Ffmpeg-devel-irc

Download

ffmpeg-devel-irc@ffmpeg.org

July 2013

1 participants
62 discussions

ffmpeg-devel.log.20130705
by burek 06 Jul '13

06 Jul '13

[00:43] Action: durandal_1707 forked FFmpeg [01:05] <cone-411> ffmpeg.git 03Stefano Sabatini 07master:41fe750f4b13: lavc/bitstream_filter: do not crash in case the argument of av_bitstream_filter_close() is NULL [01:05] <cone-411> ffmpeg.git 03Stefano Sabatini 07master:d06ae0d791cd: lavf: fix documentation for avoid_negative_ts [04:09] <cone-566> ffmpeg.git 03Michael Niedermayer 07master:b009267910df: mpegts: only reopen pmt_cb filter if its different from the previous. [09:42] <saste> You can use libavcodec or libavformat in your commercial program, but any patch you make must be published. [09:42] <saste> this is BS [10:11] <burek> <@michaelni> burek, btw, how is the fate server rewrite going ? [10:11] <burek> well i didnt rewrite the server, i just wrote another set of scripts to display what's already uploaded [10:12] <burek> i.e. i've avoided to constantly decompress things on each page view [10:12] <burek> shortly, i just added a different view to logs that already exist on the fate server [10:12] <burek> and btw, i was ready a couple of months ago to test it on the fate server itself [10:13] <burek> so, i'm kinda almost finished with it, just we need to find some time to test all that on a real fate server [10:13] <burek> but i'll need a day/two to remember what was i doing in those php scripts, before we decide to test in a real environment [10:14] <burek> just let me know when is the good time for testing, so i can prepare before it [11:05] <cone-688> ffmpeg.git 03Stefano Sabatini 07master:0f85c96091e9: lavc/avcodec.h: document bitstream filter API [11:37] <michaelni_> burek, dont know when theres a good time, whats the difference between good and bad time for this ? [11:43] <saste_> michaelni_, difference between good and bad time -> consult your favorite astrologist [11:46] <cone-688> ffmpeg.git 03Kostya Shishkov 07master:410066986f44: aic: use chroma scan tables while decoding luma component in progressive mode [11:46] <cone-688> ffmpeg.git 03Jason Garrett-Glaser 07master:d222f6e39e26: cabac: x86 version of get_cabac_bypass [11:46] <cone-688> ffmpeg.git 03Luca Barbato 07master:e6d8acf6a8fb: indeo: use a typedef for the mc function pointer [11:46] <cone-688> ffmpeg.git 03Michael Niedermayer 07master:cced6f4d5816: Merge commit 'e6d8acf6a8fba4743eb56eabe72a741d1bbee3cb' [11:58] <cone-688> ffmpeg.git 03Luca Barbato 07master:f6f36ca8ca1b: indeo: Refactor ff_ivi_dec_huff_desc [11:58] <cone-688> ffmpeg.git 03Michael Niedermayer 07master:0e2056056fc8: Merge commit 'f6f36ca8ca1b2526d3abff7d7c627322d3bce912' [12:08] <cone-688> ffmpeg.git 03Luca Barbato 07master:62256010e9bc: indeo: Refactor ff_ivi_init_tiles and ivi_decode_blocks [12:08] <cone-688> ffmpeg.git 03Michael Niedermayer 07master:9b10440dcd1a: Merge commit '62256010e9bc8879e2bf7f3b94af8ff85e239082' [12:13] <cone-688> ffmpeg.git 03Luca Barbato 07master:6dfacd7ab126: indeo: Cosmetic formatting [12:13] <cone-688> ffmpeg.git 03Michael Niedermayer 07master:7d3b55c6a427: Merge commit '6dfacd7ab126aea1392949d1aa10fdc3d3eeb911' [12:33] <cone-688> ffmpeg.git 03Luca Barbato 07master:6a10142faa1c: indeo: reject negative array indexes [12:33] <cone-688> ffmpeg.git 03Michael Niedermayer 07master:9f89bebb7bff: Merge commit '6a10142faa1cca8ba2bfe51b970754f62d60f320' [12:42] <cone-688> ffmpeg.git 03Stefano Sabatini 07master:d880b2bdfff6: doc/filters: clarify documentation for overlay repeatlast option [12:49] <cone-688> ffmpeg.git 03Luca Barbato 07master:0420c810ceb4: log: pass the correct parameters to missing_feature_sample [12:49] <cone-688> ffmpeg.git 03Michael Niedermayer 07master:8fdec02fd195: Merge commit '0420c810ceb430003f9f2793c5cfa1fe84657f3d' [12:59] <kierank> does 24-bit wav exist [13:17] <cone-688> ffmpeg.git 03Luca Barbato 07master:d3635f3ab0a2: compat: wrap math.h to avoid AIX-specific clashes [13:17] <cone-688> ffmpeg.git 03Michael Niedermayer 07master:47ca9f041cde: Merge remote-tracking branch 'qatar/master' [13:50] <cone-688> ffmpeg.git 03Peter Große 07master:c6c2231227c1: libavcodec: use 64bit counter in deprecated audio encoder API to prevent problems regarding negative dts [14:02] <cone-688> ffmpeg.git 03Jean Delvare 07master:6fc8c4cc0ef6: lavfi/delogo: don't recompute the same difference again and again [14:02] <cone-688> ffmpeg.git 03Jean Delvare 07master:4e8d6b315c8d: lavfi/delogo: remember left and right samples when interpolating [15:21] <durandal_1707> there is no Firefox OS for fate [15:24] <khali> doh [15:24] <khali> Icleanedupmykeyboardandnowthespacebarnolongerworks [15:24] <khali> thissucks! [15:25] <iive> it this joke or is it for real? [15:26] <durandal_1707> space bar issue? [15:26] <saste_> khali, so you'll be saving space since now [15:28] <saste_> michaelni_, could you mentor a socis JPEG2000 project? [15:28] <durandal_1707> why? someone want to work on j2k? [15:28] <saste_> durandal_1707, i have a mail on my inbox from a potential student [15:29] <saste_> SOCIS accepted organizations list is still waited for almost always a week since the official announcement date [15:30] <saste_> BuxiNess is not sure he will be able to mentor in the program period (August - October) [15:30] <saste_> i don't even know what it is still missing... [15:30] <durandal_1707> proper encoder [15:31] <durandal_1707> saste_: did you checked your spam box? [15:31] <saste_> DEVILS jpeg2000 [15:31] <saste_> so we miss features? [15:32] <durandal_1707> well native one supports only rgb [15:33] <saste_> is it enough we lack to cover an entire socis project? [15:33] <khali> iive: for_real,_sadly_:( [15:34] <khali> iive: I_tried_fixing_it_but_could_not [15:34] <durandal_1707> you could hack your irc client to replace _ with " " [15:34] <khali> I_guess_my_work_day_is_over_and_I_can_look_for_a_replacement_keyboard_right_away [15:35] <saste_> khali, try to dump some dirt on your keyboard, to restore the state before you cleaned it up [15:35] <durandal_1707> install virual keyboard [15:35] <iive> khali: did you open the keyboard, or you cleaned just the space under the keys? [15:35] <khali> durandal_1707: I_could_but_IRC_is_not_my_main_concern_;) [15:35] <ubitux> copy a space, use ^K to put it in the buffer [15:35] <ubitux> and then use ^Y to add it again [15:35] <saste_> ubitux, yes that's handy [15:35] <khali> doh,_right_up_and_down_keys_are_dead_too [15:36] <khali> iive: I_opened_it [15:36] <ubitux> ^A ^E for start/end buffer [15:36] <saste_> who needs to go up and down anyway, just go left and right [15:36] <khali> I_guess_I_shouldn't_have_:( [15:36] <BuxiNess> jpeg2000 needs speed ! :-) [15:36] Action: saste_ stops saying silly things [15:36] <khali> saste_:right_is_dead_too [15:37] <iive> khali: then open it again, probably the line that goes to the micro-controller have been moved a little. [15:37] <saste_> khali, i suppose it is a laptop? [15:37] <saste_> yeah checking the physical connections may help [15:38] <khali> no_its_a_regular_keyboard [15:38] <khali> Key_Tronic_from_1999_or_so [15:38] <saste_> well at least it is easy to replace [15:38] <khali> saste_: true [15:39] <khali> only_I'd_love_to_find_a_Key_Tronic_again...but_they_seem_to_no_longer_exist_:( [15:40] <iive> khali: the micro controller is the chip under the led. [15:43] <durandal_1707> are there .jpx samples? [15:51] <BuxiNess> durandal_1707, I didn't know any [15:52] <durandal_1707> nothing can create them? [15:53] <BuxiNess> With START/STOP_TIMER, in a function who is supposed to be constant. If the number of decicycles inrceases. Its not normal, right? [15:53] <khali> hallelujah! [15:53] <khali> space bar is back :) [15:54] <khali> I cleaned the conductive sheets with 90° alcool [15:54] <durandal_1707> you broke % key [15:54] <khali> %no I did not% :) [15:54] <BuxiNess> Mu nb of decicycles increase afters each frame decocding 2 runs < 4 runs < 16 runs < 32 runs ... [15:55] <durandal_1707> what function? [15:55] <khali> durandal_1707: alcool concentrations are commonly counted in degrees, at least here [15:55] <BuxiNess> in jpeg2000dec [15:55] <khali> (don't ask me why, it makes little sense I admit) [15:55] <BuxiNess> decode_frame [15:55] <Skyler_> typically you want to ignore the first few lines; only the results after a large number of iterations (10000, 100000, 1000000) matter [15:55] <durandal_1707> yes, but what function? [15:57] <BuxiNess> START_TIMER after get_buffer; STOP timer after jpeg2000_decode_tile [15:57] <BuxiNess> in function jpeg2000_decode_frame [15:58] <BuxiNess> Skyler_, yes but its for one frame, i decode just 240 frames at it takes 10 to 15 seconds [15:58] <Skyler_> you might want to time a smaller section of the code then (that runs many more times) unless that's impossible [15:59] <Skyler_> START_TIMER is not really designed for accurate/consistent timing of very large code blocks [15:59] <Skyler_> e.g. you want to time something that takes on the order of 1 nanosecond to 100 microseconds [15:59] <Skyler_> not 0.1 seconds [16:01] <xlinkz0> is copy_tb set when setting -c copy? [16:07] <BuxiNess> Skyler_, and I wonder how it works in multi-threaded env ... [16:07] <BuxiNess> is there other helper functions for that ? [16:07] <BuxiNess> for measuring 0.1 seconds time [16:08] <Skyler_> it works by using the per-core timer [16:08] <Skyler_> generally, it will give suboptimal results if the length of time you're timing is close to or larger than one context switch (e.g. 1ms or whatever your kernel uses) [16:09] <Skyler_> it attempts to ignore runs on which there was a context switch [16:10] <BuxiNess> okay this why some times there are some non 0 skips [16:11] <Skyler_> I suggest using gettimeofday or something for really really long timers [16:25] <Skyler_> basically START_TIMER "skips" the first two or so, plus anything with way longer timing results than the average so far [16:25] <Skyler_> this is designed to catch context switches [16:25] <Skyler_> naturally this doesn't really work quite as expected with runs that are 100ms long and will have 10 context switches in them [16:30] <xlinkz0> do i need to open codecs when stream copying? [16:32] <BuxiNess> Is there a wrapper function to get the the thread Id ? [16:34] <durandal_1707> usually you have it when using slice threads [16:48] <BuxiNess> And also how to get the frame ID in the decoder. [16:48] <BuxiNess> ie which frame is it, 1st, 2nd, .... [16:51] <xlinkz0> what does this mean? Codec for stream 0 does not use global headers but container format requires global headers [16:52] <BuxiNess> If I can Have in the decoder the frame position or ID, who is currently decoded [16:55] <durandal_1707> why you need it: [16:55] <BuxiNess> for debug pupose, I want to know whick frame is decoded in which thread, [16:56] <BuxiNess> and how log it takes to decode the frame [16:59] <BuxiNess> pkt_pts is fine! [17:09] <cone-352> ffmpeg.git 03Paul B Mahol 07master:9dd14477884c: lavfi/cropdetect: export cropdetect info to frame metadata [17:31] <cone-352> ffmpeg.git 03Paul B Mahol 07master:413f865a9aa0: lavfi/blackdetect: support 2 more pixels formats [18:16] <cone-352> ffmpeg.git 03Jean Delvare 07master:72abce3c4672: lavfi/delogo: option show shouldn't affect band [18:16] <cone-352> ffmpeg.git 03Jean Delvare 07master:4e10d87f384a: lavfi/delogo: band width must be at least 1 [18:44] <durandal_1707> buxiness: rgb48 in jpeg2000 doesn't work at all [18:44] <durandal_1707> michaelni_: & [20:33] <cone-352> ffmpeg.git 03Paul B Mahol 07master:48f2750de8a4: get_bits: add get_bits_le() [20:33] <cone-352> ffmpeg.git 03Paul B Mahol 07master:6516a25f04ed: ADPCM IMA WAV 2, 3 and 5 bits decoder [20:43] <BBB> what's get_bits_le that BITSTREAM_READER_LE doesn't do? [20:47] <Daemon404> BBB, work easily in the same file that also needs BE reader [20:47] <Daemon404> without ugly define/undef all over the place [20:51] Action: BBB won't object b/c he doesn't care, but raises eyebrows anyway [20:53] <kierank> BBB: let's you mix and match? [20:53] <kierank> lets* [20:56] <durandal_1707> Daemon404: define undef doesn't work, only moving code into separate file ... [20:57] <Daemon404> right... [20:57] <Daemon404> i forgot it relies on the include stuff [20:59] <BBB> kierank: putting a hack in get_bits.h for a design disaster in one obscure shitty codec seems questionable [20:59] <BBB> maybe put that hack inside that codec's .c file? [21:00] <Compn> Daemon404 : somehow i dont think your email re yadif will encourage anyone to rejoin the team :P [21:04] <Daemon404> i dont recall recall that being such a goal [21:04] <Daemon404> s/recall recall/recall/ [21:05] <durandal_1707> BBB: what is hack in get_bits.h? [21:11] <durandal_1707> i could move get_bits_le out of get_bits.h, if that is what you're looking for [22:14] <cone-352> ffmpeg.git 03Nigel Touati-Evans 07master:ef8cc06d6e50: avformat/avidec: Fix incorrect detection of badly interleaved avi [22:14] <durandal_1707> why you bumped lavc when you changed lavf? [22:28] <durandal_1707> what about adding generic graph filter which would draw points over time from some frame metadata? [23:04] <michaelni_> durandal_1707, sloppy mistake :( lucky its not of any practical relevance, ill be more carefull [23:04] <michaelni_> and such graph filter sounds quite interresting [23:09] <durandal_1707> not if there is nothing that will use it [23:34] <durandal_1707> so the only reason psnr filter is not in lavfi is because nobody cares? [23:35] <nevcairiel> its because psnr is a terrible metric [23:35] <nevcairiel> well for video anyway [23:38] <durandal_1707> but when new video codec comes around everybody and he's dog use it [00:00] --- Sat Jul 6 2013

1 0

ffmpeg.log.20130705
by burek 06 Jul '13

06 Jul '13

[00:45] <pfs35> Hello all. I'm trying to reencode a video I had problems playing using my Raspberry Pi yesterday, and I have a read error (most likely what caused the problems in the playback also). Is there any way to force ffmpeg to skip the problematic frames? I've tried using "-ec:v guess_mvs," but it hasn't solved the problem. [04:11] <diego> hi [04:11] <diego> how do i encode a video and discard the audio from the encoding? [04:12] <diego> -an i think [04:12] <diego> sorry [04:35] <diego> http://pastie.org/8111621 [04:35] <diego> i'm trying to encode some video but i get this [04:35] <diego> any ideas please? [04:36] <diego> it doesn't write anything to the file [04:36] <diego> except 4kb or something [04:37] <diego> if i remove -ss the encoding works [04:38] <diego> but i'm trying to tell ffmpeg "encode only from minute 24 until the end of the video" [04:38] <diego> i wonder what i'm doing wrong [04:47] <osmodiar> The file says: Duration: 00:01:30.18 [04:47] <diego> oh [04:47] <osmodiar> So you can't start from 24 minutes into it. [04:47] <diego> right [04:47] <diego> sorry [04:47] <diego> i just noticed that [04:47] <diego> i'm an idiot [04:47] <diego> sorry [04:47] <diego> it should be 00:00:24 i think [04:48] <diego> yep that's it, thanks [04:49] <diego> ffmpeg -i VID_20130704_224456.mp4 -b:v 1000k -an -ss 00:00:24 out.webm [04:51] <diego> this works, thanks [04:52] <the1_> hello, im restreaming an RTSP stream to 3 bitrates and 1 audio only. my command looks like this: [04:52] <the1_> ffmpeg -i rtsp://<rtsp.ip>:1935/live/aksyon3 -i /tmp/logo.jpg -vcodec libx264 -vprofile baseline -level 3 -b:v 250000 -g 45 -s 320x240 -acodec libfdk_aac -ar 44100 -ab 48k -ac 1 -r 15 -f flv rtmp://localhost/live/high -vcodec libx264 -vprofile baseline -level 3 -b:v 200000 -g 45 -s 320x240 -acodec libfdk_aac -ar 44100 -ab 48k -ac 1 -r 15 -f flv rtmp://localhost/live/medium -vcodec lib [04:53] <the1_> my questions is, how do i add logo.jpg to audio stream so that it will have a static image [04:56] <osmodiar> I think you'd have to encode it as a video using -loop_input... [04:57] <the1_> and mapping? [04:57] <mark4o> add a logo to the audio stream? [04:57] <the1_> yes thats correct [04:57] <mark4o> text to speech? [04:57] <the1_> no just a logo [04:58] <mark4o> so you want to add a video stream [05:00] <mark4o> http://ffmpeg.org/ffmpeg-filters.html#Examples-49 [05:00] <the1_> im not sure if we call that a "video stream" if i add just a logo to the audio stream [05:03] <the1_> would it still be considered as video by ffmpeg? [05:04] <mark4o> do you have an audio file, and you want to create a video file? [05:04] <the1_> the audio will be coming from an rtsp source [05:05] <mark4o> and you are recording it to a file? [05:05] <the1_> and im publishing to 3 bitrates and 1 audio-only [05:05] <the1_> but i need to have the audio-only stream with a logo [05:05] <the1_> im recording it just for HLS purposes, it gets purged after a short time [05:06] <mark4o> oh you are sending rtsp, not receiving? [05:07] <the1_> receiving rtsp and sending rtmp [05:08] <the1_> :) [05:08] <mark4o> I don't know rtsp, but I would imagine that you'd send it as a video stream if you want images [05:08] <the1_> ic [05:11] <mark4o> the1_: for a file it would be something like ffmpeg -loop 1 -i image.jpg -i audio -map 0:v -map 1:a -pix_fmt yuv420p output [05:11] <the1_> in my example above where publishing to 4 outputs, how do i make the logo go to the last output [05:12] <mark4o> the1_: sorry, don't know how to do it for rtsp [05:14] <mark4o> oh and -c:a copy if you want it to copy the audio and not transcode it [05:14] <the1_> i think its the same as file input, i just need to map the inputs to outputs [05:17] <mark4o> ok then maybe replace "output" with something like -c:a copy -f rtsp rtsp://server/video.mp4 [05:18] <mark4o> and replace audio with your input audio source [19:46] <enBrianNano> hey, what preset for libx264 do you guys think give you the best combination of speed of size? [19:47] <Mavrik> um [19:47] <Mavrik> probably default ;) [19:49] <sdl420> depends on source? [19:50] <enBrianNano> x264 source, but I'm burning in the subtitles [19:50] <atiti> hey guys, I'm trying to decode single h264 NALUs with avcodec [19:50] <sdl420> why do you want to burn in subtitles? [19:50] <osmodiar> enBrianNano: Use the slowest you can bare to use. You only have to encode once but you keep the file for longer. [19:50] <enBrianNano> so I can watch the videos on my ps3 [19:51] <sdl420> burn in means hardcoeded, right? [19:51] <enBrianNano> yep [19:51] <sdl420> which container? [19:51] <enBrianNano> mkv to mp4 [19:52] <sdl420> so you need mk4 for ps3 i guess. i cant say any about mp4:/ [19:52] <enBrianNano> I've done a bunch already; now, I'm just messing with the preset to see if I can get smaller files [19:52] <sdl420> but i would try just du demux the mkv and mux it into a mp4 [19:53] <Mavrik> enBrianNano, preset doesn't affet filesize [19:53] <osmodiar> Mavrik: They do a bit... [19:53] <Mavrik> sdl420, what are you talking about? PS3 doesn't support soft subs so you have to reencode the video. [19:54] <sdl420> ok [19:54] <sware> anyone have a windows config.h ? [19:54] <Mavrik> osmodiar, unmeasurable amount when using crf, a very tiny amount when using bitrate. they do affect quality at that bitate, yes. [19:54] <sware> from latest version [20:00] <rager> so like... I just want to do some very basic reencoding with libavcodec and such [20:01] <rager> but I don't know the first place to start - any good reading I should do? [20:01] <Mavrik> well :) [20:01] <Mavrik> rager, doc/examples and the doxygen documentation is pretty much the only place with documentation for that :/ [20:01] <rager> my issue is partially that I haven't done a single practical thing in C [20:02] <Mavrik> then you're f'ed. [20:02] <Mavrik> using libav* libraries requires you to at least be basically competent in C development since they're written to be fast, not pretty [20:03] <rager> alternately [20:05] <rager> is there a way to just invoke main() from another C file rather than calling a binary file? [20:05] <rager> I'm really just trying to do some relatively simple stuff with JNI in android [20:06] <Mavrik> no, because it would collide with your own main() [20:13] <rager> I saw one attempt using dlsym and dlopen [20:14] <Mavrik> rager, that's just doing dumb stuff from ignorance [20:14] <Mavrik> instead of just renaming main() to a function name and linking it to your binary. [20:14] <rager> I can see some people not wanting to alter ffmpeg for licensing reasons [20:15] <rager> though I guess I really should just rename main() to whatever JNI likes with... configure? make? [20:15] <rager> you can rename main() without actually renaming it in ffmpeg.c/.h, right? [20:46] <downwater> hello [20:46] Action: rager waves [20:49] <downwater> i'm looking for relevant param to convert into wmv format which fits Windows Media Player "video for local playback (pal 2,1 Mbps/s)" ... i tried : -vcodec wmv2 -acodec wmav2 -b:v 1400k -r 25 [20:49] <downwater> but i'm not sure it matches [20:49] <downwater> err s/1400/2100/ [20:50] <Mavrik> rager, no. [20:50] <rager> ok [20:50] <Mavrik> you cannot rename function without actually renaming it. [21:25] <MachinaeWolf> How did audacity capture the right tempo of the system sound but ffmpeg didn't??? How do I fix that? [21:36] <durandal_1707> MachinaeWolf: alsa? [21:36] <MachinaeWolf> Yeah I have alsa and pulse because I don't think the card I have supports the mixing deal to be able to just run with alsa [21:41] <durandal_1707> MachinaeWolf: what ffmpeg version? [21:42] <MachinaeWolf> 1.2.1 [21:43] <durandal_1707> and what samplerate is reported? [21:44] <MachinaeWolf> hmm multimedia repo tookover there, maybe I need to fix that [21:44] <MachinaeWolf> um don't know how to check that [21:57] <rager> Mavrik: what is the dlopen and why is it just stupid? [22:07] <MachinaeWolf> durandal_1707? [22:09] <rager> Mavrik: it looks to just be used to load the ffmpeg lib, then call the function "main" with the given parameters [22:19] <Chat9999> Hi [22:21] <durandal_1707> MachinaeWolf: what multimedia repo? [22:26] <MachinaeWolf> durandal_1707, the http://www.deb-multimedia.org repo [22:28] <MachinaeWolf> durandal_1707, I didn't do a giant paste of anything...??? [22:28] <MachinaeWolf> wtf? [22:28] <MachinaeWolf> I typed out that url that's it... [22:29] <durandal_1707> give exact command you use for recording [22:29] <durandal_1707> and its output [22:29] <MachinaeWolf> oh [22:32] <MachinaeWolf> http://pastebin.com/ijUWA25i [22:35] <durandal_1707> hmm, and whats wrong with sound? [22:36] <MachinaeWolf> it's distorted or at least tempo is way slower [22:48] <durandal_1707> well its either wrong samplerate or number of channels [22:52] <llogan> MachinaeWolf: does an output by arecord also sound weird? [22:52] <MachinaeWolf> idk, never recorded with that [23:01] <MachinaeWolf> So how do I do that llogan? [23:07] <MachinaeWolf> well idk how to do it properly so it didn't record anything... [23:11] <MachinaeWolf> I'm about to just rid myself of the debian multimedia repo and just use the regular versions [23:21] <llogan> MachinaeWolf: i don't know. i've never used it. [00:00] --- Sat Jul 6 2013

1 0

ffmpeg-devel.log.20130704
by burek 05 Jul '13

05 Jul '13

[00:09] <durandal_1707> this is gonna be funny [00:34] <Kovensky> <Kovensky> http://www.reddit.com/r/AskWomen/comments/1hkp8b/ladies_what_would_you_thin… [00:34] <Kovensky> <durandal_1707> Kovensky: post on #ffmpeg-devel too [00:34] <durandal_1707> :) [00:46] <durandal_1707> lol, i pinged patch 8 months old [03:07] <cone-384> ffmpeg.git 03Carl Eugen Hoyos 07master:565140da178e: avcodec/rawdec: Fix 2bpp and 4bpp rawvideo in mov [10:27] <BuxiNess> hello [10:31] <burek> <michaelni> burek, can you make fflogger use trac.ffmpeg.org instead of ffmpeg.org/trac/ffmpeg [10:31] <burek> you mean for the trac events [10:31] <burek> like this one <fflogger> [editedticket] MrNice: Ticket #2504 (Audio glitches and distortion when recording alsa) updated https://ffmpeg.org/trac/ffmpeg/ticket/2504#comment:108 [10:56] <michaelni> BuxiNess, hello [10:57] <michaelni> burek, yes thanks! [11:25] <michaelni> burek, btw, how is the fate server rewrite going ? [12:58] <nevcairiel> Daemon404: did MSE get its file system lock fixed? I'm building with msvc right now on a system with MSE installed and it goes to 100% cpu just fine [13:00] <nevcairiel> of course this laptop has a SSD and maybe the filesystem action is just fast enough not to notice [13:15] <JEEBsv> nevcairiel: I don't think the general speed of it has become better. At least it seems like compiling/linking ffmpeg doesn't crash MSE itself any more [13:16] <nevcairiel> never did that for me [13:16] <nevcairiel> except the short time [13:17] <JEEBsv> yeah, it IIRC got fixed quite some time ago [13:18] <nevcairiel> was only broken for a couple weeks [13:30] <nevcairiel> but before it still managed to lock the filesystem so that you could usually only get one process compiling [13:30] <nevcairiel> now it gives me 100% [13:30] <nevcairiel> oh well [13:31] <JEEBsv> SSD magic++ [13:33] <nevcairiel> my development PC has the code on a SSD as well now, i got a new one and used the old one for my code, guess i should test performance again there too [13:34] <wm4> yeah could see this coming [13:34] <wm4> who cares about bugs, it's not even my issue [13:34] <nevcairiel> genpts is stupid anyway [13:34] Action: nevcairiel never uses it [13:35] <wm4> so how should one determine good PTS instead? [13:36] <nevcairiel> genpts never really produces "good" PTS in my experience anyway [13:36] <wm4> libavcodec has its own code for this (setting AVFrame.best_effort_timestamp), but it's even worse [13:38] <nevcairiel> formats without any timestamps or any other timing info will most likely not play properly in my setup, but thats usually not all that common [13:40] <cone-411> ffmpeg.git 03Paul B Mahol 07master:d2ce3b3857b2: riff: remove invalid fourcc 'exr ' [13:41] <cone-411> ffmpeg.git 03Michael Niedermayer 07master:df6acc81a851: avcodec/ra288: use init_get_bits8() [13:41] <cone-411> ffmpeg.git 03Michael Niedermayer 07master:8c707a129a0f: avcodec/svq1dec: use init_get_bits8() [13:41] <cone-411> ffmpeg.git 03Michael Niedermayer 07master:b237e6282ef5: avformat/h261dec: use init_get_bits8() [13:41] <cone-411> ffmpeg.git 03Michael Niedermayer 07master:6d05039c7ecb: avcodec/sonic: Fix usage of init_get_bits() and use init_get_bits8() [14:14] <cone-411> ffmpeg.git 03Paul B Mahol 07master:253e155251ec: lavfi/crop: support more pixel formats [14:37] <wm4> JEEB: not sure if you know, but what software produces mkv files that uses Seek elements for indexing all clusters? [14:37] <JEEBsv> hmm [14:37] <JEEBsv> does mkvmerge do that? [14:38] <wm4> actually yes [14:38] <wm4> forgot to look at the muxer field [14:38] <wm4> "mkvmerge v2.9.8" [15:03] <durandal_1707> ubitux: have looked at #1618? [15:04] <ubitux> nope [15:06] <durandal_1707> ubitux: you seems inactive [15:07] <ubitux> yes i'm busy with some other stuff currently [15:33] <saste> ubitux, are you fine with my avfilter_graph_parse() fix? [15:34] <ubitux> i don't really have time to look closely, sorry [15:34] <ubitux> i'll be back on ffmpeg starting next month [15:35] <saste> ubitux, ok, just wondering since you didn't comment on list [15:35] <ubitux> yeah, sorry [15:36] <saste> who is our MPEG-TS expert? [15:36] <saste> JEEB? [15:36] <saste> http://blog.zencoder.com/2011/12/08/announcing-the-clouds-most-efficient-ht… [15:37] <JEEBsv> saste: more like kierank lol [15:37] <kierank> saste: i spoke to them about that [15:37] <kierank> basically they hacked mpeg-ts to save bits [15:37] <kierank> but be out of spec [15:37] <kierank> but work on mobile phones [15:38] <saste> kierank, do you know how they did? [15:38] <kierank> saste: lots of packets per PES [15:38] <saste> we could add an MPEG-TS option for emulating that behavior [15:38] <kierank> and at the end of the PES don't write stuffing [15:38] <saste> also, IIRC the Apple streamer implements the same feature [15:39] <saste> kierank, so why nobody posted a patch for that yet? [15:39] <kierank> because it's evil [15:40] <saste> kierank, what's the problem with that? we're evil as well [15:40] <saste> as long as it is not the default and players are happy with it [15:41] <nevcairiel> does the hls segmenter in avformat work decently anyway? I never tested, but i was kinda planning to use HLS for some things down the line [15:41] <wm4> saste: did you see my comments about the inconsistent defines for the avfilter_graph_parse fix? [15:50] <saste> wm4: no [15:50] <saste> wm4, post or irc? [15:52] <wm4> irc [15:52] <saste> i was not logged [15:53] <saste> so what's the problem? [15:53] <wm4> <wm4> what defines AV_HAVE_INCOMPATIBLE_LIBAV_ABI? I get warnings like include/libavfilter/avfilter.h:1314:5: warning: "HAVE_INCOMPATIBLE_LIBAV_ABI" is not defined [-Wundef] [15:53] <wm4> <wm4> installation problem? [15:53] <wm4> <wm4> also I see both HAVE_INCOMPATIBLE_LIBAV_ABI and AV_HAVE_INCOMPATIBLE_LIBAV_ABI in the ffmpeg sources [15:53] <wm4> it doesn't cause an actual problem currently, though [15:53] <saste> uhm... i screwed [15:53] <saste> up [15:53] <saste> once [15:53] <saste> again [15:54] <wm4> it's not so bad, it didn't break compilation [15:56] <saste> but... the problem was pre-existing [15:58] <saste> uhm so we have the HAVE_ defined in config.h, and the corresponding AV_HAVE defines in avconfig.h [15:59] <saste> michaelni, what should we use? why do we have both variants? [15:59] <saste> and yes it should work as is, althought the double defines are a bit confusing [16:00] <saste> nevcairiel, there are still a few compatibility issues, but mostly working [16:01] <wm4> michaelni: is there anything planned that could make seeking in files with TS resets make sane? or is everyone stuck with byte seeks? if so, is there any sane way to map time based seeks to byte seeks? it'd require a good bitrate estimate [16:02] <av500> wm4: read the whole file [16:02] <av500> keep a mapping in memory [16:03] <wm4> av500: yeah, but that's often not feasible... but for example an API to do relative time based seeks could help [16:03] <av500> patches welcome I guess [16:05] <wm4> I guess I can try with AVFormatContext.bit_rate [16:05] Action: wm4 wonders how hard this will fail [16:05] <av500> I doubt bit_rate is sane for TS [16:06] <av500> you can do things like seek around in the file a bit, grab byte pos and time stamps [16:06] <av500> and then deduct the bitrate and where it wraps [16:06] <wm4> and libavformat doesn't do that? [16:06] <kierank> bit rate won't be sane for vbr [16:07] <av500> wm4: no [16:07] <wm4> at least libavformat could try to use the same segment when doing ab absolute time based seek [16:07] <av500> it could [16:07] <kierank> segment? [16:07] <av500> if somebody wrote the code [16:07] <wm4> kierank: I don't know the proper term [16:07] <av500> kierank: wrt timestamp wrap around [16:07] <av500> I guess [16:07] <kierank> ah oik [16:08] <av500> but I guess broadcast TS can have TS wraps at any time [16:08] <kierank> anyway the proper way is indexing :) [16:08] <av500> not just when the counter runs over [16:08] <saste> michaelni, durandal_1707, ping for muxing resampling patch [16:08] <saste> also, there is some plan for a more high level resampling API? [16:09] <wm4> saste: isn't it af_aresample? [16:09] <kierank> av500: yes, some places splice for local news [16:09] <michaelni> saste, with av prefix its public without its not [16:09] <kierank> and other junk [16:09] <michaelni> wm4, planed yes since many many years [16:09] <wm4> ok [16:10] <saste> wm4, if you use libavfilter [16:10] <saste> and a filtergraph is still not trivial to setup [16:10] <wm4> saste: yes, I thought libavfilter was supposed to be the high level API [16:10] <saste> wm4, but you may want to resample without relying on libavfilter [16:11] <saste> maybe a sort of push/pop samples API could be useful [16:11] <saste> you buffer samples, and then call pop to get the converted samples (only if you get at least N samples) [16:12] <av500> you dont need to resample [16:13] <av500> android audioflinger does that for you :) [16:23] <cone-411> ffmpeg.git 03Zhang Rui 07master:4a4c93cb3f2f: avformat/http: support relative url redirection [17:16] <superware> michaelni: hi. do you have a few minutes to test the mpegts stream? [18:57] <Daemon404> nevcairiel, could be either [19:13] <cone-411> ffmpeg.git 03Carl Eugen Hoyos 07master:f32b8130f434: Fix opacity and increase colour dynamics of initial vmd palette. [19:13] <cone-411> ffmpeg.git 03Sean McGovern 07master:bf18abb2eb79: Rename "AVClass class" as "AVClass component_class" for external codecs. [19:23] <cone-411> ffmpeg.git 03Michael Niedermayer 07master:2ca48e466675: avformat: Append data in fill_buffer() when possible [19:23] <cone-411> ffmpeg.git 03Michael Niedermayer 07master:186ec1784336: avformat/aviobuf: Add ffio_ensure_seekback() [19:23] <cone-411> ffmpeg.git 03Michael Niedermayer 07master:8a8d9a738599: mpegts: use ffio_ensure_seekback() [19:23] <michaelni> superware, you can test the fix tomorrow with zeranoes binaries [19:56] <cone-411> ffmpeg.git 03Matthieu Bouron 07master:621ab4e4ef69: lavf/movenc: check ff_mov_init_hinting() return [20:09] <superware> michaelni, great! thanks [20:29] <saste> avoid_negative_ts integer (output) Shift timestamps to make them positive. [20:29] <saste> michaelni, shouldn't it be "non-negative"? [21:33] <durandal_1707> i demand creating release branch [21:39] <durandal_1707> michaelni: for tickets are blocking release? [21:42] <durandal_1707> s/for/what [21:43] <Daemon404> 9 concurrent release. yep. nothing insane about that. [21:43] <Daemon404> totally normal [21:46] <Daemon404> damn webgit has blame disabled again [21:46] <Daemon404> what a pain [21:51] <durandal11707> did i missed somthing? (last thing i got is: what a pain) [21:51] <Daemon404> no [21:51] <Daemon404> [15:46] <@Daemon404> damn webgit has blame disabled again [21:51] <Daemon404> [15:46] <@Daemon404> what a pain [21:53] <michaelni> durandal11707, the tickets marked as regression (minus irrelevant things) should be fixed before the release. That doesnt mean they have to be fixed just that they should [22:00] <Daemon404> hey ubitux [22:00] <Daemon404> dctdnoiz should be fine for sitll images [22:00] <Daemon404> no? [22:00] <Daemon404> (i am too lazy to check if it relies on temporal info) [22:03] <ubitux> no temporal info [22:03] <ubitux> still image denoising [22:04] <Daemon404> oic [22:04] <Daemon404> im FINALLY testing it [22:04] <Daemon404> as soon as configure finishes taking its sweet time [22:04] <ubitux> dctdnoiz will be slower [22:04] <Daemon404> i used to encode using frequency denoisers at 0.05fps [22:05] <Daemon404> on an athlon xp [22:05] <Daemon404> im used to it [22:05] <ubitux> that's pretty optimist for that filter [22:05] <Daemon404> lol [22:06] <ubitux> but i believe it can be simplified and optimized with a 3d fdct call [22:06] <ubitux> like we have in the dsputils [22:07] <Daemon404> depends if its any good [22:07] <Daemon404> ;) [22:07] <ubitux> btw, yet another denoiser was submitted to ipol.im [22:08] <ubitux> possibly a better one [22:08] <Daemon404> im pretty skeptical about "paper' denoisers [22:08] <Daemon404> usually avs has ha something better for N years [22:08] <Daemon404> than the latest papers, seemingly [22:08] <Daemon404> tl;dr DSP papers tend to be full of crap [22:10] <cone-411> ffmpeg.git 03Paul B Mahol 07master:dda8afc39179: libstagefright: unbreak compilation [22:19] <cone-411> ffmpeg.git 03Paul B Mahol 07master:d1c96b28d7b6: libstagefright: port to refcounted frames [22:21] <Daemon404> ubitux, wowwwwwwww [22:21] <Daemon404> you werent kidding [22:21] <Daemon404> took liek 30 sec for a 640x400 png [22:21] <ubitux> :D [22:21] <JEEB> :D [22:21] <ubitux> you can possibly make it slower with a complex expression [22:22] <Daemon404> ubitux, that was wth '-vf dctdnoiz' [22:22] <Daemon404> which seemingly did nothing [22:22] <durandal11707> try -vf owdenoise [22:22] <ubitux> sigma is 0 by default, so yeah [22:23] <Daemon404> ah ok [22:23] <ubitux> Daemon404: feel free to set an overlap btw [22:23] <ubitux> to make it faster [22:23] <Daemon404> sigma3 did barely anythng xD [22:23] <ubitux> like overlap=10 or something [22:23] <Daemon404> i shouldnt test it on pc98 screenshots... [22:23] <Daemon404> 30kb->290kb [22:23] <Daemon404> w/ sigma=3 [22:31] <Daemon404> note to self: 8k photos + dctdenoiz = lol [22:31] <ubitux> :D [22:31] <ubitux> just set some overlap value :p [22:32] <Daemon404> it's interesting how it seems to blur more [22:32] <Daemon404> in some areas [22:32] <ubitux> denoising is a blur op actually :p [22:32] <durandal11707> you want unsharp [22:33] <durandal11707> it can sharp [22:33] <durandal11707> and blur [22:33] <ubitux> :) [22:33] <Daemon404> ubitux, im merely interested in how it has localized blur [22:33] <Daemon404> i mean it's obv, but it looks funny [22:34] <cone-411> ffmpeg.git 03Michael Niedermayer 07master:cb678cc2cf46: avcodec/svq1enc: fix frame rotation code [22:35] <Daemon404> in: https://dl.dropboxusercontent.com/u/77349181/01970024.JPG [22:35] <Daemon404> out: https://dl.dropboxusercontent.com/u/77349181/test.png [22:36] <ubitux> high sigma :p [22:36] <Daemon404> indeed [22:36] <ubitux> how much did you set? [22:36] <Daemon404> thats 30 [22:36] <Daemon404> ;) [22:36] <ubitux> :p [22:36] <ubitux> KILL ALL THE NOIZ [22:37] <Daemon404> e.g. some rocks survive a lot better [22:37] <ubitux> doesn't it look nice with sigma 10? [22:38] <Daemon404> no [22:39] <Daemon404> some of the trees still get chewed up [22:39] <Daemon404> wayyyyyyy more than everything else [22:39] <ubitux> the dark area looks more blurry [22:39] <ubitux> or maybe that's how i perceive it :p [22:40] <ubitux> Daemon404: feel free to try a better expression [22:40] <ubitux> (and add it to the documentation if it's cool :P) [22:41] <Daemon404> https://dl.dropboxusercontent.com/u/77349181/2test.png [22:41] <Daemon404> look at the tree above the waterfall [22:41] <Daemon404> it gets chewed up far more than most anything else [22:43] <cone-411> ffmpeg.git 03Paul B Mahol 07master:35b02732b98c: configure: fix webp decoder dependency [22:51] <durandal11707> why ffmpeg eats 330M when i run owdenoise on this image? [22:53] <durandal11707> does its algo need that much memory? [22:55] <durandal11707> unsharp=luma_msize_x=7:luma_msize_y=7:luma_amount=2.5 looks best [23:03] <durandal11707> buxiness: do you plan to add yuv pixel format support to jpeg2000 ? [23:04] <buxiness> durandal11707: In hat direction encoding or decoding? [23:06] <buxiness> And JPEG2000 is more RGB like, then YUV, even if all the decoding part is done in a YUV domain [23:06] <durandal11707> decoding obviously [23:06] <buxiness> I didn't find any matrix for that. [23:07] <durandal11707> you are not making any sense [23:07] <buxiness> But at the end of decoding the MCT do a YUV --> RGB conversion. May be just test to avoid it [23:07] <durandal11707> why would yuv need special matrix, and what kind of matrix? [23:07] <buxiness> for collorspace conversion, the matrix [23:08] <durandal11707> and where is that matrix in decoder? [23:08] <buxiness> in decoder is the mct, in joeg2000dec.c [23:09] <durandal11707> and why would that be rgb specific? [23:10] <buxiness> The standard output of jpeg2000 dec is RGB. Hisorricaly is made for images [23:10] <durandal_1707> nonsense [23:10] <buxiness> Maybe [23:11] <durandal_1707> jpeg2000 can store any format [23:11] <Daemon404> "any"? [23:11] <Daemon404> so it can store LAB now? [23:11] <Daemon404> or HSV? [23:11] <durandal_1707> really anything [23:12] <durandal_1707> except that it may not be efficient (compression wise....) [23:12] <buxiness> jpeg2000 is supposed to be colorspace free, that's true [23:12] <Daemon404> ah ok [23:13] <buxiness> If we have jpeg2000 bitsream + jpeg2000 format( or container), we can have all kind of colorspace [23:14] <durandal_1707> in bitstream, subsampling are defined [23:14] <buxiness> yep [23:14] <durandal_1707> so i did not checked decoder yet, to find how hard would be to add yuv 420 support for example [23:15] <Daemon404> i think 4:2:2 or 4:4:4 would be more useful [23:15] <Daemon404> and far more likely to be used where j2k is actually used [23:15] <Daemon404> (broadcast & editing [23:15] <cone-411> ffmpeg.git 03Michael Niedermayer 07master:65abce67b52c: avformat/movenc: allow negative TS for the ipod muxer [23:15] <Daemon404> nobody uses 4:2:0 there [23:16] <durandal_1707> Daemon404: when you get one, you gen all of them [23:16] <durandal_1707> usually, but code may be full of little tricks ... [23:16] <Daemon404> "little tricks" is not how i would describe swscale :P [23:17] <durandal_1707> swscale is not that much obfuscated [23:18] <Daemon404> look in libswscale/x86/*template.c [23:19] <Daemon404> http://git.videolan.org/?p=ffmpeg.git;a=blob;f=libswscale/x86/swscale_templ… [23:19] <durandal_1707> bunch of deep macros, ease cake [23:19] <Daemon404> i dont think you quite understand what obfuscated means [23:19] <Daemon404> or have a concept of easily followable code [23:20] <Daemon404> also iirc [23:20] <Daemon404> some of that mmx [23:20] <Daemon404> is self modifying [23:20] <durandal_1707> its normal if there is C variant [23:20] <durandal_1707> its assembly after all [23:21] <durandal_1707> anyway, how would you improve it? [23:21] <buxiness> If your jp2 headers specifies no mct, you have direct YUV output [23:23] <buxiness> 4:4:4, is supported in current dec. 4:2:2 and 4:2:0 should be [23:23] <durandal_1707> does decoder actually parse block that says what colorspace is in use? [23:24] <buxiness> But so far, I never use chroma subsampling [23:24] <durandal_1707> with bayer rgb [23:24] <buxiness> durandal_1707: no, not at code stream file [23:24] <buxiness> s/code stream level [23:25] <durandal_1707> jp2 have it [23:26] <buxiness> https://en.wikipedia.org/wiki/JPEG_2000#File_format_and_code_stream [23:26] <buxiness> the file format have it, not the codestream [23:26] <durandal_1707> so you telling one need to write jp2 demuxer? [23:27] <buxiness> yes :-) [23:28] <durandal_1707> that is suboptimal [23:28] <buxiness> It depends if input is a video or just a file [23:28] <durandal_1707> well jp2 usually have single image [23:29] <durandal_1707> and looks like nobody use .jpx [23:58] <cone-411> ffmpeg.git 03Michael Niedermayer 07master:46ad287a2a87: avutil/rational: avoid llrint() and rint() [00:00] --- Fri Jul 5 2013

1 0

ffmpeg.log.20130704
by burek 05 Jul '13

05 Jul '13

[00:04] <LedM2> Hello! Its about these mp3 with ilegal header, it is not ilegal header, the id3 tagger that is out of time, i revised the mp3 it says Lavf8.564 its aac+ compression! like i do on svn Lavf5.64 any idea?. :-| [00:07] <durandal_1707> file? [00:10] <LedM2> i am using two web servers to convert mp3, and the formule is legal, and the output of mp3 files are new. the id3 tagger dont permits to edit the header of the file, its old, is a new header. [00:11] <LedM2> The comom header it opens, but the new, freezes the software id3 tagger, and if i do a reconversion, the svn ffmeg 5.64 says that the header is missing, (unrecognized). is aac+ compression. :-| [00:12] <durandal_1707> sorry, but you don't make any sense [00:12] <durandal_1707> and svn ffmpeg is pre historic and unsupported [00:13] <LedM2> you will make sense, now , a little time ago, youtube was closed various web server converters, why? why this was discovered! [00:13] <LedM2> all the players havin troubles about tags, [00:16] <LedM2> the conversion direct from a site, was censored and prohibited, they uses a newest version of ffmpeg in secret... TOP SECRET [00:17] <llogan> LedM2: maybe if you use português we can translate it and understand you better. [00:18] <LedM2> imagine you using your 128kb 44100hz normally, 5 minutes of music is, aroun 5 MB of space ocupated [00:18] <LedM2> These new mp3 "with illegal header" are 32kbps, 32000 HZ WITH A HYPER QUALITY CRASHING YOUR 128 KBPS 44KHZ HA?.:-O [00:19] <durandal_1707> nothing improved, i'm still confused [00:19] <LedM2> old ?? we are old :-P [00:19] <durandal_1707> svn ffmpeg is not supported here [00:20] <durandal_1707> ffmpeg switched to git long ago [00:22] <LedM2> and 5 minutes of music, in 32kbps/32000Hz, using these formules, READ 5 LITTLE MINUTES, the file is only around 1MB ha? and the quality is HYPER BEST OF ALL THAT I WAS LISTENED. [00:24] <LedM2> Ok! eu estou usando dois sites de conversâo de audio, porém um deles proibiu a conversão por causa de um tipo de mp3 que acusa nos id3taggers de "cabeçalho ilegal". [00:26] <llogan> LedM2: are the files with "illegal header" made by ffmpeg? [00:27] <LedM2> Não é cabeçalho ilegal, e sim o id3tagger que não tem a livraria dinamica pra editar, (é antigo), a saida de conversão de videos de urls do youtube era mp3 32kbps/32000Hz [00:27] <LedM2> Não durou muito tempo, o youtube resolveu fechar varios sites de conversão de videos no ano passado, devido a esse tipo novo espetacular de mp3 [00:28] <LedM2> Eu copiei as formulas pra fazer usando o ffmpeg versão 5.64, e até hoje nunca deu problema, mas quando baixava o arquivo direto do site, era impossivel adicionar o titulo no id3tagger, corrompendo o arquivo... [00:29] <LedM2> The files are not with ilegal headers, when i re-convert normally, the promt says "missing header" but the header exxists is here, is a other new type of frame... [00:29] <llogan> use pastebin.com to show the output of "ffmpeg -version" [00:30] <LedM2> The id3 tagger that says : ArrayIndexOutOfBounds: 0 [00:31] <durandal_1707> no file - never happened [00:31] <llogan> this is a user support channel for ffmpeg. we do not support products that are not from the FFmpeg project. [00:31] <LedM2> the version of my ffmpeg is 5.64, the versio of these strange files, i dont know, its impossible to read. do not appear on the re-conversion. [00:32] <durandal_1707> ffmpeg never had 5.64 version [00:32] <llogan> i don't understand 5.64. the ffmpeg output will show the correct version. [00:32] <LedM2> UH! OK! Lavf5.64 [00:33] <durandal_1707> that is still meaningless [00:34] <durandal_1707> 5.64 if ever existed, was even more pre historic [00:34] <durandal_1707> as current one is 55.11.100 [00:36] <LedM2> I readed very well today [00:36] <LedM2> pay attemption, why youtube was closed servers to conversion?. did you thinked in this today? [00:36] <LedM2> I heve various mp3 with these headers strong headers, are impressionantly good the quality of the sound. [00:36] <LedM2> But its impossible to edit tha id3tag. [00:37] <durandal_1707> no file - never happened [00:47] <LedM2> I have two versions of these mp3 files, 32kbps/16000Hz~32000Hz [00:49] <LedM2> The quality of the audio, is incredible, and is ffmpeg, today, is impossible to download a new file from youtube, because they closed the convesion to the site... [00:50] <durandal_1707> do you have link to failing file? [00:50] <LedM2> cloncused ? why youtube closed servers to convesion?. these files kill every type of mp3 knowed by everybody untill today, one music of 5 minutes is arond 1Mb of space used [00:51] <LedM2> is not a "failing file" is a POCKET OF BILIONAIRES FAILER! :-P [00:52] <durandal_1707> can someone op me? [00:52] <LedM2> I have today, on my 2Gb Sd card, only 1350 music! are you allright?. [00:54] <LedM2> And these mp3 with the strong header, are confusing the player because the tags of these files...some files are padron, and others are direct from these servers closed... [01:11] <LedM2> Back! here is one! : http://files.myopera.com/le78dm/musicas/twisted-sisters_dont-let-me-down.mp3 [01:11] <LedM2> access and click the blue button download... [01:13] <LedM2> FIXXING! SORRY! http://files.myopera.com/le78dm/musicas/twisted-sister_dont-let-me-down.mp3 [01:18] <LedM2> If this file is normal, i will upload others from a PC and re-give the links, but i believe that this file was uploaded to show like an example... [01:19] <LedM2> I am on mobile, its impossible to upload from here ! but i will upload more to stay on here to remember, if these type of files are wrong, or are new types... [06:17] <praveenmarkandu> anyone know how to add id3 tags to a m3u8 file? [09:55] <jeje2> Hi to All. I use FFMPEG 1.2 in my software to decompress H264 video packet from IP camera. I have on some error in the log like "Frame Gap" or "Num ref frame".I see there is a new version of FFMPEG, with avcodec55.dll, ... which fix somes bugs on h264.But my code work perfect with FFMPEG1.2 and doesn"t work with the last dll files of FFMPEG ,I get them on Latest Zeranoe FFmpeg Build Version: git-e0be3cb (2013-07-03).Function [09:55] <jeje2> avcodec_decode_video2 always return me 0 and in the log I have "no frame"... my decompression code is here http://pastebin.com/34U9SRLN. I format my AVPacket from the camera Stream with my own NALParser berfore send to FFMPEG always a whole frame [09:55] <jeje2> I can't understand the difference between FFMPEG1.2 and the lastest one [09:55] <jeje2> if someone can help me... [10:42] <burek> jeje2, you might check for api changes [10:42] <burek> in changelog [10:43] <burek> also, you might try asking in the libav-user mailing list [10:43] <burek> https://lists.ffmpeg.org/mailman/listinfo/libav-user/ [13:22] <xlinkz0> what can be changed while stream copying? [13:23] <xlinkz0> from what i gather, framerite and aspect ratio? [13:23] <xlinkz0> framrate * [13:23] <xlinkz0> framerate * [13:55] <burek> xlinkz0, only bitstream related data [13:56] <burek> i.e. whatever doesn't require decoding of the stream [13:56] <xlinkz0> and changing fps doesn't require decoding? [13:57] <burek> you don't actually change fps, but rather PTS [13:58] <burek> btw, im not sure that is possible also without re-encoding [13:58] <burek> do you have any example cmd line? [14:01] <xlinkz0> i was just reading the ffmpeg.c code and i come across line 2212 [14:01] <xlinkz0> i didn't actually test it [14:12] <Mavrik> you certanly have to reencode to change FPS [14:12] <Mavrik> unless you have a really unusual I-frame based format [14:14] <JEEBsv> depends on if you mean ConvertFPS() and AssumeFPS() [14:14] <JEEBsv> s/and/or/ [17:18] <superware> can someone here build win32 ffmpeg? I need to test a small patchset :| [17:31] <Fjorgynn> hej [17:31] <Fjorgynn> can I dump a specific timestamp with ffmpeg from a stream? [20:44] <xlinkz0> why do the streaming examples ues -threads 0 ? [20:52] <BrianNano> two quick questions: what's the status of openCL in ffmpeg? since I have hyperthreads, should I default my threads to actual cores? thanks [21:12] <xlinkz0> BrianNano: i think you're looking for #x264 [00:00] --- Fri Jul 5 2013

1 0

ffmpeg-devel.log.20130703
by burek 04 Jul '13

04 Jul '13

[00:03] <saste> any MP4 experts around? [00:04] <saste> I'm getting "[iso file] Box "avcC" size 8 invalid (read 909)" with MP4Box [00:07] <llogan> saste: i recall you wanting a PPC. i have acquired a powerbook G4. still interested? [00:08] <llogan> shipping cost may be prohibitive. i haven't checked. [00:11] <durandal_1707> saste: you asked that already, i lost my magic wand [00:12] <durandal_1707> first thing is to get mp4box source and locate that printf... [00:12] <cone-60> ffmpeg.git 03Tudor Suciu 07master:820ffaed0f64: avformat/matroskaenc: Allow VP9 and Opus in webm [00:45] <mark4o> llogan: http://ffmpeg.org/download.html now links to a non-functional site http://ffmpegmac.net/ ? [00:47] <llogan> mark4o: non-functional? worksforme [00:47] <llogan> or am i missing something? [00:47] <mark4o> says "This site is currently under construction. [00:47] <mark4o> Please try again later." [00:48] <llogan> i can't duplicate the issue. anyone else? [00:54] <michaelni> bcoudurier, do you have time to review the remaining gxf patches ? (0617 22:19 Reuben Martin (1.5K) [FFmpeg-devel] [RFC] GXF Patches - HD) [00:55] <mark4o> llogan: aha, it works on IPv4. It seems ffmpegmac.net has an IPv6 address and an IPv4 address, and the IPv4 address works and the IPv6 address doesn't. [01:00] <llogan> mark4o: ah, i see. (reminds me to check my upcoming new server for IPv6 stuff soon). [01:04] <Compn> mark4o : what web browser do you have that defaults to ipv6 first ? [01:05] <mark4o> I tried chrome, firefox, and safari and they all use ipv6 first if you have it. [01:05] <mark4o> In firefox if I set network.dns.disableIPv6 in about:config it will use ipv4 only and then ffmpegmac.net works. [01:08] <Compn> wow thats terrible [01:08] <Compn> ipv6 first ? so many broken sites [01:09] <Compn> good to know, thanks [01:09] <mark4o> broken sites? that is the first broken site i've encountered since enabling ipv6 in my router a few months ago [01:10] <Compn> sometimes i make stuff up. [01:10] <Compn> :) [01:10] <mark4o> most sites do not have an ipv6 address, or wouldn't publish one if it didn't work [02:32] <cone-60> ffmpeg.git 03Michael Niedermayer 07master:e0be3cbd2ba7: MAINTAINERS: add address to contact "AvxSynth Team" [09:43] <khali> hi [09:43] <khali> can anyone explain to me what "av_cold" means/does? [09:44] <nevcairiel> it informs the compiler about functions that are unlikely to be executed [09:44] <ubitux> http://gcc.gnu.org/onlinedocs/gcc/Function-Attributes.html [09:45] <nevcairiel> which means its used for functions that are rarely called [09:45] <nevcairiel> like init/close [09:46] <khali> ah, so "av_cold" resolves to "cold" if the compiler supports it, and nothing if it doesn't? [09:46] <nevcairiel> __attribute__((cold)) if you want to be precise, but sure [09:47] <khali> oh, and there is "hot" as well [09:48] <khali> but apparently we don't use that [09:48] <nevcairiel> i havent seen that in use anywhere [09:48] <khali> nevcairiel, ubitux: thanks! [10:43] <cone-5> ffmpeg.git 03Stefano Sabatini 07master:b2498dcf5ca9: lavc/avcodec.h: fix prolems -> problems typo [11:25] <cone-5> ffmpeg.git 03Luca Barbato 07master:5650e331a733: jpeg2000: Validate resolution levels [11:25] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:37a970bbb668: Merge commit '5650e331a733711b809bc27f968500f0f5530126' [11:28] <cone-5> ffmpeg.git 03Stefano Sabatini 07master:628ceac65267: doc: add script to compute texi files dependencies [11:28] <cone-5> ffmpeg.git 03Timothy Gu 07master:4ec46b1160eb: doc/muxers: Add AIFF doc [11:33] <cone-5> ffmpeg.git 03Stefano Sabatini 07master:759dfd56845d: examples/muxing: make more compact audio/video_time computation [11:46] <michaelni> saste, Cannot open file 'doc/avtools-common-opts.texi': No such file or directory at /home/michael/ffmpeg-git/ffmpeg/doc/texidep.pl line 12, <$fh> line 217. [11:47] <michaelni> mkdir whatever ; cd whatever ; ../configure --cc='ccache gcc' && make -j12 [11:50] <saste> michaelni, revert [11:51] <saste> wait [12:02] <nevcairiel> quite a bunch of fate boxes are also failing to build now [12:03] Action: michaelni suspects all will fail [13:08] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:fbcc03db8f39: jpeg2000: Check code-block size [13:08] <cone-5> ffmpeg.git 03Luca Barbato 07master:be3271009ef8: jpeg2000: Define the maximum decomposition levels [13:08] <cone-5> ffmpeg.git 03Luca Barbato 07master:5efadcb8cdef: jpeg2000: Clean up return paths and error messages [13:08] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:1a3598aae768: jpeg2000: Use bytestream2 [13:08] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:3f0364327c32: Merge commit 'fbcc03db8f3919e2ea46d6ad11a00906b1c0ef0e' [13:08] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:b0448e512357: Merge commit 'be3271009ef893fbed9640e8edbd771685fd957b' [13:08] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:26f6acc66b66: Merge commit '5efadcb8cdef6fc5ca5f1f72e45ef3b23016b1c4' [13:08] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:57bc64e235fd: Merge commit '1a3598aae768465a8efc8475b6df5a8261bc62fc' [13:08] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:ffed7227c331: Revert "doc: add script to compute texi files dependencies" [13:08] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:7c71ef4f1727: Merge commit '57bc64e' [13:09] Action: michaelni looks around and sees no saste [13:09] <michaelni> saste, ive reverted it [13:10] <durandal_1707> why? [13:10] <michaelni> durandal_1707, because it broke build [13:11] <durandal_1707> i read that in commit log [13:12] <durandal_1707> it broke normal build, or something special? [13:13] <durandal_1707> aw, it made fate red [13:14] <saste> michaelni, ok thanks [13:23] <cone-5> ffmpeg.git 03Stefano Sabatini 07master:5efbeae38cdc: configure: rename have_incompatible_fork_abi -> have_incompatible_libav_abi [13:23] <cone-5> ffmpeg.git 03Stefano Sabatini 07master:838bd731393a: lavfi: create Libav-API compatibility layer for avfilter_graph_parse() at the next bump [13:24] <nevcairiel> does it really make all that much sense to try to restore compat month (if not years) later? :d [13:26] <ubitux> allow apps to build with both projects with a single code path [13:26] <saste> http://thread.gmane.org/gmane.comp.video.ffmpeg.libav.user/11317 [13:26] <saste> michaelni, do you have any hint? [13:27] <saste> why the profile information is not propagated? [13:27] <khali> when working with yuv420p pixel format, each Y, U and V sample is 8-bit, isn't it? [13:27] <saste> or better, how should I copy the profile information? [13:27] <nevcairiel> khali: yes [13:33] <saste> michaelni, plans for the next release? [13:33] <durandal_1707> yes, really, this becoming tedious [13:37] <michaelni> durandal_1707, help fix some bugs [13:39] <michaelni> saste, yes but my todo list is long [13:39] <michaelni> redoing merges because someone commited in the meantime [13:40] <michaelni> reverting commits that broke build [13:41] <michaelni> and above all, merging my own commits that libav decided to rebase and randomly change [13:41] <Compn> how many breakages have we had in the past month? ;\ [13:41] <michaelni> too many ... [13:41] <michaelni> i mean 1 is too many already IMHO [13:42] <Compn> is there anyway to have a pre-commit hook that tries to build ffmpeg? :P [13:43] <michaelni> yes we have (sort of), just dont commit to the main tree and instead commit to your private one and ask me to merge, i test it before push [13:43] <Compn> pre-commit hook aka michael :) [13:44] <michaelni> exactly :) [13:47] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:d3cb302b8850: jpeg2000: Validate SOT parsing [13:47] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:2e85737b0e0b: Merge commit 'd3cb302b88503c3111e25add196622110c056188' [13:51] <wm4> uh so how does avfilter_graph_parse work [13:52] <wm4> is there an API that actually works on both libs? [13:52] <nevcairiel> in this case probably not since libav changed theirs and ffmpeg kept the old one, or something like that [13:53] <nevcairiel> not sure libav has the "2" variant [13:53] <wm4> I think it was actually ffmpeg which changed it [13:54] <wm4> ah libav has the 2 variant [13:56] <xlinkz0> why are there no examples of using av_write_frame? [13:56] <xlinkz0> i found a seemingly pretty well written example on the web but it's from 0.49 :( [13:57] <wm4> so I guess users can actually use the 2 variant with no problems... except that I don't quite understand what the 2 variant does [13:57] Action: wm4 looks it up [14:06] <wm4> ok avfilter_graph_parse2 is oddly different [14:20] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:278a923c5118: jpeg2000: Validate SIZ parsing [14:20] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:f0358dc1d30c: Merge commit '278a923c51187d37445d88a6b21082036ec9568d' [14:29] <cone-5> ffmpeg.git 03Luca Barbato 07master:7e201d575dc4: jpeg2000: Validate block lengthinc [14:30] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:db5c93a1a97d: Merge commit '7e201d575dc4385eb67314b0419d4d77185e65f4' [14:36] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:17e5d614a864: jpeg2000: Check zero bit-plane validity [14:36] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:2cad1ea29e4d: Merge commit '17e5d614a8647d51b9795cb8bccf97ee33ca2d58' [14:37] <superware> michaelni: hi. I couldn't manage to build ffmpeg, spent more than a few hours on this, too advanced for me. Is there any other way to test the mpegts patchset? do you know of someone that might be able to build related dll for me to try w/ a udp stream? [14:38] <michaelni> superware, i suspect zeranoe could do that for you but probably not for free [14:39] <wm4> superware: building ffmpeg is quite simply on other OSes... now I don't want to say "install Linux", but... [14:39] <superware> wm4: can you do a cross build? [14:41] <wm4> cross compiling from Linux to Windows isn't that hard either, but requires the use of mingw [14:43] <superware> wm4: if I'll send the UDP stream to your IP, can you please try the patchset? [14:44] <wm4> no, I'm behind a router anyway [14:45] <superware> I see [14:45] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:eae63e3c156f: jpeg2000: Check component number in get_coc() and get_qcc() [14:45] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:fd54dd028bc9: jpeg2000: check len before parsing header [14:45] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:a4a16821d732: Merge commit 'fd54dd028bc9f7bfb80ebf823a533dc84b73f936' [14:48] <michaelni> superware, btw if you manage to build ffmpeg yourself you also would have learnt alot and this may be usefull to you in the future as you then can build your own custom versions and dont depend on what binaries you find on the web [14:49] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:ac921338a498: jpeg2000: Correctly calculate sgnd [14:49] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:86a2602e1617: jpeg2000: Drop unused and writeonly fields [14:49] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:9dfa5ea5dcc8: Merge commit '86a2602e16171ddc10c3ec655c23b16ffa595059' [14:51] <superware> michaelni: it's not that big advantage. It's was just pure luck that I happend to stumble upon such a bug, I mean, it's quite severe IMHO. [14:52] <superware> or at least should have been spotted by other users a long time ago [14:52] <michaelni> it probably was but i guess they didnt report it [14:56] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:78962d3df49a: jpeg2000: Check that nreslevels2decode has been initialized before use [14:56] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:ef35d6dbc6c3: jpeg2000: Propagate error code from get_cox() [14:56] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:ac3ca3d311bb: Merge commit 'ef35d6dbc6c3b7ba6b13ac13fc8e797cc1268c8f' [14:56] <superware> well, I'll be here if you have a few minutes to try the stream out. thanks. [15:01] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:2c3901b2c34b: jpeg2000: Remove unneeded returns [15:01] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:b564784a207b: jpeg2000: Check that there is a SOT before SOD [15:01] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:273f600b037c: Merge commit 'b564784a207b1395d2b5a41e580539df04651096' [15:13] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:cf04af2086be: jpeg2000: Check that we have enough components for MCT [15:13] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:b44925ae6b4b: jpeg2000: Initialize code blocks structures in precincts to 0 [15:13] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:4cbd5ed11b01: jpeg2000: Fix compute precedence error in lut_gain index [15:13] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:c7a092cb2d57: Merge commit '4cbd5ed11b0161a6a192174934b241e0cc0599a9' [15:26] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:4e11b1553451: jpeg2000: Calculate code-block coord in ff_jpeg2000_init_component() [15:26] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:952f7ed3c0b0: jpeg2000: Compute quantization for 'scalar derived' in the correct case. [15:26] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:f9581f1414ec: jpeg2000: Improve reduced resolution decoding [15:26] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:11cf4b72475f: Merge commit 'f9581f1414ec8e3e7d0868793cda4f2e4fa113ea' [15:39] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:a458b91cf435: jpeg2000: Split int/float codepaths depending on the DWT [15:39] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:5bf208f65970: jpeg2000: Use separate fields for int and float codepaths [15:39] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:2328e7e86521: Merge commit '5bf208f659703895df7926238dcfa8a8175de36b' [15:43] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:690c8af97bb1: jpeg2000: Remove unneeded variable initialization [15:43] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:db13e02c6f03: jpeg2000: Compute corrections of sub-band coordinates [15:43] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:44741113257b: Merge commit 'db13e02c6f0355fabd540e77b61ade127d6411ff' [16:08] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:33e665a6abf1: jpeg2000: Update pixel format support [16:08] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:5b73916d3452: jpeg2000: Simplify init_tile() [16:08] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:589e5b52f634: jpeg2000: Use the correct sizeof in memset for T1 data [16:08] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:ee7da5baccc3: Merge commit '589e5b52f634f6b2d307a167a19eef7e7328cb08' [16:18] <cone-5> ffmpeg.git 03Luca Barbato 07master:daeb4e3042f2: jpeg2000: Proper cleanup on failure in decode_frame() [16:18] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:1abb875c980a: Merge commit 'daeb4e3042f2ecae2d41aaa4cae0bed932539788' [16:32] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:aa16bbaf9bca: jpeg2000: Refactor SOT marker parsing [16:32] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:09d5929f3721: jpeg2000: Do not crash on NULL node in tag_tree_decode [16:32] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:d57c737ac30c: jpeg2000: Simplify jpeg2000_decode_packets() [16:32] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:19ff6f55e800: Merge commit 'd57c737ac30cf079a342fa649fd4888c2d059dd9' [16:42] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:64f6570c6e2c: jpeg2000: Use EBCOT's CAUSAL and BYPASS mode in decode_cblk() [16:42] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:53d5d89c1b75: jpeg2000: Speed up jpeg2000_decode_tile() [16:42] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:c1dcbc590d90: jpeg2000: Do not assume a single tile [16:42] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:e11099db203c: jpeg2000: Optimize dequantization [16:42] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:f0552e63a669: jpeg2000: Reset s->numX/Ytiles on tile deallocation [16:42] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:22e18ea39e37: jpeg2000: Optimize output sample conversion [16:42] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:5eb23e6b0ae2: Merge commit '22e18ea39e371030cc78973d1b46aae45a7ea215' [16:48] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:c4604b8792ca: jpeg2000: Merge rescaling with interleaving in 9/7 IDWT [16:48] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:95a57d26d865: jpeg2000: Use the matching coding style struct for transform selection [16:48] <cone-5> ffmpeg.git 03Michael Niedermayer 07master:13f86baa1d98: Merge remote-tracking branch 'qatar/master' [16:50] <durandal_1707> michaelni: why you merge stuff? [16:52] <Compn> why not ? [16:54] <durandal_1707> then why complaining? [16:55] <durandal_1707> michaelni: what bugs needs to be fixed? [16:55] <durandal_1707> i posted patches for two bug fix and nobody cares [16:55] <michaelni> the ones marked as imortant on trac should be an approximate list [16:55] <michaelni> which patches ? [16:57] <Compn> there arent many people reviewing patches lately, remember 2 weeks is the time limit so far :) [17:03] <durandal_1707> looks like it extending to 2 years [17:04] <durandal_1707> michaelni: i think more than half of important tickets are not that [17:05] <michaelni> thats why i said approximate :) [17:06] <michaelni> also dont hesitate to update priority where its bad [17:07] <durandal_1707> i hestiate to touch trac because cehoyos may undo it at random [17:22] <cone-5> ffmpeg.git 03Matthieu Bouron 07master:4943eff2cd28: lavf/movenc: fix writing of some iTunes metadata tag [17:29] <wm4> what defines AV_HAVE_INCOMPATIBLE_LIBAV_ABI? I get warnings like include/libavfilter/avfilter.h:1314:5: warning: "HAVE_INCOMPATIBLE_LIBAV_ABI" is not defined [-Wundef] [17:30] <wm4> installation problem? [17:30] <nevcairiel> configure should set it [17:30] <nevcairiel> but config.h is usually not included in public API [17:30] <nevcairiel> so someone messed up [17:32] <wm4> also I see both HAVE_INCOMPATIBLE_LIBAV_ABI and AV_HAVE_INCOMPATIBLE_LIBAV_ABI in the ffmpeg sources [18:57] <Daemon404> lul [21:50] <durandal_1707> michaelni: what's your opinion on adpcm ima wav in separate file? [22:03] <durandal_1707> no more comments for "default" opt patch? [22:04] <durandal_1707> michaelni: are you forcing another policy? that only you commit stuff directly? [22:07] <michaelni> no that would be really wrong [22:07] <michaelni> IMHO best is if maintainers commit in their repository and that gets merged [22:08] <Daemon404> the only annoyance with that is 9000 merge commit messages [22:08] <Daemon404> (even more than now) [22:08] <michaelni> yes [22:09] <michaelni> that problem is a bit specific to merging 2 year diverged trees though [22:09] <Daemon404> well it would be nice of it the branch could be rebased on master before merge [22:10] <Daemon404> and thus no crappy message [22:10] <michaelni> about spliting adpcm, i have no objections if that makes technical sense [22:11] <michaelni> rebases can lead to "darn i tested that and it worked why does it now fail on every single checkout i try from the past" [22:12] <Daemon404> heh [22:12] <michaelni> iam not 100% sure but possibly that avisynth 2.5.? bug was of that category or at least someone said it worked and now it doesnt [22:13] <Daemon404> it's probably another biproduct of support for the peice of crap avxsynth [22:13] <Daemon404> <_< [22:13] <durandal_1707> michaelni: its not splitting. it needs little endian bitreader and one already used in adpcm.c is bigendian [22:14] <michaelni> durandal_1707, ah, thats a mess, i guess if puting it in a seperate files cleanly solves it then that should be fine [22:17] <durandal_1707> and why redoing merges is that hard? [22:19] <durandal_1707> alternatively i could move swf adpcm to separate file [22:23] <michaelni> i have no real preferrace on the adpcm issue, its hard to guess ahead of time which would be better [22:29] <durandal_1707> but why one can't use different bitreaders at same time? [22:30] <durandal_1707> can i push something now? [22:31] <durandal_1707> michaelni: i would like to avoid adding new files with copyright spam [22:33] <michaelni> sure you can push if you have some commit that you want to push [22:34] <michaelni> fixing the bitreader so 2 can be used at the same time is of course also an option [22:37] <durandal_1707> how? [22:38] <durandal_1707> adding get_bits_le/be ? [22:38] <durandal_1707> or get_be_bits/get_le_bits ? [22:42] <michaelni> yes, something like that [22:47] <cone-568> ffmpeg.git 03Paul B Mahol 07master:a64248d1ec15: lavu/opt: show default values [23:06] <mateo`> michaelni: i'm not familiar with chapters at all. Is there an option to create them ? [23:06] <durandal_1707> yes, .ffmeta [23:07] <durandal_1707> with a text editor [23:07] <Daemon404> do we not support ogg-style chapter text files? [23:08] <Daemon404> thats what basically every other app on earth uses [23:08] <durandal_1707> i know nothing about ogg [23:08] <mateo`> durandal_1707: thanks [23:09] <Daemon404> http://wiki.xiph.org/Chapter_Extension#Examples [23:09] <Daemon404> it's exactly what the example shows [23:09] <Daemon404> super simple [23:09] <Daemon404> (and why its popular) [23:09] <durandal_1707> i think there is open bug report for that [23:19] <cone-568> ffmpeg.git 03Jean Delvare 07master:99f1d7493349: lavfi/delogo: take SAR into account [23:19] <cone-568> ffmpeg.git 03Jean Delvare 07master:f0bcb13aedcf: lavfi/delogo: avoid propagation of rounding errors in chroma planes [00:00] --- Thu Jul 4 2013

1 0

ffmpeg.log.20130703
by burek 04 Jul '13

04 Jul '13

[01:53] <BrianNano> quick question, does ffmpeg get its fonts for subtitles from fontconfig? [02:03] <relaxed> BrianNano: I believe so. [02:28] <BrianNano> thanks [02:28] <BrianNano> I'm extracting fonts from mkvs and dumping them in ~/.fonts that I'm hoping ffmpeg is using when I hard core my subs. [02:40] <SUNgh0st> hey ffmpeg is giving me some trouble :( [02:42] <klaxa> ok [02:42] <SUNgh0st> when i do a ffmpeg screencast and this script usually works on all the other systems i used it will like freebsd and pretty much all linux distros... but when i do it on OpenIndiana Solaris system it works good the first time , lets say i want to do another screencast it will load up ffmpeg and freeze my entire pc and i tried this alot of times and it works the first time and if i want to do another one it freezes my entire system .. what can be wro [02:44] <SUNgh0st> is there a way to stop the screencasting instead of using ctrl c in the terminal or thats how you stop it ? [02:44] <klaxa> you use "q" generally [02:44] <SUNgh0st> just q [02:45] <klaxa> but i think SIGINT and SIGTERM will most likely have the same outcome [02:45] <klaxa> (ctrl+c and pkill ffmpeg) [02:45] <SUNgh0st> yeah i tried that too [02:46] <SUNgh0st> because i thought the second time i say well do a test screencast ok first time works ok.. pkill ffmpeg well ctrl c than i said hmm maybe pkill ffmpeg and i load up the script again to do a 2 screencast test and it loads up and Freeze the system [02:46] <SUNgh0st> i have no idea what is causing this issue as this is the first time i experience this [02:46] <klaxa> have you checked your system logs? [02:47] <SUNgh0st> no [02:48] <SUNgh0st> hmm what log should look at in /var/log [02:50] <SUNgh0st> keep in mind this is a OpenIndiana system which is Solaris [02:54] <SUNgh0st> no idea how to check a system log on this system [02:54] <SUNgh0st> the version of the solaris build of ffmpeg is 1.0 [02:54] <SUNgh0st> my friend uses thesame system and doesn't get freeze up like me [02:55] <SUNgh0st> well i will give the "q" a go see if that does anything [02:55] <SUNgh0st> thanks [02:55] <llogan> he leaves just when i was about to reply [03:36] <SUNgh0st> install: The -c, -f, -n options each require a directory following! [03:36] <SUNgh0st> make: *** [install-progs] Error 2 [03:36] <SUNgh0st> why doesn't ffmpeg compiles proper anymore [04:11] <jonathan_> can ffmpeg take my 4:3 video and convert it to 16:9 WITHOUT CROPPING IT? I'd like to add black bars to the left & right sides of the video but can't get it to work properly [04:18] <relaxed> jonathan_: yes. pastie.org what you're doing. [04:25] <jonathan_> http://pastie.org/private/f7lhxmwthmbexqer81qxtw [04:25] <jonathan_> obviously I'm using the pad filter incorrectly but I'm really confused about how it's supposed to work (and if it's even the right option) [04:28] <relaxed> man ffmpeg-filters | less +/^' pad' [04:28] <relaxed> read up ^^ [04:28] <jonathan_> I don't have man pages for ffmpeg [04:28] <jonathan_> :( [04:29] <jonathan_> can you pastie that for me? [04:29] <relaxed> http://ffmpeg.org/ffmpeg-filters.html#pad [04:29] <jonathan_> thanks, I'll try again [04:32] <llogan> jonathan_: pad="ih*(16/9):ih:(ow-iw)/2:(oh-ih)/2" [04:33] <llogan> pillarboxing [04:34] <jonathan_> llogan: thanks, but that didn't work (although I'll point out my input is not exactly 16:9, I doubt that should cause an error) [04:35] <jonathan_> llogan: ok, even with the right aspect ratio, it still doesn't work [04:36] <llogan> "doesn't work" is, unfortunately, a very common error here. [04:36] <jonathan_> sorry - it says "Unable to parse option value "0x0" as image size" [04:37] <llogan> my input was 640x480 4:3 [04:40] <jonathan_> mine is too. http://pastie.org/private/z9q3a2geb24nof9qxui7uw [04:41] <relaxed> why doesn't it list the SAR/PAR ? [04:41] <jonathan_> sorry, I don't know what that means :/ [04:42] <relaxed> That was directed at llogan. [04:43] <llogan> jonathan_: you modified my example [04:43] <llogan> although now i think it won't give you what you want [04:43] <jonathan_> yes, my desired aspect ratio is different, but the same error appears with your example too - I'll paste that for you in just a sec [04:45] <llogan> relaxed: i'm not sure. [04:45] <jonathan_> even with your example, it still gets an error: http://pastie.org/private/ayw5awvc0e67wtnccio6iq [04:45] <jonathan_> wait that was the wrong one [04:46] <llogan> try a more recent build. http://www.evermeet.cx/ffmpeg/snapshots/ [04:46] <jonathan_> here: http://pastie.org/private/9vunspfyfyurmsbtzilqva [04:47] <mark4o> change -filter to -filter:v [04:47] <llogan> that should probably do it [04:49] <llogan> also add: -codec:a copy -codec:s copy [04:49] <jonathan_> it worked without the codec options [04:50] <jonathan_> thanks for all your help, guys! [04:50] <llogan> those are just extra. "-codec:a copy" is so you don't re-encode audio [04:50] <jonathan_> ah, cool, thanks :) [07:53] <SUNgh0st> can someone help me to explain why ffmpeg would freeze up my computer ? i use openindiana (opensolaris) system and it will work once though , but if i want to do another screencast again it will start loading the ffmpeg and boom my system just freezes well the mouse works but other than that everything else doesnt and i can't even go to my other tty consoles to just pkill x .. i tried the dev source which i was thinking maybe it was the openindiana v [07:53] <SUNgh0st> save ** [07:54] <SUNgh0st> blah same [07:55] <SUNgh0st> im using nvidia graphics and the graphic drivers is Solaris nVidia offical drivers [07:57] <SUNgh0st> hmm i could be my driver [07:57] <SUNgh0st> yeah hmm i think i may just figured it out its using 295.75 drivers [07:58] <SUNgh0st> i found a 319 [07:58] <SUNgh0st> will install that one see if that helps any [07:58] <SUNgh0st> Fixed a bug that could cause X to crash when using Vertex Buffer Objects (VBOs) with indirect rendering. [07:58] <SUNgh0st> hmm [14:38] <superware> can someone please build shared libraries with a specific (mpegts) patchset so I can test it against a udp stream? I gave up trying to build it myself... [15:18] <skurmedel> Yoooo. [15:18] <skurmedel> Any ffmpeg compilation pro's in here? [15:18] <skurmedel> :) [15:32] <superware> skurmedel: I need to test a patchset, so tell me if you find one :) [15:45] <skurmedel> superware hehe [15:45] <skurmedel> I can't even compile ffmpeg :( [15:45] <kode54> I did it real lazy like [15:46] <kode54> brew install ffmpeg [15:46] <skurmedel> I'm certain my pkg-config paths are set up correctly, but ffmpegs configure doesn't seem to use em [15:46] <skurmedel> I wish I could, but I'm on CentOS 6.4 and that ffmpeg is older than time itself. We need a newer version for our pipeline. [17:06] <jeje2> Hi to All. I use FFMPEG 1.2 in my software to decompress H264 video packet from IP camera. I have on some error in the log like "Frame Gap" or "Num ref frame". [17:07] <jeje2> I see there is a new version of FFMPEG, with avcodec55.dll, ... which fix somes bugs on h264 [17:08] <jeje2> But my code work perfect with FFMPEG1.2 and doesn"t work with the last dll files of FFMPEG ,I get them on Latest Zeranoe FFmpeg Build Version: git-e0be3cb (2013-07-03) [17:09] <jeje2> it seems with the last avcodec file, the function avcodec_decode_video2 return me 0... [17:09] <jeje2> and in the log I have "no frame" [17:10] <jeje2> I can pastebin my code if someone can have a look [17:19] <braincracker> somebody know how is audio multiplexed with sound in new flv? [17:19] <braincracker> chromecache has audio and video in different files [17:19] <jeje2> http://pastebin.com/34U9SRLN [17:20] <jeje2> I'm lloking the difference in the FFMPEG source but I can't find the difference making it not work with the last FFMPEG and work with FFMPEG 1.2 [18:24] <DylanCl> Hello. I'm using FFMPEG to stream me playing the game 'Minecraf' to www.twitch.tv but my screen keeps "rolling back". [18:24] <DylanCl> I'm really not sure if it's the script or just me, but I want to get it fixed :/ [19:45] <superware> can someone build win32 ffmpeg? I need to test a small patchset [20:11] <Lunixed> Hey Guys , I am trying to convert a mov file to mp4 file with h264 video and aacv2 audio encoding. [20:11] <Lunixed> ffmpeg.exe" -i INPUT.MOV OUTPUT.mp4 [20:12] <Lunixed> works but decreases quality ,also audio not converted , what should be my parameters , also is there any command I can see the type of audio encoding present in the .mov file ? [20:15] <Mavrik> ffprobe for output of what the file is [20:15] <LithosLaptop> is the audio already in he-aac? [20:16] <Mavrik> Lunixed, something like ffmpeg -i INPUT.MOV -codec copy OUTPUT.MP4 [20:16] <Mavrik> if you don't have to transcode audio [20:16] <Mavrik> ffmpeg -i INPUT.MOV -codec:v copy -codec:a libfdk_aac -b:a 128k OUTPUT.MP4 [20:16] <Mavrik> if you have to transcode audio [20:18] <Mavrik> oh, add -profile:a he-aacv2 for aacv2 [20:22] <Lunixed> Oh Thanks [20:23] <Lunixed> for lossess I founfsomething online , currently I am doing -vcodec copy and no -acodec option while converting from mov to mp4 [20:25] <mark4o> with no audio codec specified it depends on what encoders you have compiled in [20:25] <mark4o> check https://ffmpeg.org/trac/ffmpeg/wiki/AACEncodingGuide for info on the aac encoders that can be used with ffmpeg [20:26] <Mavrik> Lunixed, don't leave out configuration parameters. [20:27] <Mavrik> it'll destroy your quality. [20:30] <LithosLaptop> Lunixed: what audio codec is the source mov file using and what aac audio encoders does your ffmpeg build include? [20:37] <Lunixed> Maverick|MSG: what exactly are configuration params ? Thanks for the fast response btw :) [20:37] <Maverick|MSG> Mavrik ^^ [20:38] <Lunixed> >.< [20:38] <Lunixed> sorry Mavrik [20:39] <Mavrik> Lunixed, um, the ones I've written few lines up? [20:39] <Lunixed> Oh sorry I missed that , what does this do ? -b:a 128k bitrate ? [20:40] <Mavrik> yep [20:40] <Mavrik> if you don't pass it ffmpeg will choose something for you [20:40] <Mavrik> probably something that'll sound terrible [20:40] <Mavrik> same for video parameters ;) [20:41] <mark4o> 128k is extremely high for AAC-HEv2 [20:41] <Mavrik> yp. [20:41] <Mavrik> (hence no he-aacv2 parameter in that command line) [20:43] <LithosLaptop> AAC-HEv2 only usefull up to about 80Kbps [20:44] <Mavrik> 64k is the highest I usually go [21:02] <superware> can someone here build win32 ffmpeg? I need to test a small patchset [21:09] <joltman> I need to trim x amount of minutes and seconds off an H264 file. Some from the front, some from the back. I'm sure there are ways to do this in ffmpeg, but I'm not sure what switches I should be looking at. Anyone have some advice? [21:09] <Mavrik> hmm, -t and -ss [21:10] <joltman> -t is just how long to run [21:10] <joltman> isn't it? [21:10] <joltman> ohhh [21:10] <Mavrik> yep, and -ss is start [21:10] <joltman> i see what you're saying [21:11] <Mavrik> sadly, I think you can't say "5 seconds off end" [21:11] <joltman> so i know i want to start at hh:mm:ss [21:11] <Mavrik> so you'll have to say "1:43:50" if you video is "1:43:55" long. [21:11] <joltman> and i know i want to run for x amount of seconds after that point [21:11] <joltman> -codec copy command [21:13] <joltman> so...something like... [21:13] <joltman> ffmpeg -ss 01:00:00.0 -t 05:00:00.0 -i input.mkv -codec copy outfile.mkv [21:14] <joltman> seek to 1 hour into the video, then time for 5 hours and copy to outfile.mkv [21:14] <Mavrik> yep [21:14] <joltman> sweet [21:14] <joltman> thnx! [21:14] <Mavrik> joltman, small detail: there's a difference if you have -ss before or after -i [21:14] <Mavrik> keep it before -i and move it after if you have problems [21:15] <joltman> i read somewhere that having it before -i meant it would seek to that point. after -i means it would wait -ss time and then begin the ffmpeg process [21:16] <Mavrik> um, it's different actually [21:16] <Mavrik> it doesn't matter if you use "-codec copy" [21:16] <Mavrik> but if you're transcoding, -ss after -i will cause ffmpeg to decode whole video up to the ss point and then start encoding [21:17] <joltman> ok, i'm just doing a codec=copy [21:28] <joltman> Mavrik: Thanks!!! it worked! [21:28] <Mavrik> np ^^ [22:29] <superware> Mavrik: can you build ffmpeg? [22:32] <superware> can someone here build win32 ffmpeg? please, I need to test a small patchset [22:40] <llogan> superware: you may want to try Zeranoe's blog: http://ffmpeg.zeranoe.com/blog/?p=187 [22:41] <llogan> or the related forum: http://ffmpeg.zeranoe.com/forum/viewforum.php?f=5 [22:43] <llogan> or ask for assistance (or offer to pay someone) on ffmpeg-user mailing list [23:04] <ThiefMaster> hey, when using -target pal-dvd to convert e.g. a mkv to mpeg for a dvd.. is there a way to reduce the quality? it's just a little big too big to burn it onto a normal dvd5 [23:06] <superware> llogan: too complex for me [23:06] <Mavrik> ThiefMaster, try setting bitrate [23:07] <Mavrik> -b:v [23:08] <llogan> superware: then offer to pay someone [23:08] <superware> I'm in Windows [23:09] <superware> no budget [23:09] <llogan> oh, i see. you spent it all on Windows. [23:10] <llogan> i guess you'll just have to wait for the patches to be committed. [23:10] <durandal_1707> llogan: LOL [23:10] <superware> :) I can stream the mpegts via the Internet, if you want to help me verify http://ffmpeg.org/pipermail/ffmpeg-devel/2013-June/145286.html [23:11] <llogan> superware: i'd help you, but (un)fortunately I'm a linux neckbeard. [23:11] <ubitux> just boot a linux, clone the source, run the configure and make, and in about 2 hours you can try that [23:11] <superware> llogan: no problem, the patchset also affects linux :) [23:12] <superware> llogan: we can try it in 5 minutes [23:12] <ubitux> it's been days you're asking for that, get a grip on yourself :/ [23:12] <superware> ubitux: yeah, 2 hours, without any unix/linux experience [23:12] <ubitux> ubuntu doesn't require any particular skill [23:13] <superware> :| [23:13] <durandal_1707> superware: is there link to stream for that mpegts so i can reproduce/test ? [23:13] <superware> durandal_1707: it's a udp unicast, I need to setup the remote IP [23:14] <superware> durandal_1707: can I please /msg you? [23:14] <durandal_1707> first put few bucks on may paypal [23:15] <durandal_1707> can I ask to ask to ask can i ask [23:15] <superware> my PayPal balance was just zeroed, sorry. come on man. [23:15] <llogan> ThiefMaster: i think the target will use ~5500-6000k, IIRC, so you can try something lower than that with Mavrik's suggestion. [23:16] <durandal_1707> superware: ahh, than you need to tell me a joke [23:16] <llogan> What did the leper say to the prostitute? [23:16] <llogan> "Keep the tip." [23:17] <brontosaurusrex> superware, tell a joke man, comeon man! [23:17] <superware> durandal_1707: you WILL laugh when you'll see the video stream ;) [23:17] <superware> joke joke, ok [23:18] <durandal_1707> superware: you can msg me, but if you become boring i will ignore you.... [23:18] <brontosaurusrex> wheres the joke? [23:18] <superware> writing... [23:18] <durandal_1707> so you gonna setup link? [23:23] <superware> An old man gets on a crowded bus and no one gives him a seat. As the bus shakes and rattles, the old man's cane slips on the floor and he falls. As he gets up, a seven-year-old kid, sitting nearby, turns to him and says, "If you put a little rubber thingy on the end of your stick, it wouldn't slip." The old man snaps back, "Well, if your daddy did the same thing seven years ago, I would have [23:23] <superware> a seat today." [23:23] <superware> you asked for it [23:24] <llogan> acceptable. you may continue. [23:27] <superware> llogan: can your machine receive packets UDP packets? (IP to IP) [23:27] <superware> -packets [23:29] <durandal_1707> why it can be only unicast? [23:29] <superware> it's hardware, that's the output [23:30] <durandal_1707> could you record it? [23:30] <superware> "only" unicast? unicast will somehow survive over the net, multicast won't [23:30] <durandal_1707> and eg, just streamcopy [23:30] <superware> no because UDP is datagrams, not a stream [23:31] <superware> the original issue was with datagrams sizes (not standard, but valid) [23:31] <durandal_1707> ahh so its udp protocol issue in lavf? [23:32] <superware> http://ffmpeg.org/pipermail/ffmpeg-devel/2013-June/145286.html and the two following [23:33] <superware> there was a udp.c fix which michaelni already commited, but it wasn't the only issue :) [23:33] <superware> https://github.com/FFmpeg/FFmpeg/commit/804c7b2c62a6e7b596d84609c1707fd383a… [23:35] <superware> port forwarding can bring the stream to any intranet machine [23:35] <brontosaurusrex> i liked llogan's joke btw [23:37] <brontosaurusrex> here is mine, from irc, from today: someguy: doesnt mean Im always write [23:38] <superware> come on brontosaurusrex, you just had to ruin, write? [23:40] <durandal_1707> superware: port forwarding what? [23:41] <superware> a specific port at the router to a specific internal IP [23:42] <superware> http://en.wikipedia.org/wiki/Port_forwarding [23:44] <durandal_1707> there are 10 routers behind me [23:45] <superware> 10 different networks? [23:46] <superware> I guess the router with the internet IP can port-forward to your intranet IP [23:50] <superware> later people [00:00] --- Thu Jul 4 2013

1 0

ffmpeg-devel.log.20130702
by burek 03 Jul '13

03 Jul '13

[00:01] <cone-241> ffmpeg.git 03Paul B Mahol 07master:51f1bf330969: bfi: check return value of av_malloc() [00:07] <cone-241> ffmpeg.git 03Paul B Mahol 07master:16e0416fa47c: idcin: check return value of av_malloc() [00:19] <cone-241> ffmpeg.git 03Paul B Mahol 07master:e1b0d3a389ff: westwood_vqa: check return value of av_malloc() [00:46] <cone-241> ffmpeg.git 03Michael Niedermayer 07master:bf87908cd8da: rmdec: Pass AVIOContext to rm_read_metadata() [01:48] <wm4> saste: so is the icy patch ok? [01:49] <saste> wm4: yes [01:49] <saste> should I push it? [01:49] <wm4> nice [01:49] <wm4> sure [01:49] <saste> i don't know if you have commit rights [01:49] <wm4> I regret that it's so... specific though [01:49] <wm4> no I don't [01:56] <saste> wm4: for "inline" I meant the inline keyword [01:56] <wm4> saste: I think in this case, the compiler knows better anyway [01:56] <wm4> it's a static function [01:58] <saste> wm4, i don't mind (it's a nit and you can ignore it altogether) [02:20] <cone-241> ffmpeg.git 03wm4 07master:a92fbe16f2dc: lavf/http: add support for reading streamcast metadata [02:20] <wm4> saste: thanks [03:28] <cone-241> ffmpeg.git 03Michael Niedermayer 07master:f27b22b4974c: h264: move 444 border xchg under if (deblock_top) [03:28] <cone-241> ffmpeg.git 03Michael Niedermayer 07master:a35494c56691: h264: fix indention of xchg_mb_border() [10:23] <saste> michaelni, what's the use case for dump_extradata (bitstream filter)? [10:25] <j-b> good morning [10:29] <saste> j-b, mood gorning [10:31] <j-b> :) [10:31] <j-b> come stai? [11:12] <saste> hi j-b, tutto bene grazie :) [11:32] <JEEB> hmm, I guess I'll make an issue on the tracker for the ogm seeking thing [11:32] <JEEB> since it seems I've registered onto the tracker at some point o_O [11:32] Action: JEEB has no recollection altogether of it [11:35] <JEEB> ahaha [11:35] <JEEB> that didn't take a while at all [11:35] <JEEB> to find out what uname/pw combo I was using [11:36] <wm4> was the password "password"? [11:38] <JEEB> I have a password that contains 'password' on a certain service because they kept sending me my password back in plain text [11:41] <cone-468> ffmpeg.git 03Luca Barbato 07master:8bd9039900e6: Revert "indeo5: reject negative motion vectors" [11:41] <cone-468> ffmpeg.git 03Michael Niedermayer 07master:bc62b24d113e: Merge commit '8bd9039900e67a633cc838d691c449e71c060553' [11:47] <cone-468> ffmpeg.git 03Loren Merritt 07master:c8b920a9b7fa: lls/x86: use 3-operator vaddpd in ADDPD_MEM [11:47] <cone-468> ffmpeg.git 03Michael Niedermayer 07master:d1fa6718952c: Merge commit 'c8b920a9b7fa534a6141695ace4e8c2dfcd56cee' [11:56] <cone-468> ffmpeg.git 03Anton Khirnov 07master:8ad3267ce3a1: oggdec: do not fall back on binary search in the generic code. [11:56] <cone-468> ffmpeg.git 03Michael Niedermayer 07master:951a36550480: Merge remote-tracking branch 'qatar/master' [12:13] <JEEB> do I have to provide the text file when uploading to the samples ftp in case I've made a trac issue? [12:14] <cone-468> ffmpeg.git 03Michael Niedermayer 07master:a478e99a603f: avutil/x86: reenable ff_update_lls_avx() [13:26] <Compn> JEEB : no, but its nice if you use filename like issue2711.avi or so :P [13:30] <JEEB> Compn, that's a given :) [14:04] <durandal_1707> michaelni: why av_parser_init returns NULL when error happens when initializating parser? [14:04] <durandal_1707> NULL is returned when no parser is found [14:04] <durandal_1707> no parser is found != error when initializing parser [14:05] <nevcairiel> what should it return otherwise? :p [14:05] <durandal_1707> its just stupid design [14:06] <durandal_1707> now i should need to introduce av_parser_init2 to fix this incosistency [14:06] <nevcairiel> its not like parser init can really fail [14:06] <nevcairiel> only in OOM [14:06] <durandal_1707> and than it fails [14:07] <durandal_1707> and says that it cant find parser, where there is parser [14:07] <nevcairiel> feel free to log an extra message in the error case in init so you know that it failed [14:07] <nevcairiel> adding a init2 just for that is stupid API bloat [14:11] <JEEB> ok, sample is uploading and https://trac.ffmpeg.org/ticket/2739 is done [14:16] <wm4> durandal_1707: libavformat doesn't even read mid-stream ogg tags [14:17] <wm4> it tries to read them as packets instead [14:17] <wm4> printing funny error messages [14:26] <durandal_1707> wm4: i don't want to touch that shit [14:31] <JEEB> and sample is fully uploaded [15:02] <xlinkz0> has anyone used the ffmpeg4android build scripts? [15:09] <michaelni> xlinkz0, silly awnser but i guess its author would have used it [15:09] <xlinkz0> i would guess so too lol [15:10] <durandal_1707> ubitux: writing AV_OPT_BOOLEAN? [15:14] <cone-60> ffmpeg.git 03Carl Eugen Hoyos 07master:2492f7774f6c: avformat: Print "Consider increasing the value for analzeduration and probesize" even more often [15:14] <cone-60> ffmpeg.git 03Michael Niedermayer 07master:ccba9535dc35: av_find_stream_info: initialize ret [15:14] <cone-60> ffmpeg.git 03Michael Niedermayer 07master:c826c5698b9c: avformat_find_stream_info: try to preserve the error code [15:14] <cone-60> ffmpeg.git 03Michael Niedermayer 07master:c5ddd753549c: avformat_find_stream_info: Do not consider no streams an error in flush codecs [15:14] <cone-60> ffmpeg.git 03Michael Niedermayer 07master:d59ec0adbbc4: avformat_find_stream_info: move ret init down from flush codecs. [16:41] <cone-60> ffmpeg.git 03Jean Delvare 07master:de8d56c6e3f0: MAINTAINERS: clarify the expected format in the document header [16:41] <cone-60> ffmpeg.git 03Jean Delvare 07master:0472f64c5527: Add a maintainer for the delogo filter [16:49] <michaelni> saste, do you want to do another review on the delogo SAR or ok if i apply it ? [17:03] <t4nk188> Hi. I hope this is the right place to ask: Is there a way how I can control FFmpeg to write the same (ideally exactly *one*) number of AAC samples/frames into each "chunk" of a MP4 file? [17:03] <durandal_1707> t4nk188: i already told you there is no way [17:03] <durandal_1707> the only question is: is it possible at all... [17:04] <durandal_1707> anyway you can try implementing it ... [17:05] <t4nk188> What about -chunk_size and -chunk_duration options? [17:06] <t4nk188> I see no reason why it should not be possible, in theory. Also, I wonder, how does FFmpeg's MP4 muxer decide it currently? [17:06] <t4nk188> The chunk sizes are varying, but what is the reasoning? [17:07] <t4nk188> Implementing it in FFmpeg certainly would be *much* work for us. And currently we want the evaluate different possibilities and FFmpeg is one.... [17:07] <durandal_1707> t4nk188: the chunk_* thing is not used in mp4 for whatever reason [17:07] <Paranoialmaniac> -chunk_* option is global option for libavformat [17:08] <durandal_1707> but its not used by mov muxer [17:08] <Paranoialmaniac> not specific for mov/mp4 muxer [17:08] <t4nk188> okay, thanks for confirming [17:08] <t4nk188> there goes that idea... [17:08] <t4nk188> So, currently FFmpeg's MP4 muxer has no option for the chunk size, right? Only for "fragment" size, right? [17:09] <Paranoialmaniac> t4nk188: you try just muxing? [17:09] <Paranoialmaniac> or encoding and muxing simultaneously? [17:09] <t4nk188> Nope, I do conversion from WMA to AAC in MP4 [17:10] <Paranoialmaniac> so simultaneously [17:10] <durandal_1707> chunk_size/duration have nothing to do with whay you are talking about... [17:11] <t4nk188> I do not talk about duration, sorry [17:11] <durandal_1707> so private mov muxer options have more sense [17:11] <t4nk188> I talk about number of samples per chunk [17:11] <t4nk188> Obviously it is NOT constant, but varries [17:11] <durandal_1707> then write it or ask someone to write it... [17:11] <t4nk188> But why? And what is it controlled by? [17:12] <t4nk188> There must be a reason why 3221 samples are in the first chunk and 3343 in the second [17:12] <durandal_1707> look at code [17:12] <t4nk188> I don't think it's arbitary... [17:12] <durandal_1707> usually there are no reason at all [17:12] <t4nk188> huh? [17:13] <t4nk188> FFmpeg is not a small project, so "look at the code" is not really helpful [17:13] <t4nk188> Especially before I know what I am looking for exactly [17:14] <t4nk188> I do not want to spend a lot of time on this, just to see it doesn't work [17:14] <t4nk188> Our time is limited and FFmpeg just one option we are evaluating [17:14] <durandal_1707> i told 10 times already that it can be done with current code [17:14] <durandal_1707> *can NOT [17:15] <t4nk188> I think I got that, but still I'd like to know how it is controlled currently [17:15] <durandal_1707> for some codecs its fixed and for others its not [17:15] <Paranoialmaniac> ask who knows the control [17:16] <t4nk188> So the encoder tells the MP4 muxer when the "start" the next chunk? [17:16] <durandal_1707> nope [17:17] <t4nk188> Then, can you explain "for some codecs its fixed and for others its not" a bit more, please? [17:25] <t4nk188> Okay, then, could you at least give me pointer where I have to look? At the AAC encoder (e.g. faac) or at the MP4 muxer (libavformat, I guess)? [17:25] <durandal_1707> why would you look if you are not interested in changing it? [17:25] <t4nk188> I did not say I'm not interested in changing it [17:25] <durandal_1707> obviosly code that covers it is in libavformat/movenc.c [17:25] <t4nk188> But I cannot decide before I have understood what the matter is [17:26] <t4nk188> You see, before we go on with the project, we need to decide the direction [17:26] <t4nk188> If FFmpeg could can be changed in 1-2 work days, that would be okay [17:27] <t4nk188> But currently I have no idea how much work it would be (at least I now know FFmpeg can not do it as-is) [17:27] <t4nk188> I will investigate movenc.c, thanks... [17:29] <durandal_1707> see mov_write_stsc_tag in movenc.c [17:30] <Paranoialmaniac> mov_write_stsc_tag does nothing about samples per chunk [17:30] <Paranoialmaniac> it just writes stsc atom [17:31] <t4nk188> But I needs to have the info "sample per chunk" for each chunk [17:31] <t4nk188> it [17:31] <t4nk188> so I have to check where it gathers that info... [17:31] <durandal_1707> but it actually writes samples per chunk [17:31] <Paranoialmaniac> mov_write_stsc_tag does not decide samples per chunk [17:32] <durandal_1707> i did not said it does [18:07] <ubitux> durandal_1707: no i'm not [18:54] <cone-60> ffmpeg.git 03wm4 07master:c3316ddb83e7: lavf: add support for libgme [19:02] <saste> who can help with the MP4 format? [19:02] <saste> i have a custom libavf application, and the resulting MP4 can be played by ffmpeg and vlc, but mp4info complains with: [19:03] <saste> ReadProperties: atom 'avcC' is too small; overrun at property: configurationVersion (src/mp4atom.cpp,386) [19:03] <saste> the file is not playable by Quicktime [19:04] <saste> with MP4Box -info: [iso file] Box "avcC" size 8 invalid (read 909) [19:05] <durandal_1707> its bad file [19:05] <JEEBsv> sounds like you are failing to create a correct avcC structure [19:05] <JEEBsv> time to go read 14496-15 [19:06] <JEEBsv> I mean, you can even put Annex B into "MP4" and ffmpeg and friends will play it :P Doesn't mean it's valid tho [19:07] <JEEBsv> in other words, testing with ffmpeg-using things only is not a good idea [19:07] <durandal_1707> because ffmpeg devs are usually lazy [21:20] Action: durandal_1707 why this crap doesn't work!!!!! [21:50] <superware> michaelni: hi man. any ETA for http://ffmpeg.org/pipermail/ffmpeg-devel/2013-June/145286.html ? :| [21:55] <michaelni> superware, if you test it, it will be sooner than if not [21:56] <michaelni> burek, can you make fflogger use trac.ffmpeg.org instead of ffmpeg.org/trac/ffmpeg [21:56] <superware> I'm worried I might spend hours before I can compile the working win32 libraries [21:58] <superware> michaelni, can you please compile the related win32 dll? [21:59] <superware> if it's not too much trouble [22:00] <michaelni> superware, you misunderstand something i think [22:00] <michaelni> we are all volunteers [22:00] <michaelni> its a issue you want fixed [22:01] <michaelni> so you do the work or you pay someone to do it for you or you wait [22:01] <michaelni> the patchset should be tested & reviewed ideally [22:02] <superware> ok, I want to test it [22:02] <michaelni> ok, so get ffmpeg git, apply the patchset and build it [22:02] <michaelni> see our documentation and wiki on how to do all that [22:03] <michaelni> and please improve the wiki where its not clear enough [22:03] <superware> ok, but I'm afraid the process might turn out too challenging for me [22:04] <michaelni> https://trac.ffmpeg.org/wiki/CompilationGuide [22:05] <superware> ok, thanks [22:33] <durandal_1707> if bit reader is little endian for specific crap, and others crap in same file use big endian bit reader is placing function in separate file/header best option? [22:33] <durandal_1707> bswaping buffer is silly [22:54] <superware> michaelni, building win32 ffmpeg turns out to be too complex for me :( [22:54] <superware> I'll ask for help in #ffmpeg [23:24] <durandal_1707> finally bug fixed [23:32] <durandal_1707> argh, recompiling again.... [23:38] <durandal_1707> wm4: gonna add libmng? [00:00] --- Wed Jul 3 2013

1 0

ffmpeg.log.20130702
by burek 03 Jul '13

03 Jul '13

[02:38] <BrianNano> I have a question about hard coding subtitles. If I don't have a version of ffmpeg with libass enabled, is my best approach to use mencoder, dump the file into an avi with hardcoded subs, and then use ffmpeg to make a mp4. I'm concerned about the b_frames in the mkv that I'm converting. [02:39] <klaxa> woah wait what? before converting 20 times, either compile ffmpeg yourself, or grab a static link [02:40] <BrianNano> thanks, of course, to make things super difficult, I'm on obsd. [02:41] <klaxa> oh haha... [02:41] <BrianNano> I asked the port maintainer to add libass support to the next updated package, but if that doesn't work, I'll start hacking on the port. I just want a solution that is reasonable in the mean time. [02:41] <klaxa> mmhh... i think compiling yourself might be the best approach, if you can't do that, maybe use mplayer for decoding and applying video filters [02:42] <klaxa> BrianNano: http://klaxa1337.blogspot.de/2012/09/streaming-videos-over-http-while.html [02:42] <klaxa> that's totally horrid and i wouldn't do it like that anymore [02:42] <klaxa> but the piping i used might be useful for you [02:43] <klaxa> what it does is decode with mplayer2, apply subtitles, output as yuv4mpeg to a pipe, which then gets read by ffmpeg as yuv4mpeg input then encoded to whatever [02:44] <BrianNano> thanks for the link [02:45] <BrianNano> I just have a bunch of mkvs that I'd like to stream over rygel to my ps3. Yeah, I know, I should probably just get a media player for my TV, but it's just convenient [02:46] <klaxa> you can play mkv files with the ps3 afaik [02:46] <BrianNano> check out tmux as an alternative to screen: http://tmux.sourceforge.net/ [02:46] <klaxa> yeah i uh... tried out tmux [02:46] <BrianNano> klaxa, yeah, rygel will transcode on the fly, but my subtitles go poof:) [02:46] <klaxa> ah heh [02:47] <BrianNano> and ass subtitles are so much nicer than srt; I tried converting to srt for rygel, but it doesn't look good at all since srt is super limited [02:47] <klaxa> mhm [02:48] <klaxa> well i guess compiling ffmpeg from source might be the "cleanest" solution [02:49] <BrianNano> yeah, I agree, but the existing port patches around configuration are ugly: http://www.openbsd.org/cgi-bin/cvsweb/ports/graphics/ffmpeg/patches/ ,so I'll have to hack on that. Let me give you a pastebin of what I'm trying to do, and you can tell me if I'm crazy. [02:50] <BrianNano> http://pastebin.com/fKhq1hxu [02:51] <klaxa> can avi even contain h264 video? [02:51] <klaxa> last command might need -c copy [02:51] <BrianNano> yep, it just can't handle bframes [02:51] <BrianNano> I changed the last to ffmpeg -i out.avi out.mp4; do I need a codec all copy? [02:52] <BrianNano> now, my theory is that since I'm muxing to mp4, I should be good with the bframes set to two, but I don't really know for sure. [02:52] <klaxa> i would do that, otherwise ffmpeg will encode it again [02:53] <klaxa> i'm not too sure about mp4's limitations, but i think just about any h264 stream should be okay [02:53] <BrianNano> the last piece is just muxing from avi to mp4 [02:53] <klaxa> no it's actually encoding [02:53] <klaxa> if you don't specify to copy the codecs, ffmpeg will encode with default settings for the specified container [02:54] <BrianNano> so is this better: ffmpeg -i out.avi -c copy out.mp4 [02:54] <klaxa> also, do you really want pcm audio? can't you just copy the audio stream at least? [02:54] <klaxa> yes [02:54] <BrianNano> I don't want pcm; I want to stick with acc [02:55] <klaxa> mencoder <input file> -ass -sub subtitles.ass -ovc x264 -x264encopts subq=4:bframes=2:b_pyramid=normal:weight_b -oac pcm -o <avi output file> [02:55] <klaxa> change -oac pcm to -oac copy then [02:56] <BrianNano> Audio format 0x4134504d is incompatible with '-oac copy', please try '-oac pcm' instead or use '-fafmttag' to override it. [02:57] <klaxa> haha what [02:57] <klaxa> skip audio copying then [02:57] <klaxa> and copy it from the original mkv [02:58] <klaxa> to do that: ffmpeg -fix_sub_duration -i input.mkv -codec copy -map 0 -map -0:a out.mkv [02:59] <klaxa> then: ffmpeg -fix_sub_duration -i out.avi -i input.mkv -map 0 -map 1 -c copy out.mp4 [03:01] <BrianNano> so I'm grabbing just the audio from the new out.mkv with map? [03:04] <klaxa> oh wait... [03:04] <klaxa> ffmpeg -fix_sub_duration -i out.avi -i input.mkv -map 0 -map 1:a -c copy out.mp4 [03:04] <klaxa> like that [03:05] <klaxa> first you are mapping everything, BUT audio [03:05] <klaxa> then you use mencoder to encode the video + subtitles [03:05] <klaxa> then you mux the video with the original audio [03:06] <Lyude> hi klaxa [03:06] <klaxa> hi Lyude [03:11] <BrianNano> so I create a mkv with only audio and subtitles first? [03:12] <klaxa> no you are creating an mkv with only video and subtitles [03:12] <klaxa> we're not touching the audio [03:12] <klaxa> with -map -<input number>[:stream specifier] you unmap streams [03:12] <klaxa> as opposed to -map <input number>[:stream specifier] [03:14] <BrianNano> now I'm really confused on map [03:14] <klaxa> if you put a "-" before the thing you are mapping, you unmap it [03:14] <klaxa> http://ffmpeg.org/trac/ffmpeg/wiki/How%20to%20use%20-map%20option [03:16] <BrianNano> reading it now [03:18] <BrianNano> great, I have 21 streams because of all the fonts for ass subtitles [03:20] <klaxa> heh [03:21] <klaxa> also, why exactly are you applying -fix_sub_duration? [03:21] <klaxa> .ass is sophisticated enough to not mess up timing [03:23] <BrianNano> early efforts of getting this work, I ended up with soft subs in a mp4, but to get that to work, I had to fix sub duration. I found the fix via googling a lot, but then I remembered that rygel can't handle soft subs that are ass. [03:23] <BrianNano> https://ffmpeg.org/trac/ffmpeg/ticket/2212 [03:24] <BrianNano> so I have 21 streams, and I want all but one, is there a way to do that with out a very, very long commandline of mappings? [03:24] <klaxa> is that bug relevant though? [03:24] <BrianNano> my ffmpeg is from March [03:25] <klaxa> i just provided it: ffmpeg -i out.avi -i input.mkv -map 0 -map -1:a -c copy out.mp4 [03:25] <klaxa> ah wait [03:25] <klaxa> no [03:25] <klaxa> ffmpeg -i input.mkv -codec copy -map 0 -map -0:a out.mkv [03:25] <klaxa> that one [03:25] <klaxa> map everything from 0 (-map 0), unmap all audio streams from 0 (-map -0:a) [03:38] <BrianNano> so like this: http://pastebin.com/KmdHbdFT [03:39] <BrianNano> this cleans up the names: http://pastebin.com/iRYyqtdV [03:42] <klaxa> not exactly but we're close! [03:42] <klaxa> you can combine the -fix_sub_duration and the unmapping of the audio, thus merging the first two ffmpeg commands [03:43] <BrianNano> great [03:43] <klaxa> you *have* to combine the last two ffmpeg commands because the first one's syntax is not entirely correct [03:43] <klaxa> ffmpeg -i out.mkv -map 0 -map 1:a -c copy out_audio.mkv [03:43] <klaxa> there is no input with id 1 ;) [03:43] <klaxa> ffmpeg -i out_vs.avi -i input.mkv -map 0 -map 1:a -c copy out.mp4 will most likely work [03:44] <klaxa> we are taking the original source and take the audio straight from there [03:47] <BrianNano> gotcha [03:47] <BrianNano> I'm running the long mencode part right now; I'll stop by later after going through these steps to let you know what happened. *fingers crossed* [03:48] <klaxa> good luck! [03:48] <klaxa> depending on how long it takes i might already be asleep though [03:48] <BrianNano> the file is only 24 minutes long, so it shouldn't take forever [03:49] <relaxed> why not use ffmpeg's ass filter to hardsub? [03:49] <BrianNano> because I have a version that doesn't have it enabled, and I'm not setup to hack on the port right now. [03:50] <klaxa> openbsd madness [03:50] <BrianNano> I asked the port maintainer to add it [03:50] <BrianNano> but I don't have a local ports tree setup to hack on it, so I just needed a work around for the time being [03:50] <relaxed> at least there's no remote hole in your ffmpeg for heck of a long time. [03:50] <BrianNano> obsd is secure as long as you don't add software to it:) [03:51] <BrianNano> and I've been using it for years, so I'm just used to it [03:51] <relaxed> uou should learn to compile software on it locally :-P [03:52] <relaxed> you* [03:52] <BrianNano> I know how to; I just don't have space to pull down the ports tree locally [03:53] <relaxed> Install the precompiled deps via pkg_add and grab ffmpeg's source. [03:56] <BrianNano> it's not quite that simple; check out the current patches on ffmpeg: http://marc.info/?l=openbsd-ports&m=137125315310155&w=2 [03:57] <BrianNano> for now, I'm good; I already asked the ports maintainer to add support. If that doesn't happen in the next package release for ffmpeg, then I'll setup a build box and start hacking on it. But this is getting way off topic now. [04:52] <cortexman> i'm using cinelerra which I guess is using ffmpeg, and when I open a video taken on my phone, I see this repeatedly on my console: [04:52] <cortexman> [h264 @ 0x7f187ae76ae0]brainfart cropping not supported, this could look slightly wrong ... [04:52] <cortexman> i know this isn't a cinelerra support channel.. but i'm curious what this means. i'm really curious as to how this could ever be a good error message also [04:53] <cortexman> also the video is totally cropped and messed up. for some reason the software doesn't leave the aspect ratio alone. [09:41] <mrAlmond> /msg NickServ identify gianluca [09:41] <mrAlmond> whoops [10:41] <BrianNano> So I ended up having to rebuild my ffmpeg to support libass for hardsubs, and I tested it out and it works great on my PS3. However, I have 50 files that I want to re-encode. Does anybody have tips on how best to automate this process? I'm comfortable using ksh, perl, and c. [10:47] <BrianNano> let me clarify, I re-encoded 10bit MKVs to 8bit MP4 with hardsubs, and it worked great. I even ended up with a smaller file size than the original mkv, and I used crf 18. [10:49] <JEEB> BrianNano, keep resolution within whatever limits you want, level within 4.1 and keep a note on the refs (reference frame count) -- neither ffmpeg or libx264 will limit them to match the level you set. [10:49] <JEEB> and that's more or less it [10:49] <JEEB> keeping the reference frame count within the level limits is the hardest part of that, in my humble opinion [10:51] <BrianNano> is there anyway to speed up adding the hard subs? that was my bottleneck [10:51] <relaxed> BrianNano: Use a for loop. I'm not sure of the syntax for ksh, but for bash it would be- for i in *mkv; do ffmpeg -i "4i [10:51] <relaxed> er, mistyped :/ [10:52] <JEEB> I don't think you can make the libass rendering/overlay any faster, and I'm surprised *that* is your bottleneck [10:52] <BrianNano> so am I [10:52] <relaxed> for i in *mkv; do ffmpeg -i "$i" <encoding options here> "${i%.*}".mp4; done [10:52] <JEEB> usually it's the video encoder that's the bottleneck [10:52] <JEEB> (in this case libx264) [10:52] <BrianNano> I don't understand why it took longer for me to hardsub it than it did for me convert the 10bit to 8bit [10:53] <BrianNano> yep [10:53] <relaxed> Now you're doing both? [10:54] <mrAlmond> Paranoialmaniac are you there? [10:55] <BrianNano> here's a pastebin of what ended up working for me: http://pastebin.com/drs4CuuZ [10:56] <BrianNano> it's just manual [10:57] <BrianNano> and I had to spend an hour figuring out the Makefile for my openBSD port of ffmpeg to enable libass, but now, I have a working patch I can submit for review tomorrow [10:58] <relaxed> I don't see where you're downconverting to yuv420p [10:59] <BrianNano> I can't play high10 h264 files on my ps3 [11:00] <JEEB> relaxed, default output colorspace for 10bit 4:2:0 and with 8bit libx264 is yuv420p [11:01] <relaxed> Are you sure it wouldn't be yuv420p10le or uv420p10be ? [11:01] <relaxed> which is different than yuv420p [11:02] <BrianNano> what is yuv420p? any suggestions for what I can do better? I'm new at this, and after fighting with mencoder all weekend, I'm liking ffmpeg a lot now. [11:02] <JEEB> you shouldn't use mencoder in general [11:02] <JEEB> relaxed, you can't output that with a standard 8bit libx264 [11:02] <BrianNano> I learned that the hard way [11:02] <JEEB> also it wouldn't play on you-know-what [11:03] <JEEB> because the 10bit AVC/H.264 hw decoders cost four digit sums at the very least [11:03] <JEEB> yuv420p10le would be the default for *10bit* libx264 [11:03] <JEEB> naturally [11:08] <BrianNano> so my subtitles are encoding at about 8fps, but my video encoded around 16fps. Is this because of the aac audio that I had to use -strict -2 for? [11:21] <mrAlmond> JEEB : About the issue I'm having with that MP4...using "mediainfo -f" it says codec is avc1 and the working one says codec is "AVC" what's the difference? [11:21] <JEEB> just read 14496-15 and make sure your AVCC creation is correct :P [11:23] <mrAlmond> ook [11:23] <mrAlmond> so that means that my avcc creation is not perfect? [11:24] <JEEB> not necessarily, but that's the first thing I remember from your yesterday's thing [11:25] <JEEB> mediainfo is a crapshoot of a library/app, but if it can't find your profile/level you're doing something wrong [11:29] <mrAlmond> ok tnx [11:30] <JEEB> 14496-15 should be your primary thing that you're following [11:31] <JEEB> but if there's something in that that you'd like an alternative implementation of visible, try L-SMASH's source code or so [11:51] <makara> hi. I'm encoding a guy speaking. Its a lecture. The background doesn't change and the camera doesn't move, but in the encoded file the pixels for the background are flying around much better. I'm concerned this is adding unnecessary weight to the file. When I encode at RF 25 it happens a lot less, but then I want to use RF20 so that his face/body is clear. What should I do? [11:55] <zap0> build a mask. [11:55] <zap0> no 2D algo understands what a 'background' is. [11:56] <zap0> can you tell it what a background is? by capturing a reference background when the speaker is out of shot [12:42] <makara> zap0, no [12:44] <makara> ok, but the thing is, it doesn't sparkle in the MTS original which I understand is already in h264, so why should it come to life in the resultant? [13:38] <makara> helloo? [13:43] <paprika> Hi [13:43] <paprika> I was wonder how to set the muxrate for a ts using this example https://github.com/FFmpeg/FFmpeg/blob/master/doc/examples/muxing.c [13:43] <paprika> *wondering [13:44] <paprika> I've tried passing av_dict_set(&dict, "muxrate", "20000000", 0); into avformat_write_header [13:44] <paprika> but I get a dts < pcr warning [13:59] <t4nk832> Hi! I need to reduce the frame rate of a network video stream but avoid frame duplication due to network delays. How can I do this? [14:08] <zap0> where do i get a list of the depth sizes? s16le etc.. what is u8 ? [14:12] <zap0> nevermind its just u8 [14:13] <durandal_1707> u8 is unsinged char [14:18] <zap0> yeah, thanks. i know. i meant what is the symobl for unsigned char. [14:18] <zap0> its 'u8' [14:48] <shocksock> What does av_opt_set(codec_context->priv_data, "muxrate", "20000000", 0); do? What is a codecs muxrate? [14:49] <JEEB> a muxrate is the rate of the output container, aka padding etc. possibly added to make the mux be of a certain rate [14:49] <JEEB> I think pretty much only MPEG-TS and maybe PS has that [14:50] <shocksock> So in order to specify it I would have to apply the above command to each codec context? [14:51] <bencc> encoding/decoding a video frame takes much longer than an audio frame? [14:52] <JEEB> well, audio generally tends to be simpler and have less data in a single "frame" :P [14:55] <bencc> JEEB: can I assume that audio takes roughly 1/10 to encode or decode? [14:55] <JEEB> nfi [14:56] <bencc> ok [16:01] <bunniefoofoo> When should avcodec_free_frame() be used? It doesn't seem to free much. Should I always use av_frame_free ? [16:24] <bunniefoofoo> I think I may have found a memory leak with avio's dyn_buf. I've looked at ffmpeg source and it seems like everything is being freed. http://pastebin.com/6czZyhx9 [16:31] <t4nk188> Hi, I got a qestion regarding the muximing of AAC into MP4 via FFmpeg [16:31] <durandal_1707> so what? [16:32] <t4nk188> How does FFmpeg decide how many AAC frames or "samples" (in MP4 jargon) get muxed into one "chunk" ??? [16:32] <t4nk188> I see no pattern [16:33] <t4nk188> Ideally, I would like to force each sample into its own chunk [16:33] <t4nk188> if *all* sampels go into a single chunk, that would be okay too [16:33] <t4nk188> Any ideas? [16:34] <t4nk188> BTW: I already tryied -chunk_size and -chunk_duration, but they have no effect :-( [16:37] <t4nk188> So how can I set the chunk size in ffmpeg? Is there any way? [16:39] <b_jonas> t4nk188: what is the full command you are using? [16:39] <t4nk188> ffmpeg -i input.wav -c:a libfaac -b:a 128k output.m4a [16:40] <klaxa> i would like to advise you to use libfdk-aac https://ffmpeg.org/trac/ffmpeg/wiki/AACEncodingGuide [16:42] <t4nk188> does it have an option to define the numbe of samples per Chunk? [16:42] <t4nk188> Shouldn't that be a *muxer* option? [16:42] <durandal_1707> nope [16:42] <durandal_1707> why you need exact number of samples per "chunk"? [16:42] <t4nk188> sorry, "nope" what? [16:43] <durandal_1707> google it [16:43] <t4nk188> I need to create multiple files and they need to have the *identical* chunk layout [16:43] <t4nk188> I googled for hours, but no help [16:43] <durandal_1707> what is "chunk layout" ? [16:44] <t4nk188> file #1 and file #2 need to have the exactly same number of samples in each chunk [16:44] <t4nk188> I mean, the first chunk of file #1 needs to have the same number of chunks as the first chunk of file #2 [16:44] <t4nk188> And so on... [16:44] <durandal_1707> define "chunk" [16:44] <t4nk188> chunk as in MP4 [16:45] <t4nk188> MP4 container stores "samples" and groups them to "chunks" [16:45] <durandal_1707> i have no idea what are you talking about [16:45] <t4nk188> that is basic MP4 [16:45] <t4nk188> MP4 stores data as "samples" [16:45] <t4nk188> for each "sample" we have a size [16:45] <t4nk188> but we do NOT have an offset [16:45] <t4nk188> Offsets are given for "chunks" [16:46] <t4nk188> also there is a table that maps "samples" to "chunks" [16:46] <t4nk188> you can find all that in the MP4/ISOBaseMedia specs [16:47] <durandal_1707> usually there are different "chunks" and each have distinct name and function [16:47] <t4nk188> Now, when I check the M4A file that FFmpeg is creating, I see the number of samplesPerChunk is fluctuating [16:47] <t4nk188> I have audio-only files, so there only is a single (audio) track. And thus there only are audio chunks [16:48] <durandal_1707> what codec? [16:48] <durandal_1707> are codecs alwayws same? as number of channels and sample rate? [16:49] <zap0> anyone know about AAC vs MP3 audio encoding? [16:49] <t4nk188> AAC, encoded with faac [16:49] <t4nk188> Yes, all Stereo @ 44 KHz [16:50] <t4nk188> Please see here: [16:50] <t4nk188> http://i.imgur.com/9pLHDrK.png [16:51] <t4nk188> As you can see, "samplesPerChunk" is always different :-O [16:51] <t4nk188> But what is the reasoning behind that ??? [16:52] <t4nk188> And how can I force my own settings, if possible at all ??? [16:53] <t4nk188> Is this something to ask in the "dev" channel? [16:56] <t4nk188> Nobody knows an answer? [16:58] <durandal_1707> i see nowhere in code that this can be controlled [16:58] <durandal_1707> and i don't know if its allowed at all [16:58] <durandal_1707> why you need it anyway? [16:58] <Paranoialmaniac> if constant sampling rate, -movflags frag_duration? [16:59] <Paranoialmaniac> oops [16:59] <samguyjones> I'm trying to extract EXIF data from PNGs and GIFs. Does anyone know if that's something ffmpeg can do? [16:59] <Paranoialmaniac> this is fragment thing [16:59] <durandal_1707> t4nk188: anyway feel free to ask on -devel and/or open bug report [16:59] <Paranoialmaniac> it seems ffmpeg libavformat has no option to specify samples per chunk [17:00] <t4nk188> It's for a project I work on [17:00] <t4nk188> I cannot go too much into detail, but we need the same number of samples per chunk in different files [17:00] <t4nk188> Ideally, each sample would get its own chunk [17:00] <t4nk188> The overhead would be okay [17:01] <Paranoialmaniac> how about the last chunk in a file? [17:01] <t4nk188> what is with the last chunk? [17:01] <Paranoialmaniac> the same number of samples per chunk < [17:01] <t4nk188> the last one too [17:02] <durandal_1707> samguyjones: nope, EXIF extraction of any kind and any container is not implemented, if you need it open feature request [17:03] <Paranoialmaniac> t4nk188: even if non-mod samples per chunk in the source material? [17:04] <t4nk188> Well, if we force a chunk-size of N and the total sample count is not mod-N, the last chunk would be smaller than N [17:04] <t4nk188> But as long as that is consistent between different files, it would be okay [17:04] <t4nk188> and if N=1, there is no problem anyway [17:04] <samguyjones> durandal: Thanks. I'm checking because another developer suggested it. Might it be something the dev team would do, or is it too far off-mission? [17:05] <durandal_1707> samguyjones: there is nothing off-mission, people works on what they have time/motivation [17:05] <durandal_1707> if there is something not done, it means nobody yet come with patch [17:06] <durandal_1707> usually because nobody cares [17:06] <durandal_1707> and user either as they do not push for it [17:58] <mark4o> t4nk188: ffmpeg does not do it but you can rechunk it with mp4box after encoding. e.g. mp4box -inter 500 file.mp4 will rechunk it into 500ms chunks [17:59] <mark4o> or .m4a - works fine with just audio [18:04] <grr> When I use ffmpeg on Windows and it runs into a lot of errors during transcoding, it almost freezes the whole system for me, making it hard to impossible to kill the process. How can I work around that, for instance, is there an option to make it quit after encountering too many errors? [18:37] <grr> found http://superuser.com/questions/588147/can-i-make-ffmpeg-stop-if-integrity-c… [22:31] <Macey> Hi All, [22:31] <Macey> -i dd-htngu.mp4 -ss 00:00:00 -t 00:03:45 -acodec mp2 -b:a 128k -vcodec mpeg4 -b:v 3500k -r 25 -s 1024x768 -aspect 16:9 -flags +aic+mv4 -vf "drawtext=fontfile='C\:\\Windows\\Fonts\\arial.ttf': timecode='10\:00\:00\:00': r=25: x=(w-tw)/2: y=h-(2*lh): fontcolor=white: fontsize=72 : box=1: boxcolor=0x00000000@1" -y ZX12345.mp4 [22:31] <Macey> is my command line... i'm not seeing errors but, the Burntin T/C is a good seconds behind [22:31] <Macey> any idea? [22:43] <Macey> *good 10 seconds behind [22:59] <superware> I badly need to test the patchset http://ffmpeg.org/pipermail/ffmpeg-devel/2013-June/145286.html (and two following), can someone please help me build win32 avformat.dll so I can verify the patch works? [22:59] <durandal_1707> what exactly you need? [23:00] <superware> a compiled win32 avformat.dll with the above 3-parts patch (very small) [23:01] <superware> durandal_1707: can you build win32? [23:01] <durandal_1707> superware: no [23:02] <durandal_1707> i was thinking i could help you building ffmpeg [23:02] <durandal_1707> but looks like you are not interested in that [23:03] <superware> I'm afraid I have no past experience with that, I've spend the last hour trying to build with MSVS but couldn't [23:03] <superware> if you can guide me through the process, I'll really appreciate it [23:08] <durandal_1707> superware: there is guide for building with MSVC [23:08] <durandal_1707> there is documentation [23:09] <durandal_1707> so read it and ask what you do not understand [23:09] <durandal_1707> i cant waste my time for free giving you exact instruction by instruction [23:09] <durandal_1707> even i'm now wasting my precious time replying to you [23:13] <superware> I appreciate your intentions, but I'm afraid we both might waste our time just to check a single patchset. An alternative might be to provide someone with the "problematic" stream (udp unicast over the net) so we can verify the patchset. [23:47] <superware> Mavrik: hi [00:00] --- Wed Jul 3 2013

1 0

ffmpeg-devel.log.20130701
by burek 02 Jul '13

02 Jul '13

[00:26] <cone-914> ffmpeg.git 03Carl Eugen Hoyos 07master:f4aa8085f23c: Avoid a null pointer dereference on oom in the aac encoder. [00:26] <cone-914> ffmpeg.git 03Carl Eugen Hoyos 07master:712ef25116b4: Avoid a null pointer dereference after oom on frame size change in mpegvideo.c. [01:02] <wm4> wow kfreebsd [01:14] <durandal11707> wm4: there is fear that multithread audio decoding will eat betteries life [01:15] <durandal11707> and its really not that important, as audio codecs are usually sane enough to be decoded in real time [01:17] <durandal11707> exluding extreme cases like als [01:20] <pengvado> Daemon404: fixed [01:20] <Daemon404> i saw [01:30] <durandal11707> wm4: also there are some small issues with it (same with video) that i'm not really motiviated to fix right now [01:30] <wm4> what issues [01:31] <durandal11707> if error happens when decoding some frame it may not be reported at all [01:31] <durandal11707> also pts/dts may be wrong for broken files [01:32] <durandal11707> same issues are present for video decoding too [01:32] <wm4> that kind of sucks [02:23] <cone-914> ffmpeg.git 03Michael Niedermayer 07master:ae038c091446: vorbisdec: propagate errors from setup_classifs() [02:23] <cone-914> ffmpeg.git 03Michael Niedermayer 07master:709cae2bcbc0: vorbisdec: Check VLC tables during use instead of setup [02:34] <cone-914> ffmpeg.git 03Loren Merritt 07master:1221bb623978: x86: lpc: fix a segfault in av_evaluate_lls_sse2() [02:34] <cone-914> ffmpeg.git 03Michael Niedermayer 07master:4e488ac5f5c0: Merge remote-tracking branch 'qatar/master' [02:34] <cone-914> ffmpeg.git 03Michael Niedermayer 07master:a6e46ed51ab9: Revert "avutil/x86: disable ff_evaluate_lls_sse2() for 32bit" [02:47] <cone-914> ffmpeg.git 03Marton Balint 07master:447d2e31d167: ffplay: use frame->pkt_pts instead of pkt->pts in audio pts calculation [02:48] <cone-914> ffmpeg.git 03Marton Balint 07master:02fc61a5a683: ffplay: always send zero packets to flush audio decoders [02:48] <cone-914> ffmpeg.git 03Michael Niedermayer 07master:ec837a08993d: Merge remote-tracking branch 'cus/stable' [03:31] <cone-914> ffmpeg.git 03Carl Eugen Hoyos 07master:b42bcaef29e6: Avoid a null pointer dereference in avcodec_decode_audio4(). [04:09] <ubitux> http://lucy.pkh.me/test.swf generated with ffmpeg \o/ [04:17] <Daemon404> interesting choice of video [04:18] <ubitux> :) [04:18] <Daemon404> somewhat mesmerizing [04:18] <ubitux> i'm actually porting my "html loops" into swf [04:19] <Daemon404> to be less portable? [04:19] <ubitux> well my html loops were/are buggy with firefox [04:19] <ubitux> anyway, i wasn't able to find decent tools to generate swf, so i had to patch ffmpeg :P [04:19] <Daemon404> o [04:20] <ubitux> portability is overrated anyway [04:20] <ubitux> people stupid enough to want to watch braindead loops will make the effort to install a flash plugin :) [04:21] <ubitux> Daemon404: you recognize the source of the video? [04:21] <Daemon404> no [04:28] <ubitux> you disappoint me [09:37] <cone-792> ffmpeg.git 03Jean Delvare 07master:16fd75ceec1d: lavfi/delogo: use weighted interpolation [10:42] <JEEB> michaelni, thanks for fixing 2720 :) [10:46] <wm4> meanwhile we're discussing about removing mplayer's demux_ts from mpv, but we can't because it's better than ffmpeg's [10:53] <JEEB> switch to VLC's instead nîn [11:23] <nevcairiel> wm4: define "better"? I've had no real issues with the ffmpeg ts demuxer [11:26] <wm4> nevcairiel: timestamp problems with tv caps [11:26] <wm4> no proper playback is possible [11:27] <nevcairiel> curious that i haven't heard anything, you wouldn't happen to have a file handy would you? [11:37] <cone-792> ffmpeg.git 03Luca Barbato 07master:7388c0c58601: indeo: Properly forward the error codes [11:37] <cone-792> ffmpeg.git 03Michael Niedermayer 07master:43229d609a17: Merge commit '7388c0c58601477db076e2e74e8b11f8a644384a' [11:38] <wm4> nevcairiel: he says later [11:45] <cone-792> ffmpeg.git 03Luca Barbato 07master:dd3754a48854: indeo: use proper error code [11:45] <cone-792> ffmpeg.git 03Michael Niedermayer 07master:280afa40bcf5: Merge commit 'dd3754a48854cd570d38db72394491aab0f36570' [12:03] <cone-792> ffmpeg.git 03Luca Barbato 07master:1194a410807b: indeo5: reject negative motion vectors [12:03] <cone-792> ffmpeg.git 03Michael Niedermayer 07master:cc53c28fdc2b: Merge commit '1194a410807bac3eafbeb632578b937656d273e7' [12:12] <cone-792> ffmpeg.git 03Luca Barbato 07master:b36e1893ef34: indeo: check for reference when inheriting mvs [12:12] <cone-792> ffmpeg.git 03Michael Niedermayer 07master:38f1d5676918: Merge remote-tracking branch 'qatar/master' [16:01] <cone-608> ffmpeg.git 03Stefano Sabatini 07master:7ebab4159a69: doc/encoders: apply various fixes to libmp3lame documentation [16:42] <saste> wm4: libgme => fuc***ng awesome [18:05] <saste> ubitux: ping [18:07] <saste> uh? and now why avfiltergraph.h was merged into avfilter.h? [18:07] <saste> i wish people would be doing real work rather than shuffling things around [18:08] <ubitux> saste: pong [18:08] <saste> ubitux: about the avfiltergraph_ thing [18:10] <saste> so i propose this: is avfilter_graph_parse() deprecated in l.? [18:10] <saste> discard the first part above [18:11] <saste> we can restore lavfi compatibility, but at the next bump [18:12] <ubitux> wasn't avfilter_graph_parse() version supposed to be useful in some circumstances? [18:12] <ubitux> but well yeah i guess that's ok [18:12] <saste> ubitux, no [18:12] <saste> the new function (avfilter_graph_parse2()) is more versatile [18:13] <ubitux> i dont think the fork deprecated the v1 [18:13] <saste> indeed it superceded avfilter_graph_parse(), then l. reimplemented it, but without a wheel (no way to set log_ctx) [18:15] <saste> do you see better solutions? [18:15] <saste> do you propose to bump libavfilter just after release? [18:16] <ubitux> carl was proposing to use the api incompatible fork macro [18:16] <ubitux> abi* [18:16] <saste> ETOOBORINGWORK [18:17] <ubitux> well, using that macro make the code immediately compatible [18:17] <ubitux> and then we can consider a slow api break for the function [18:17] <ubitux> slow and proper* [18:17] <saste> note that i personally don't care about fork compatibility, so i'm going to do my proposal and just let people who want it fixed to improve it [18:17] <ubitux> don't ask me then ;) [18:19] <saste> ubitux, how can an API break be "slow"?? [18:21] <ubitux> long deprecated phase [18:21] <nevcairiel> deprecate now, delete in 10 years [18:23] <saste> socis folks still didn't publish accepted organizations [18:25] <saste> does someone know if nicolas bertrand is available on irc? [18:27] <ubitux> BuxiNess? [18:27] <Compn> saste : shuffling things around is real work. how else are you going to increase commit numbers? :P [18:28] <saste> Compn, why? can't you just reindent stuff here and there? [18:30] <Compn> cosmetics dont count! [18:30] <BuxiNess> ubitux, yes and hello [18:31] <ubitux> hello, it seems saste want you [18:31] <BuxiNess> yes [18:33] <BuxiNess> And btw , made ffmpeg comparsion of native decoder and libopenjpeg [18:34] <BuxiNess> libopenjpeg is faster, with threading enabled. I have much more context switch in native decoder than libopenjpeg. This it can be realted? [18:35] <BuxiNess> realted/related [19:21] <saste> wow avfilter_parse_graph2() docs is *confusing* [19:25] <Compn> BuxiNess : how much faster ? [19:27] <durandal_1707> +1 [19:31] <saste> ubitux, and it turns out that avfilter_parse_graph2() adopts a different logic with respects to our avfilter_parse_graph() [19:34] <saste> ubitux, help me to find an alternative name for our modified avfilter_graph_parse() [19:35] <ubitux> avfilter_graph_parsep [19:35] <saste> avfilter_graph_parse_ptr()? [19:35] <ubitux> whatever yeah [19:35] <ubitux> i think it was suggested to simplify that whole interface anyway [19:35] <ubitux> at least wm4 was looking forward this [19:36] <saste> one problem at a time [19:36] <saste> let's first restore binary compatibility [19:36] <saste> after that we'll have three functions, of which we really need just one [19:37] <saste> the other ones being libav-compatibility cruft [20:19] <Daemon404> hey nevcairiel [20:20] <Daemon404> why do you label dirac as VC-2 in lav? [20:20] <Daemon404> vc-2 is just intra only dirac... [20:20] <nevcairiel> thats the name field of the AVCodec [20:20] <Daemon404> O.o [20:20] <Daemon404> weird. [20:21] <nevcairiel> .long_name = NULL_IF_CONFIG_SMALL("BBC Dirac VC-2"), [20:21] <JEEBsv> yup [20:46] Action: Daemon404 wonders why AV_QSORT exists [20:57] <cone-965> ffmpeg.git 03Michael Niedermayer 07master:b791a0831b0a: avcodec/x86/dsputil_init: only use xvid idct for lowres=0 [21:00] <Daemon404> ..... [21:00] <Daemon404> b791a0831b0a027e7ba4eb6961cc0180472ac603 is fuckign bullshit [21:00] <Daemon404> and is api breakage [21:00] <Daemon404> youre changing behavior of a public api [21:01] <Daemon404> and thats not fix, thats a shitty hack [21:01] <michaelni> Daemon404, ? [21:01] <Daemon404> people who set idct algo ti xvid's will not get the same behavior as before [21:01] <Daemon404> it i settable via api [21:02] <michaelni> behavior before was to crash [21:02] <michaelni> xvid idct is 8x8 [21:02] <michaelni> for lowres > 0 you need 4x4 or smaller [21:02] <Daemon404> er... fuuuuuuuuck [21:02] <Daemon404> is misread that as lowres!=0 [21:02] <Daemon404> carry on [21:02] <Daemon404> im retarded [21:02] <Daemon404> sorry [21:03] <Daemon404> (although i dont know of anyone actually using lowres) [21:13] <superware> hi michaelni [21:13] <michaelni> hi superware [21:19] <superware> michaelni: I was wondering whether you'll like to test my challenging mpegts stream :) to verify there aren't any further issues. [21:38] <michaelni> superware, you can apply the patchset, compile and test it yourself if you want or wait until it is reviewed & applies [21:38] <michaelni> applieD [21:53] <superware> michaelni: but I saw the patch was applied, wasn't it? [21:54] <durandal_1707> wan't it on mailig list? [21:54] <superware> it was [21:55] <durandal_1707> than its easy to check if its applied [21:56] <Daemon404> https://github.com/openembedded/oe-core/blob/master/meta/recipes-multimedia… [21:56] <Daemon404> does this seem useful to us? [21:57] <Daemon404> wait im dumb [21:57] <Daemon404> thats for gstreamer [21:57] <Daemon404> derp [21:57] <superware> durandal_1707: ok, I see mpegts.c was applied, but aviobuf.c and avio_internal.h not [21:58] <superware> durandal_1707: hmmm maybe none was applied [21:59] <superware> durandal_1707: can you help me test it? [21:59] <superware> (the stream) [21:59] <durandal_1707> i have paypal [22:00] <superware> heh, with enough credit, I hope [22:02] <superware> durandal_1707: what's the avg review & apply time from patchset? [22:03] <Daemon404> pretty quick [22:03] <Daemon404> assuming responses [22:04] <superware> http://ffmpeg.org/pipermail/ffmpeg-devel/2013-June/145286.html and the two following had no responses, right? [22:16] <superware> thanks, bye [23:25] <Rathann> hm apparently nobody tested --enable-libcdio with libcdio not installed [23:26] <Rathann> the check doesn't bail out of configure if it fails [23:26] <durandal_1707> EPATCHWELCOME [23:27] <Rathann> sure [23:27] <Rathann> :) [23:32] <Rathann> correction, it's fixed in master already [23:36] <Compn> ah yes, the uav bugreport i trolled [00:00] --- Tue Jul 2 2013

1 0

ffmpeg.log.20130701
by burek 02 Jul '13

02 Jul '13

[11:57] <mrAlmond> Hi everyone [11:58] <mrAlmond> Does anybody can tell me why this file is not playing on Windows Media Player (it works on mplayer and VLC!)? [11:58] <mrAlmond> https://dl.dropboxusercontent.com/u/14905859/wmp_error_vlc_ok.mpg [11:58] <mrAlmond> it's an avc base profile 640x480 MP4 container [11:58] <mrAlmond> no audio, just video track [12:10] <relaxed> mrAlmond: why is the extension .mpg? [12:11] <relaxed> I think that alone would trip up wmp. [12:15] <mrAlmond> relaxed : it's mp4 [12:16] <mrAlmond> my fault, rename it mp4 [12:18] <mrAlmond> relaxed here's the new link, https://dl.dropboxusercontent.com/u/14905859/wmp_error_vlc_ok.mp4 [12:29] <relaxed> did you try playing it after renaming it? [12:29] <relaxed> which version of Windows is this? XP can't play AVC be default. [12:29] <relaxed> by* [12:39] <xlinkz0> has anyone used ffmpeg on android? [12:40] <xlinkz0> i need a wrapper for ffmpeg's main function since i can't chmod stuff from an android app on most phones [13:44] <natrixnatrix89> I have a problem recording a rtsp stream with ffmpeg or playing with ffplay.. [13:44] <natrixnatrix89> If I open the url in vlc, it plays just fine.., but in ffmpeg it's lagging and lots of frames are lost.. [13:44] <natrixnatrix89> and I get lots of these errors: [13:44] <natrixnatrix89> [h264 @ 0x19273e0] RTP: missed 25 packets [13:44] <natrixnatrix89> [h264 @ 0x1913e60] error while decoding MB 1 20, bytestream (-6) [13:44] <natrixnatrix89> [h264 @ 0x1913e60] Cannot use next picture in error concealment [13:44] <natrixnatrix89> [h264 @ 0x1913e60] concealing 2048 DC, 2048 AC, 2048 MV errors in P frame [13:45] <natrixnatrix89> Is there some flag or argument to add, in order to improve the playback? [13:59] <Ph4nt0mBSD> hello im having so strange freeze up with using ffmpeg on my OpenIndiana system.. as i never expeirence this on BSD systems as ffmpeg works very good... but when i'm trying to do a screencast it does however work sometimes but sometimes when i press ctrl c to stop the encoding or whatever screencasting it just sometimes will keep saying ^C alot of times and just keep doing it like if i were holding the c key but im not and it will just say c and un [14:00] <Ph4nt0mBSD> ffmpeg version 1.0 Copyright (c) 2000-2012 the FFmpeg developers built on Feb 9 2013 13:18:58 with gcc 4.6.3 (GCC) [14:01] <Ph4nt0mBSD> any idea why it does this ? [14:02] <Ph4nt0mBSD> i do the screencast normal like i do with other systems which always work fine.. but on openindiana when i press ctrl c to stop it it doesnt really stop it and just freezes the system and i can do anything cant go to the tty virtual console to pkill x i can do nothing just freezes [14:03] <Ph4nt0mBSD> im out of luck i tried my bestest to resolved this but i can i dont know anymore its doing my head in to what is causing this issues [15:15] <xlinkz0> natrixnatrix89: use -rtp_transport tcp [15:15] <xlinkz0> before the -i option [15:15] <natrixnatrix89> I was trying with ?tcp in url.. [15:15] <xlinkz0> doesn't work [15:15] <xlinkz0> actually it is -rtsp_transport tcp [15:15] <natrixnatrix89> let's try that! [15:17] <natrixnatrix89> Unrecognized option 'rtp_transport' [15:17] <natrixnatrix89> now rtsp [15:17] <natrixnatrix89> wow.. [15:18] <natrixnatrix89> seems to be working [15:18] <natrixnatrix89> amazing [15:18] <natrixnatrix89> awesome! [15:19] <natrixnatrix89> Also.. is there a way to ignore broken frames? [15:19] <natrixnatrix89> Just in case my rtsp stream uses more bandwidth than is available? [15:20] <natrixnatrix89> xlinkz0: Do you have some experience with ffserver + libx264? [15:20] <xlinkz0> don't have experience with ffserver sorry [15:21] <xlinkz0> but ask and maybe someone else will [15:21] <natrixnatrix89> Usually everyone agrees that ffserver is hell :/ [15:21] <natrixnatrix89> The problem is that you can't use libx264 presets with ffserver.. [15:22] <natrixnatrix89> I was hoping someone knows how to set preset = ultrafast for libx264 in ffserver.. [15:23] <natrixnatrix89> does anyone know a free/good alternative for wowza? [15:23] <natrixnatrix89> If I want to restream a rtsp stream.. [15:25] <xlinkz0> crtmpserver [15:25] <xlinkz0> but i haven't managed to get it to work yet :\ [15:28] <natrixnatrix89> hmm. I'm guessing it won't support hls and webm? [15:34] <natrixnatrix89> I don't know why I'm so attracted to ffserver although it doesn't work.. [15:41] <torbjorn> is there a way to determine video and audio encoding of a given file? without starting to play it? [15:43] <natrixnatrix89> tobjorn: how about ffprobe? [15:43] <torbjorn> great, thanks [15:44] <torbjorn> also is there a way to auto-determine/suggest containre for an audiofile? [15:44] <torbjorn> ie vorbis should be an ogg file etc. [15:45] <natrixnatrix89> dunno.. usually there can be various containers for codecs.. but I agree that for audio there can be just one.. [16:07] <mrAlmond> How can I convert a variable frame rate H264/MP4 in an equivalent file at constant frame rate. [16:10] <Mavrik> mrAlmond, reencode with fps filter. [16:10] <mrAlmond> Mavrik do you have an example cmd line? [16:11] <Mavrik> ffmpeg -i blahblah.mp4 -codec:v libx264 -vf fps=25 -crf 23 -preset slow -codec:a copy out.mp4 [16:11] <Mavrik> crf being the desired output quality. [16:11] <mrAlmond> I'm asking this because I have a vfr h264 MP4 that does not play in windows media player (but it plays in VLC and mplayer) [16:12] <mrAlmond> and I think it could be a problem of VFR [16:12] <mrAlmond> that maybe is not supported by Windows 7 WMP. [16:12] <mrAlmond> this is the file [16:13] <mrAlmond> https://dl.dropboxusercontent.com/u/14905859/wmp_error_vlc_ok.mp4 [16:14] <JEEB> I'm pretty sure it's not the VFR aspect per se [16:15] <JEEB> since all MP4 files depend fully on timestamps [16:15] <JEEB> not on any kind of "frame rate" [16:16] <mrAlmond> I hope but I can't see many differences between a working MP4 and the one I'm generating [16:17] <mrAlmond> Stream #0.0(und): Video: h264 (Baseline), yuv420p, 640x480, 1267 kb/s, 17.19 fps, 1k tbr, 18k tbn, 36k tbc [16:17] <JEEB> have you looked at the construction of your "mp4" file at all in, say, L-SMASH's boxdumper? [16:17] <mrAlmond> yes [16:17] <mrAlmond> I've compared with a working one [16:18] <mrAlmond> and maybe the main difference is that in my file there are a lot of entries for the "decoding time to sample box" [16:18] <mrAlmond> and in the working one just one entry [16:18] <mrAlmond> and this is problably due to the vfr [16:19] <mrAlmond> just because my file has variable pts [16:22] <JEEB> waait a bit [16:25] <JEEB> also that sample you linked only has one 'stts' atom [16:25] <JEEB> s/atom/box/ [16:26] <JEEB> oh, entries [16:27] <JEEB> and no, having a lot of those is completely ok [16:27] <JEEB> every sample after all has to have its own stts entry that is on the timeline [16:28] <JEEB> of course I have in some cases multiple samples with the same delta, but in general I have just one sample per entry [16:38] <mrAlmond> JEEB : so you says that it "should" be ok? [16:38] <mrAlmond> Do you see something strange on it? [16:42] <JEEB> I would probably wait until someone more proficient in "mp4" comes around [16:42] <JEEB> like Paranoialmaniac [16:42] <JEEB> I've only helped people read the spec at points, and I know some off-shoot knowledge, but I wouldn't be able to spot derpy things [16:59] <mrAlmond> JEEB thank you so much [17:00] <mrAlmond> Paranoialmanic...if you are here...please help :-) [17:35] <JEEBsv> mrAlmond: he went away from home some two hours ago to work, so I would say that calling him right now won't help :) [17:38] <mrAlmond> JEEBsv : doh! Thanks. [17:41] <mrAlmond> JEEBsv : using mediainfo I can see that some info are missing [17:41] <mrAlmond> for example [17:41] <mrAlmond> Format profile : Baseline(a)L3.0 [17:41] <mrAlmond> Format settings, CABAC : No [17:41] <mrAlmond> Format settings, ReFrames : 1 frame [17:41] <mrAlmond> Format settings, GOP : M=1, N=60 [17:42] <mrAlmond> these are in an MP4 that is working [17:42] <mrAlmond> and are missing in the broken one [17:42] <JEEBsv> uh-oh [17:43] <JEEBsv> sounds like you are not giving it a full-blown AVCc then [17:43] <JEEBsv> I mean, mediainfo is far from perfect [17:43] <JEEBsv> but if it can't find profile and friends from that file something's really ba [17:43] <JEEBsv> *bad [17:43] <JEEBsv> profile/level [18:14] <mrAlmond> JEEB : ok so probably I'm missing something in the conversion between annexb and avcc [18:15] <mrAlmond> I will look at mediainfo source code [18:15] <mrAlmond> maybe it will give me some hints [18:15] <mrAlmond> thank you...I'm going home [18:29] <JEEBsv> ugh @ looking at mediainfo source code [18:29] <JEEBsv> whyh not read 14496-15 :< [19:07] <Avon> Has anyone been successful using MacCaption/CaptionMaker with a file encoded using ffmpeg? I'm having trouble. [19:08] <Avon> Here is my ffmpeg output: http://pastebin.com/xJsE4xfJ [19:09] <relaxed> I see a semicolon [19:09] <Avon> Semicolon? [19:10] <relaxed> ; <-- in your timecode [19:10] <relaxed> was that intended? [19:10] <Avon> Yes, the original source file had a semicolon for the timecode as well [19:11] <relaxed> try with the timecode before the input [19:11] <Avon> Alright sure. [19:12] <Avon> Option timecode (set initial TimeCode value.) cannot be applied to input file 1.mxf -- you are trying to apply an input option to an output file or vice versa. Move this option before the file it belongs to. [19:12] <Avon> Error parsing options for input file 1.mxf. [19:13] <relaxed> does it work if you omit timecode? [19:13] <Avon> Do you mean if the video plays or if the thing encodes? [19:13] <relaxed> the latter [19:14] <Avon> It will encode if i omit the timecode. [19:15] <Avon> http://pastebin.com/vxiFab8L [19:15] <relaxed> try single quoting the timecode and remove the backslashes. [19:15] <Avon> Sure thing. [19:16] <digidog> Hi. Short question: after compiling and installing ffmpeg dev snapshot from sources it works for a moment, but after closing terminal and reopening it, it can't be found no more. Ot says ffmpeg is currently not installed. Does anyone ha da similar situation or knows how to fix this ? [19:16] <Avon> http://pastebin.com/9DQuGLx3 [19:17] <Avon> The thing is, im just getting errors when running it through MacCaption/CaptionMaker. :( [19:18] <Avon> Using MXFScope to analyze the file also throws me an error and says it cant be read. But the file plays just fine? [19:20] <digidog> ffmpeg seems to be installed into home/my_account/bin/ but I wonder if it should rather be under usr/local/bin ? Not sure ... :( [19:24] <JEEBsv> digidog: nothing is more or less invalid, the installation under your home folder is something that works even if you have no root [19:28] <digidog> @JEEBsv: Thank you. Sorry but I have to ask again: I followed this instruction https://ffmpeg.org/trac/ffmpeg/wiki/UbuntuCompilationGuide carefully but ffmpeg has started only once directly after installing it. When I close the terminal and open it again, typing "ffmpeg" causes info message that ffmpeg is not installed. navigating to /home/my_account/bin/ where ffmpeg is inside makes no difference [19:29] <JEEBsv> digidog: just add to your dot-bashrc or something export PATH=${HOME}/bin:${PATH} [19:29] <JEEBsv> in your home folder [19:29] <JEEBsv> as in, you just have to add that directory to your PATH [19:30] <JEEBsv> or you just call the whole path for the binary [19:30] <JEEBsv> if you're in the same directory you can do dot-slash-ffmpeg [19:32] <JEEBsv> (Windows by default has dot in the PATH so you can run stuff in the current directory without the dot-slash (or backslash) [19:32] <digidog> JEEBsv: PATH! Arrg. That's what I forgot to check twice! Thank you so much! I was only reading log output while it installs, and there it says, that it already did this. But I forgot to check, if it really was done correctly. Thank you very much. I am sure that's the point ... [20:05] <Avon> Having trouble running a file through maccaption/captionmaker Here is the output: http://pastebin.com/hLQRj2H2 [20:06] <Avon> MXFScope is telling me KLV is not aligned to KAG. [20:06] <Avon> Any ideas? [20:13] <digidog> JEEBsv: thanks a million for reminding me! PATH was the issue. fixed. thank you very much. Another short question: any change to log time postion by hitting a key o.s. while ffplay runs a video file from terminal? [20:15] <digidog> in terminal we can create cshorts for logging like in .bashrc something like: alias logme='echo $1 >> my.log' ... so I thought there is maybe a change to make it possible to log a certain paused postion while ffplay runs a file ... [20:15] <digidog> that would make creating position lists a very fast and smooth workflow [20:18] <digidog> I had the idea: that while ffplay runs we get output on terminal and could route this nto the log file like we do by installing software with >> my.log [20:19] <digidog> but it sadly doesn't write it int o the log. I have tried userterminal$ ffplay filename.mov >> my.log [20:20] <digidog> it creates the file but doesn't writes into it [20:23] <digidog> ok, moment. I think it my fault that it doesn't writes to log. needs further testing ... [20:26] <relaxed> digidog: the message are over stderr [20:26] <relaxed> messages* [20:27] <relaxed> ffmpeg > blah.txt 2>&1 [20:28] <relaxed> You could have awk pull out exactly what you need while running or do it later. [20:29] <relaxed> Everyone should know a little (g)awk. [20:31] <digidog> :D [20:33] <digidog> Great! you are right. this way it logs like expected. but how can I trigger an additional info? lets say I press space to pause and another key to get an event listet in the log. isi this possible? then I would only need to write a little paser to get my log time positions ... [20:44] <digidog> changing postiton of movie file with mouse for example causes a log entry which differs from the normal time position log list. that'S the way I thought it could be used another way with certain commands ... well ... just a thought [20:54] <digidog> I other words I need an option to get a timestamp of actual position in a log file. MOmentary I check the ffmpeg docs online and the most key commands available for ffplay only change somethng on the viewing window. Only a position change caused by mouse will be registered in log file so far I can see in the moment ... [23:31] <ismaell> hi [23:31] <ismaell> hope there's someone around who can answer this one... what does it mean "end of stream wanted"? [23:32] <ismaell> it's on the api doc [23:33] <durandal_1707> link? [23:35] <ismaell> durandal_1707: nevermind, looking at old docs :/, well, in the new ones it says av_write_frame return 1 "if flushed and there is no more data to flush" [23:35] <ismaell> returns* [23:37] <ismaell> durandal_1707: that's not clear either :/ [23:38] <ismaell> now, it seems it would return 1 if I run out of space, right? [23:38] <ismaell> (I'm writing to disk) [23:38] <ismaell> I guess the same applies to lost tcp connections [23:38] <durandal_1707> i don't think so [23:39] <ismaell> I was expecting a negative number actually, but when the disk ran out of space (well, actually not 100%, but about 94%, the rest is reserved) [23:39] <ismaell> it started to return 1 [23:40] <ismaell> or I think that is the issue... [23:40] <ismaell> kinda odd [23:40] <durandal_1707> it have nothing to do with disk being full or not [00:00] --- Tue Jul 2 2013

1 0