[Ffmpeg-devel-irc] ffmpeg.log.20130925

[burek]

[00:08] <BallsDeep> I am trying two different commands to transcode DTS to AC-3.
[00:08] <BallsDeep> Note:
[00:08] <BallsDeep> I know I can use one "ffmpeg" command to get the end result but I want to try the "dcadec" command.
[00:09] <BallsDeep> Commands I am using:
[00:09] <BallsDeep> dcadec -o float filename.dts > filename.raw
[00:09] <BallsDeep> ffmpeg -ar 48k -ac 6 -f f32le -i filename.raw filename.ac3
[00:09] <BallsDeep> The end result: The AC-3 track is created and it sounds like shit.
[00:11] <BallsDeep> Any ideas?
[00:38] <LithosLaptop> have you tried adding a bitrate parameter?
[00:38] <LithosLaptop> -ab 384k
[00:43] <BallsDeep> yes, 448k and 640k. same problem occurs (static in the background and the dialogue sounds like a bunch of chipmunks talking)
[00:44] <LithosLaptop> oh ok, maybe ffmpeg is incorrectly picking up the input file's sampling rate
[00:45] <LithosLaptop> oh
[00:45] <LithosLaptop> nm
[00:45] <LithosLaptop> I see you specify it
[00:46] <LithosLaptop> are you sure the raw file is sampled at 48k?
[00:48] <LithosLaptop> also make sure each sample really is f32le
[00:57] <BallsDeep> When I use the dcadec command, I don't know what "-o float" is creating for a raw file. I'm assuming "f32le". How would I know? After all, it's a raw file so how could I tell? Is there a piece of software out there that can tell me what a raw file contains?
[01:01] <Jan-> is the argument "-ac 2" still the preferred way to just dump two audio channels?
[01:01] <Jan-> Assuming we had four to begin with, that is?
[04:11] <Zutara> Hi. I'm trying to batch convert MKV's to AVI's but when I run the command it asks me if I want to replace the MKV's. What am i doing wrong? Pastebin: http://pastebin.com/t7rbBufY
[04:13] <sacarasc> avconv help is in #libav.
[04:16] <Zutara> I get the same output when I do ffmpeg.
[04:17] <sacarasc> Does it also say that it is deprecated? Because that's avconv too.
[04:17] <Zutara> Yeah.
[04:23] <Zutara> Why is it asking me to overwrite it?
[04:58] <grepper> Zutara: for this you should ask in #bash. You can't use wildcards the way you are using them.
[04:59] <grepper> you could use a loop: for vid in *.mkv; do  avconv "$vid" ...
[04:59] <grepper> anyway, neither an ffmpeg or avconv question really
[05:00] <Zutara> Okie dokie. :)
[05:10] <grepper> Zutara:  for vid in *.mkv; do  avconv "$vid" -vcodec copy -acodec copy -y "converted-$vid" ; done
[05:11] <Zutara> I had that sent to me in #libav. What do I put inside the quotes?
[05:11] <grepper> what quotes
[05:11] <Zutara> in the command you just sent.
[05:11] <Zutara> Or do I leave the quotes out/
[05:11] <Zutara> *?
[05:11] <klaxa> i suggest you read some stuff about bash
[05:11] <grepper> type it exactly as shown
[05:11] <grepper> heh
[05:12] <klaxa> $vid in this case is a variable
[05:12] <klaxa> which gets expanded to its value when interpreted by bash
[05:12] <klaxa> the values are the files that match *.mkv in terms of bash expansion
[05:13] <defaultro> how come, when I use -crf 0, the video gets some macro blocking. That's why I just always use -crf 15 but I'm not sure if this is the highest quality
[05:14] <klaxa> bad source?
[05:14] <klaxa> weird filters?
[05:14] <klaxa> :x
[05:14] <defaultro> k
[05:15] <klaxa> -crf 0 should produce lossless, at least if we are talking about libx264
[05:15] <klaxa> maybe ask in #x264 they might know more
[05:17] <defaultro> yup, but everytime i use -crf 0, video will definitely get macro blocking
[05:17] <defaultro> k
[05:20] <theholyduck> defaultro, well then, either something is wrong with your ffmpeg, or ffmpeg doesnt support it right
[05:20] <theholyduck> i dunno
[05:20] <theholyduck> i almost always use x264 directly
[05:20] <defaultro> got it
[05:20] <defaultro> i've never tried x264 directly
[05:20] <elBradford> HI! Looking for some assistance with compiling a static build of ffmpeg and librtmp for Android. Does anyone have experience in this area?
[05:20] <defaultro> i wasn't aware we can do that :)
[05:26] <elBradford> I just need some help with the ffmpeg build finding the rtmpdump library
[05:28] <defaultro> hey folks, is cqp an old option? my ffmpeg doesn't know it
[05:31] <klaxa> defaultro: try -qp 0 instead
[05:32] <defaultro> yup, it's running
[05:32] <defaultro> found it here, http://trac.ffmpeg.org/wiki/x264EncodingGuide#LosslessH.264
[05:32] <defaultro> does it matter if my extension is either mkv or mp4?
[05:32] <klaxa> it changes the container
[05:33] <defaultro> but quality is the same right?
[05:33] <klaxa> matroska is probably the most sophisticated container
[05:33] <defaultro> k
[05:33] <klaxa> yeah since the video stream is independent from the container
[05:33] <defaultro> what;s the reason for choosing a container?
[05:34] <klaxa> compatibility vs. features
[05:34] <defaultro> is it because of the number of supported video players?
[05:34] <defaultro> got it
[05:34] <defaultro> mp4 is famous
[05:34] <defaultro> and has more support. Am i right?
[05:34] <klaxa> mkv is too, you can put anything into mkv
[05:34] <defaultro> k
[05:34] <klaxa> i'd think so
[05:34] <defaultro> that's great info i learned tonight :)
[05:35] <defaultro> wow, my bitrate is 239932 kb/s, lol
[05:35] <klaxa> well it's lossless
[05:36] <klaxa> wait what?
[05:36] <defaultro> not sure how youtube will treat this kinds of upload
[05:36] <klaxa> 240 mbps?
[05:36] <defaultro> yes, it's lossless
[05:36] <defaultro> yeah
[05:36] <defaultro> that's what ffmpeg -i wrote
[05:36] <klaxa> well youtube should be able to deal with it
[05:36] <klaxa> because a) they use ffmpeg themselves
[05:36] <defaultro> k
[05:36] <klaxa> and b) i uploaded lossless videos and they could deal with it
[05:36] <defaultro> i just hate youtube because they reencode it again
[05:37] <defaultro> our video loses the nice quality of the image
[05:38] <elBradford> Is this the right room to ask questions regarding building ffmpeg?
[05:38] <defaultro> it is ok elBradford
[05:38] <defaultro> if someone knows, they will help you
[05:39] <elBradford> Thanks defaultro
[05:39] <defaultro> but remember in irc, you have to be patient :)
[05:39] <klaxa> elBradford: get the sources for librtmp, compile static, get sources for ffmpeg, compile static and link against static librtmp
[05:39] <defaultro> sometimes you won't even get an answer
[05:39] <klaxa> did that make sense to you?
[05:40] <elBradford> klaxa - it seems like ffmpeg is looking at pkg config for librtmp - if I include it using a -I flag, will that work?
[05:40] <defaultro> i'm reencoding a 60fps video and I'm using -qp 0. Bitrate is showing 487210 kbits/s
[05:40] <klaxa> do you have a static version of librtmp?
[05:41] <elBradford> Yeah, I can build that.
[05:41] <elBradford> I mean, I have built that already
[05:41] <klaxa> you will need static versions of all libraries you want to include
[05:42] <klaxa> did you read the compilation guide?
[05:42] <klaxa> https://trac.ffmpeg.org/wiki/CompilationGuide
[05:43] <elBradford> I'll read through that, I hadn't seen that. I'm doing it against the Android toolchain, but that will be helpful. I'll come back if I get stuck. Thanks Klaxa
[08:24] <yusari> Hi all. I'm looking for a CLI command to convert a 23.98fps mov file to 24fps without losing any quality. Can anyone help?
[08:25] <relaxed> why is that necessary?
[08:27] <yusari> Hi relaxed. It's for digital cinema conversion.
[08:30] <relaxed> what format does it need to be in?
[08:31] <yusari> input - 23.98fps mov, output 24fps mov
[08:34] <relaxed> ffmpeg -i input.mov -c:v libx264 -crf 15 -r 24 -c:a pcm_s16le output.mov
[08:36] <yusari> Thanks a lot, relaxed!
[08:36] <yusari> would the same command work if I wanted to convert a 25fps mov file to a 24fps mov file as well?
[08:37] <relaxed> yes, ffmpeg will duplicate or drop frames to achieve your desired frame rate.
[08:38] <yusari> and it will do so for both the image as well as audio?
[08:40] <relaxed> I'm not sure about the audio.
[08:40] <relaxed> It should maintain sync if that's your concern.
[08:46] <yusari> wouldnt the libx264 option re-encode the video, thus losing some of its quality?
[08:47] <yusari> my input file is apple prores 442, and I'd like to keep the output file the same
[08:48] <relaxed> you will have to re-encode to change the frame rate.
[08:50] <yusari> no way to change the frame rate without re-encoding, thus losing quality?
[08:50] <relaxed> No
[08:51] <relaxed> ffmpeg prores enocoder only supports yuv422p10le
[08:51] <relaxed> did you mean 422?
[08:52] <yusari> yes, im sorry
[08:52] <yusari> I meant 422
[08:53] <relaxed> then remove "-c:v libx264 -crf 15" and use -c:v prores
[08:53] <yusari> ffmpeg -i input.mov -c:v prores -r 24 -c:a pcm_s16le output.mov
[08:54] <yusari> so that would change the existing frame rate (whether its 23.98 or 25) to 24fps
[08:55] <yusari> wrt audio, I require 24bps for cinema assets, so I would use pcm_s24le instead, correct?
[08:55] <relaxed> yeah, there's more info prores here https://trac.ffmpeg.org/wiki/vfxEncodingGuide
[08:55] <relaxed> for audio try -c:a copy
[08:58] <yusari> would -c:a copy change the audio tempo as well tosync with the new image frame rate?
[08:58] <relaxed> no, it would copy the stream
[08:58] <yusari> or would you suggest I use something like Audacity for that?
[08:59] <yusari> right, so ffmpeg can't be used to change the tempo of the audio to match the changed frame rate of the image
[09:02] <relaxed> maybe there's a filter for it, check "ffmpeg -filters | less"
[09:11] <yusari> thers's an atempo switch to adjust tempo
[12:26] <aleray> hi, I have a bunch a hd MOV files which I would like to concatenate without reencoding (to then convert to webm or ogg). How can I do this? The machine I will use for that is a mac
[12:43] <aleray> I tried: ffmpeg -i "concat:input1.mpg|input2.mpg|input3.mpg" -c copy output.mpg but th equality is altered
[12:43] <aleray> the quality
[12:43] <aleray> I would like to not reencode it
[12:44] <viric> that can be done with very specific input formats, isn't it? I guess.
[12:47] <aleray> viric, it seems to be mpeg 2 inside the mov
[12:48] <viric> I don't know the details
[12:53] <aleray> viric, would an output of ffprobe or so help?
[12:56] <viric> aleray: no, I don't know the details of ffmpeg :)
[14:06] <chrrbss> can anyone help me with itoffset? I'm overlaying a bunch of images onto a video and want to offset the appearance of the image by xxx seconds
[14:06] <chrrbss> currently got: ffmpeg -i ./../../../sourcevideo/test.mov -f image2 -itoffset 0:0:20 -i %2d.png -s 1280x720 -strict -2 -filter_complex 'overlay=0:0;' combined.mp4
[14:10] <chrrbss> sorry; pastebin is here: http://pastebin.com/h7nMQhNT
[14:15] <aleray> back
[14:16] <chrrbss> and pastebin with output: http://pastebin.com/z7Jwxm1D
[14:16] <aleray> viric, I was away, I may have miss your answer if any
[14:16] <chrrbss> sorry, dont use irc that much :)
[14:20] <aleray> "ffmpeg -i file.mov" gives me http://dpaste.com/1395871
[14:20] <aleray> how can I do to concatenate several of those mov files
[14:20] <aleray> ?
[14:25] <viric> no answer
[14:29] <Bombo> aleray: i did that with mencoder once
[14:29] <Bombo> but i found this https://trac.ffmpeg.org/wiki/How%20to%20concatenate%20%28join,%20merge%29%20media%20files
[14:29] <aleray> Bombo, I tried but somehow the file quality is reduced
[14:29] <aleray> it seems to reencode the thing
[14:30] <Bombo> aleray: you tried what exactly?
[14:30] <Bombo> did you chose 'copy' as codec?
[14:30] <aleray> Bombo, ffmpeg -f concat -i 1.MOV 2.MOV -c copy output.mpg
[14:31] <aleray> Bombo, and "ffmpeg -i "concat|1.MOV|2.MOV" output.mpg"
[14:31] <Bombo> that would need to reencode from h264 to mpeg2
[14:31] <Bombo> try output.mov
[14:32] <aleray> Bombo, ok
[14:32] <aleray> the first command or the second?
[14:32] <aleray> boiled_sugar,
[14:32] <aleray> oups
[14:32] <Bombo> try both
[14:33] <Bombo> for lossless copy, stay with the same codec/container
[14:34] <aleray> Bombo, could you write down the exact command you would try?
[14:38] <Bombo> ffmpeg -i "concat:1.mov|2.mov" -c copy output.mov
[14:39] <Bombo> try this
[14:39] <aleray> Bombo,
[14:39] <aleray> ok
[14:39] <aleray> thanks
[14:39] <Bombo> does it work? i didn't try ;)
[14:40] <aleray> Bombo, I think this is what I'm running now, but it takes time so I suspect it reencodes the video
[14:40] <aleray> I'm waiting for it to be done to see how it looks like
[14:40] <aleray> 5 minutes to go or so
[14:41] <Bombo> or this one with demuxing: ffmpeg -f concat -i <(find . -name '*.mov' -printf "file '%p'\n") -c copy output.mov
[15:11] <aleray> Bombo,seems to be working with ~/bin/ffmpeg -f concat -i files.txt -c copy output6.mov
[15:12] <chrrbss> in case anyone is interested; i found the solution by moving itoffset in front of the input of the images like so: http://pastebin.com/ZedxVnY3
[15:23] <aleray> Bombo, it turned out that some lecture we have recorded overlaps (end of lecture A and beginning of lecrue B on the same file). Is it possible to concat only part of a file?
[15:43] <ayaka> I want to embed a live stream into web, could ffsever do it and which output format shall I choice
[15:45] <Bombo> aleray: not sure, but you could cut the movs before cat
[15:47] <Bombo> aleray: copy from timestamp to timestamp... with -ss 00:00:00 (start time) and -t 00:01:00 (duration, 1min)
[18:44] <version2beta> I am having issues getting closed captions (608/708) out of ffmpeg using ccextractor when the filter chain includes yadif. Everything works fine when yadif isn't in the chain, but with it, ccextractor complains about "Error: Broken AVC stream - forbidden_zero_bit not zero ..."
[18:46] <version2beta> my ffmpeg command is basically `ffmpeg -i udp://239.10.10.1:1234 -an -sn -map 0:v -c copy -f mpegts pipe:3 -map 0:v -vf yadif=0:-1:1,scale=720:-1 -c:v mjpeg -r 1 -f image2pipe pipe:1`
[18:47] <version2beta> Both pipes are streamed into Nodejs where code either pipes it back out to ccextractor (pipe:3) or splits out jpegs and sends them off to a better life.
[18:48] <version2beta> btw, the error occurs only about once every three minutes, but consistently with yadif and not yet once without yadif.
[19:35] <Niatross> I want to transcode a DTS track to AC-3. I want to do it in three steps.
[19:35] <Niatross> BTW: I know you can do a DTS to AC-3 transcode with one "ffmpeg" command, but I want to break it down into 3 steps (DTS to RAW&&.RAW to WAV&&.WAV to AC-3)
[19:36] <Niatross> Are the following steps correct:
[19:36] <Niatross> 1 ) DTS to RAW:
[19:36] <Niatross> ffmpeg -i input.dts -f f32le output.raw
[19:36] <Niatross> 2) RAW to WAV:
[19:36] <Niatross> ffmpeg -ar 48k -ac 6 -f f32le -i input.raw -acodec pcm_f32le output.wav
[19:36] <Niatross> or
[19:37] <Niatross> ffmpeg -ar 48k -ac 6 -f f32le -i input.raw -acodec pcm_s32le output.wav
[19:37] <Niatross> 3) WAV to AC-3:
[19:37] <Niatross> ffmpeg -i input.wav -ab 640k output.ac3
[19:37] <Niatross> My questions are:
[19:37] <Niatross> Which command in Step 2 should I use (should I keep it floating-point all the way until it reaches the AC-3 encoder or should I switch it to signed-integer)?
[19:37] <Niatross> Should I keep every input and output 32-bit (instead of 16-bit) before it reaches the AC-3 encoder?
[20:45] <Niatross> anyone?
[20:48] <durandal_1707> Niatross: keep it float
[20:49] <durandal_1707> ffmpeg -h encoder=ac3
[20:50] <iive> btw, why raw? wav pcm IS raw.
[20:51] <Niatross> i want no header on it
[21:11] <Niatross> member:durandal_1707 On step 2, should I leave "-acodec pcm_f32le" or let it default to 16 bit? What I mean by this is: Should I keep every input and output 32-bit before it reaches the AC-3 encoder?)?
[21:13] <durandal_1707> Niatross: converting float to int is pointless as int would be converted back to float
[21:13] <durandal_1707> same apply to 32bit to non32bit
[21:15] <durandal_1707> if you do not care at all about audio quality and transcoding speed (the pointless 3 commands that already points out) you can do whatever you like
[21:15] <Niatross> ok, so I'll leave the  "-acodec pcm_f32le", so it doesn't default to 16-bit.
[21:15] <durandal_1707> if you want
[21:16] <Niatross> well maybe it wouldn't default to 16 bit, since it will take the "-f f32le" from the raw input
[21:18] <durandal_1707> nope it will default to 16
[21:19] <Niatross> ok, so it's best to leave the "-acodec pcm_f32le" to keep it 32-bit until the ac-3 encoder gets a hold of it
[21:20] <durandal_1707> yes
[21:20] <Niatross> thanks for helping me!!!!!!!!!!!!
[22:14] <kolosos> http://pastebin.com/2A5ardCU   "buffer underflow" error in converting youtube[480].flv video to ntsc-dvd .mpg
[22:14] <kolosos> Any tips?
[22:19] <kolosos> (h264 -> mpeg2video)   (aac -> ac3)     [dvd @ 0x9b3720]   buffer underflow i=1 bufi=1550 size=1792
[22:20] <viric> kolosos: it can be a pts problem
[22:20] <viric> is it really a so simple conversion?
[22:20] <kolosos> viric: http://pastebin.com/2A5ardCU
[22:21] <viric> ahm looks like so, yes
[22:21] <viric> no idea.
[22:23] <kolosos> viric, thanks for the help. i'll try a different youtube video & see what happens
[22:25] <kolosos> going to try jump480.webm instead of jump480.flv
[22:27] <JEEB> kolosos, buffer underflows mean that the mpeg-2 encoder wasn't able to keep the rate within the constraints set
[22:29] <JEEB> at least that's what VBV underflow means, not exactly sure if this is the same thing :P
[22:35] <kolosos> JEEB - isn't -target NTSC-DVD supposed to take care of that?
[22:36] <JEEB> it should just set the limit
[22:36] <JEEB> if the encoder cannot cope with it...
[22:36] <JEEB> :P
[22:38] <JEEB> ugh
[22:38] <JEEB> mpegenc.c
[22:38] <JEEB> it's a weird mix of things
[22:40] <kolosos> JEEB: for instance, the .flv video was AAC at 44100 Hz  and the .mpg output was AC2 at 48000 Hz.  Is ffmpeg telling me that it can't figure out how to upsample?
[22:40] <JEEB> no
[22:40] <JEEB> the warning is coming from mpegenc.c
[22:40] <JEEB> or wait... mpegenc.c is from libavformat
[22:40] <JEEB> asdf
[22:40] <JEEB> it's a muxer
[22:41] <JEEB> so it's saying that a buffer underflow happens during muxing
[22:42] <JEEB> but I'm still not sure if it's saying that packets are too big or too small...
[22:42] <JEEB>             if(stream->buffer_index < pkt_desc->size ||
[22:42] <JEEB>                stream->predecode_packet == stream->premux_packet){
[22:42] <JEEB> buffer index is smaller than the size of the packet?
[22:42] <JEEB> or
[22:42] <JEEB> predecode_packet equals premux_packet
[22:43] <JEEB> sorry I can't make heads or tails of this :P
[22:43] <JEEB> but if it doesn't error out
[22:43] <JEEB> I guess it's OK?
[22:43] <kolosos> np, i was hoping for an answer like "you are doing something totally idiotic, use the -XYZ switch"
[22:45] <kolosos> next I'm going to try the 720p MP4 encoded youtube vid
[22:46] <JEEB> anyways, the actual encode seems to be going on OK, but the muxer then is telling that there are buffer underflows
[22:46] <JEEB> muxer puts streams into a container (in this case, MPEG-PS), encoders take in raw data and create streams
[22:49] <kolosos> guessing that it's the 44100 vs 48000 audio, can i create a -target NTSC-DVD with an AC2 rate of 44100 s/s ?
[22:49] <JEEB> it's not that
[22:50] <JEEB> the audio rate has to be 48kHz in DVDs
[22:50] <kolosos> ah ok
[22:50] <kolosos> so you think it's a real software issue
[22:50] <kolosos> some internal buffer not being ready
[22:50] <JEEB> no
[22:50] <kolosos> ^_^
[22:51] <kolosos> the "muxer" is not happy... why
[22:51] <JEEB> I don't know what exactly it's trying to say, but it doesn't happen in the video/audio encoders, it's just warning you that the packetizing (?) buffer is either underused or overused, I just can't read from the code which it is :P
[22:51] <JEEB> also as far as I can see, it's not an error
[22:51] <JEEB> just a warning
[22:51] <kolosos> OH
[22:51] <kolosos> OOOOH
[22:52] <kolosos> ok let me see if i can play the output...
[22:52] <JEEB> it should play just fine methinks
[22:54] <kolosos> ffprobe jump480m.mpg looks good :
[22:54] <kolosos> mpeg2video yuv420p, 720x480 [SAR 32:27 DAR 16:9], 29.97 fps, 29.97 tbr, 90k tbn, 59.94 tbc  Audio: ac3, 48000 Hz, stereo, fltp, 448 kb/s
[22:56] <JEEB> also, d'oh... it is an AV_LOG_ERROR :/ then why did you get multiple of them...
[22:56] Action: JEEB pokes the -devel channel
[22:57] <kolosos> mpeg2video 720x480 plays perfectly in totem
[22:57] <JEEB> yes, the encoding doesn't fail as such, the video and audio should be just fine and the packets written should be just fine
[22:59] <kolosos> ok, in that case let me try do dvdauthor from the mpeg
[22:59] <JEEB> but if it's an error then some packets didn't get written... or what the flying fuck happened
[22:59] <JEEB> join #ffmpeg-devel to see if I get a response
[22:59] <JEEB> I'm going to sleep soon :P
[22:59] <JEEB> <JEEB> avformat/mpegenc.c:remove_decoded_packets() -- what does the buffer underflow error actually mean?
[22:59] <JEEB> <JEEB> is it a VBV underflow, aka bits are going out too fast? or something else
[23:00] <JEEB> I asked this
[23:02] <kolosos> dvdauthor seems to agree: lots of    WARN: Discontinuity of 71 in audio channel 0; please remultiplex input.
[23:03] <JEEB> there's basically a counter in MPEG-TS (and I think in MPEG-PS as well)
[23:03] <JEEB> in each packet
[23:03] <JEEB> so that just means that there was a discontinuity of X packets in there :P
[23:03] <kolosos> ah, so smart software can see there are samples missing
[23:04] <JEEB> In MPEG-TS it was just 0-15 so not sure if you could count a discontinuity of 71 :D
[23:05] <kolosos> but, again, just a warning (i hope) so i SHOULD have a playable DVD. any idea how to test that in fedora?
[23:05] <JEEB> no, as I just said it is AV_LOG_ERROR
[23:05] <JEEB> I thought it was a warning because you got it multiple times
[23:06] Action: kolosos doesn't know what AV_LOG_ERROR is and how do you know that?
[23:06] <JEEB>                 av_log(ctx, AV_LOG_ERROR,
[23:06] <JEEB>                        "buffer underflow i=%d bufi=%d size=%d\n",
[23:06] <JEEB>                        i, stream->buffer_index, pkt_desc->size);
[23:06] <JEEB> that's the message you're getting
[23:06] <JEEB> although it only breaks from that loop
[23:07] <JEEB> it doesn't return nonzero or whatever
[23:07] <JEEB> argh
[23:07] <JEEB> anyways, join #ffmpeg-devel and see if anyone answers what I asked and possibly you then ask extra questions :P
[23:13] <kolosos> JEEB: thanks for help
[23:35] <kolosos> JEEB - it seems to be related to the input file - I tried another youtube vid which did not have the same problem
[23:35] <JEEBsv> uhh
[23:35] <JEEBsv> after you decode the input
[23:35] <JEEBsv> it's all raw data in both cases
[00:00] --- Thu Sep 26 2013