[Ffmpeg-devel-irc] ffmpeg.log.20120303

[burek]

[02:06] <wsd> Hi, I think I am misunderstanding how ffprobe -print_format is supposed to work or something. I just want to extract the fps from a videoclip in a script, but I am getting this output from ffprobe -print_format xml videoclip: http://pastebin.com/tUtJdP5P
[02:06] <wsd> Is this the expected functionality?
[02:08] <wsd_> (it's easy enough to work around this, but when I saw xml/json output was possible I thought it might make my script less fragile)
[02:11] <theBlah> heya :) i'm trying to combine an audio.wma with a single still image for x frames (length of audio) with an mp4 output - anyone happen to be able to help with that?
[02:11] <theBlah> i'm trying to convert old code that used to use -loop_input "image.jpg" , but i understand that's no longer acceptable
[03:12] <grepper> theBlah: try -loop 1
[03:17] <Ultimatespirit> Is it possible to tell ffmpeg to NOT include a file? Like copy everything but one of two audio streams?
[03:22] <grepper> Ultimatespirit: you are looping for the -map option ( not sure what you mean by 'not include a file' )
[03:22] <grepper> er, looking
[03:23] <Ultimatespirit> Basically I have two audio streams and i want to re-encode the file to not have one of the audio streams, while making the other AC3
[03:23] <Ultimatespirit> when I use map it does both, so chances are i'm not using it right...
[03:27] <grepper> -map 0:1 for example
[03:27] <Ultimatespirit> I'm currently doing something like this: -map 0 -vcodec copy -c:a:1 ac3 -ab 192k -c:s copy -c:t copy
[03:28] <Ultimatespirit> should i declare -map 0:1 before -c:a:1?
[03:29] <grepper> after afaik
[03:30] <Ultimatespirit> afaik?
[03:30] <grepper> "As Far As I Know"
[03:31] <grepper> if it was an older ffmpeg I'd just say 'yes', but I haven't used the new syntax much
[03:31] <Ultimatespirit> the new syntax is basically just a bunch of aliases of the old stuff, so it should work...
[03:33] <grepper> and its not ?
[03:34] <grepper> use both -map calls just before the output file
[03:35] <Ultimatespirit> oh, that makes sense...
[03:35] <grepper> hm, no the new docs say differently ... http://ffmpeg.org/ffmpeg.html#Stream-selection
[03:35] <grepper> the way I said is how I was used to doing it ...
[03:36] <grepper> let me know what you experiments yield :)
[03:38] <Ultimatespirit> it worked!
[03:38] <Ultimatespirit> as long as the map is declared somewhere in there i think it'll work
[03:39] <Ultimatespirit> thanks grepper :D
[03:39] <grepper> np
[05:30] <rickk> hi
[05:31] <rickk> can anyone tell me which field in mpeg ts is requied to be set?
[05:31] <rickk> or just url for reference
[09:27] <TACPILOT> my camera shoots in avchd (YV12) 4:2:0 color space is there any point in converting it to YUV 4:2:2 ??
[09:34] <rickk> there are plenty of sources
[09:40] <zap0> TACPILOT, for what reason do you want to convert it?
[09:42] <TACPILOT> will have several steps in editing work flow and want to ensure I not loose any quality along the way until its actually convrted to a lossy format
[09:43] <TACPILOT> I've read yuv is a bit more acurate than yv12 but not sure if it makes any diff once recorded
[09:44] <Haferstroh> Hello, I am looking for documentation of codec options in ffmpeg, like the option for -c:v libx264 -crf 0  -preset ultrafast, where can I get those Information?
[09:44] <Haferstroh> Iwas looking into the ffmpeg documentation, but it is confusing to be honest
[09:46] <rickk> yuv and yv12?
[09:47] <TACPILOT> yuv is color space 4:2:2 yv12 is color space 4:2:0
[09:48] <rickk> do you know that codec use yv12
[09:48] <rickk> anyway
[09:48] <TACPILOT> avchd
[09:48] <TACPILOT> h.264
[09:48] <rickk> whatever
[09:49] <rickk> they conver it to yv12
[09:49] <rickk> when it encode .
[09:49] <TACPILOT> its recorded in that in the camera for better compression
[09:49] <rickk> decode output also yv12
[09:50] <rickk> which one?
[09:50] <rickk> yuyv?
[09:50] <TACPILOT> ?
[09:50] <Haferstroh> In the fmmpeg AVCodec-Context AVOptions there is a command:
[09:51] <Haferstroh> -flags and then e.g. mv4
[09:51] <Haferstroh> how do I use this command? Is it this way : -flags mv4 for example ? or -mv4 ?
[09:52] <rickk> what is recorded in the camera??
[09:53] <TACPILOT> being so new to this type stuff I recommend looking at examples from others ... let me find u some links
[09:54] <TACPILOT> http://forum.videohelp.com/threads/277807-Useful-FFmpeg-Syntax-Examples
[09:54] <TACPILOT> http://rhughes2832.hubpages.com/hub/ffmpeg
[09:54] <TACPILOT> they were just some google searches
[09:55] <TACPILOT>  Haferstroh: that may help u see the command sets a bit more clearly
[09:57] <Haferstroh> thank you
[09:57] <TACPILOT> :)
[09:58] <Haferstroh> Two things I am always stumbling upon are "demuxer" and "muxer" what does these things mean?
[10:00] <rickk> muxer is in encoder, demuxer is in decoder
[10:01] <Haferstroh> encoder and decoder for what sort of things ?
[10:01] <rickk> mux the audio , video , subtitles and etc..
[10:01] <Haferstroh> than its just another word for encoder and decoder ?
[10:02] <rickk> it is one of encoder procedure
[10:02] <Haferstroh> okay thx I got it
[10:02] <TACPILOT> basically combining multiple things into one and then splitting one into multiple things
[10:04] <Haferstroh> I did a search for it and muxer is short for multiplexer
[10:05] <TACPILOT> :)
[10:12] <Haferstroh> If I wanna change the ratio of a video maybe from 320x240 and bitrate 690k to 1024x768. The mulitplier from 1st to 2nd is 3,2. Does I have to multiply the bitrate with the same multiplier from 690k to 2,21m ?
[10:15] <zap0> Haferstroh, imagine a 1x1 square...  now you enlarge to 2x2...   how big is the result, in area?   compared with the original
[10:18] <Haferstroh> A(2x2) = 4*A(1x)
[10:18] <Haferstroh> A(1x1)
[10:19] <Haferstroh> no you talked about a square ^^ not a cyrcle
[10:19] <Haferstroh> but its the same I quess
[10:19] <Haferstroh> yes it is
[10:34] <cbsrobot> Haferstroh: a lot of tutorials on the net are outdated ... but some still work
[10:36] <TACPILOT> once you start understandng the fundamentals .. then the doc's will make more sense to you
[10:39] <Haferstroh> Hello cbsrobot
[10:52] <Prottey> ffserver 0.10: cannot handle vorbis audio streaming (WebM)
[10:53] <Prottey> when I feed a stream to ffm file, it says "Unable to set CBR to [foo]: not supported"
[10:53] <Prottey> this issue shows with both libvorbis and internal vorbis
[14:09] <perrth> hello, I'm using ffmpeg 0.10 to play midi file, timidity.cfg got parsed correctly but sounds merely hiss and buzzing. Does midi supported via modplug? note: mod file played well
[14:16] <ubitux> perrth: is it possible to have a sample?
[14:17] <perrth> it's from windows bundled midi file
[14:17] <perrth> can I send that?
[14:20] <zap0> what makes you think it conforms to the MIDI standard?
[14:21] <perrth> hmm, yeah i remmember it's sometime refered as rmi instea of midi. well I will look for sample on the net
[14:23] <ubitux> "cfg" is not in the extension list of the libmodplug wrapper
[14:23] <ubitux> try renaming it to .mid or something
[14:23] <ubitux> i'm not sure what you mean by "parsed correctly" btw
[14:24] <perrth> no timidity.cfg is GUS pat mapping
[14:24] <perrth> the virtual synth
[14:24] <perrth> try this file http://ambassadormusic.com/music/samp-choochoo.mid, it also sounds strange
[14:27] <perrth> here is what I meant by "parsed correctly": http://pastebin.com/S65ba8Da
[14:28] <perrth> if timidity.cfg not parsed, the numbered list will be blank
[14:29] <ubitux> mmh did you try to play it with something also using libmodplug?
[14:29] <perrth> yes mod file played fine
[14:30] <perrth> maybe because it has its own wave/synth embedded
[14:30] <perrth> unlike midi
[14:32] <ubitux> i don't understand what you are using to play it
[14:32] <perrth> ?
[14:32] <perrth> ffplay
[14:32] <ubitux> you said "mod file"
[14:32] <ubitux> the software playing it well
[14:32] <ubitux> through libmodplug
[14:32] <ubitux> do you have one?
[14:33] <perrth> its ffplay
[14:33] <perrth> ffplay with libmodplug enabled
[14:33] <ubitux> what i'm asking is if you have a software playing samp-choochoo.mid the correct way, and if possible a software using libmodplug too
[14:34] <ubitux> iirc, vlc has a libmodplug support too
[14:34] <perrth> ah sorry, windows media player and winamp
[14:34] <ubitux> are you able to play that file with it?
[14:34] <ubitux> mmh
[14:34] <perrth> winamp played both mod and mid, wmp just mid
[14:34] <ubitux> isn't winamp using libmikmod?
[14:35] <perrth> it said Nullsoft MIDI player
[14:35] <ubitux> ok
[14:36] <ubitux> ffmpeg is using libmodplug to play these files; what i need to know is if it's a libmodplug issue, or an API misusage in the wrapper
[14:36] <ubitux> in the later case, it needs to be fixed
[14:36] <ubitux> otherwise, it has to be fixed in libmodplug
[14:37] <perrth> But in linux, does it played? just in case
[14:38] <ubitux> it plays, sounds a bit weird
[14:38] <ubitux> i can't compare with a "working" one though
[14:38] <perrth> how weird :D
[14:38] <perrth> ?
[14:39] <ubitux> long notes
[14:40] <perrth> that's much better then. Here it almost like those pats not decoded properly sounds merely buzzing
[14:42] <ubitux> perrth: http://blackhole.ubitux.fr/samp-choochoo.mp3
[14:42] <ubitux> sounds like this.
[14:48] <perrth> I think mine is worse
[14:54] <perrth> anyway why ffmpeg no longer accept negative value for -aq?
[15:05] <scravag> I need a howto in order to decode live MPEGTS data embedding H264 (code not commandline)
[15:05] <scravag> Any help would appreciated.
[15:07] <ubitux> perrth: why do you want a negative value?
[15:07] <ubitux> perrth: i'm using libmodplug 0.8.8.4
[15:11] <perrth> ubitux: I rely on libvorbis autov for streaming which use -1 and -2
[15:12] <perrth> can I make it accept negative value?
[15:12] <perrth> I also use libmodplug 0.8.8.4
[15:13] <perrth> using gravis GUS pat
[15:19] <ubitux> it seems libvorbis indeed allows negative quality (from -0.1 to 1.0)
[15:20] <perrth> in aotuv it goes far to -2
[15:20] <perrth> around 32kbps
[15:21] <perrth> great for streaming
[15:40] <cbsrobot> ubitux: odd qustion
[15:40] <cbsrobot> "question
[15:40] <cbsrobot> can I concatenate subs ?
[15:40] <ubitux> i'd say it might be possible
[15:41] <ubitux> did you try ffmpeg -i a.srt -i b.srt out.srt?
[15:41] <cbsrobot> nah - I mean like this: http://ffmpeg.org/faq.html#How-can-I-join-video-files_003f
[15:41] <cbsrobot> join subs and the tc will be updated
[15:42] <ubitux> just try :p
[15:42] <cbsrobot> heh - ok
[15:55] <cbsrobot> ubitux: lol the subtitle number increses but the starttime and endtime stay the same
[15:55] <cbsrobot> http://pastie.org/3511862
[15:55] <cbsrobot> thats where they concatenate
[15:56] <cbsrobot> and I get a lot of : [srt @ 0x102009800] Application provided invalid, non monotonically increasing dts to muxer in stream 0: 650780100 >= 36295200
[15:56] <cbsrobot> I used the mkfifo way ...
[15:57] <cbsrobot> ffmpeg -i a.srt -i b.srt out.srt did not work, out.srt was the same as a.srt
[16:00] <theholyduck> cbsrobot, tip
[16:00] <theholyduck> mkvmerge CAN merge subtitles
[16:00] <theholyduck> and split them
[16:00] <cbsrobot> ubitux: I guess it will not be easy to support it, because the srt does not contain any information of the total length of the movie - (or a sort of an end tag)
[16:00] <theholyduck> atleast. i've used the append function for entire files, including embedded subtitles.
[16:00] <theholyduck> and its fixed the subtitles right.
[16:00] <theholyduck> and i've cut, and it cut the subtitles right
[16:01] <cbsrobot> theholyduck: ok thanks for the tip
[16:01] <theholyduck> so, you could try using the srts and appending them in mkvmerge.
[16:01] <theholyduck> the mkvmerge gui is a easy way to do this
[16:01] <cbsrobot> I'll try it out ...
[16:03] <ubitux> gaupol can merge the subtitles btw
[16:03] <ubitux> iirc
[16:04] <cbsrobot> I was just wondering when I author a bluray and have 10 files to concatenate - and only some of them have subtitle what would be the best way to do it ...
[16:07] <cbsrobot> maybe encode the original with ffmpeg and x264 to an mkv, then mkvmerge them together and then author the bluray with tsmuxr ?
[23:23] <TACPILOT> any one have a suggestion for a post production image stabalizer ??
[00:00] --- Sun Mar  4 2012