[Ffmpeg-devel-irc] ffmpeg.log.20121101

burek burek021 at gmail.com
Fri Nov 2 02:05:01 CET 2012


[01:24] <burzum> hi there
[01:25] <burzum> im trying to figure out how to fix a video that has video and audio out of sync
[01:25] <burzum> ive read about async and asyncts but I could not get it to work using one of both
[01:25] <burzum> i read this http://www.kkoncepts.net/node/69
[01:26] <burzum> http://pastebin.com/PMz127xs
[01:26] <burzum> i dont see the offset for the 2nd stream
[01:27] <burzum> so i dont know what value I should use for itsoffset
[05:01] <kevin11> anyone? having trouble updating to ffmpeg 1.0 and newest x264 on Ubuntu Lucid Lynx following these instructions -> https://ffmpeg.org/trac/ffmpeg/wiki/UbuntuCompilationGuideLucid#libvpx
[05:02] <llogan> kevin11: the guide uses ffmpeg git, not ffmpeg 1.0
[05:02] <llogan> what's the problem, exactly?
[05:02] <kevin11> it seems all goes well. Instructions here use git, and so am I. sorry if that wasn't clear
[05:02] <kevin11> until
[05:03] Action: llogan has to leave in 2 minutes for 90 minutes
[05:03] <kevin11>  /home/kolds/ffmpeg/ffmpeg_5:N-38636-g84833b0-1_amd64.deb
[05:03] <kevin11> kolds at ubuntujeos:~/ffmpeg$ ffmpeg -version
[05:03] <kevin11> The program 'ffmpeg' is currently not installed.  You can install it by typing:
[05:03] <kevin11> sudo apt-get install ffmpeg
[05:03] <llogan> try "hash -r"
[05:03] <llogan> then "ffmpeg"
[05:05] <kevin11> no dice. same msg - not installed
[05:05] <llogan> then try: cd ~/ffmpeg && sudo dpkg -i (or is it -I) ffmpeg_5:N-38636-g84833b0-1_amd64.deb
[05:05] <llogan> apparently the deb file was created but i don't know why it did not install.
[05:06] <llogan> i have to go now, but if you're patient i will be back in ~2 hours.
[05:06] <kevin11> I may be asleep by then but can come on tomorrow. Thanks for listening
[05:06] <llogan> ok. i'll probably be around here tomorrow.
[05:27] <burek> burzum, see which stream starts first and then adjust the other stream (that starts later) by moving it towards the beginning
[06:21] <odigem> why when i make shots with -vframes 3 -r 1/60 ffmpeg always create two identical shots?
[06:45] <llogan> odigem: i don't understand what you're asking
[07:11] <fling> Hello!
[07:11] <fling> how to use http://my-webcam/mjpeg.cgi as a source for ffserver?
[07:51] <fling> Launch ffmpeg -f mjpeg -r 20 -i http://&
[07:51] <fling> it works
[07:51] <fling> now how to connect to ffserver using jwplayer or other thing I may use on a webpage?
[08:28] <fling> ffplay -f rtp -i http://myffserver:8090/live.rtsp
[08:28] <fling> [rtp @ 0x7fc58c0008c0] Unsupported RTP version packet received
[08:28] <fling> [rtp @ 0x7fc58c0008c0] Received too short packet
[08:37] <fling> Format flv
[08:37] <fling> VideoCodec flv
[08:37] <fling> Thu Nov  1 14:36:24 2012 [NULL @ 0x21e6a00]video codec not compatible with flv
[08:37] <fling> how may this be?
[08:50] <impulze> hm is there a proper way to find out which ffmpeg version first provided a specific version of libav*?
[08:52] <fling> VideoCodec libvpx
[08:52] <fling> Thu Nov  1 14:50:37 2012 Only VP8 video and Vorbis audio are supported for WebM.
[08:53] <fling> how to make a playable stream with ffserver?
[09:38] <fling> no error message but stops encoding somewhy > http://dpaste.com/822001/
[09:43] <fling> ok, it works ^ after adding -an -vcodec libx264 -b 100k
[09:43] <fling> what are names of default vpre presets?
[09:49] <fling> I need a preset that will use less cpu
[09:50] <zap0> why?   why is using CPU an issue?
[09:56] <mudkipz> I think -vpre is depreciated maybe
[09:57] <mudkipz> I use -preset
[09:57] <mudkipz> and the options are: ultrafast superfast veryfast faster fast medium slow slower veryslow placebo
[11:11] <Lennon> Hi guys
[11:12] <Lennon> I'll back in few minutes
[11:12] <Lennon> have a problem compiling ffmpeg with option --disable-swscale
[11:12] <Lennon> cmdutils.c: In function opt_default: cmdutils.c:444:20: warning: unused variable sc [-Wunused-variable] cmdutils.c: In function show_pix_fmts: cmdutils.c:1139:28: warning: unused variable pix_fmt [-Wunused-variable] LD	ffprobe_g libavfilter/libavfilter.so: undefined reference to `sws_isSupportedInput' libavfilter/libavfilter.so: undefined reference to `sws_isSupportedOutput' libavfilter/libavfilter.so: undefined 
[11:13] <Lennon> libavfilter/libavfilter.so: undefined reference to `sws_get_class' libavfilter/libavfilter.so: undefined reference to `sws_scale' collect2: ld returned 1 exit status make: *** [ffprobe_g] Error 1
[11:13] <Lennon> any idea?
[11:19] <saste> Lennon, you found a dependency problem in the build system
[11:20] <saste> the best way to fix it is to avoid to --disable-swscale, or in that case disable also libavfilter
[11:20] <saste> if you want to track the issue please file a bug report
[11:49] <Lennon> @saste, thanks for answer. Do you mean that i haven't met some dependencies on my sestem?
[11:50] <saste> Lennon, no, it's a problem of internal FFmpeg libraries dependencies
[11:50] <Lennon> I need the option --diable-swscale, because I would like to use ffmpeg-php and it's needed for work with GD image
[11:50] <saste> in theory it should be possible to build libavfilter even without libswscale
[11:51] <Lennon> ok, do I need libavfilter for work with videos? probably yes, right?
[11:51] <saste> Lennon, why GD images conflict with libswscale?
[11:51] <Lennon> only if you are using ffmpeg-php
[11:51] <saste> and yes, IIRC libavfilter is an hard dependency for ffmpeg the tool
[11:52] <saste> but libavfilter without libswscale is not that useful (especially for video)
[11:53] <Lennon> ok, so could you help me please point me how to bouild libavfilter without swscale?
[11:53] <Lennon> i dont know how, but many people compile ffmpeg without swscale with no problem ..
[12:05] <saste> Lennon, something is fishy with your build, if you --disable-swscale the scale filter should not be built at all, and your build should not fail
[12:05] <saste> but again, an ffmpeg with no libswscale is not that useful
[12:06] <saste> you still didn't tell because GD images conflict with libswscale
[12:06] <saste> makes no sense to me
[12:07] <Mavrik> he's having a problem BECAUSE swscale isn't built
[12:07] <Mavrik> libavfilter tries to link against it
[12:10] <saste> Malmis, swscale is not built because --disable-swscale
[12:10] <saste> also i tried it locally, libavfilter can be build without libswscale, but ffmpeg is not build at all
[12:11] <saste> libavfilter can be compiled without libswscale, but is not that useful (especially for video)
[12:11] <Mavrik> saste: the point is his build fails because of missing swscale, not that it isn't useful
[12:11] <Mavrik> slight difference :)
[12:12] <saste> Mavrik, if you --disable-swscale, it will hardly build libswscale
[12:13] <Mavrik> yes. however it should finalize ffmpeg build right?
[12:13] <Mavrik> or configure script should explicitly tell the user that he can't build ffmpeg without swscale
[12:13] <saste> Mavrik, ffmpeg_deps="avcodec avfilter avformat swscale swresample"
[12:14] <saste> that is ffmpeg won't be build unless all these libs are built
[12:14] <saste> no warning is created at the present time
[12:14] <Mavrik> hm, so why does the "--disable-swscale" option exist in configure script?
[12:15] <saste> for lot of reasons
[12:15] <saste> suppose you only care about libavutil
[12:15] <saste> then you --disable-everything --enable-avutil
[12:15] <Mavrik> ah, I see
[12:15] <saste> or you may only be interested in building the libraries, and not the tools
[12:16] <Lennon> so if I use --disable-everything --enable-avutil it should work
[12:16] <Lennon> ?
[12:16] <Mavrik> so probably the best solution is to fix configure to clearly tell the user libavfilter can't be built without building libswscale
[12:17] <saste> Mavrik, it can
[12:18] <saste> you don't always need scale
[12:18] <saste> for example if you only want to filter *audio*
[12:18] <Mavrik> ah
[12:18] <Lennon> @saste, you ask about the reason why do I need option --disable-swscale:
[12:18] <Lennon> ffmpeg-php is not compatible with ffmpeg's swscaler. Enabling swscaler in ffmpeg disables the img_convert functions that ffmpeg-php needs to do its thing.
[12:18] <Lennon> http://ffmpeg-php.sourceforge.net/
[12:18] <saste> Lennon, argh
[12:19] <Mavrik> argh.
[12:19] <Mavrik> I wonder if that will even work on 1.0
[12:19] <saste> Lennon, img_convert was deprecated and removed something like two years ago (or even more)
[12:19] <saste> the best bet is to find an old ffmpeg version, e.g. 0.5, and pray that it still works
[12:20] <Mavrik> 0.7 still has img_convert I think
[12:20] <Mavrik> lemme go check logs :)
[12:20] <saste> but really, you shouldn't rely on outdated and unmaintained packages
[12:20] <Lennon> aaa, ok, thanks for that point
[12:20] <saste> so I assume ffmpeg-php is unmaintained?
[12:21] <Lennon> yes, I think ffmpeg-php is no more developed :-/
[12:21] <Lennon> or shoul i say developing
[12:25] <DonGnom> hi folks, im currently writing an bash wrapper script for automated encoding of mpegts to x264 when i recall correct i had several month ago a little problem with ac3 coded streams which only have stereo channels enabled (like stuttering audio when switching to that channel) but it was solved when i reencoded it with the ac3 channel option set to the amount of channels used (2 on that stereo ac3 stream) so as i want to do it automated via my wrapper
[12:25] <DonGnom> script is it possible to get the used channel count from the input file within the output of ffmpeg -i <infile>?
[12:25] <kode54> I wish there were some visual editor for avisynth scripts
[12:26] <kode54> I'm currently using virtualdub with ffmpeg input plugin, to chop out frame ranges, then serve the remainder of the video to avisynth, which then performs deinterlacing, back to ffmpeg.exe
[12:27] <mudkipz> I think ffmpeg can deinterlace
[12:27] <kode54> ffmpeg has a fancy motion compensating / frame doubling deinterlace filter?
[12:29] <kode54> running QTGMC script with avisynth, distributed to four cores, runs a hair short of real time
[12:30] <kode54> but then it seems the virtualdub ffmpeg input freezes as it seeks past the second cut
[12:31] <kode54> hmm, maybe not
[12:31] <kode54> maybe it just froze after so many sessions of restarting as I tweaked the avisynth settings
[12:32] <kode54> huh, nope, froze up again
[13:06] <Lennon> @saste, may I ask you? I succesfully compiled ffmpeg 0.7.13 with option --disable-swscale, but when I try command ffmpeg it says: The program 'ffmpeg' is currently not installed.  You can install it by typing: apt-get install libav-tools
[13:06] <saste> Lennon, make install
[13:06] <saste> be sure to know what you're doing
[13:06] <Lennon> but when I compile ffmpeg and then install it using make install i saw it installed liavutils
[13:06] <saste> by default make install will install in /usr/local
[13:07] <Lennon> yes ... i know that .. i did that
[13:07] <saste> you're also supposed to update the PATH
[13:07] <Lennon> ./configure --disable-swscale && make && make install
[13:07] <Lennon> yep, that could be the problem
[13:08] <Lennon> but I nedd to tell i'm not to familiar with paths ... what should I do?
[13:17] <Lennon> @saste but I nedd to tell i'm not to familiar with paths ... what should I do? first time i compile ffmpeg i didnt need to wxport path
[13:18] <saste> Lennon, !google PATH
[13:18] <saste> it's really not ffmpeg-specific
[13:21] <Lennon> ok, first time I installed ffmpeg, I had the binary file at /usr/local/bin which is in my PATH var, but when I try to locate ffmpeg binary file now I could not locate it ...
[13:23] <Lennon> any idea?
[13:28] <relaxed> Lennon: the default prefix is /usr/local/bin, use the absolute path.
[13:30] <Lennon> relaxed thanks, but /usr/local/bin is my PATH var, and there is no ffmpeg binary file ... when I try command locate ffmpeg I couldn't find ffmpeg binary file
[13:30] <Lennon> # /usr/local/bin/ffmpeg -su: /usr/local/bin/ffmpeg: No such file or directory
[13:31] <relaxed> bash shell?
[13:31] <relaxed> run rehash
[13:31] <Lennon> bash
[13:32] <Lennon> No command 'rehash' found, did you mean:  Command 'rhash' from package 'rhash' (universe)  Command 'rahash' from package 'radare-common' (universe) rehash: command not found
[13:32] <Lennon> should i install rhash?
[13:32] <relaxed> ok, ok. If make failed then it wouldn't install it.
[13:32] <Lennon> but make didn't failed
[13:32] <relaxed> is the ffmpeg binary in your src dir
[13:33] <relaxed> if it is, run 'sudo make install'
[13:36] <burek> Lennon, why exactly do you need ffmpeg-php
[13:36] <burek> hash -r
[13:36] <Lennon> relaxed i did that, but in source dir i find only ffmpeg.c
[13:37] <relaxed> then make failed
[13:37] <Lennon> burek, I? trying to use it on web site to upload and manipulate with videos and save information about videos to db
[13:37] <relaxed> or else you ran make clean- it's on of the two
[13:37] <relaxed> one*
[13:37] <burek> Lennon, can you please take some time first to read this: http://ffmpeg.org/trac/ffmpeg/wiki/Using%20FFmpeg%20from%20PHP%20scripts
[13:38] <burek> it will help in further discussion
[13:50] <Lennon> burek, ok, thnaks for that, it could be maybe better for manipulating videos, but it can recive information about videos for saving it in db?
[13:51] <burek> sure
[13:51] <burek> use ffprobe
[13:55] <Lennon> ok, thanks ... so I will try to use ffmpeg directly ... and leave ffmpeg-php behind
[13:55] <burek> it would be a smart choice on the long run
[13:55] <Lennon> yep, you're right
[14:19] <kode54> okay, I'm giving up on ffms2
[14:19] <kode54> which options would I use to specify multiple frame ranges from the source video?
[14:30] <kode54> ie. can I use -ss and -t multiple times
[14:53] <kode54> never mind, I just used single ranges and encoded to .ts files and appended them together
[15:02] <burek> kode54, maybe you could use segments
[15:02] <burek> http://ffmpeg.org/ffmpeg.html#segment_002c-stream_005fsegment_002c-ssegment
[15:02] <burek> im not sure if that's muxer only
[15:02] <burek> or it can be a demuxer too
[15:07] <DonGnom> is there any way to get the audio channel count from the output of ffmpeg -i <infile> ?
[15:08] <burek> DonGnom, did you try ffprobe?
[15:08] <DonGnom> burek: as i can see its the same output
[15:08] <burek> man ffprobe
[15:08] <burek> it has much more output options
[15:08] <burek> even csv, json, etc
[15:11] <DonGnom> burek: hm yes i tried that but i can't se any change when using -print_format csv (and some others)
[15:12] <burek> can you pastebin some example output from ffproba
[15:12] <burek> ffprobe*
[15:13] <DonGnom> burek: sure here you go http://sebsauvage.net/paste/?949f4efcff23138e#PU8XdlOUEtYajT0mdIKcmbSCrUsTBvdwpji4QUldB2A=
[15:15] <DonGnom> burek: when using json there are {\n} at the end of the output
[15:16] <burek> try this
[15:16] <burek> ffprobe -show_streams -i input 2>/dev/null
[15:17] <DonGnom> burek: nice :) thank you
[15:17] <burek> :beer: :)
[15:18] <DonGnom> so now i think i can overwork my bash script to parse that output and gather all needed informations
[15:18] <burek> yes, also
[15:18] <burek> ffprobe -print_format csv -show_streams -i ...
[15:18] <burek> will give you just 2 lines
[15:19] <DonGnom> btw is it possible to embedd the dvb_teletex "subtitle" in some way into the file?
[15:19] <burek> if you just want to count it
[15:19] <burek> etc
[15:19] <DonGnom> burek: its 5 lines and no i don't want to count i want to parse out the data to generate an recoding command line for ffmpeg based on the exisiting streams.
[15:19] <DonGnom> my last encode failed with async audio & video
[15:20] <burek> oh ok
[15:20] <burek> it is possible to embed anythin
[15:20] <burek> either like an a/v/s stream
[15:20] <burek> or like a data
[15:20] <burek> attachments etc
[15:20] <DonGnom> burek: yes but i like to embedd it in that way that it can be used as teletex (as it is) and not simple subtitle
[15:21] <burek> oh, well, i don't know much about that, so hang around until someone else comes up with some ideas :)
[15:21] <DonGnom> like when i use xbmc on the .ts file i could display and navigat through the teletex. would be fine to have that in the mkv too but is not "absolutely" nesessary
[16:05] <DonGnom> hmm is the "fastfirstpass" removed?
[16:13] <DonGnom> is there any replacement for -vpre fastfirstpass and -vpre hq?
[16:26] <JEEBsv> DonGnom: -pass 1 automagically sets fast first pass settings, unless you use the placebo -preset. -preset sets libx264's internal presets that set speed vs compression settings. see mewiki's x264 settings page for details
[16:27] <DonGnom> JEEBsv: so i could possibly drob that options?
[16:28] <JEEBsv> yes, and just use the slowest preset that is 'fast enough' for you
[16:29] <DonGnom> JEEBsv: hm i only have a handfull of presets and all are for HD resolutions but i have dvd resolution so i think i just drop all -vpre options without replacement hsould be ok right?
[16:30] <JEEBsv> those are not the presets you are looking for
[16:30] <JEEBsv> -preset sets libx264's INTERNAL presets
[16:31] <DonGnom> JEEBsv: so not the ffpreset files ok which one is the one of choice?
[16:31] <JEEBsv> see mewiki's x264 settings page for a listing
[16:31] <DonGnom> JEEBsv: that page? http://mewiki.project357.com/wiki/X264_Settings
[16:31] <JEEBsv> yes
[16:31] <DonGnom> and how do i figure out which one i woant? :D
[16:31] <DonGnom> -o
[16:32] <JEEBsv> pick the slowest that is still fast enough for you
[16:32] <JEEBsv> that's all
[16:32] <DonGnom> and with "fast enough for me" you mean by realtime?
[16:32] <JEEBsv> whatever is your needed speed
[16:33] <DonGnom> JEEBsv: i think i have no "needed" speed.
[16:33] <DonGnom> but i think ill start with medium and look what happens :)
[16:33] <divVerent> that's the idea
[16:34] <JEEBsv> medium is default
[16:34] <divVerent> basically, start with medium, go to a faster one if you find it too slow, or to a slower one if you think it's ok if it is slower
[16:34] <divVerent> medium is probably the most tested one though
[16:35] <divVerent> but x264 generally is quite good quality, so I don't expect bugs in the more obscure modes
[16:35] <DonGnom> i think i just prefer good quality as endresult and low size the machine would encode some time without me watching it so i don't really care about the time it takes.
[16:36] <DonGnom> (i just want to lower the size of my mpegts recodrings from my settopbox without quality loss)
[16:36] <DonGnom> *recordings
[16:38] <divVerent> if you have infinite time, use placebo :P
[16:39] <DonGnom> divVerent: :-D what does it do?
[16:39] <divVerent> it is the slowest and best quality mode there is
[16:39] <divVerent> it's a lot slower than veryslow
[16:40] <divVerent> but only slightly better quality (did anyone ever notice a difference?)
[16:40] <DonGnom> probably you could only notice it on 4k videos :D
[16:40] <divVerent> for optimum quality, basically do this
[16:40] <divVerent> calculate how much time you have for one encode
[16:40] <divVerent> and the fps from that
[16:41] <divVerent> like, if you want to encode one 90min 25fps movie in 8 hours
[16:41] <DonGnom> im currently using 1500k bitrate in x624 agains 15000k in the mpeg2ts and i did not really notice any difference between them so...
[16:41] <divVerent> then you have 135000 frames
[16:41] <divVerent> to do this in 8 hours, you need at least 3.6 fps
[16:41] <divVerent> then do some test encodes of the same res as you need, and measure the fps of these in the different modes
[16:42] <divVerent> and pick one around 4 to 7 fps
[16:42] <divVerent> to measure the fps, you need to let the encoder run for at least 1, 2 minutes
[16:42] <divVerent> for the number to stabilize
[16:42] <divVerent> that way, you'll basically find "the slowest mode you can afford"
[16:43] <DonGnom> divVerent: i think medium is enough for me :)
[16:44] <divVerent> what's kinda more important for quality, is the right tune= mode
[16:44] <divVerent> basically: film or animation or grain
[16:45] <DonGnom> yeah but i think it doesnt matter i have limited quality as base (mpeg2ts from dvb-s/c) and i think i can't improve the quality when the source is not "perfect" :)
[16:45] <divVerent> sure
[16:45] <divVerent> well, the thing is
[16:45] <divVerent> the better you set this up
[16:45] <divVerent> the lower bitrate you can encode to
[16:45] <DonGnom> hmm thats true.
[16:46] <divVerent> but finding the right tuning mode is somewhat simple
[16:46] <divVerent> is it animation (large one colored surfaces and such stuff)? then -tune animation
[16:46] <DonGnom> i just hope my current setting does not again end up in async video/audio :(
[16:47] <divVerent> is it real recorded film with a little grain? then -tune film
[16:47] <divVerent> is it old looking recorded film with lots of grain? then -tune grain
[16:47] <divVerent> is it none of that (e.g. CGI scenes)? then no tune mode
[16:47] <DonGnom> divVerent: problem is that are lots of files and i just want to start a batch job and leave it alone.
[16:47] <klaxa> while speaking about this, is there any way to effectively remove banding by dithering or something?
[16:47] <divVerent> also, if it's sometimes this sometimes that, no tune option is also best
[16:47] <divVerent> yes, then use no tune mode
[16:55] <DonGnom> so lets look how bitrate 1250k with bt 250k looks :)
[17:10] <DonGnom> oh i almost forgot to say thankyou so... Thank you!
[17:22] <grepper> I've been using 'ffmpeg -h' in my script to find out which loop option the user's version supports (-loop_input vs -loop 0).  Its worked till recently, but now find that there is no '*loop*' in the output at all.  Any suggestions as to how to check this now ?
[17:22] <relaxed> -h full
[17:23] <grepper> Currently testing with version 1.0
[17:23] <grepper> ah
[17:23] <grepper> is -h full backwards compatible ?
[17:24] <relaxed> I'm not sure
[17:25] <relaxed> I should send a patch for: ffmpeg -h lets_make_it_hard_to_get_documentation
[17:25] <grepper> heh
[17:25] <DonGnom> grepper: tested it on 0.10.3 nothing changed in output and no error
[17:25] <grepper> okay, thanks for the help guys
[17:27] <lucas^> hello. I'm trying to use FFmpeg shared DLLs within a Windows 8 Metro app
[17:28] <lucas^> does libavformat / libavcodec make use of the Win32 API at any point?
[17:31] <grepper> relaxed, DonGnom: just tried -h full with 0.7.13 and it works fine as well, so I'll guess I'm good to go for most if not all versions :)
[17:37] <ellipsis753> Hello. How can I have ffmpeg convert my vob file losslessly to another format? Will -c copy work? Thanks.
[17:38] <sacarasc> ellipsis753: Sure.
[17:41] <ellipsis753> Basically I've got a scratched by totally readable dvd that I'm trying to reburn without the dataloss from the scratches. I'm fairly sure that the vobs I copied off the dvd have dataloss so I'm thinking if I convert them with ffmpeg losslessly this would fix scratches and give me another original quality copy. (dvds use error correction data so this is possible I think)
[17:44] <ellipsis753> I just used ffmpeg -i VTS_01_1.VOB -c copy foo.vob. This should restore the error-correcting data in foo.vob for when I reburn it right?
[17:45] <sacarasc> That is the same as doing cp VTS_01_1.VOB foo.vob, pretty much.
[17:46] <ellipsis753> sacarasc, Yeah. That's what I was wondering.
[17:47] <ellipsis753> sacarasc, It's kind of confusing. The thing is that part of the error-correcting data in my original is corrupt. I'd like to replace it.
[17:48] <ellipsis753> Strangly the vob looses about 10mb when I do that though.
[17:51] <ellipsis753> Basically cp would copy the vob with the bugs in that file. I want to fix the bugs in the original. Can ffmpeg do this?
[19:51] <Jester01> hi folks! anybody could offer some quick advice, I am trying to concatenate 4 files and used -i "concat:file1|file2|file3|file4" but ffmpeg prints the duration of the first one only and the output only contains that too
[19:52] <sacarasc> What kind of files are they?
[19:53] <Jester01> I have looked at the faq, not sure what I should see there except for the command I used
[19:53] <Jester01> ah, I might have been mistaken by that
[19:54] <Jester01> so concat itself only works for files that can be simpy concatenated, I get it
[19:55] <Jester01> and I found some more interesting concat example too let me try
[19:55] <Jester01> sacarasc: they are webm files from youtube
[20:00] <relaxed> Jester01: you can concat them with mkvmerge
[20:06] <Jester01> thanks that did it
[20:09] <joelataylor> Hi, I'm trying to use FFmpeg via CLI, and I'm getting this error: Unable to parse option value "0x0" as image size
[20:10] <joelataylor> running this command: ffmpeg -i w_0.mp3 -acodec libvorbis w_0.ogg
[20:13] <joelataylor> Error opening filters!
[20:14] <relaxed> add -vn
[20:16] <joelataylor> relaxed: as a global option?
[20:17] <joelataylor> disable video, ok, I'll try it
[20:18] <joelataylor> didn't work relaxed & http://pastie.org/5168125
[20:20] <ubitux> try to move -vn as output option
[20:21] <relaxed> yes, after the input
[21:00] <joelataylor> ok - it successfully convert to wav
[21:00] <joelataylor> and then to ogg
[21:05] <joelataylor> ubitux: relaxed that did it! Except: ffmpeg -i w_1.mp3 -vn -b:a 128k w_1.ogg  - still creates a 22mb file from a 4mb file
[21:05] <joelataylor> it seems to be ignoring my bitrate
[21:07] <joelataylor> ah ha! this worked: ffmpeg -i w_0.mp3 -vn -acodec libvorbis -b:a 128k w_0.ogg
[21:12] <KING_LEE> hello! I am still working on a script to convert folders of .MTS files to .mov or .avi files in order to play and edit them in cinelerra. i had some help from burek and relaxed with the commands and my script looks like this now  http://pastie.org/5168327  but when I run the script the output files are the size of 0 kb  http://pastie.org/5168319
[21:14] <ubitux> joelataylor: by default it encodes in lossless, so flac is selected
[21:14] <relaxed> KING_LEE: "No such filter: 'yadif'"
[21:14] <joelataylor> ubitux: gotcha, thanks!
[21:15] <ubitux> -c:v mjpeg -q:v 1 -c:a pcm_s16le
[21:15] <ubitux> wat.
[21:15] <llogan> KING_LEE: yadif requires --enable-gpl in your ./configure
[21:15] <ubitux> KING_LEE: maybe you forgot --enable-gpl at build time.
[21:16] <llogan> shit. we cancelled eachother out
[21:16] <relaxed> ubitux: we tried other codecs in order to import. do you have a better suggestion?
[21:16] <relaxed> even rawvideo, I believe
[21:17] <llogan> i'd expect huffyuv to work with it.
[21:17] <llogan> but i've never used cinnelelawahever
[21:18] <relaxed> KING_LEE: replace '-c:v mjpeg -q:v 1' with '-c:v huffyuv'
[21:19] <KING_LEE> yeah cinelerra is a bit of work to get it going but i cant really get used to kdenlive
[21:19] <llogan> the output will be large
[21:19] <relaxed> it's built against ffmpeg's libs so I can't understand why it's so picky.
[21:19] <llogan> also scale=1920:1080 is superfluous since it appears to be the same size of your source
[21:20] <llogan> i'm not sure how well mov handles huffyuv
[21:20] <llogan> perhaps mkv or avi
[21:22] <llogan> anyone try this yet: http://www.shotcutapp.com/
[21:23] <llogan> made by dan dennedy, the author of dvgrab and kino
[21:23] <KING_LEE> hmm,... i didnt like kino that much
[21:24] <llogan> so don't even try
[21:24] <KING_LEE> i am kinda used to cinelerra now and i would like to stick with it
[21:24] <llogan> i'm not saying your should switch
[21:25] <KING_LEE> i will give it a try, i had a look at openshot too
[21:26] <KING_LEE> but i still cant let go of that script, i want to get it going
[21:27] <relaxed> KING_LEE: I'll install cinelerra and see what formats it takes.
[21:27] <relaxed> I've been meaning to play with it.
[21:28] <relaxed> Sometime this weekend, check back.
[21:29] <KING_LEE> it's nice
[21:30] <KING_LEE> i think you gave me this link last time http://ffmpeg.gusari.org/static/
[21:31] <KING_LEE> I think I did not understand that fully... when i run ./ffmpeg  how can i run my script with it?
[21:44] <llogan> KING_LEE: "path/to/ffmpeg -i ..."
[21:47] <KING_LEE> sorry, what?
[21:48] <KING_LEE> i used it like the other script last time but it didnt work
[21:48] <KING_LEE> http://pastie.org/5047656
[21:49] <llogan> i can't give you a copy and paste answer because i don't know where you put the ffmpeg binary from gusari
[21:55] <KING_LEE> in /home/robert   i ectracted it and copied it into the folder of MTS files I want to recode
[22:02] <KING_LEE> sorry, maybe I handle things awfully wrong, I just find it kinda hart to get behind ffmpeg
[22:07] <llogan> ~/ffmpeg
[22:12] <KING_LEE> llogan i am confused
[22:19] <relaxed> damn it, KING_LEE, learn some basic scripting.
[22:20] <relaxed> replace ffmpeg in the script with the abosulate path to the static binary. If you don't know what that means, google until you do.
[22:20] <relaxed> absolute*
[22:21] <cbsrobot> or RTFM
[22:21] <cbsrobot> (had to be said)
[22:39] <KING_LEE> yeah, you are right. sorry, I just want to get my editing going, I am not much of a terminal user but I'll try. thanks for your suppurt and for being honest
[22:55] <mcepl> I have here dump of DVB stream (via mplayer -dumpstream) which I can play perfectly both with ffplay and mplayer, but when I try to convert it to .avi file (via mencoder -ovc copy -oac copy) it completely looses sync of audio and video.
[22:55] <mcepl> some tips how to fix it?
[22:57] <mcepl> (ffmpeg crashes with "Application provided invalid, non monotonically increasing dts to muxer in stream")
[23:15] <kode54> ugh!
[23:15] <kode54> damn vlc
[23:15] <kode54> I somehow didn't set the capture frame rate on the video capture device the last time I recorded
[23:16] <kode54> and it recorded an unlimited frame rate h.264 clip
[23:16] <kode54> thousands of dupe frames
[23:25] <dv_> is it possible to flip color channels? I have input data that has its R and B channels in the wrong order
[23:26] <saste> dv_: Cb/Cr?
[23:26] <dv_> no, input data is a set of PNG files
[23:52] <pzich> it seems like it'd be simple to write
[00:00] --- Fri Nov  2 2012


More information about the Ffmpeg-devel-irc mailing list