[Ffmpeg-devel-irc] ffmpeg.log.20151217

[burek]

[02:27:00 CET] <zumba_addict> anyone here has tried slowmovideo?
[03:18:11 CET] <jigg> What's the lowest framerate possible for FFmpeg output? assume mkv or mp4 container.
[03:19:53 CET] <c_14> matroska presumably 1 frame every millisecond
[03:19:56 CET] <c_14> don't know about mp4
[03:21:10 CET] <c_14> Probably something like 1/int_max
[03:22:15 CET] <c_14> Since matroska has millisecond resolution of timestamps and I think you can set any timebase you want for mp4
[03:55:17 CET] <DHE> you might want a sane number though. players might have a problem if frames are too far apart and you seek around
[05:09:36 CET] <jigg> I mean the slowest. Could I do 1 frame per day?
[05:11:30 CET] <jigg> No seeking needed.
[08:23:47 CET] <blackerking> Hello which option display all recognized information(Codecs/Streams...) of a container?
[08:25:11 CET] <cbsrobot_> blackerking: try ffprobe
[08:56:29 CET] <waressearcher2> blackerking: hallo
[08:56:36 CET] <waressearcher2> ffmpeg -codecs ?
[08:56:59 CET] <blackerking> Okay probier ich gleich
[08:57:31 CET] <blackerking> Die Dateien sind von einer HDD aus einer Skybox und die Dateien lassen sich nicht mit Windows kopieren, probiere erst einmal ein linux um sie zu verschieben
[08:58:50 CET] <blackerking> und dann den Befehl :)
[09:09:39 CET] <jigg> simplest way to get the timestamp of frame x?
[09:12:50 CET] <waressearcher2> ffmpeg zu benutzen ?
[09:14:25 CET] <jigg> yes
[09:14:52 CET] <jigg> but which filter or option
[09:19:57 CET] <waressearcher2> ffmpeg -i input.mp4 -t 10 -filter:v "fps=fps=25, showinfo" -f null - 2>&1 grep pts_time | awk '{print $6}' | cut -d: -f2
[09:24:21 CET] <jigg> does that get stamps for all the frames in the 10 second interval?
[09:32:08 CET] <waressearcher2> jigg: ja
[09:32:55 CET] <jigg> cool thanks ;)
[10:35:26 CET] <xintox_> hello
[11:05:45 CET] <sitoumbaz> Dear all I am using ffmpeg for decoding x265_nal unit but when I compile my program I get  this error : libswresample.a(soxr_resample.o): In function `create':
[11:05:45 CET] <sitoumbaz> (.text+0x354): undefined reference to `soxr_create'
[11:06:30 CET] <sitoumbaz> I include all libs listed below : LIBS=-lpthread -lavcodec -lavformat -lswscale -lswresample -lavutil -lx265 -lm -ldl -lz
[11:07:19 CET] <sitoumbaz> Is there some one who Has  idea about this error ?
[11:28:30 CET] <jigg> when you compile? if using the libav library you should visit the devel channel.
[11:33:11 CET] <bencoh> hmm, afaict library users should still ask here, not in the -devel chan. unless it's related to ffmpeg development itself
[11:53:53 CET] <waressearcher2> xintox: hallo
[11:54:03 CET] <sitoumbaz> of course the devel channel is related only on the development of ffmpeg not for use of ffmpeg
[13:33:43 CET] <fritsch> when reading jpg / png / tiff whatever images via an avio context, it might happen that avformat_open_input returns empty as it does not know what it is reading ... now helping this function by specifiing an AVInputFormat* helps.
[13:34:00 CET] <fritsch> the question now is: what to specify in av_find_intput_format(string)
[13:34:17 CET] <fritsch> I succesfully used: jpeg_pipe, png_pipe, tiff_pipe, webp_pie
[13:34:35 CET] <fritsch> which all work correctly - but there is very little documentation on what is expected as parameter
[13:34:53 CET] <fritsch> https://github.com/fritsch/xbmc/commit/1d38509b680b5a6020abc3dffc7a5d5ccc7c4ac2#diff-6c48c7b2be31f1d5311113860b1a10ddR170 <- code
[16:05:28 CET] <kevinranks> when using the ffmpeg -ss option to seek to a point in the movie, how does ffmpeg determine what bytes to request?
[16:10:50 CET] <kevinranks> running the command with log level debug, i can see it requesting partial content of the file and it gets the total bytes....but then it sends back a request for a byte range that's bigger than the original file and it becomes "416 Unsatisfiable Request" by the remote server
[16:34:06 CET] <kevinranks> I've posted my question in the forum as well as stack overflow for anyone interested. Would love to learn more about how this works: http://ffmpeg.gusari.org/viewtopic.php?f=11&t=2530
[16:34:46 CET] <waressearcher2> kevinranks: hallo, wie geth's wie steht's ?
[16:36:21 CET] <kevinranks> Hi waressearcher2, do you speak English?
[16:37:37 CET] <waressearcher2> kevinranks: manchmal
[16:38:06 CET] <kevinranks> hehe
[17:03:30 CET] <termos> I'm getting the message: /usr/bin/ld: /usr/local/lib/libavformat.a(matroskadec.o): undefined reference to symbol 'BZ2_bzDecompressEnd' when linking against libffmpeg from my C application, this error has not occured before with older versions. Any idea what's wrong? I tried moving the -lbz2 linking flag around when compiling
[17:07:32 CET] <termos> Oh, it worked when I set the libraries correctly in my CMake...
[17:08:04 CET] <waressearcher2> klaxa: was fur "Excess Flood" eigentlich ?
[17:11:00 CET] <G3ntoo> is it possible to output to an icecast:// that's https? even though it's silly, I tried icecasts://
[17:27:18 CET] <shincodex> avcodec_decode_video2(codecContext, frameNative, &frameFinished, &packet); returns frame and gets converted by sws scale to rgb then written ppm
[17:27:45 CET] <shincodex> But images dumped are doing image wrap effect indicating something is screwed up by stride somewhere.
[17:32:49 CET] <Mavrik> And? :)
[17:33:25 CET] <waressearcher2> shincodex: was ist passiert ? geht's es dir gut ?
[17:34:11 CET] <shincodex> nien
[17:34:37 CET] <waressearcher2> was, bist du Wahnsinn ?
[17:44:40 CET] <shincodex> middle finger to typedef standard types uint8_t unsigned char
[17:49:11 CET] <waressearcher2> böse böse böse
[17:50:14 CET] <shincodex> akuma akuma akuma
[21:23:53 CET] <Lac3rat3d> anyone know how to rip a BluRay disc to mkv or mp4?
[21:28:04 CET] <kepstin> an unencrypted one? sure, that's easy, there's some mpeg-ts files that you can just use as input to ffmpeg (might need to concatenate them in some cases).
[21:28:50 CET] <Lac3rat3d> ya it's not encrypted
[21:28:52 CET] <Lac3rat3d> i have the .iso
[21:28:59 CET] <c_14> If you build with libbluray, there's a bluray demuxer
[21:29:02 CET] <Lac3rat3d> just wanna rip it to mp4/x264/ac3
[21:31:04 CET] <furq> handbrake will do it if it's decrypted
[21:31:12 CET] <Lac3rat3d> pfft
[21:31:13 CET] <Lac3rat3d> handbrake
[21:31:17 CET] <Lac3rat3d> just uses ffmpeg underneath
[21:31:24 CET] <Lac3rat3d> i just wanna learn how to do it with ffmpeg
[21:31:27 CET] <furq> it doesn't just use ffmpeg
[21:31:33 CET] <Lac3rat3d> mostly
[21:31:43 CET] <furq> otherwise it wouldn't be able to preserve chapter markers and whatnot
[21:31:56 CET] <Lac3rat3d> don't really care about chapters
[21:32:01 CET] <Lac3rat3d> just wanna rip it to mp4/x264/ac3
[21:32:16 CET] <Lac3rat3d> how do i tell if it's encrypted or not
[21:36:43 CET] <Lac3rat3d> I figured it out
[21:36:46 CET] <Lac3rat3d> mount the iso
[21:37:28 CET] <Lac3rat3d> ffmpeg -i path/to/iso/BDMV/STREAM/00300.m2ts -c:v copy -c:a copy -c:s copy output.mkv
[22:23:49 CET] <spiderkeys> I'm using libavformat to mux some raw h264 packets into fragmented mp4 and I'm running into a weird issue where if I specify "empty_moov" as one of the options to feed to avformat_write_header(), the program gets stuck in some kind of infinite loop. The setup in code: http://pastie.org/10638776
[22:23:53 CET] <spiderkeys> Any ideas?
[22:24:20 CET] <spiderkeys> If I just use "frag_keyframe" it goes through fine.
[22:25:09 CET] <spiderkeys> I've got log level set to trace, but it doesn't create any output in that step, so I have no idea where it's having a problem, so currently I'm just digging through movenc.c
[22:41:45 CET] <pkeuter> i am trying to cut a fragmented mp4 file (frag_keyframe+empty_moov) by seeking and setting a duration. I´m trying to pipe this to a stream, but this doesn´t work because the file is not a streamable format. Obviously it works when I re-set the flags, but the whole ´cutting´ takes kind of long. Is there a way I could achieve a bit of a performance-gain? I.e. Is using a filter faster? Is it faster to first save to a file and then p
[22:41:45 CET] <pkeuter> ass that file? Or other things I might need to consider?
[22:47:50 CET] <spiderkeys> Nevermind! I tracked down what the problem was for my issue. I didn't set the buffer to writable when I allocated the custom avioContext for my output, so movenc.c's write_header couldn't perform any tell/read/write ops
[23:04:02 CET] <Mista_D> just cloned git, trying to extract r25887     http://pastebin.ca/3290582   Plz assit
[23:05:11 CET] <Mista_D> using manual on ffmpeg.org/git-howto.html
[23:09:24 CET] <c_14> Mista_D: git doesn't have revisions
[23:10:21 CET] <Mista_D> c_14: is there older SVN repo up somwhere?
[23:10:59 CET] <jcay> do you want to get old sources?
[23:12:14 CET] <llogan> Mista_D: why?
[23:13:09 CET] <Mista_D> looking for SVN r25887 to test older processing module built by somone specifically for thqat version.
[23:13:59 CET] <jcay> how do you know it's this revision?
[23:14:14 CET] <llogan> what about the 51458 newer commits you'd be missing out on?
[23:16:11 CET] <Mista_D> Is that wrong info ? : https://ffmpeg.org/git-howto.html#toc-Finding-a-specific-svn-revision
[23:17:54 CET] <jcay> for me this command works "git show :/'as revision 23456'"
[23:18:58 CET] <llogan> Mista_D: works for me. your git is probably too old
[23:19:12 CET] <Mista_D> 1.7.1
[23:19:31 CET] <llogan> 35b0169
[23:19:35 CET] <jcay> mine is 1.9.1
[23:20:30 CET] <furq> 1.7.1 is ancient
[23:27:26 CET] <jcay> maybe the command part is right, but the earliest git version part is wrong...
[23:29:16 CET] <jcay> https://github.com/git/git ---> github of git :D I should clone that :D
[23:31:24 CET] <kepstin> that's just a mirror, the official git git is at git://git.kernel.org/pub/scm/git/git.git
[23:31:34 CET] <jcay> oh
[23:31:43 CET] <jcay> thanks
[23:31:47 CET] <kepstin> but for most uses, probably doesn't make a difference
[23:32:00 CET] <jcay> I guess not
[23:32:02 CET] <furq> either way 1.7.1 is over five years old
[23:32:18 CET] <furq> it's so old that ubuntu 12.04 is using a newer version
[23:34:58 CET] <jcay> yeah, v1.7.1, last commit Date:   Fri Apr 23 18:27:17 2010 -0700
[23:41:12 CET] <jcay> I don't quite understand git branching, they are all local even if you pull?
[23:49:27 CET] <JEEB> jcay: git branch shows the branches you have checked out at some point locally, git branch -r shows all of your locally fetched states of remote branches
[23:50:38 CET] <JEEB> when you do "git checkout branch_name" git goes look for such a branch in the remotes you have fetched (the default cloned one usually is called "origin"), and if it finds one, it creates a new local branch that follows that remote branch
[23:51:04 CET] <JEEB> or well, "track" is the proper verb I guess rather than "follow"
[23:51:45 CET] <jcay> and when I push my local branch, it becomes public?
[23:53:00 CET] <JEEB> yes, when you push something to a remote, then it becomes available there
[23:54:16 CET] <jcay> I see, so you can't make a branch like "test" you need a unique name for that
[23:54:43 CET] <JEEB> well, "test" might be unique enough
[23:54:52 CET] <jcay> unlikely :D
[23:55:01 CET] <JEEB> if it's your repo, you should know :D
[23:55:49 CET] <jcay> :)
[23:55:52 CET] <JEEB> git push remote HEAD:test tries to push current HEAD (short-hand for current state) to remote'
[23:55:56 CET] <JEEB> s test
[23:56:17 CET] <jcay> oh cool
[23:56:37 CET] <JEEB> git push remote :test then proceeds to push nothingness into remote's test
[23:56:46 CET] <JEEB> (aka "removing branches")
[23:58:00 CET] <JEEB> also if what you are trying to push is not a follow-up in history for the current state on the remote that you're trying to push to
[23:58:15 CET] <JEEB> you will get a fat warning
[23:58:52 CET] <jcay> what suxx about SVN merging, you do it yourself locally and then commiting ;/ that's quite a big disadvantage
[23:59:03 CET] <jcay> oh, that's smart :)
[23:59:21 CET] <JEEB> svn merges also lose any resemblance of history
[23:59:31 CET] <jcay> also, yes
[23:59:35 CET] <JEEB> you just throw up a lot of bits
[23:59:47 CET] <JEEB> and call the commit "merge beliyaal's branch"
[00:00:00 CET] --- Fri Dec 18 2015