[Ffmpeg-devel-irc] ffmpeg.log.20131217

burek burek021 at gmail.com
Wed Dec 18 02:05:01 CET 2013

[00:29] <bri> does anyone know if features from ffmbc would ever be merged into ffmpeg?
[00:31] <saste> bri, ffmbc is GPL, so we can't merge its features in ffmbc without relicensing
[00:31] <saste> in practice it means that all feature should be implemented in ffmpeg from scratch
[00:31] <iive> only baptiste could say that.
[00:31] <bri> i noticed that. (at least in comparing the mxf demux/mux)
[00:32] <bri> is it permissible to change the license like that?
[00:32] <saste> not without original author explicit consent
[00:33] <bri> weird.
[00:37] <bri> i know there's definitely some good features in there, like proper support for reading opatom mxf
[00:42] <klaxa> licensing is weird, right? :)
[00:44] <bri> klaxa: yes. i don't like dealing with it ;(
[04:01] <llogan> DeadSix27: -g
[06:00] <SLAYBoz> i was just told that faac license is not GPL compatible  so why does ffmpeg include faac
[06:18] <grepper> it and various other libraries are included if you use --enable-nonfree
[06:33] <allengreen> how can I force the rtsp's rtp use tcp only?
[15:21] <eckesicle> Hello, I'm getting an "Operation Not permitted" error when calling avformat_open_input on an mp4 file. I do not believe it is a permissions error because I am able to read and write to the same file using stdio. What could cause this?
[15:22] <ubitux> return -1 in the code
[15:22] <eckesicle> Yes the method returns -1
[15:23] <ubitux> it's often mapped to ENOPERM
[15:23] <ubitux> but it was just meant to be a negative return value originally
[15:23] <ubitux> can you share a sample?
[15:24] <ubitux> does it happen with ffmpeg/ffprobe?
[15:24] <eckesicle> i am developing using libavforamt
[15:26] <eckesicle> https://gist.github.com/cfeckardt/0877ce3b15c26fc06e00
[15:27] <koljjhi> In trancoding a DV file to MPEG the resulting left audio channel has a buzz. How do I fix that?
[15:27] <koljjhi> Here's the code I use- $ffmpeg -i Input.DV -target ntsc-dvd Output.mpeg
[15:28] <eckesicle> ubitux: it could well be that I am missing some library that needs to be linked with libavformat to read mp4 files
[15:28] <ubitux> libavformat has native support for mp4, that's not it
[15:29] <ubitux> your ffmpeg should use the same libavformat, that's why i'm asking if the probing work with it
[15:29] <ubitux> it's just a first step to figure out if it's a problem with the sample, or with your code
[15:29] <eckesicle> the file's on an embedded device
[15:29] <eckesicle> a camera to be specific
[15:30] <eckesicle> but ill see if i can extract a sample file
[15:30] <eckesicle> ffmpeg is compiled with minimum libraries
[15:31] <ubitux> check if it has mp4/mov demuxer then
[15:31] <eckesicle> --enable-decoder=mpeg4 \
[15:31] <ubitux> demuxer, not decoder
[15:31] <JEEB> if you have a minimum config within libav* then make sure you have all of the protocols you need enabled, file among others, as well as mov demuxer and various other parsers you might need
[15:31] <eckesicle> no I don't see it in the compilation options
[15:32] <ubitux> --enable-demuxer=mov
[15:33] <JEEB> eckesicle, just see the list you get after you run configure with your options, if it starts lacking stuff then you know you're disabling too much :P
[15:33] <JEEB> also this is one of the problems I see the most
[15:33] <JEEB> you start minimizing your build BEFORE you actually have the code working
[15:33] <eckesicle> hmm we have to
[15:34] <eckesicle> have very little memory to spare
[15:34] <JEEB> that makes it more understandable but you should really keep an eye on what's actually enabled in your builds :P
[15:34] <eckesicle> :D
[15:35] <JEEB> you have no idea how many people have had similar problems because of just disabling features they actually needed :D
[15:35] <eckesicle> :)
[15:35] <eckesicle> anything else i need for mp4 files with x264?
[15:35] <JEEB> the h264 parser and decoder, mp4 and mov demuxers, file and other protocols
[15:37] <eckesicle> what's the command for parser
[15:37] <eckesicle> enable-parse=h264 _
[15:37] <eckesicle> ?
[15:37] <eckesicle> enable-parser=h264
[15:37] <JEEB> just look at the configure line when you run it without any options :P
[15:37] <JEEB> as in the output of it
[15:37] <JEEB> it lists all the things buildable with LGPLv2
[15:38] <eckesicle> ah ok
[16:11] <eckesicle> Nope
[16:11] <eckesicle> didn't work
[16:57] <eckesicle> Okay so I enabled all relevant, demuxers, decoders and parsers
[16:58] <eckesicle> still getting -1 returned when calling avformat_open_input
[16:58] <eckesicle> https://gist.github.com/cfeckardt/6cac7cd854c098654cd2
[17:21] <eckesicle> Hmm
[17:21] <eckesicle> it segfaults on av_probe_input_format2 too.
[18:04] <DrSlony> Hi, I want to transcode a "-preset ultrafast" screencast recording to "-preset slower" and I want to set the framerate at 15fps, where do I put the -r? before the -i or after? does it make a difference?
[18:04] <DrSlony> code: http://paste2.org/YEAVY7zD
[18:04] <sacarasc> It's an output option, so you put it after.
[18:12] <DrSlony> thank you
[18:13] <DrSlony> Is there anyone here who would like to take a shot at stabilizing (deshaking) a video clip?
[20:32] <caribou|> hello
[20:33] <caribou|> i'm trying to make a video out of a sequence of images and i have troubls matching files like screen_1387009721.5.jpg screen_1387009731.51.jpg screen_1387009741.51.jpg screen_1387009751.5.jpg
[20:33] <caribou|> (this is the sequence)
[20:34] <caribou|> i think i can't use the sequence pattern so probably with glob_sequence ?
[20:34] <caribou|> (as i can't use glob on windows)
[20:34] <caribou|> again i tried stuff but it doesn't seem to work
[20:36] <Compn> start renaming it :)
[20:36] <Compn> if you are on windows, i recommend oscars' renamer program
[20:36] <Compn> fairly easy to use
[20:39] <caribou|> i wish i could have done it with ffmpeg but yeah i guess you removed the potential hassle of finding a proper rename software
[20:39] <caribou|> i'll try it, thanks :)
[20:41] <Compn> but uh
[20:41] <Compn> what is your command line caribou| ?
[20:41] <Compn> there is a special image input option to fix image names like that
[20:42] <Compn>  %3d or something
[21:20] <llogan> caribou|: just use the glob pattern
[21:20] <llogan> http://trac.ffmpeg.org/wiki/Create%20a%20video%20slideshow%20from%20images
[21:20] <llogan> http://ffmpeg.org/ffmpeg-formats.html#image2-2
[21:21] <llogan> no need to rename
[22:04] <caribou|> Compn, isn't %3d just 3d digits ?
[22:04] <caribou|> i have [fixed digits number].[one|two(digits)]
[22:05] <caribou|> llogan, i don't have the glob pattern
[22:05] <caribou|> only glob_sequence :/
[22:24] <klaxa> is %3d even a valid strint format?
[22:25] <klaxa> it is
[22:25] <klaxa> *string even
[22:25] <DeadSix27> yep
[22:26] <DeadSix27> Mainly used for image sequences (gets replaced by an integer), in fact i don't know any other usage.
[22:29] <klaxa> well it paddes with spaces so that's bleh
[22:29] <DeadSix27> shouldnt
[22:29] <klaxa> padding with zeros is smarter and that would be %03d
[22:29] <DeadSix27> the 3d, makes it zero appended
[22:30] <DeadSix27> i mean
[22:30] <klaxa> %03d prepends zeros %3d prepends spaces
[22:30] <DeadSix27> yes.
[22:30] <DeadSix27> theres also the option, -start_number
[22:30] <DeadSix27> to define from where it starts.
[22:30] <DeadSix27> counting
[22:31] <DeadSix27> in case you dont want 0.
[22:31] <DeadSix27> or want 0 w/e
[22:48] <Overflip> Unknown option "--enable-libdirac". using the current ffmpeg version. Does anyone know the last version that his library is supported?
[22:50] <braincracker> lemon juice is good with salt ?
[22:53] <Compn> Overflip : current, you mean git ?
[22:53] <Compn> dirac should still be there
[22:53] <Overflip> ffmpeg-2.1.1
[22:53] <Overflip> not from git
[22:54] <Compn> oh nevermind
[22:54] <Overflip> I'm about to try 1.2.4
[22:54] <Compn> it was removed by libav (and remove-merged to ffmpeg) in 2012
[22:54] <Compn> 2012-04-24
[22:54] <Compn> get an ffmpeg before that date
[22:55] <Compn> or maybe michaelni will undo that merge ?
[22:55] <Overflip> ty =)
[23:06] <Overflip> what about libx264?
[23:08] <Overflip> ah nm
[23:20] <Overflip> libavcodec/libavcodec.a(libx264.o): In function `X264_init'
[23:20] <Overflip> ffmpeg-2.1.1/libavcodec/libx264.c:569: undefined reference to `x264_encoder_open_138'
[23:20] <sacarasc> That sounds like you have 2 copies of x264 installed.
[23:29] <Overflip> well I'm getting ERROR: libx264 not found
[23:29] <Overflip> with running ./configure
[23:29] <Overflip> which x264
[23:30] <Overflip> shows it in /usr/local/bin/x264
[23:32] <klaxa> you need libx264 in a place where configure finds it or you tell configure where to look for it
[23:32] <klaxa> regardless you need libx264 (apparently libx264.so.138)
[23:33] <Overflip> how do I tell configure where it is?
[23:34] <klaxa> --extra-ldflags=-L/path/to/lib/dir --extra-cflags=-I/path/to/include/dir
[23:34] <klaxa> those directories hold libx264.so and x264.h respectivley
[23:35] <Overflip> thanks I'll give that a try
[23:45] <Vandalite> has anyone ever run into a situation where ffmpeg works perfectly fine when running as the root user, but fails with no clear error when run as another user on the same system?
[23:47] <Overflip> I can only find x264.o and x264.h, do I need to rename thos?
[23:48] <Vandalite> I mean, ffmpeg runs, it appears to be able to read the file, but breaks setting up the encoders with: Error while opening encoder for output stream #0:0 - maybe incorrect parameters such as bit_rate, rate, width or height
[23:51] <Overflip> ya I'm confused
[23:53] <Overflip> How do I tell ffmpeg where x264 is installed?
[00:00] --- Wed Dec 18 2013

More information about the Ffmpeg-devel-irc mailing list