[Ffmpeg-devel-irc] ffmpeg.log.20130416

[burek]

[00:03] <Ed|Laptop> hey guys, my friend uses WinFF and I am trying to mimic his encoding flags on my Linux Server's ffmpeg.  Would anybody be able to give me a hand with translating the flags from WinFF to ffmpeg?  All of my initial attempts have failed for various reasons
[01:29] <gregoiregentil> I have an h264 file slightly damaged, because of missing frames. ffplay doesn't play it well as it jumps over the missing frames. But if I add "-fflags +igndts", it works well.
[01:29] <gregoiregentil>  I want to re-encode the file: with the command "ffmpeg -i a.mpg -fflags +igndts -an -vcodec mpeg4 -r 65535/2733 greg.mp4"
[01:30] <gregoiregentil> it doesn't work well like ffplay WITHOUT igndts. I see that ffmpeg drops frames
[01:30] <gregoiregentil> Why igndts doesn't seem to work with ffmpeg while it's working with ffplay?
[03:53] <praveenmarkandu> Hi, would anyone know until what protocol version FFMPEG is compatible to encode HLS
[03:54] <praveenmarkandu> for example, newer versions of HLS are able to split the audio and video into different files & streams
[03:54] <praveenmarkandu> thus you can easily cater for dubbed content
[03:54] <praveenmarkandu> anyone know this?
[03:56] <praveenmarkandu> from what I understand FFMPEG produces a HLS compatible format, not exactly HLS itself
[04:51] <DooMMasteR> is there a way to use the parametric stereo feature of libaacplus via ffmpeg?
[04:51] <DooMMasteR> and why is this mode not supported? libaacplus: bad aac setting: br:32000, AACch:1, AACsr:24000
[05:07] <DooMMasteR> using anything else aside from -ac 2 -ar 44100 will break
[06:26] <DooMMasteR> moin
[06:49] <spyro> how do I reduce the bitrate of an m4v, but preserve the audio as-is and keep the same codec for the video?
[06:49] <spyro> also, how do I determine the bitrate of the original version?
[07:07] <djdduty> Hey guys
[07:08] <djdduty> I am trying to screentshot capture only a region of my screen every x seconds, so I made a script that scrot's, then uses that image as input for ffmpeg, but I cannot figure out how to define the xoffset, yoffset, and width/height without using x11grab
[07:08] <djdduty> but if I use x11grab it turns into a video capture.
[07:23] <spyro> how do I encode using h264?  I've tried both h264 and x264 but both give me "unknown encoder"
[07:24] <djdduty> I got it.
[07:24] <djdduty> I just did an x11grab, but using -t 1 it worked fine for my purpose
[07:24] <djdduty> I am worried of cpu/ram usage though
[07:24] <djdduty> x11grab is very cpu intensive
[07:29] <spyro> ouch
[07:29] <spyro> libx264 segfault
[07:42] <LithosLaptop> spyro: libx264
[07:42] <LithosLaptop> oh nm
[07:43] <spyro> yeah, I tried it and I got a segfault
[07:54] <spyro> wtf?
[08:00] <fling> so tell me about autocrop someone please
[08:01] <fling> I want so autocrop videostream while remuxing into mkv without any reencoding
[08:13] <DooMMasteR> fling: how would you change the video content without reencoding?
[08:14] <fling> DooMMasteR: mkv has a crop feature in addition to aspect ratio
[08:14] <DooMMasteR> yeah some containers
[08:14] <fling> DooMMasteR: and a lot of other fancy things
[08:14] <DooMMasteR> and then also just some players :P
[08:14] <fling> DooMMasteR: ok, so I want to autocrop somehow, I'm tired of black borders :D
[08:15] <DooMMasteR> then my hint would go towards mkvtools
[08:15] <fling> yeah
[08:15] <fling> woot is next? I want it to be automagical
[08:15] <fling> is not possible?
[08:15] <DooMMasteR> hmm
[08:16] <fling> should I create some script?
[08:16] <DooMMasteR> mkvtools offer only manual crop tagging, I see
[08:16] <fling> that will detect first and mlvtool second?
[08:16] <fling> s/mlv/mkv/
[08:16] <DooMMasteR> yeah I think that would be the way to go
[08:17] <DooMMasteR> and then& who encodes with bars&
[08:17] <fling> how to detect? :P
[08:17] <fling> these people are strange
[08:19] <DooMMasteR> mencoder hat cropdetect back in the days :P
[08:19] <DooMMasteR> and still
[08:20] <DooMMasteR> fling: ffplay -vf cropdetect
[08:20] <fling> DooMMasteR: thanks!
[08:21] <fling> should not I reencode if I'm only removing black borders?
[08:21] <fling> and fixing aspect ratio
[08:21] <DooMMasteR> fling: that will only give you the crop values
[08:21] <fling> I know
[08:22] <DooMMasteR> you can then feed them into mkvtools :)
[08:22] <DooMMasteR> and it will be fine with any player supporting the feature
[08:22] <fling> DooMMasteR: I need to decide, mkvtools or reencode :P
[08:22] <fling> idk what is better
[08:22] <fling> hmm hmm
[08:22] <DooMMasteR> check if your players support the mkv feature
[08:22] <fling> is it supported in s/mplayer2?
[08:22] <fling> ok
[08:25] <fling> [cropdetect @ 0x7f6c04001010] x1:40 x2:1239 y1:0 y2:719 w:1200 h:720 x:40 y:0 pos:11379201 pts:1669668 t:27.827800 crop=1200:720:40:0
[08:25] <fling> is not it working without DISPLAY variable?
[08:25] <fling> I want to run it without X
[08:59] <kaffeebohne> Hi. I wanted to record my desktop using ffmpeg but somehow, it wont work (ubuntu 13.04 64 bit) http://paste42.de/5328/&linenr
[09:07] <tckb> Any experiences using ffserver?
[09:20] <DooMMasteR> lol
[09:20] <DooMMasteR> only bad experience :P
[09:20] <DooMMasteR> I usually stream from ffmpeg to VLC via local UDP :P
[10:05] <jsea> I'm developing an embedded linux mobile device and want to play mp4 directly on its frame buffer (it doesn't have x11) but "ffplay -i [source]" complains "Could not initialize SDL - Unable to open mouse (Did you set the DISPLAY variable?)". Could I get any advice??
[10:07] <theholyduck> jsea, i know mplayer has support for using some pc framebuffers atleast
[10:07] <theholyduck> you might want to start looking there
[10:08] <jsea> theholyduck: cool, i'll look at that right now. thanks a lot!
[10:33] <jsea> theholyduck; I just succeeded to play mp4 on fb using mplayer, thanks again.
[10:35] <jsea> theholyduck; it looks like using ffmpeg as decoder and my tiny problem is that it doesn't give me a colorspace converter to change yuv420p to bgra so its color is weird now
[10:39] <relaxed> jsea: look at mplayer's "format" video filter
[12:36] <alpha_one_x86> Hello, I try encode video to mp4, but it's very dirty, who can help me?
[12:37] <spaam> only if you tell us what you tried with
[12:38] <alpha_one_x86> ffmpeg -i origin.mp4 -vcodec libx264 -vprofile high -preset slow -b:v 500k -maxrate 500k -bufsize 1000k -threads 4 -b:a 128k  10014.mp4
[12:43] <spaam> alpha_one_x86: dirty how?
[12:43] <spaam> alpha_one_x86: sure. 500k might be low for 720p things
[12:45] <alpha_one_x86> it's for 768x432, http://chatounix.com/stp/testflv/flv3.php (warning: adult content) -> give square effect
[12:56] <zap0> flv is over.  consider mp4
[12:57] <alpha_one_x86> zap0: it's mp4 used
[13:36] <jimi_> Can someone help with this error? trying to converg OGV to MPEG http://pastie.org/7598650
[13:36] <jimi_> convert ^
[13:50] <JEEB> line 22 is your answer jimi_
[13:50] <jimi_> oh
[14:02] <newbi22> anybody here
[14:03] <newbi22> i need a little help
[14:03] <newbi22> anybody?
[14:07] <newbi22> plz somebody help me
[14:10] <newbi22> hi
[14:12] <newbi22> hello?
[14:19] <newbi22> what is diffrence between PCM and PCM Planar ?
[14:20] <JEEB> you mean between interleaved and planar?
[14:20] <JEEB> interleaved has the data of different channels in one "line" after another in the data. For example with stereo L->R->L->R->...
[14:21] <newbi22> is that pcm_s16le  interleaved ?
[14:21] <newbi22> and?
[14:21] <durandal_1707> pcm_s16le is not planar
[14:22] <JEEB> planar has the data in multiple planes: so you'd basically have one line of memory for each channel
[14:23] <newbi22> it means ... [                            L                      ]  [                       R                      ]   ?
[14:24] <JEEB> two separate memory planes for stereo, for example, yes
[14:24] <newbi22> thank you very much
[14:25] <newbi22> one more dumb question..
[14:27] <newbi22> i wanna convert (AV_SAMPLE_FMT_S16P)  to (AV_SAMPLE_FMT_S16).
[14:27] <JEEB> with ffmpeg you then use libswresample
[14:29] <newbi22> Is it impossble that just simply copy the block?
[14:29] <durandal_1707> to where?
[14:30] <newbi22> to output file..
[14:31] <newbi22> i think there are no planar fmt  old version
[14:32] <durandal_1707> and where you gonna use that file?
[14:33] <newbi22> my program use it  to calc
[14:34] <newbi22> in pcm_s16le format..
[14:36] <newbi22> old version always gives me that in  pcm_s16le. even if input was planar
[14:36] <newbi22> there are no fmt  about planar
[14:36] <newbi22> but newest is not.
[14:37] <JEEB> just use libswresample to convert from planar to interleaved, that most probably will be simpler than you making your own interleaver
[14:37] <newbi22> ok i see
[14:37] <newbi22> thanks very much
[14:58] <UnknownzD> hey everyone
[14:59] <UnknownzD> can I ask a  question regarding mp4 conversion here?
[14:59] <UnknownzD> I was trying to convert a video from wmv format to mp4 format
[15:00] <durandal_1707> and?
[15:00] <UnknownzD> and this is the detail regarding the conversion
[15:00] <UnknownzD> Stream mapping:
[15:00] <UnknownzD>   Stream #0:1 -> #0:0 (vc1 -> libx264)
[15:00] <UnknownzD>   Stream #0:0 -> #0:1 (wmapro -> libfaac)
[15:00] <UnknownzD> problem is that the program crashed
[15:00] <UnknownzD> just at the start point
[15:00] <UnknownzD> Multiple frames in a packet from stream 0
[15:00] <UnknownzD> :\ k
[15:01] <UnknownzD> https://privatepaste.com/2063d605f6
[15:01] <UnknownzD> was compiled with latest snapshot of x264
[15:02] <UnknownzD> and faac 1.28
[15:02] <UnknownzD> I don't know if it is a problem of MSYS / MinGW or not
[15:02] <UnknownzD> and I can't recompile the whole thing in debian atm
[15:03] <relaxed> UnknownzD: try the windows build listed on the download page
[15:06] <UnknownzD> ok still showing Multiple frames in a packet from stream 0 this time
[15:06] <UnknownzD> but it is not crashing any more
[15:06] <UnknownzD> so whats the reason?
[15:07] <UnknownzD> https://privatepaste.com/d46c27eb2f
[15:13] <UnknownzD> ok relaxed I have found that the program only crashes
[15:13] <UnknownzD> when libfaac is used
[15:13] <UnknownzD> so should I report it?
[15:14] <UnknownzD> because when I try to do the same thing on another computer in my home
[15:14] <UnknownzD> it automatically selected libfaac instead of the libvo_aacenc
[15:15] <UnknownzD> actually I didn't choose to use libvo_aacenc at all on this box :\
[15:51] <MozartsGhost> Hi Guys, I have a few questions, I hope someone is interested to lend some ideas. I want to switch over to using FFmpeg. I currently stream about 80 tv channels from DVB-s cards into VLC and then from vlc to WowZa with rtp. With vlc I specify 3 duplicate outputs, two rtp outputs going unicast to my two Wowza servers, and one http access output for testing and stream analysis.
[15:51] <MozartsGhost> 1. I have a stream analyzer analyzing video and doing silence detection, it connects to the vlc encoder directly with http
[15:51] <MozartsGhost>    and grabs the stream and does its thing..
[15:51] <MozartsGhost> 2. What is the best way to setup FFmpeg to be able to accept an http connections from my stream analyzer ?
[15:51] <MozartsGhost>    can it do it directly as an ouput ? .. or do I have to use some intermediary software for it ?
[15:51] <MozartsGhost> basically I have been testing with FFmpeg, taking DVB-S input and piping it to ffmpeg and then streaming it to wowza via udp, and it seems to be doing a MUCH MUCH better job, no artifacting or twitching, VLC is driving me CRAZY and I want to switch all my encoders over to FFmpeg
[15:51] <MozartsGhost> what would you suggest ?
[15:52] <Sashmo_> Hey guys, is there a way to add an external time sync/clock to ffmpeg encoding?
[15:52] <MozartsGhost> oh, using dvbstream output and pipe'ing it to ffmpeg as input, if u were wondering.
[15:54] <Sashmo_> MozartsGhost: for me?
[15:54] <MozartsGhost> oh, sorry Sashmo, no :) that was part of a question I asked about two seconds before you joined ;)
[15:54] <Sashmo_> MozartsGhost: thanks
[15:55] <durandal_1707> MozartsGhost: ffmpeg supports http ...
[15:56] <MozartsGhost> durandal_1707: thanks, yes, but I have only seen examples of it sending TO an server .. is it also possible for ffmpeg to accept connections and act as an http server ?
[15:58] <Oele> hi.. does anyone know which program is using "Lavf52.111.0" as its http user agent?
[15:59] <ubitux> any program using lavf to fetch data from your server
[15:59] <ubitux> but an old one though
[16:00] <Oele> hm... so any program that uses lavf identifies itself as 'lavf'? :(
[16:01] <durandal_1707> MozartsGhost: yes, see ffserver
[16:02] <MozartsGhost> durandal_1707: ok, thanks ;) I was going to use ffserver .. just wanted to see if it could send directly
[17:16] <MozartsGhost> is anyone here familiar with the -f tee muxer ?
[17:22] <ubitux> the doc is not enough?
[17:43] <MozartsGhost> @ubitux: yea, eventually found the right port of the documentation .. lol, asked a bit pre-maturely ;|
[17:43] <Sashmo_> Hey guys, is there a way to add an external time sync/clock to ffmpeg encoding?
[17:45] <divVerent> Fun abuse: get this file: http://rm.sudo.rm-f.org/~xonotic/spectrum.mp3, then run this command:
[17:45] <divVerent> ffplay -f lavfi "amovie=spectrum.mp3, asplit [out0] [x], [x] showspectrum=size=640x256:slide=1 [out1]"
[17:45] <Fjorgynn> :o
[17:46] <MozartsGhost> now only struggling to get get ffserver to actually accept connections from my ffmpeg, ffmpeg keeps exiting with "connection refused" even tho ffserver gets the POST from it, with no error
[17:46] <MozartsGhost> abit of teething problems I guess ..
[17:46] <divVerent> better command: ffplay -f lavfi 'amovie=http\\://rm.sudo.rm-f.org/~xonotic/spectrum.mp3, asplit [out0] [x], [x] showspectrum=size=640x256:slide=1 [out1]'
[17:46] <ubitux> divVerent: haha nice :)
[17:46] <divVerent> hehe
[17:46] <divVerent> wrote a tool to edit spectrum that way ;)
[17:47] <ubitux> nice :)
[17:47] <divVerent> REALLY should have gotten that idea three weeks earlier ;)
[17:47] <divVerent> then this command could have been used as "ffplay now shows lyrics"
[17:47] <ubitux> divVerent: can you try to a large energy instead of zero?
[17:47] <divVerent> that sounds bad
[17:48] <divVerent> could, but my program doesn't do that yet
[17:48] <divVerent> it is basically an image-driven EQ
[17:48] <divVerent> for similar effect, I could though run my program with white noise as input, then overlay ;)
[17:48] <ubitux> i've seen various exploitation of this already, but this one is nice
[17:48] <divVerent> this one is nice because it's sneaky
[17:48] <divVerent> the others I know are all additive, and thus appear as weird noise
[17:48] <divVerent> like that Aphex face
[17:49] <ubitux> aphex music is all weird noise so it's ok
[17:49] <divVerent> true :P
[17:49] <ubitux> i've seen some text already
[17:49] <divVerent> yes
[17:49] <ubitux> i don't remember how it was affecting the sound
[17:49] <divVerent> http://rm.sudo.rm-f.org/~xonotic/temp/what.wav is BTW a typical result when using white noise and a mostly-black image
[17:49] <divVerent> the tool I have seen is basically many sine synths, one for each pixel row
[17:50] <ubitux> hehe
[17:50] <ubitux> anyway, so you say using high freq would sound bad?
[17:50] <ubitux> high energy*
[17:50] <divVerent> yes, would annoy me personally :P
[17:50] <divVerent> I also hate those feeping power supplies
[17:50] <divVerent> and putting text as high energy in high freqs would do the same
[17:51] <divVerent> on the other hand, losing some of the higher freqs isn't as noticeable
[17:52] <ubitux> it looks better with cbrt btw :)
[17:52] <ubitux> ./ffplay -f lavfi "amovie=spectrum.mp3, asplit [out0] [x], [x] showspectrum=size=640x256:slide=1:scale=cbrt [out1]"
[17:52] <ubitux> color=intensity doesn't help though
[17:53] <ubitux> with saturation we might make it better
[17:55] <durandal_1707> hah, that would be nice audio filter ...
[17:57] <divVerent> ffplay -f lavfi 'amovie=http\\://rm.sudo.rm-f.org/~xonotic/spectrum2.mp3, asplit [out0] [x], [x] showspectrum=size=640x256:slide=1 [out1]'
[17:57] <divVerent> additive version :P
[17:57] <divVerent> argh, why not
[17:58] <ubitux> you can drop the '[x], [x]' btw
[17:58] <divVerent> works now
[17:58] <divVerent> even for asplit? nice
[17:59] <divVerent> for this one I BTW used white noise as input to my tool, and audacity to merge them
[17:59] <ubitux> indeed we hear it
[18:00] <ubitux> divVerent: don't want to write a filter in ffmpeg to do that? :)
[18:00] <divVerent> where should it get the lyrics from?
[18:01] <ubitux> mmh i guess there is a text sync?
[18:01] <ubitux> or not at all?
[18:01] <ubitux> if not... as a text parameter
[18:02] <divVerent> oh, my tool is made for images :P
[18:03] <ubitux> what do you mean?
[18:03] <divVerent> it uses an image file as input
[18:03] <divVerent> not limited to text
[18:04] <ubitux> you send the text as bitmap?
[18:04] <divVerent> yes
[18:04] <ubitux> ow.
[18:05] <ubitux> well you could use freetype to rasterize
[18:05] <divVerent> http://rm.sudo.rm-f.org/~xonotic/temp/out.wav
[18:05] <divVerent> this too :P
[18:05] <ubitux> lavfi already links against it
[18:05] <ubitux> is your tool opensource?
[18:05] <divVerent> sort of, not released yet, but will be
[18:06] <ubitux> ok :)
[18:06] <ubitux> i'm interested :D
[18:06] <durandal_1707> well that filter would not be that useful for audiophiles ....
[18:06] <divVerent> hehe
[18:06] <divVerent> sure
[18:06] <ubitux> divVerent: removelogo loads a bitmap
[18:06] <divVerent> yes
[18:06] <ubitux> or you can use a video stream
[18:06] <ubitux> so you could do the same :)
[18:06] <divVerent> the catch here would be having to know the audio length in advance, to properly map the bitmap
[18:07] <ubitux> how does your tool usage looks like?
[18:07] <divVerent> sh fftimgeq.sh infile.wav image.png outfile.wav
[18:08] <ubitux> .sh... wat
[18:08] <divVerent> my script
[18:08] <divVerent> I will upload it soon
[18:08] <ubitux> please share when ready, it looks awesome :)
[18:08] <durandal_1707> you could use subtitle too...
[18:09] <durandal_1707> AS->A
[18:09] <ubitux> durandal_1707: wanna help with subtitles in lavfi? :)
[18:09] <durandal_1707> or much better: S->A
[18:09] <divVerent> https://github.com/divVerent/fftdiff
[18:09] <divVerent> and yes, it is a mess
[18:09] <divVerent> fftimgeq.c is the tool, it works on raw files though
[18:09] <divVerent> and the .sh is a wrapper to make it usable
[18:10] <divVerent> I derived it from "fftdiff" which makes some sort of difference between two audio files
[18:10] <ubitux> awesome :)
[18:10] <divVerent> take original and karaoke version, get voice only
[18:10] <divVerent> hardest part just now was removing the f-word from the license ;)
[18:11] <ubitux> you write awesome stuff divVerent :)
[18:11] <divVerent> and vocalizer.c is an equivalent tool to fftdiff I once wrote because I forgot where I had put fftdiff :P
[18:11] <divVerent> but then found it later again
[18:11] <divVerent> the history is mostly from an old cvs
[18:11] <divVerent> thus the "9 years ago"
[18:11] <durandal_1707> fftdiff? sound interesting
[18:12] <divVerent> too bad fftdiff has some bugs - mainly: output is shifted by about 0.4 sec
[18:12] <divVerent> I fixed the bugs in fftimgeq, need to eventually backport the fixes to fftdiff
[18:24] <divVerent> as for making a filter...let's see, maybe
[18:25] <divVerent> could do the redsizing with libswscale so that is no issue either
[18:28] <tlhiv_work> hi folks ... i'm having trouble getting ffmpeg to encode ... it seems to be complaining about my audio --> http://pastebin.tlhiv.org/XaQ5QBje
[18:30] <durandal_1707> thegeek: everything is written and explained in output
[18:30] <durandal_1707> thegeek: ^
[18:30] <durandal_1707> tlhiv_work: ^
[18:31] <tlhiv_work> i'm not seeing the problem
[18:31] <durandal_1707> tlhiv_work: [NULL @ 0x665b90] Codec is experimental but experimental codecs are not enabled, see -strict -2
[18:31] <divVerent> better make libfdk_aac work by installing it then recompiling ffmpeg
[18:32] <divVerent> the aac encoder is really not very good
[18:33] <tlhiv_work> i'm using libfaac on another machine and it seems fine ... i'm not sure how to incorporate that into this particular ffmpeg
[18:35] <dagerik> hey im trying to extract 150 seconds fro ma video here. using this; ffmpeg -ss 00:10:30.00 -i Scanners.mp4 -t 150 -c copy test.avi
[18:35] <dagerik> however when playing test.avi mplayer just chrashes with lots of FAAD: Failed to decode frame
[18:39] <Znurre> hi, I was wondering if there's any way to convert a batch of MP3 files to Ogg Vorbis and preserve the gapless playback information? basically, the tracks were ripped and encoded to MP3 using LAME, so the LAME tags for gapless playback is in the files, and I would like to transfer these to vorbis tags when encoding
[18:40] <durandal_1707> first transcoding is bad and ugly idea
[18:41] <Znurre> my mobile doesn't support gapless playback for MP3 files, but it does for Ogg Vorbis, hence the question
[18:41] <durandal_1707> iirc gapless playback information is usually stored as metadata
[18:42] <Znurre> I am aware that I will get a quality loss
[18:46] <durandal_1707> Znurre: how are gapless tags in mp3 named?
[18:46] <durandal_1707> and how in vorbis?
[18:46] <durandal_1707> then when transcoding, you would just need to make sure that tags are properly named ....
[18:47] <Znurre> MCDI and TLEN in MP3, not sure about Vorbis... I don't see any such tags in my gapless Ogg Vorbis files
[18:47] <Znurre> weird, I will have to look that up
[18:56] <Znurre> seems like in the ogg vorbis case it's built-in to the container rather than being a tag, not much info to find about the subject
[18:56] <Znurre> but I guess simply mapping tags wouldn't be a solution
[18:58] <durandal_1707> so why not try it and report outcome? and if it is not good, report feature request on bug tracker?
[18:58] <Znurre> I tried doing a simple find exec, the result was no gapless playback in the generated ogg files
[18:59] <Znurre> thought I would ask here first before filing a feature request, since I am no ffmpeg guru :)
[18:59] <durandal_1707> i see no point in coming here to ask if one can report feature request or bug reports ....
[19:00] <Znurre> since ffmpeg is such a competent software I figured there might be a way I did not know of
[19:00] <Znurre> which is why I came here, since this is the user support channel
[19:00] <durandal_1707> if you want such feature: report it on bug tracker or add it yourself or find someone to do it ....
[19:01] <Znurre> I will file a feature request for it though
[19:01] <Znurre> thanks for taking your time
[20:33] <divVerent> ubitux: know if this is intended?
[20:33] <divVerent> ffmpeg -i in.wav out.ogg
[20:34] <divVerent> will use flac, not vorbis, as codec
[20:34] <ubitux> yes
[20:34] <ubitux> lossless prefered
[20:34] <divVerent> this is intended?
[20:34] <divVerent> still wrong :P
[20:34] <divVerent> not DWIM at all
[20:34] <divVerent> normally it makes sense
[20:34] <ubitux> that's a generic behaviour
[20:34] <divVerent> for the special case of .ogg, not so much
[20:34] <divVerent> but for .m4a it chooses aac
[20:34] <divVerent> and not alac
[20:35] <ubitux> also i'm not sure we have a native vorbis encoder
[20:35] <ubitux> but well yeah that might be improved
[20:37] <slimjimflim> hi i'm running ffmpeg from drupal (phpvideotoolkit) and it keeps erroring out about having the wrong codec  can anyone help?  here's the raw cli output:  http://hpaste.org/85929
[20:38] <slimjimflim> ffmpeg version N-52119-gde656ea cloned from git yesterday
[20:40] <divVerent> [mp4 @ 0xadd5d60] track 0: could not find tag, codec not currently supported in container
[20:40] <divVerent> this sounds pretty critical
[20:40] <divVerent> does flv really go into mp4?
[20:40] <ubitux> flv into mp4? wat
[20:41] <slimjimflim> hm yea that's probably a bad idea
[20:41] <divVerent> yes, this message is from movenc.c, so it is the encoding
[20:41] <slimjimflim> i had errors when i tried to go from avi to flv too, but it was setup different
[20:41] <slimjimflim> let me retest and see if i can do it logically this time
[20:42] <divVerent> e.g. just try some, any all-capable container like mkv :P
[21:25] <ubitux> maruthi:
[21:25] <ubitux> - grep -i stereo3d configure
[21:25] <ubitux> stereo3d_filter_deps="gpl"
[21:25] <ubitux> you need to --enable-gpl if you want the filter
[21:26] <maruthi> trying with it...now..
[21:26] <maruthi> ubitux...thank u very much.... it worked
[21:27] <maruthi> i have one more doubt...
[21:28] <maruthi> some online links suggested options like -c:a  -b:v
[21:28] <maruthi> ffmpeg says unrecognized options !!
[21:29] <ubitux> what command line?
[21:31] <maruthi> ffmpeg -i sbs.mp4 -vf mp=stereo3d -acodec copy -threads 10 -b:v 10000k -preset ultrafast -vcodec libx264 ~/test2.mkv
[21:31] <ubitux> use pastebin with the failing command
[21:33] <maruthi> http://pastebin.com/3EHjVMwD
[21:36] <klaxa> >ffmpeg version 0.7.12, Copyright (c) 2000-2011 the FFmpeg developers
[21:36] <klaxa> that version is quite old
[21:36] <klaxa> command line options were different back then
[21:37] <maruthi> ok klaxa...i will try with the latest one
[21:38] <ubitux> maruthi: ./ffmpeg
[21:38] <ubitux> no need to install it, since you just built it
[21:39] <klaxa> ah heh right
[21:45] <maruthi> still building ...  :(
[21:47] <maruthi> ok built it ..this time new error reg. cmd line options
[21:48] <maruthi> http://pastebin.com/rwBDw8c7
[21:48] <maruthi> please tell  the equivalent cmd line option in link above as per latest ffmpeg
[22:02] <klaxa> not sure if this solves it, but try to put the -preset option behind -vcodec
[22:02] <klaxa> and behind libx264 of course
[22:02] <klaxa> so -vcodec libx264 -preset ultrafast
[22:03] <maruthi> i got it now.. preset is changed to pre now
[22:22] <Fjorgynn> good night
[22:41] <doub> I'm running ffmpeg to stream a webcam image to an ffserver process. I've got a lot of "rc buffer underflow", and the reported fps in the ffmpeg console is 7.5, while my command line asked for 30 (and the input stream is: Stream #0:0: Video: rawvideo (I420 / 0x30323449), yuv420p, 1280x720, 331776 kb/s, 30 fps, 30 tbr, 1000k tbn, 1000k tbc). What could be wrong?
[22:42] <doub> I've looked at my webcam lsusb, and it supports 720p at 7.5 and 10fps in raw, and 30fps in mjpeg. Is it possible ffmpeg failed to choose the right format?
[00:00] --- Wed Apr 17 2013