> try your luck in #ffmpeg-devel [20:45] <tempusfol> ehm I started here, went there, and then durandal_1707 replied to me here [20:46] <durandal_1707> instead of pinging there, ping here [20:46] <durandal_1707> but whatever, what recently changed that caused such breakage? [20:46] <tempusfol> (that's what I'm doing) [20:47] <tempusfol> nothing peculiar, it's a script I've used for over two years [20:47] <durandal_1707> so nothing? did not installed new x264, instealled something other that changed those libs or .... [20:47] <tempusfol> today, it has recompiled x264, mplayer, frei0r, x264, ffmpeg is the last one...and ffmpeg failed [20:48] <tempusfol> when the scripts starts, it changes libs of *everything* [20:48] <tempusfol> ...start fresh. [20:48] <durandal_1707> reboot [20:48] <tempusfol> d'oh [20:48] <cbsrobot> tempusfol: maybe yyou have 2 libx264 - in different places [20:49] <durandal_1707> disable ccache [20:49] <tempusfol> ...I could try that as well [21:18] <tempusfol> durandal_1707: reboot didn't help, ccache disabled didn't help, obliterating all the local mirrors and re-git cloning everything solved. I don't know why it was solved though. [21:20] <durandal_1707> re-git clonining should really not be needed [21:20] <tempusfol> I think so [21:49] <JodaZ> is there something like a -re * some_factor [21:53] <JodaZ> does anyone know how/if i can produce individual HLS segments from input video on demand ? [21:58] <JodaZ> :/ i used to get answers here [22:01] <llogan> JodaZ: wot's your question? i probably just missed it. [22:01] <JodaZ> llogan, its 3 lines up [22:02] <JodaZ> o wait, you joined after *_* [22:02] <JodaZ> does anyone know how/if i can produce individual HLS segments from input video on demand ? [22:09] <hughmanwho> Hello! Anyone on here know how to stream rtsp via ffmpeg and save it to a file? Here is a pastebin of what I've done and my results: http://pastebin.com/j6N0WVX0 [22:10] <tmatth> hughmanwho: try saving it to another container, maybe mp4 can't store your audio [22:14] <JodaZ> audio: none ?_? [22:15] <JodaZ> hughmanwho, is there audio ? [22:17] <hughmanwho> no audio [22:20] <hughmanwho> Seems to be doing something now, or at least it's missing packets which I assume means it's also getting some. What format would you recommend [22:20] <hughmanwho> ? [22:24] <hughmanwho> Switched over to mjpg and am now getting a bunch of this: http://pastebin.com/j6N0WVX0 [22:27] <tmatth> hughmanwho: can you pastebin the sdp? [22:29] <hughmanwho> ffmpeg -i rtsp://192.168.14.212/live.sdp -t 10 outputFile.mjpeg [22:31] <hughmanwho> That one did produce a short clip.. only it seemed like it was only a frame or two, it flashed something then disappeared [22:31] <hughmanwho> Is mjpeg a valid extension or only valid for streaming or something like that? [22:35] <Rajeev> what is the significance in a MP4 files non-standard atom 'yqoo' [22:39] <Rajeev> is this a standard type of atom now? [22:45] <hughmanwho> What would be a good file type to use for saving videos? [22:55] <trose> I'm trying to pipe images directly into ffmpeg from python as in this example http://stackoverflow.com/questions/13294919/can-you-stream-images-to-ffmpeg… [22:56] <trose> I'd like to use png instead of jpeg for better quality though [22:56] <trose> doesn't seem like ffmpeg supports png for pipe2image [22:56] <trose> is there a work around? [23:18] <hughmanwho> Thanks tmatth and JodaZ, for some reason I don't have any problem saving it as a .gif which probably does what I need it to do for now [23:23] <hughmanwho> Any idea why I keep getting 'non-existing PPS referenced' errors? [23:23] <hughmanwho> Here's a pastebin: http://pastebin.com/j6N0WVX0 [23:45] <hughmanwho> How can I change the 'analyzeduration'? I'm getting a 'max_analyze_duration 5000000 reached at 5000000 microseconds error.. I call [23:45] <hughmanwho> C:\Tools\ffmpeg-20130428\bin>ffmpeg -i rtsp://admin:admin@192.168.14.25/defaultPrimary?streamType=u -t 60 -analyzeduration 10000000 outputFile.gif [23:45] <hughmanwho> yet still get the same error without the max_analyze_duration having been changed [23:46] <relaxed> Are you actually getting an error? [23:48] <relaxed> hughmanwho: That message meant to be informative. It's not an error message. [23:49] <hughmanwho> I'm getting a bunch of 'non-existing PPS referenced', 'non-existing PPS 0 referenced', 'decode_slice_header error [23:49] <hughmanwho> ' and 'no frame!' errors [23:49] <hughmanwho> see this pastebin to see the entire output: http://pastebin.com/j6N0WVX0 [23:49] <relaxed> does it stop encoding? [23:50] <hughmanwho> Good to know regarding that msg [23:50] <hughmanwho> I also get a 'Output file #0 does not contain any stream' error [23:50] <relaxed> where is that in the pastebin? [23:50] <hughmanwho> Very bottom [23:51] <hughmanwho> Never mind.. that got cut off [23:51] <hughmanwho> just a sec [23:52] <Jorky> Hello Guys [23:52] <Jorky> anyone here at the moment? [23:52] <sacarasc> Yo. [23:52] <Jorky> I just want to know what is the command to convert whole album to another format [23:53] <sacarasc> What OS? [23:53] <Jorky> for instance album of mp3 songs to wma at the same time [23:53] <hughmanwho> http://pastebin.com/j6N0WVX0 [23:53] <Jorky> linux [23:53] <Jorky> archlinux [23:53] <sacarasc> for music in *.mp3; do ffmpeg -i "$music" "$music".wmv; done [23:53] <sacarasc> Or some variation upon that. [23:54] <Jorky> with music you mean name of the album? [23:54] <sacarasc> No, that's just a variable name. [23:54] <sacarasc> Do that in the directory with the files. [23:54] <hughmanwho> I have to run.. relaxed I appreciate it, if you have any thoughts, I'll keep this open and check on your reply or you can email me at hughmanwho(a)gmail.com Thanks again relaxed!! [23:55] <Jorky> yes but this means that I must do it for every song in album? [23:55] <relaxed> Jorky: for i in *mp3; do ffmpeg -i "$i" -c:a wmav2 "${i%.*}".wma; done [23:55] <sacarasc> No. [23:55] <sacarasc> That's an even better one! [23:55] <sacarasc> That will do it for each file called whatever.mp3. [23:56] <sacarasc> I really need to remember the ${i%.*} thing... I never can. [23:56] <relaxed> Jorky: study this http://mywiki.wooledge.org/BashGuide [23:57] <Jorky> hm [23:57] <Jorky> I still don't get it [23:58] <relaxed> It's a for loop that takes each mp3 in the current dir and encodes them one at a time using ffmpeg. [23:59] <Jorky> can you type me more specific script [00:00] --- Sat May 4 2013

1 0

ffmpeg-devel.log.20130502
by burek 03 May '13

03 May '13

[00:32] <cone-76> ffmpeg.git 03Michael Niedermayer 07master:c1f2c4c3b492: vmdav: Try to fix unpack_rle() [01:23] <cehoyos> michaelni: Did you check http://samples.ffmpeg.org/game-formats/sierra-vmd/lastdynasty/ ? [01:24] <cehoyos> HG060808.VMD does not play, worked with 0.8, HR061502.VMD seems to have never played. [02:21] <michaelni> cehoyos, i checked all the game-formats/sierra-vmd files that i had on my disk, which should have been all that are on the server, but ill recheck [02:51] <cone-76> ffmpeg.git 03Michael Niedermayer 07master:e91fd754c63d: avcodec/ansi: Check x/y [02:51] <cone-76> ffmpeg.git 03Michael Niedermayer 07master:6021615bbe39: avcodec/ansi: Fix right/bottom end checks [10:41] <highgod> Hi, all, I want to ask a configure quesion. I want to use xxxx/cl/cl.h to detect whether the system is mac os or others, so 1.I add a check_header cl/cl.h,2.I add cl_cl_h to HAVE_LIST, it works fine on mingw, but when I use ubuntu to configure, the HAVE_CL_CL_H always zero.The path of cl.h is /opt/AMDAPP/include/cl/cl.h [11:05] <durandal_1707> saste: gonna remove dupe documentation for asetpts? [11:06] <durandal_1707> michaelni: you tried all vmdav samples? [11:40] <smj> how do I negate an overlayed input? [11:40] <smj> wrong channel, sorry [12:14] <michaelni> durandal_1707, all i had, yes [12:14] <durandal_1707> michaelni: in samples.mplayerhq.hu/game-formats/sierr.. ? [12:16] <michaelni> i downloaded all the sierr.. samples some while ago and tested all i had on my disk [12:17] <durandal_1707> because i'm found yet another sample that worked previously fine, but broke after "fix" from qatar [12:17] <durandal_1707> michaelni: all samples you have play fine? [12:17] <michaelni> no [12:17] <michaelni> none used the code that i tried to fix [12:17] <michaelni> i put an abort() in there and it didnt trigger unless i made a mistake [12:18] <durandal_1707> ok so what samples do not play and are not on server? [12:18] <michaelni> iam not aware of any [13:01] <cone-25> ffmpeg.git 03Diego Biurrun 07master:a650c906cbdb: ppc: Only compile AltiVec FFT assembly when AltiVec is enabled [13:01] <cone-25> ffmpeg.git 03Luca Barbato 07master:0671adbb18f7: id3v2: K&R formatting cosmetics [13:01] <cone-25> ffmpeg.git 03Michael Niedermayer 07master:0beff4282322: Merge remote-tracking branch 'qatar/master' [13:06] <cone-25> ffmpeg.git 03Carl Eugen Hoyos 07master:d0a661d48a2b: Add wmv2 dependency to vc1 compilation. [13:06] <cone-25> ffmpeg.git 03Michael Niedermayer 07master:4d4ec18b2142: Merge remote-tracking branch 'cehoyos/master' [13:43] <cone-25> ffmpeg.git 03Paul B Mahol 07master:cdada4f6390a: vmdav: return meaningful error codes [13:43] <cone-25> ffmpeg.git 03Paul B Mahol 07master:663794f63da0: vmdav: unbreak decoding of samples from game The Last Dynasty [13:51] <cone-25> ffmpeg.git 03Michael Niedermayer 07master:89b5039ff265: avutil/mem: add av_memdup() [14:19] <durandal_1707> michaelni: two pal8 inputs used with blend filter gives gray output. intead of picking rgb24 [14:20] <durandal_1707> blend supports gray but not pal8 [14:22] <cone-25> ffmpeg.git 03Paul B Mahol 07master:990b13806d38: lavfi/blend: use correct way to check number of planes [14:22] <cone-25> ffmpeg.git 03Paul B Mahol 07master:a64c9afe3de3: lavfi/blend: merge config_input_top into config_output [14:47] <cone-25> ffmpeg.git 03highgod0401 07master:6354a6ce5271: lavu/opencl: fix bug of opencl options [15:35] <durandal_1707> michaelni: ffmpeg -i random_valid_input random segv here [15:36] <durandal_1707> where random is anything not recognised as suitable output format [16:06] <cone-25> ffmpeg.git 03Michael Niedermayer 07master:49ec4c7e49a4: ffmpeg: fix segfault with random output [16:51] <durandal_1707> http://www.ohloh.net/p/ffmpeg/managers [17:32] <cone-25> ffmpeg.git 03Michael Niedermayer 07master:8937230719ad: ape_decode_value_3900: check tmpk [17:32] <cone-25> ffmpeg.git 03Michael Niedermayer 07master:60ab44804237: ape_decode_value_3860: check k before using it in get_bits() [17:42] <durandal_1707> michaelni: nut duration bug is still present [18:14] <durandal_1707> is this: link->src->outputs[link->srcpad - link->src->output_pads] = NULL; always safe? [18:28] <durandal_1707> michaelni: lavu/common.h and others use UINT64_C but the only place its defined is in lzma.h [19:01] <cone-25> ffmpeg.git 03Michael Niedermayer 07master:fd68371b0728: nutdec: Implement duration parsing for indexed nuts [19:02] <durandal_1707> michaelni: thanks [19:03] <michaelni> np [21:27] <michaelni> durandal_1707, (U)INT*_C is C99 [21:27] <michaelni> but if some major platformss lack it we might want to consider reverting 738940e802667f3f661d0ae1ce2f755a9f5cba36 [21:35] <durandal_1707> well it looks its defined on FreeBSD only if #if !defined(__cplusplus) || defined(__STDC_CONSTANT_MACROS) [21:36] <durandal_1707> hmm actually i'm compiling cpp thing ... [21:36] <durandal_1707> so that is reason why its not defined ... [22:16] <cone-25> ffmpeg.git 03Michael Niedermayer 07master:8d3c99e82531: mmvideo/mm_decode_inter: check horizontal coordinate too [22:16] <cone-25> ffmpeg.git 03Michael Niedermayer 07master:ae2132ac90f0: mmvideo/mm_decode_intra: check horizontal coordinate too [00:00] --- Fri May 3 2013

1 0

ffmpeg.log.20130502
by burek 03 May '13

03 May '13

[00:10] <tonsofpcs> is there a way to tell ffmpeg / ffmbc to use multiple threads for MPEG2 (IMX or XDCAM HD 422) encodes? It seems to be only utilizing one CPU in my 8 core system. [00:24] <cbsrobot> tonsofpcs: mpeg2 should be multithreaded [00:24] <cbsrobot> what command line are you using ? [00:25] <tonsofpcs> cbsrobot: I'm using ffmbc. Command lines ripped rom the google code wiki: [00:25] <tonsofpcs> IMX 50: https://code.google.com/p/ffmbc/wiki/IMXD10Encoding [00:25] <tonsofpcs> XDCAM HD 50: https://code.google.com/p/ffmbc/wiki/XDCAMHD422Encoding [00:25] <tonsofpcs> err, IMX 30, sorry [00:32] <cbsrobot> what codec is the source file ? [00:33] <tonsofpcs> cbsrobot: Long GOP MPEG2 video, TWOS audio inside AIFF, externally referenced by an mov container [00:37] <tonsofpcs> Windows Server 2008r2 if it matters. [00:39] <tonsofpcs> honestly, in the near term the multi-threading won't matter too much as I'll be transcoding a bulk of content so if 8 files get done in 4 hours or if 8 files get done sequentially in 30 minutes each doesn't really matter. In the long run, I think people will expect "I put a file there it should be done by now" as the CPU won't be under high load [00:39] <tonsofpcs> I'm also kinda surprised that the RAM in this system is barely being touched. [00:40] <tonsofpcs> hmm, just saw a spot where it used two cores at once but by quick glance it seems that core 3 usage + core 7 usage (the two used) <= 100% at all times... like the core 3 graph looks like someone flipped the core 7 graph [00:41] <tonsofpcs> right now I'm just rendering one XDCAM 50 HD 422. overall shows 12-13% CPU, 67 MB RAM used by ffmbc [00:42] <cbsrobot> do you have ht ? [00:43] <cbsrobot> I guess it won't help much [00:43] <cbsrobot> cpu is the key for fast transcoding [00:43] <tonsofpcs> it's a Xeon E3-1240 V2. http://ark.intel.com/products/65730 says yes. [00:44] <tonsofpcs> 4 core/8 threads [00:44] <tonsofpcs> so i guess that's thread 3/thread 7 [00:44] <cbsrobot> I made up a command line similar to the ffmbc imx50 and it uses all the cores [00:44] <cbsrobot> do you have the latest ffmpeg installed ? [00:44] <cbsrobot> or at least compiled ? [00:46] <tonsofpcs> I have a binary.... ffmpeg-20130424gitaa96439-win64-static. ffmbc is 0.7rc8-win64 [00:47] <cbsrobot> well try: [00:47] <cbsrobot> ffmpeg input.mov -vf scale=720:608 -c:v mpeg2video -c:a pcm_s16le -intra_vlc 1 -non_linear_quant 1 -flags +ildct+low_delay -ps 1 -qmin 1 -qmax 12 -rc_max_vbv_use 1 -rc_min_vbv_use 1 -g 1 -q:v 1 -pix_fmt yuv422p -b:v 50000000 -maxrate 50000000 -minrate 50000000 -bufsize 2000000 -rc_init_occupancy 2000000 -vtag mx5p output.mov [00:48] <cbsrobot> i'm not sure the -qmax 12 is correct [00:48] <tonsofpcs> 720:608? [00:49] <cbsrobot> I get: [mov @ 0x102066600] D-10/IMX must use 720x608 or 720x512 video resolution [00:49] <cbsrobot> so i just picked one [00:49] <tonsofpcs> oh, 512 is appropraite here. NTSC land ;) [00:49] <cbsrobot> heh I'm in pal [00:49] <tonsofpcs> (525 lines, a bunch of them are useless, hence 512, at least that's my guess) [00:49] <tonsofpcs> (608 would be 625) [00:50] <tonsofpcs> oo, i'm seeing a third thread now (haven't tried your line yet, will try it tomorrow as I'm about to leave for the evening) [00:50] <cbsrobot> ok [00:55] <tonsofpcs> ah, there's a -threads option in ffmbc (thanks to kierank for pointing it out) [00:56] <tonsofpcs> it seems to help a lot [00:56] <cbsrobot> did you set it ? [00:56] <tonsofpcs> cheers cbsrobot. I'll be playing aaround more in the morning, need to head to a meeting now. [00:56] <cbsrobot> sure [00:56] <tonsofpcs> yea, just ran a quick test with it. Leaving the one that I have going going since 3 hours or 1 hour makes no difference when the system is in testing by me only and I'm gone for the next 12 hours ;) [00:57] <diegoviola> hi [00:57] <diegoviola> is libav a fork of ffmpeg? [00:58] <intracube> diegoviola: yes [00:58] <diegoviola> ok [00:58] <diegoviola> i'm currently having a problem with VLC, i think VLC uses ffmpeg or libav, i filled a bug report on their issue tracker, and they mention a feature in VLC works with libav but not with ffmpeg [00:58] <diegoviola> https://trac.videolan.org/vlc/ticket/8544 [00:58] <diegoviola> i was wondering if someone here could help [01:00] <cbsrobot> diegoviola: cehoyos is on cc [01:00] <cbsrobot> maybe he's working on a solution [01:00] <diegoviola> oh ok, sorry [01:01] <diegoviola> is he a ffmpeg developer? [01:01] <cbsrobot> yes [01:01] <diegoviola> cool, thanks [01:01] <diegoviola> :) [01:02] <intracube> diegoviola: I have that same issue with VLC 2.0.5 (linux) when exporting png snapshots [01:03] <intracube> but I use mplayer to take stills [01:03] <diegoviola> intracube: ok, it was reported in the bug tracker that it works with libav head, but doesn't with ffmpeg head [01:03] <diegoviola> intracube: maybe they are working on a fix, as cbsrobot said [01:04] <intracube> diegoviola: how is this related to libav/ffmpeg? does VLC rely on their libraries or something? [01:04] <cbsrobot> I have no problem with vlc 2.0.6 on osx [01:05] <diegoviola> intracube: i think vlc uses ffmpeg or libav, depending how it was built [01:07] <WilliamWS> what [mp4 muxer @ 0x961e480] [Eval @ 0xbfb1e1e8] Undefined constant or missing '(' in 'faststart' [mp4 muxer @ 0x961e480] Unable to parse option value "faststart" [mp4 muxer @ 0x961e480] Error setting option movflags to value faststart. means? [01:07] <WilliamWS> This opition is referenced on the documentation and I am using the last stable version. [01:07] <ubitux> it's present in 1.2 [01:08] <ubitux> what's your command line? [01:08] <WilliamWS> ffmpeg -f video4linux2 -i /dev/video0 -f mp4 -movflags faststart video.mp4 [01:08] <WilliamWS> It is short, I think there is no need to pastbin it, or am I wrong? [01:09] <cbsrobot> &. and the COMPLETE console output [01:09] <intracube> diegoviola: just tried updating to 2.0.6 - no difference here [01:10] <ubitux> WilliamWS: yes, use pastebin anyway [01:11] <diegoviola> intracube: i see, i also tried compiling latest vlc from git/master and it gave me another problem, something related to: [0x7f4d18000958] main video output debug: no image encoded [01:11] <diegoviola> intracube: it didn't even produce an output/file [01:11] <intracube> diegoviola: not sure what that would be, sorry [01:11] <WilliamWS> http://pastebin.com/mkL1G1Mk [01:11] <WilliamWS> sorry [01:11] <diegoviola> intracube: np [01:13] <cbsrobot> WilliamWS: ffmpeg version 0.9.2 [01:13] <ubitux> WilliamWS: this is not the last stable [01:14] <ubitux> this is pretty old [01:14] <WilliamWS> I noticed that right now [01:14] <ubitux> it was release a bit before 2012 [01:14] <WilliamWS> the "Download" link on the site links to an old anchor [01:15] <WilliamWS> -.- [01:15] <ubitux> huh? [01:15] <ubitux> what link? [01:15] <WilliamWS> My apologizes for that [01:15] <WilliamWS> the one you find when using a search engine [01:15] <WilliamWS> example: http://search.yahoo.com/r/_ylt=A0oG7netoYFR8kcAMLdXNyoA/SIG=121t8r2k4/EXP=1… [01:15] <ubitux> ah, not a bad link on our side then [01:16] <ubitux> http://ffmpeg.org/download.html#releases :) [01:16] <ubitux> maybe you shouldn't use yahoo :) [01:16] <WilliamWS> google give the same result [01:16] <WilliamWS> and really I hate google [01:16] <WilliamWS> magi is the last? [01:17] <WilliamWS> magic* [01:17] <WilliamWS> Are you pretty, really, mega, hyper, Tera, zipper sure about this? [01:22] <ubitux> last stable [01:22] <ubitux> but you can use git as well [01:22] <ubitux> 1.2 is starting to get old too :) [01:23] <ubitux> 'should be enough for faststart though [01:24] <WilliamWS> ubitux thank you so much. I am compiling this on a Pentium MMX 266MHz with 64MB RAM [01:24] <ubitux> see you next week [01:25] <WilliamWS> see yah [01:25] <WilliamWS> ahahhahahaha [01:25] <ubitux> ccache is your friend btw [01:25] <WilliamWS> the 0.9 was compilled in 2:38h [01:25] <ubitux> 2 days and 38 hours? [01:26] <ubitux> that sounds pretty decent for a P2 [01:26] <WilliamWS> 2 hour and 38 minutes [01:49] <defaultro> hey folks, is it possible to overlay vignetting on a video? [01:50] <defaultro> interesting what I found, http://www.francois-tarlier.com/blog/ [01:51] <ubitux> i've sent a patch to do that a while ago [01:51] <ubitux> but i didn't have time to get it done [01:51] <ubitux> (got some ugly banding i needed to fix) [02:20] <LunaVorax> Hello [02:21] <LunaVorax> Is the lossless h264 encoding bug-free? [02:21] <LunaVorax> Last time I was here someone reported that ffmpeg wasn't producing lossless h264 videos using the proper settings [02:39] <iive> never heard of that. you know encoding is done by libx264. [03:24] <WilliamWS> compilation completed [03:24] <WilliamWS> \o\ [03:42] <WilliamWS> anyone recomends me some good material with examples to learn how to sue ffmpeg? [03:42] <WilliamWS> something besides the documentations. [03:43] <Plorkyeran> a lawyer would be a good start [03:44] <klaxa> haha [03:44] <WilliamWS> This specific lawyer coffin has some ffmpeg instructions written on it? [03:46] <WilliamWS> like ffmpeg -i rawlawyer -f earth /home/lawyer -f hell /home/coffin [03:46] <klaxa> hmm... i doubt that'll work [03:46] <klaxa> hmm... or does it? multiple output files, etc [03:46] <WilliamWS> hi klaxa, your nick sounds really nice [03:47] <WilliamWS> klaxa... interesting, any special meaning? [03:47] <klaxa> thanks, it got it from a random string generator [03:47] <WilliamWS> klaxa, I do not know, I am still trying to understand the documentation [03:48] <klaxa> well general syntax is: ffmpeg [-f input_format [input format options]] -i input [output options] [-f output format] output [03:49] <klaxa> output options is probably what you want to adjust most, encoding settings, filters, etc. [03:49] <WilliamWS> My problem is with streamming. I want to get frames from a webcam and send it to a folder on apache domain so it can be accessed and played in "real time" [03:50] <klaxa> do you have your webcam as /dev/video0 ? [03:50] <WilliamWS> ffmpeg -f video4linux2 -i /dev/video0 -f mp4 -movflags faststart /var/www/video.mp4 [03:51] <klaxa> looking good [03:51] <WilliamWS> more or less this [03:52] <WilliamWS> the video is created but I can not read it only if I stop the covnersion and "finalize" the file [03:53] <klaxa> hmm... [03:53] <klaxa> well you wouldn't be able to watch it live anyways [03:53] <klaxa> that's now how streaming works [04:24] <smj> Which one or what is the correct format, -vf pp=al:f/c or -vf pp=al/c:f ? [04:26] <smj> oh, nevermind [04:28] <WilliamWS> how I can do that? [04:28] <smj> me? [04:28] <WilliamWS> ahhahah [04:28] <WilliamWS> smj the correct format for what? [04:29] <smj> for all the subsubsuboptions [04:29] <smj> but yeah it's | instead of / [04:30] <WilliamWS> if it is a flag you use flag1+flag2 if it is a audio and respective video you can use audioblablabla:videoblablabal [04:30] <WilliamWS> If I understood what you said... [04:32] <WilliamWS> And sorry I am still learning how to use ffmpeg and reading this: http://ffmpeg.org/ffmpeg.html [05:33] <smj> Is there a way to make a color key transparent in when using an overlay? [06:03] <yajiv> Hi folks&I have a quick question about printing encoder params&.I have a command line configuring x264 that works well from ffmpeg&however when i pass what appears to be the same from ffserver it fails, seemingly as x4->enc = x264_encoder_open(&x4->params); [06:04] <yajiv> looking through ffserver.c i noticed that it passes some encoder defaults, which i suspect are the culprits&&so my question is: is there an easy way to print out all the param values that are passed into the x264 encoder? [06:04] <yajiv> maybe a flag that prints it? [06:06] <relaxed> yajiv: try -v debug or run $(strings input | grep x264) on its output. [06:09] <yajiv> relaxed: will give it a shot &thanks [06:14] <yajiv> relaxed: nothing that prints out all the params. It's not obvious why I couldn't open the encoder when running from ffserver. Will just manually print out the params from libx264.c i guess [06:14] <yajiv> thanks [06:16] <relaxed> the output file will always contain the encoder settings. [06:31] <yajiv> yeah, that was actually my bad&..digging a little deeper to resolve my issues ow. [06:42] <vhann> Hi, I have a video file I want to play on my phone while in the subway (noisy env) but the speech volume is very low. I am trying to raise it with "ffmpeg -i infile.flv -vcodec copy -dialnorm -10 -filter volume=20dB outfile.flv" but sound doesn't seem to change [06:43] <vhann> Is there a way to get/read the (average) volume on an absolute scale (so I can compare infile and outfile to make sure outfile is louder)? [07:13] <vhann> Oh well, I demux'ed the audio, fed into audacity, amplified there and remux'ed [07:14] <vhann> It seems there was a gunshot at some point in the video and audacity wouldn't amplify (without clipping) because of that loud sound. Could there be something similar in ffmpeg? [07:29] <yajiv> relaxed: So here's the encoder output for the version that encodes super quickly. [07:29] <yajiv> Stream #0:0, 0, 1/90000: Video: h264 (libx264), yuv444p, 480x800, 1/15, q=-1--1, 90k tbn, 15 tbc [07:29] <yajiv> and here is the case that doesn't from ffserver. [07:29] <yajiv> Stream #0:0: Video: h264 (libx264), yuv444p, 480x800, q=255-1, 1000k tbn, 15 tbc [07:29] <yajiv> i see the sign issues with qmin&but i'm not sure why the tbn is out of whack. any suggestions? [07:30] <yajiv> basically both encode now, however the line above encoders and streams with a latency of 0.5s while the one through ffserver does so at a latency of 3-4 s [07:30] <Mavrik> um [07:30] <Mavrik> ffserver buffers alot [07:31] <yajiv> at what level? is it possible to disable that? [07:33] <yajiv> also my encoder params are clearly not what they should be. not sure where the ffserver pipeline is introducing the errors either&.btw, thanks for the response mavrik [07:34] <Mavrik> yajiv, basically ffserver writes input to a temporary file and then serves it with a delay [07:34] <Mavrik> I think buffering can be set [07:35] <Mavrik> yajiv, ignore the qmin and tbc/tbn things though, I've seen them be totally off on libx264 encodes many times [07:35] <yajiv> k [07:36] <yajiv> # Set this to the number of seconds backwards in time to start. Note that [07:36] <yajiv> # most players will buffer 5-10 seconds of video, and also you need to allow [07:36] <yajiv> # for a keyframe to appear in the data stream. [07:36] <yajiv> i see this [07:36] <yajiv> but i've set the player to not buffer&.so that shouldn't be an issue. [07:36] <yajiv> the writing to a temp file will certainly add latency. [07:36] <yajiv> i'll try to see if i can keep it in memory. [07:37] <Mavrik> also GOP size has a great effect on latency :) [07:38] <yajiv> when i run the command line through ffmpeg and pipe to udp and play it..i get a latency of around 300ms [07:38] <yajiv> it's when going through ffserver with the same command line that i see the latency go up to 3s or so. [07:41] <yajiv> i'm using the x264 config for zero latency and speed (slightly modified) but works pretty well. [11:40] <smj> how do I negate overlayed input? [11:41] <durandal_1707> -vf negate? [11:41] <smj> but where do I put that? [11:42] <durandal_1707> maybe after overlay [11:43] <smj> -i :0.0+1570,143 -filter_complex 'negate,overlay=10:10,negate' this does what I want but it could be more efficient [11:44] <durandal_1707> why you add dupe overlay? [11:44] <durandal_1707> *negate [11:44] <durandal_1707> whay you need negated? [11:46] <smj> -filter_complex 'negate,overlay=10:10' negates the whole output [11:47] <smj> no, -filter_complex 'negate,overlay=10:10' negates the main image [11:48] <EvilTengil> Hi! Does anyone here have experience with real-time streaming? I am trying to stream the camera feed from an iPhone to VLC on a computer. Currently I am using MJPEG but I'm having trouble with stutter and long delays. [11:48] <smj> -filter_complex 'overlay=10:10,negate' negates the whole output [11:48] <durandal_1707> smj: and what you need negated? [11:48] <smj> I don't need anything negated, I just want, and why does that matter? [11:55] <durandal_1707> smj: see "[1]negate[a],[0][a]overlay=10:10" [11:55] <durandal_1707> only one negate [11:57] <smj> thanks, I might be getting a hang of it [11:58] <durandal_1707> about speed, without opacity, its just memcpy so it could be made faster in that case [12:09] <EvilTengil> Does MJPEG support variable frame rate? [12:10] <JEEB> variable frame rate is generally either supported or not on a *container* format level, not on the video format level [12:11] <JEEB> f.ex. MJPEG in mkv and mov should support it just fine [12:16] <EvilTengil> ok, i selected my format by calling avformat_alloc_output_context2(&oc, NULL, "mjpeg", filename); [12:17] <EvilTengil> it seems to work, but mjpeg is just the coder right? does this function actually give me some other format that happens to work or what is happening exactly? [12:17] <EvilTengil> coder = codec [12:20] <JEEB> ffmpeg -formats |grep "mjpeg" gives me "DE mjpeg raw MJPEG video" [12:20] <JEEB> so you are just setting your output libavformat format to be raw MJPEG [12:20] <JEEB> so you are not putting it into any container [12:21] <JEEB> libavcodec handles decoding and encoding, libavformat handles demuxing and muxing and general parsing [12:30] <EvilTengil> is this the correct way of doing it or should i put it into a mpegts container or something else? [12:30] <JEEB> if I were to do something with that thing that doesn't only take in raw MJPEG streams, I would probably put it into something since a raw picture stream doesn't contain ANY timing information [12:30] <JEEB> it's just a slew of pictures [12:30] <EvilTengil> i'm just trying to stream a camera feed from a phone to my computer. no audio or anything. i just want to have a smooth video with low latency. [12:31] <EvilTengil> okey. do you think mpegts would be the best choice then? [12:31] <JEEB> since most phones don't read MJPEG from videos and thus you would have to write the parser/decoder by yourself, I would guess you could pack it into mkv or so? Possibly mpeg-ts as well, but I've no idea how well VFR stuff would work with it. [12:32] <JEEB> many phones nowadays do take in H.264 in mpeg-ts for example, but MJPEG is something probably unsupported by default, so you would end up doing software on the phone side as well in any case :) [12:33] <EvilTengil> i tried MKV but i got the following in my console: [12:33] <EvilTengil> [NULL @ 0x23b2000] Requested output format 'mkv' is not a suitable output format [12:34] <klaxa> try 'matroska' [12:36] <EvilTengil> that worked! :) [13:06] <smj> I'm not sure, but is x11grab's frame rate limited to 29.97? I don't think I could get smooth video recording before editing one line in x11grab.c [13:08] <durandal_1707> smj: what you edited? [13:10] <smj> I replaced "ntsc" with "59.94" on line 605 [13:11] <durandal_1707> funny, but that is not needed, same can be done with -framerate 59.94 [13:11] <smj> is it not the same as -r? [13:11] <durandal_1707> nope [13:11] <durandal_1707> -framerate is option for x11grab only [13:34] <smj> I've spent way too much time perfecting my encoder settings [13:34] <smj> oh, wrong channel [13:36] <EvilTengil> Does anyone know what time_base is should use if I want to encode video with variable frame rate? Also, how would I go about calculating the pts? [13:43] <EvilTengil> I tried calculating my pts like: [13:43] <EvilTengil> frame->pts = av_rescale_q(1, video_st->codec->time_base, video_st->time_base) * delta_time_in_seconds [13:44] <EvilTengil> but the video freezes up after a second or so [16:21] <podman> does anyone know if iOS supports variable frame rate h264 video? [17:05] <Zeranoe> Has anyone successfully compile FFmpeg with vid.stab support? [17:06] <JEEBsv> podman: it'd be pretty hard to have a mp4 parser that doesn't support VFR [17:06] <JEEBsv> how high rates it supports is a separate problem of course [17:06] <durandal_1707> Zeranoe: tried it? [17:06] <Zeranoe> durandal_1707: You got it working? [17:06] <podman> JEEBsv: yeah, I'm just trying to figure out why a customer's video isn't playing properly in iOS [17:07] <podman> or... really play at all [17:07] <durandal_1707> Zeranoe: i never tried [17:08] <durandal_1707> Zeranoe: i just ask, if you tried did you find some bug.... [17:08] <podman> JEEBsv: the only thing I can think of that seems out of place is that the output video is VFR [17:09] <Zeranoe> durandal_1707: I'm just not sure how to compile it properly. FFmpeg's configure fails at pkg-config detection of the lib. But the lib provides no .pc files so I'm not surprised [17:10] <JEEBsv> podman: the only thing that comes to mind is that the mp4 parser could have a limited integer precision with the timstamps and certain kinds of timestamps (regardless of CFR/VFR) might make that overflow and fail [17:10] <JEEBsv> that said I don't have an iDevice so I can't really test jack shit with them :) [17:10] <JEEBsv> in any case it'd be rather hard to implement an mp4 parser that doesn't support VFR [17:10] <JEEBsv> since you are in any case parsing the timestamps [17:11] <podman> JEEBsv: actually, it looks like the encoding service decided to change the level from baseline(a)L3.0 to baseline(a)5.1 [17:11] <podman> they used to use @L3.0 [17:11] <durandal_1707> Zeranoe: but you must have pc for vid.stab instealled somewhere.. [17:13] <Zeranoe> durandal_1707: If there is a .pc file for vid.stab, it didn't come from the official source [17:15] <durandal_1707> Zeranoe: but configure check for it, so it must be somewhere [17:18] <podman> JEEBsv: anything here jump out at you? https://gist.github.com/podman/0ca4972e2dc0c06e7748 [18:14] <Plorkyeran> iirc iphone only supports 3.1 [18:14] <Plorkyeran> so 5.1 is the issue [19:57] <EvilTengil> Does anyone know which container formats that supports data (AVMEDIA_TYPE_DATA) streams? I want to embed some floating point values for each frame. [21:11] <computer_hobby> I am trying to overlay a video clip (snow.mp4) onto another video (pc2.mp4). Both clips are 6 seconds long. 2 seconds in to pc2.mp4 playing I would like snow.mp4 to appear and play for 2 seconds and then disappear. I've got this working except for the 'play for 2 seconds and disappear' part. [21:11] <computer_hobby> Heres a picture of the overlay working: http://i.imgur.com/PaOxHAy.png [21:11] <computer_hobby> But then The video continues to play into my next scene: http://i.imgur.com/P90Lp0h.png [21:12] <computer_hobby> here's my command: [21:12] <computer_hobby> ffmpeg -y -i pc2.mp4 -i snow_short.mp4 -filter_complex "[1]scale=259:150,deshake[insert]; [0][insert]overlay=x='w-1':y=0" overlay.mp4 [21:12] <computer_hobby> is there a way to clip the duration of the overlay being shown? [21:13] <computer_hobby> err sorry here is the command I used: [21:13] <computer_hobby> #ffmpeg -y -i pc2.mp4 -itsoffset 2 -i snow_short.mp4 -filter_complex "[1]scale=259:150[insert]; [0][insert]overlay=x=116:y=143" overlay.mp4 [21:13] <durandal_1707> what version? [21:13] <computer_hobby> 1.2 [21:13] <computer_hobby> I used the -tsoffset 2 param to delay the overlay, can't figure out how to stop it though. `-t` stops both clips not just the overlay [21:14] <durandal_1707> git one have enable option expression for overlay (timeline support) so one can set when overlay will be active [21:14] <computer_hobby> oh interesting [21:17] <computer_hobby> is this the feature? https://github.com/FFmpeg/FFmpeg/blob/15d2f26998909ffd7d2779946df9d8e0c07ef… [21:19] <ubitux> computer_hobby: yes, http://ffmpeg.org/ffmpeg-filters.html#Timeline-editing [21:19] <ubitux> overlay supports it [21:22] <computer_hobby> does this look like the right syntax, assuming I've got the git version installed? overlay=enable='between(t,3,5)':x=116:y=143 [21:24] <ubitux> it looks ok yes [21:25] <computer_hobby> So the docs page are for the latest in git? that explains why I get 'Option 'repeatlast' not found' with 'overlay=x=116:y=143:repeatlast=0' which would seem another way to solve it is I clipped down the video beforehand [21:25] <computer_hobby> Thanks for the help ubitux and durandal_1707 [21:25] <computer_hobby> *solve it if [21:26] <ubitux> yes, online docs are for the latest git [22:59] <computer_hobby> worked great: [22:59] <computer_hobby> ffmpeg -y -i pc2.mp4 -itsoffset 2 -i snow.mp4 -filter_complex "[1]scale=259:150[insert]; [0][insert]overlay=enable='between(t,2,5.07)':x=116:y=143" overlay.mp4 [22:59] <ubitux> :) [23:03] <computer_hobby> thanks! [23:43] <dn> hello guys, could somebody tell me what is the fatest mp4 flash and adobe cli example, i don't care about -b etc just want to encode video really fast [00:00] --- Fri May 3 2013

1 0

ffmpeg-devel.log.20130501
by burek 02 May '13

02 May '13

[00:09] <J_Darnley> seanDuncan: The wrong channel but that is because qpfile is a feature in x264.c [00:09] <J_Darnley> not in the library [00:12] <seanDuncan> darn. I'm sure I can pipe into x264 via ffmpeg to get this feature but that bugs me because it defeats the purpose behind ffmpeg's 'x264opts' argument. [00:12] <seanDuncan> do you know if ffmpeg will ever have qpfile available through x264opts? [00:16] <J_Darnley> No. [00:16] <J_Darnley> The qpfile just forces the encoder to be called with the given frame types [00:17] <J_Darnley> You could probably add this feature to ffmpeg [00:17] <J_Darnley> Or rather, one could. [00:19] <seanDuncan> yes, I fully understand the beauty of "patches welcome" in the world of ffmpeg. :) [00:19] <seanDuncan> ok, looks like piping may be my only option if I want to use this feature [00:21] <seanDuncan> In case you're curious , I'm trying to do keyframe aligned transcodes across multiple frame sizes while making use of the quality enhancements that come from MB-tree (like so: http://pastebin.com/u0FYfZnT) [00:23] <J_Darnley> You should have a look at the ffmpeg docs. I vaguely recall a feature for forcing keyframes. [00:23] <cbsrobot> seanDuncan: try: &.. :no-scenecut=1:qpfile=homeMovie1080p.stats.qp: && [00:23] <J_Darnley> Perhaps I'm imagining it though [00:26] <seanDuncan> @J_Darnley you are correct there is a force keyframes feature in ffmpeg, however that one will only force keyframes at a 'seconds' level (from the ffmpeg docs " first frames after each specified time") and that isn't precise enough for my needs [00:27] <cone-480> ffmpeg.git 03Michael Niedermayer 07master:7edb984dd051: avutil/log: Fix context pointer used for get_category() [00:27] <cone-480> ffmpeg.git 03Michael Niedermayer 07master:c3bd306e78f9: h264_ps: check croping values [00:27] <cone-480> ffmpeg.git 03Michael Niedermayer 07master:5299ee54a73f: h264: assert that croping values at slice level are consistent [00:27] <cone-480> ffmpeg.git 03Michael Niedermayer 07master:fbaf75b166cd: h264: Use the correct croping values. [00:27] <J_Darnley> Do you have varying frame rates? [00:28] <J_Darnley> Between the different encodes I mean. [00:28] <seanDuncan> no, constant frame rates across all. but scene detect is on in the first transcodes. so sometimes I won't be able to specify , with right amount of preciseness, the correct frame to put an IDR [00:29] <cbsrobot> seanDuncan: did it work ? [00:29] <seanDuncan> @cbsrobot, haven't tried it yet, but I don't think that will cause it to recognize the qpfile option [00:30] <seanDuncan> brb [01:06] <seanDuncan> @J-Darnley: Did you have a suggestion given the fact that I have constant frame rates across all bitrates? [01:07] <J_Darnley> No, I just thought that ffmpeg's feature would be accurate enough in that case [01:43] <seanDuncan> @J-Darnley ok, thanks anyways. It seems like the force_keyframe feature just wont fit my needs. I'll test if allows for something like setting keyframes at times like 1.001 or 00:05:10.042 but even if it does my command line string will become huge for anything that has more than 15 keyframes [04:58] <cone-739> ffmpeg.git 03Clément BSsch 07master:e3e66568ccee: doc/filters: replace hqdn3d with smartblur in timeline section. [04:58] <cone-739> ffmpeg.git 03Clément BSsch 07master:14f45bcbe73f: doc/filters: fix scale example equivalence. [07:53] <kierank> there used to be yuv blending support in drawtext, has that gone? [09:51] <cone-670> ffmpeg.git 03Anton Khirnov 07master:95220be1faac: indeo3: fix off by one in MV validity check [09:51] <cone-670> ffmpeg.git 03Michael Niedermayer 07master:71091698a231: Merge commit '95220be1faac628d849a004644c0d102df0aa98b' [10:00] <cone-670> ffmpeg.git 03Anton Khirnov 07master:bda9e6d923df: indeo3: use put_pixels instead of put_no_rnd_pixels in copy_cell() [10:00] <cone-670> ffmpeg.git 03Michael Niedermayer 07master:d8b6557c866d: Merge commit 'bda9e6d923df6d46cc5cafb30a59a01d2514fa61' [10:11] <cone-670> ffmpeg.git 03Anton Khirnov 07master:3d10614529b0: FATE: add an additional indeo3 test [10:11] <cone-670> ffmpeg.git 03Anton Khirnov 07master:e036bb7899d0: lavc: clear AVBuffers on decoded frames if refcounted_frames is not set [10:11] <cone-670> ffmpeg.git 03Michael Niedermayer 07master:52fa2945b594: Merge commit 'e036bb7899d0faca9159206be9bf5552e76e7633' [10:17] <cone-670> ffmpeg.git 03Anton Khirnov 07master:1012155722fc: vf_split: fix description [10:17] <cone-670> ffmpeg.git 03Anton Khirnov 07master:949c603ed921: af_channelmap: remove now unnecessary goto [10:17] <cone-670> ffmpeg.git 03Michael Niedermayer 07master:a1b887971441: Merge commit '949c603ed9210df0e1b8e1aa82c71b93543d8071' [10:23] <cone-670> ffmpeg.git 03Anton Khirnov 07master:8aaab1113c00: af_channelmap: ensure the output channel layout is valid. [10:23] <cone-670> ffmpeg.git 03Anton Khirnov 07master:c22263d3e813: graphparser: only print filter arguments if they are non-NULL [10:23] <cone-670> ffmpeg.git 03Michael Niedermayer 07master:e41d3aba0123: Merge commit 'c22263d3e813d442df8fa5f5ba8993573fe775d8' [10:31] <cone-670> ffmpeg.git 03Anton Khirnov 07master:2548834b2248: vf_setpts: cosmetics, reformat [10:31] <cone-670> ffmpeg.git 03Michael Niedermayer 07master:0061ba044a44: Merge commit '2548834b2248ad4d7da9bd0afed09d7d0cddeeb0' [11:22] <cone-670> ffmpeg.git 03Anton Khirnov 07master:b472938233b9: lavfi: add an asetpts filter [11:22] <cone-670> ffmpeg.git 03Michael Niedermayer 07master:1cba7fa364a9: Merge commit 'b472938233b98178ed6c1353c37e0dc7ab585902' [11:47] <cone-670> ffmpeg.git 03Anton Khirnov 07master:a1e05b0487a1: lavfi: add trim and atrim filters. [11:47] <cone-670> ffmpeg.git 03Michael Niedermayer 07master:3319679d0a36: Merge commit 'a1e05b0487a1939334c2920fc7f9936bc9efe876' [12:17] <cone-670> ffmpeg.git 03Anton Khirnov 07master:3d624420086e: avconv_opt: allocate an OutputFile before the streams. [12:17] <cone-670> ffmpeg.git 03Michael Niedermayer 07master:1924a61a0a39: Merge commit '3d624420086ed3cd0c74f4510f0285968e21c117' [13:39] <cone-670> ffmpeg.git 03Clément BSsch 07master:24939d567b8a: fate/pixfmts: use the appropriate variable names for temporary files. [14:02] <cone-670> ffmpeg.git 03Michael Niedermayer 07master:1a94d7c7c8ca: avfilter/trim: improve rounding precission [14:17] <ubitux> saste: i'm done with the fate-run.sh/filter*.mak thing, but the overlay and alphamerge rgb versions started to fail on msvc for some reason [14:18] <ubitux> so far, the only relevant diffs are that the tests are using the filter complex system, framecrc instead of a md5, and no fifo [14:18] <ubitux> the fifo doesn't seem to be the case [14:19] <ubitux> md5 would have mismatch also, so i doubt it's the framecrc/md5 thing [14:19] <ubitux> so possibly the problem is because of the switch to the filter complex system [14:19] <ubitux> there is some kind of chroma shift issue of some kind [14:25] <cone-670> ffmpeg.git 03Anton Khirnov 07master:a83c0da539fb: avconv: make -t insert trim/atrim filters. [14:25] <cone-670> ffmpeg.git 03Michael Niedermayer 07master:76c1f9200fa3: Merge commit 'a83c0da539fb07260310bc3b34056239d2b138b2' [14:28] <ubitux> michaelni: -s or -ss? [14:46] <cehoyos> j-b: Any news on c99-to-c89 ? Should we try to find another place to host it? [14:46] <cehoyos> (Same question for ftp...) [14:47] <ubitux> is it that much trouble to host it ourselves? [14:47] <ubitux> i mean, it's just a little tarball isn't it? [14:48] <cehoyos> Did you read the license terms? (I didn't.) [14:48] <cehoyos> And, even more important: Did you try to build it? [14:48] <ubitux> ah, right there is the auto build part [14:52] <cehoyos> nevcairiel: Do you have many (interlaced) vc1 samples? We only have one sample left that shows artefacts. (Or are you using the binary decoder?) [15:02] <michaelni> ubitux, -ss [15:32] <ubitux> i wonder if overlay shouldn't belong as a pixfmt test.. [15:36] <cehoyos> michaelni: Is the MPlayer homepage on svn? [16:52] <michaelni> cehoyos, i dont remember, i dont think i ever did anything with the mp hp [16:54] <cehoyos> michaelni: I found the repo, thank you! [17:03] <cehoyos> michaelni: A user posted an aac encoding patch on libav-user: http://thread.gmane.org/gmane.comp.video.ffmpeg.libav.user/10693/focus=10709 [17:31] <cone-151> ffmpeg.git 03Anton Khirnov 07master:8cd472d3f947: avconv: make output -ss insert trim/atrim filters. [17:31] <cone-151> ffmpeg.git 03Michael Niedermayer 07master:4818388e6c21: Merge commit '8cd472d3f947a6233e7dc628f0dc71c74e62413a' [17:52] <cone-151> ffmpeg.git 03Michael Niedermayer 07master:2514b80045d5: sws: share xyzgamma tables. [17:57] <cone-151> ffmpeg.git 03Diego Biurrun 07master:643e433bf744: mpegaudiosp: More consistent names for ppc/x86 optimization files [17:57] <cone-151> ffmpeg.git 03Diego Biurrun 07master:6b110d3a739c: ppc: More consistent names for H.264 optimizations files [17:57] <cone-151> ffmpeg.git 03Michael Niedermayer 07master:3ad5d8694c76: Merge commit '6b110d3a739c31602b59887ad65c67025df3f49d' [18:01] <cone-151> ffmpeg.git 03Diego Biurrun 07master:f5eecee865a7: ppc: util_altivec: Surround AltiVec-related code by appropriate ifdefs [18:01] <cone-151> ffmpeg.git 03Michael Niedermayer 07master:148f69eb4e4c: Merge commit 'f5eecee865a731d2412fde2f73c29f8f8115c499' [18:06] <cone-151> ffmpeg.git 03Diego Biurrun 07master:a053dbfcfb3a: ppc: Move AltiVec utility headers out of AltiVec ifdefs [18:06] <cone-151> ffmpeg.git 03Diego Biurrun 07master:42b9150b0d4f: fft: Remove pointless #ifdefs around function declarations [18:06] <cone-151> ffmpeg.git 03Michael Niedermayer 07master:ebd99ff112e8: Merge commit '42b9150b0d4f0a130c1d93dc991fd5412743a8cf' [18:11] <cone-151> ffmpeg.git 03Diego Biurrun 07master:38282149b6ce: ppc: More consistent arch initialization [18:11] <cone-151> ffmpeg.git 03Michael Niedermayer 07master:cae8f469fee7: Merge commit '38282149b6ce8f4b8361e3b84542ba9aa8a1f32f' [18:21] <cone-151> ffmpeg.git 03Diego Biurrun 07master:088f38a4f9f5: avcodec: Drop unnecessary ff_ name prefixes from static functions [18:21] <cone-151> ffmpeg.git 03Michael Niedermayer 07master:04fa81877913: Merge commit '088f38a4f9f54bb923405c67c9e72d96d90aa284' [18:25] <cone-151> ffmpeg.git 03Diego Biurrun 07master:383fd4d47807: arm: Drop unnecessary ff_ name prefixes from static functions [18:25] <cone-151> ffmpeg.git 03Diego Biurrun 07master:f2e9d44a5757: x86: Drop unnecessary ff_ name prefixes from static functions [18:25] <cone-151> ffmpeg.git 03Diego Biurrun 07master:7f75f2f2bd69: ppc: Drop unnecessary ff_ name prefixes from static functions [18:25] <cone-151> ffmpeg.git 03Michael Niedermayer 07master:dbcf7e9ef7f8: Merge commit '7f75f2f2bd692857c1c1ca7f414eb30ece3de93d' [18:32] <cone-151> ffmpeg.git 03Diego Biurrun 07master:e926b5ceb196: avformat: Drop unnecessary ff_ name prefixes from static functions [18:32] <cone-151> ffmpeg.git 03Michael Niedermayer 07master:a5ba4e186bcb: Merge commit 'e926b5ceb1962833f0c884a328382bc2eca67aff' [18:40] <cone-151> ffmpeg.git 03Derek Buitenhuis 07master:4cdb3c50b538: avcodec: doxy: Clarify what get_buffer2 must fill in buf[] [18:40] <cone-151> ffmpeg.git 03Derek Buitenhuis 07master:00b8e650e83b: avcodec: doxy: Clarify how to use buf[] when implementing get_buffer2() [18:40] <cone-151> ffmpeg.git 03Derek Buitenhuis 07master:af8a47905a3f: avutil: doxy: Small clarification for av_buffer_create() [18:40] <cone-151> ffmpeg.git 03Michael Niedermayer 07master:67731ef40f9f: Merge remote-tracking branch 'qatar/master' [19:56] <durandal_1707> is dupe asetpts gonna be removed? [19:57] <durandal_1707> *documentation [19:59] <ubitux> durandal_1707: do you plan to merge alphaextra with plane* filters? [20:00] <ubitux> alphaextract* alphamerge [20:02] <durandal_1707> well for alphaextract yes - exceppt i would need to add rgba slow code..., alphamerge - maybe if you explain what planesmerge or better name for filter would do [20:02] <durandal_1707> alphaextract would just become what negate is - alias [20:03] <durandal_1707> and i also need to add 16 bit support [20:04] <ubitux> would be nice [20:07] <durandal_1707> it would be also nice if I could know outlink number from config_output, so sscanf is not needed and I could name pads more user friendly [20:08] <durandal_1707> some pointer trickery could be done... [20:20] <gnafu> So it looks like ffmpeg has three ProRes encoders. Which one is the best at this point? [20:22] <cone-151> ffmpeg.git 03Nicolas George 07master:276fcbde6c55: lavf: data muxer and demuxer. [20:22] <cone-151> ffmpeg.git 03Nicolas George 07master:ceb33f3aee15: lavc: copy subtitle_header when copying context. [20:22] <cone-151> ffmpeg.git 03Michael Niedermayer 07master:d602f16a377a: Merge remote-tracking branch 'cigaes/master' [20:22] <durandal_1707> gnafu: really? there are 2 only with 3 names [20:23] <gnafu> Aah, okay. I was just looking at -codecs output without looking further :-P. [20:27] <gnafu> Seems the default, prores_anatoliy, is much faster than prores_kostya. [20:27] <gnafu> ~5 times as fast on my machine :-P. [20:28] <ubitux> what about the output size? [20:29] <gnafu> I didn't really pay attention to anything else at this point. I don't have any need for ProRes encoding; I just had a sudden urge to see what was there XD. [20:29] <gnafu> I'll probably compare them more when I'm more bored. [20:30] <gnafu> 15 fps vs. 3 fps was all I paid attention to so far. [20:30] <ubitux> :) [20:30] <JEEB> afaik kostya's implementation is one that sets its defaults for better compression [20:30] <gnafu> (720p video on an older AMD64.) [20:30] <JEEB> you could make it faster with settings IIRC [20:31] <gnafu> Yeah, that's what I get from the blog post I just found :-). [20:48] <smj> http://pastebin.fi/raw/vc4z05gn5v is the OpenCL detection a bit inaccurate? It worked fine a few commits ago, but sure, it might all be just my system. [20:49] <smj> I might have broken my x264 for now [20:50] <durandal_1707> is OpenCL really that good? [20:50] <JEEB> if that's x264's configure then see its configure's help about disabling opencl [20:51] <JEEB> durandal_1707, no [20:51] <JEEB> also I think x264's opencl shouldn't get enabled on !windows to begin with for now :s [20:51] <JEEB> or wait... [20:51] <JEEB> no [20:52] <JEEB> o_O [20:52] <smj> it's ffmpeg [20:52] <JEEB> try adding the ldflag to link to whatever gives you dlopen and friends [20:53] <smj> I like the idea of moving some load off my i3 to GTX650 [20:54] <JEEB> no idea what ffmpeg does with opencl, but the x264's opencl is not really worth it in general [20:54] <JEEB> it's a nice example of something actually workable being implemented, but nothing more :s [20:54] <smj> but if it's not currently in much use, I guess I compile without it [20:56] <smj> that's ffmpeg's ./configure and to me it looks like broken libx264 prevents OpenCL from being detected correctly [20:56] <JEEB> well the errors just mean that it can't find those "dl*" things from the currently linked libraries [20:57] <JEEB> it's possible that it just was linked against some other CRT or whatever if you're on one of those "on-the-edge" distros [20:57] <JEEB> but I really don't know, which is why I said you should try to add the ldflag which would link whatever provides those functions first [21:17] <smj> okay, ffmpeg finds opencl after libx264 is linked with -ldl [21:29] <durandal_1707> why http://ffmpeg.org/projects.html mentions qemacs? [23:07] <cehoyos> gnafu, JEEB: Afair, Anatolyi's prores encoder is both faster and produces smaller output, the person who tested it initially didn't realize and implemented interlaced encoding for the other encoder;-( (iirc) [00:00] --- Thu May 2 2013

1 0

ffmpeg.log.20130501
by burek 02 May '13

02 May '13

[00:00] <SeanDuncan_> I get the possibility of piping to x264. it's just that piping defeats the purpose of why x264opts exists in ffmpeg in the first place. :) That's just me complaining tho :) [00:00] <durandal11707> i'm not talking about piping [00:01] <ElMarikon> or u try out if it hides in ffmpeg's -x264-params [00:01] <ElMarikon> that is a strange other option there... [00:02] <SeanDuncan_> @durandal11707 I didn't quite know what you meant when you said "!pb" [00:03] <SeanDuncan_> cna you elaborate? [00:04] <SeanDuncan_> oh [00:04] <SeanDuncan_> oops, didn't see your question there [00:04] <SeanDuncan_> yes, I made sure it works with x264 [00:04] <iive> SeanDuncan_: !pb is command to the helper bot, to provide instructions how to use pastebin [00:05] <SeanDuncan_> @ durandal110707 yes, I made sure it works with x264 [00:05] <SeanDuncan_> thanks @iive . color me red [00:07] <SeanDuncan_> @durandal11707 my .qp file looks just like the one described (format wise) [00:08] <SeanDuncan_> at http://mewiki.project357.com/wiki/X264_Settings#qpfile [02:06] <Zarx> does ffmpeg support trimming by frame range, or only by time? [02:07] <durandal11707> Zarx: see select filter [02:07] <durandal11707> it support both by time-pts and frame number [02:10] <durandal11707> but this is not for -c copy only case ... [02:10] <Zarx> where can i find documentation for using the select filter? [02:13] <Zarx> thanks [02:20] <otters> so if ffmpeg can't do it, what IS good for editing mp4 chapters? [03:13] <plut0> any suggestions on a good crf # for encoding bluray movies? [03:14] <zap0> depends on the content [03:15] <zap0> if it were that simple to state, don't you think it'd be automated by now [03:15] <plut0> i would think so [03:28] <shakaponk> hello folks, compiling ffmpeg like in the ffmpeg ubuntu compilation guide getting error xext not found [03:28] <shakaponk> what is missing? [03:29] <klaxa> xext apparently [03:29] <klaxa> even though this is not an ffmpeg command, please paste everything regardless [03:31] <shakaponk> hold on, different machines [03:32] <shakaponk> pastie.org/7745934 [03:33] <shakaponk> doin the same steps as here: http://ffmpeg.org/trac/ffmpeg/wiki/UbuntuCompilationGuide [03:33] <shakaponk> stepinstalling ffmpeg [03:33] <klaxa> try installing libxext-dev [03:33] <klaxa> then run configure again [03:34] <klaxa> um... did you install all the packages listed at the top? [03:34] <shakaponk> yes i did... that's the package name i would like to know, couldn't find it [03:35] <klaxa> so can you install it now or is it somewhat missing? [03:35] <shakaponk> what is the package name for xfixes? quite the same error [03:35] <klaxa> libxfixes-dev [03:36] <klaxa> these files are listed in the preparation packages for compilation at the top of the page you linked [03:36] <shakaponk> seems to work, movin on, thx a lot [03:37] <shakaponk> klaxa: oh i guess i lost track and executed for servers instead... now i see it [03:37] <shakaponk> thx anyway for your input [03:37] <klaxa> heh [13:44] <test4> hi ... i need a codec for a special avi (from corean samsung camera ... ) i remember the name is something like "huiiffuijg" but i dont remember the exact name .... can U help me ? [13:45] <LithosLaptop> huffyuv? [13:45] <test4> exact mister :) [13:46] <test4> u won [13:46] <LithosLaptop> :) [13:46] <LithosLaptop> I was just about to close irc [13:46] <test4> ty ... that was all :) [13:46] <LithosLaptop> haha [13:46] <LithosLaptop> no prob [13:47] <test4> bye have a good day .. sunny :) [14:14] <chrisbuchholz> Hey guys. Can somebody tell me what the exit code 5 mean in ffmpeg? [14:35] <chrisbuchholz> I am trying to run ffmpeg from php by running exec('/usr/local/bin/ffmpeg') but i get error code 5 [14:54] <xlinkz0> chrisbuchholz: run with -report and check the log [14:55] <chrisbuchholz> im afraid it doesnt work. Just get empty output and error code 5 [14:57] <xlinkz0> if no output then you have a broken ffmpeg instalation [14:57] <xlinkz0> fix that [16:46] <bugggg> anyone here with a samsung tv have a working set of arguments for ffmpeg that will make it play 1080p full sbs 3d movie trailers? the half sbs ones work just fine [18:47] <DooMMasteR> hmm [18:47] <DooMMasteR> maybe someon can help [18:48] <DooMMasteR> ffmpeg version 1.2 Copyright (c) 2000-2013 the FFmpeg developers knows -start_number [18:51] <DooMMasteR> ffmpeg -start_number 0767 -i IMG_%04d.JPG -s 1280:-1 -r 5 -vcodec libx264 -crf 15 test.mkv [18:52] <DooMMasteR> Unrecognized option 'start_number' [18:52] <DooMMasteR> is the return for that sequence -.- [18:55] <DooMMasteR> hmm on another ffmpeg version that line works fine -.- [18:59] <klaxa> http://ffmpeg.org/ffmpeg-formats.html#image2-2 [19:33] <CalimeroTeknik> er, is that normal? ffmpeg: relocation error: ffmpeg: symbol swr_get_class, version LIBSWRESAMPLE_0 not defined in file libswresample.so.0 with link time reference [19:50] <CalimeroTeknik> is this supposed to work? I get either the error up there with the -loop 1 option, or a video that lasts for a fraction of second http://ffmpeg.org/trac/ffmpeg/wiki/Create%20a%20video%20slideshow%20from%20… [19:53] <manitu> hi ho.. i have a little question.. i recovered videos using testdisk, but ffmpeg gives me that error now: "Could not find codec parameters for stream 0 (Video: h264 (avc1 / 0x31637661), 1920x1080): unspecified pixel format" [19:53] <manitu> i have another, working files from that camera.. is there any way to manually define those parameters? [19:53] <manitu> i want to convert mov to mp4 [19:54] <CalimeroTeknik> yes indeed, you can force these [19:54] <manitu> it already writes the 1920x1080 down, but how can it not find the pixel format.. or is that something else [19:57] <CalimeroTeknik> try x264 [19:57] <CalimeroTeknik> or variants like x264rip-tmod-git [20:03] <WilliamWS> hi [20:04] <WilliamWS> how fast is ffmpeg in matter of accessing a camera and creating a image file from a single frame? Some extras, the processor is very limited and memory is very low [20:05] <pzich> I think it frequently depends on the camera drivers, actually [20:05] <pzich> in terms of getting the camera powered up and ready to go [20:05] <WilliamWS> My point is, ffmpeg is a good way to make this or it will have a inferior speed compared with other options like mplayer, etc? [20:06] <pzich> I haven't personally tried it, but I doubt ffmpeg/mplayer would make a big different in the whole process [20:06] <WilliamWS> the processor in question is a pentium MMX 266MHz with 64MB RAM under linux [20:07] <WilliamWS> I would like to make a video stream from this machine, but... yeah I know... [20:07] <bugggg> ah it looks like people are awake now, i'll re-ask something that i asked earlier... [20:07] <bugggg> anyone here with a samsung tv have a working set of arguments for ffmpeg that will make it play 1080p full sbs 3d movie trailers? the half sbs ones work just fine [20:08] <WilliamWS> ffmpeg can keep the camera running and get frames only when a particular event happens? [20:09] <WilliamWS> ffmpeg has too many options and really I took a long time on the docs to "understand" something usefull. [20:09] <WilliamWS> Sorry if I am asking something stupid. [20:09] <manitu> CalimeroTeknik, how to configure that? other video has Stream #0:0(eng): Video: h264 (Main) (avc1 / 0x31637661), yuv420p, 1920x1080 [SAR 1:1 DAR 16:9], 14729 kb/s, 29.97 fps, 29.97 tbr, 30k tbn, 59.94 tbc [20:10] <SeanDuncan_> @manitu: you may want to try adding '-probesize 5000000 -analyzeduration 500000' [20:11] <SeanDuncan_> sometimes that helps ffmpeg to figure it out [20:11] <durandal_1707> SeanDuncan_: why you keep adding @ before nick? [20:12] <SeanDuncan_> the colorspace, I mean [20:13] <manitu> SeanDuncan_, doesn't work also.. normal videos taken by the cam works, but the problem is that they are recovered (my girlfriend deleted :/ ).. looks like it couldn't recover all meta-data.. i hope they are not broken now [20:15] <WilliamWS> manitu you should delete girlfriend to avoid this in the future. [20:15] <WilliamWS> to prevent* [20:15] <manitu> WilliamWS, i know, but baby still needs the food [20:15] <WilliamWS> manitu that sad =/ [20:15] <WilliamWS> hahahah [20:16] <manitu> :D [20:16] <WilliamWS> If is from some interest I am compiling ffmpeg on the refered machine right now [20:23] <CalimeroTeknik> manitu, if I were you I'd just concat your video with some of the beginning of a file that has the same format [20:25] <CalimeroTeknik> dd if=validvideo bs=1024 count=1024 |cat brokenvideo > trythis ; mplayer trythis [20:28] <WilliamWS> 2:38h to complete ffmpeg compilation [20:28] <WilliamWS> on a 266MHz processor 64MB RAM Debian Squeeze [20:29] <WilliamWS> not that bad I thought I would take longer [20:33] <WilliamWS> surely ffmpeg can do it faster than mplayer and fswecam [20:33] <WilliamWS> cool lol [20:35] <WilliamWS> lol I can even make a "decent" movie using ffmpeg on this machine [20:35] <WilliamWS> I did not got it even with opencv or some v4l withery [20:35] <WilliamWS> witchery* [20:35] <manitu> CalimeroTeknik, i tried it and it's still broken hmm [20:37] <manitu> CalimeroTeknik, the really first line is nearly the same [20:38] <WilliamWS> what is the less resource (processor) hunger format to use on a video? [20:38] <manitu> niko^A|^M&mdat^@^@^@^B and niko^K^LfGmdat^@^@^@^B [20:38] <WilliamWS> I am using mp4 right now, but I do not know... [20:38] <manitu> is the a bit changed part [20:49] <WilliamWS> the Debian ffmpeg version returns this [video4linux2 @ 0x8748a60]Wrong size (0x0) /dev/video0: Error while opening file [20:49] <WilliamWS> but the one I jsut compiled worked fine [20:49] <WilliamWS> why the debian version of ffmpeg can not open the camera? [20:50] <WilliamWS> I did: ffmpeg -f video4linux2 -i /dev/video0 -vframes 120 output.mp4 [21:21] <durandal_1707> WilliamWS: debian version is very old fork of ffmpeg [21:22] <WilliamWS> I see, thank you [21:23] <WilliamWS> anyways I already compiled it. I compiled it because the debian pkg did not worked then I thought it was sme hardware issue. [21:23] <WilliamWS> hardware compatibility issue* [21:27] <Jordan_> So I'm going to be including ffmpeg.exe as gplv2 along side my proprietary, what all do I need to do? I just have a FFmpeg License.txt file in my install dir. http://pastebin.com/sJzi3XFe [21:27] <Jordan_> Just want to make sure I comply etc. [21:29] <Jordan_> is that all I need to do? [21:30] <Rathann> Jordan_: read the GPLv2 [21:30] <Rathann> if in doubt, ask a lawyer [21:31] <Jordan_> Ok give me 30 mins to read thouroughly [22:03] <Jordan_> the output from the Program is covered only if its contents constitute a work based on the Program (independent of having been made by running the Program). Whether that is true depends on what the Program does [22:04] <Jordan_> depends on what the program does, real explicit [22:04] <Jordan_> no one could judge that and no lawyer is going to know that either [22:23] <Jordan_> JEEB there? [22:26] <WilliamWS> Jordan_ do not take me bad but your mother are covered by some licence? [22:26] <WilliamWS> I mean, she gave birth to you and now you are her property? [22:28] <Jordan_> if i 7z ffmpeg and x264 how small can i make it cuz i'm getting 80mb? [22:28] <Jordan_> but the 7z i extracted that included binaries was 25mb [22:28] <WilliamWS> ok I will let this law stuff aside and keep quiet and as well ignored. [22:29] Action: WilliamWS makes silence [22:29] <Mavrik> Jordan_, basically you'll have to be able to show exact source of ffmpeg as you distribute it [22:30] <Mavrik> Jordan_, if you link against it your code will also have to be covered with GPL compatible license [22:30] <Jordan_> no not according to Section 2, there are 3 possibilites [22:30] <Mavrik> if you only invoke it then no [22:30] <Jordan_> I mean section 3 [22:30] <Jordan_> I don't link, but i do use its output [22:31] <Mavrik> output doesn't matter. [22:31] <Jordan_> i use the CLI [22:31] <Mavrik> GPL covers the code, not what software produces [22:31] <Jordan_> ffmpeg.exe stderr output as well as the encoded video [22:32] <Jordan_> not necessarily all I can tell you is that it depends [22:32] <Jordan_> someone was too lazy to explain it further [22:32] <Mavrik> again, GPL covers the software source code [22:32] <Mavrik> not the products of running a compiled executable [22:34] <Jordan_> Mavrik output from the Program is covered only if its contents constitute a work based on the Program [22:34] <Jordan_> and that depends [22:35] <Mavrik> "ok". [22:35] <Jordan_> so the answer is to anyone that uses output must be it depends, and no one knows [22:36] <Jordan_> What does "You must cause" [22:36] <Jordan_> what do they mean by "cuase" [22:38] <Jordan_> Actally 2 doesn't apply to me I didn't make any modifications [22:38] <Jordan_> (Section 2) [22:41] <Jordan_> wtf how is the source 87mb!!! [22:41] <Jordan_> after 7zip [22:43] <Jordan_> unless i can get it to 2-3mb range i'll just include a written offer [22:56] <Jordan_> OK so it looks like the largest usage of disk is the .git folder by far [22:56] <Jordan_> can I delete that with no problem, that's not actual source [22:58] <iive> Jordan_: that's the whole source repository :P [22:58] <Jordan_> so i can delete that right [22:59] <iive> i guess you don't want to distribute it, since people can just pull it from ffmpeg git repository. [22:59] <Jordan_> i doubt anyone would ever even look at it [22:59] <Jordan_> but that is not what GPL is referring to right [23:01] <iive> nope. I doubt that you have committed your changes in it. [23:06] <Jordan_> so now 7mb still kind of high [23:07] <Jordan_> what's the best compression available [23:26] <Jordan_> 7mb is just too much, I will include written offer [23:52] <Zarx> can the -progess option in ffmpeg be redirected to stderr, or only to a url or file? [23:53] <sacarasc> http://tldp.org/HOWTO/Bash-Prog-Intro-HOWTO-3.html assuming you're on *nix. [23:53] <Zarx> windows actually [23:54] <sacarasc> Then I don't know. [00:00] --- Thu May 2 2013

1 0