[Ffmpeg-devel-irc] ffmpeg.log.20130319

[burek]

[00:05] <f0x> Hi there, is there anyone with experience in piping output from RTMPDump -> FFMpeg using pipe:0 ?
[00:48] <iwanttobefreak> Hello! I try all the ways to add subtitles to a h264 file but i can't :(
[00:54] <klaxa> well that's because h264 is a video codec... what container are you trying to use?
[00:59] <exco> I'm trying to get some h264 stream from a webcam "restreamed" to a provider like ustream. It sort of works but it eventually dies after a while. http://pastebin.com/D6h6FvJE
[00:59] <exco> anything I can do so the missed rtp packets won't kill the stream?
[01:03] <exco> other project: I'm trying to restream an mjpeg stream to e.g. ustream and because not a lot is happening on the stream - (i-frames?) the stream stays black for about 15-30s until a picture appears.  ffmpeg foo looks like this atm: ffmpeg -re -f mjpeg -i "$INURL" -s "$INRES" -r 25 -vcodec libx264 -r "$FPS" -preset ultrafast -crf 26 -threads 0 -s "$OUTRES" -f flv "$URL"
[01:03] <exco> stream is here http://www.ustream.tv/channel/s21-instacam
[01:06] <iwanttobefreak> klaxa, with mkv
[01:06] <klaxa> have you tried mkvmerge?
[01:06] <iwanttobefreak> yes
[01:06] <iwanttobefreak> and the error is "Malformed codec id a_aac"
[01:07] <klaxa> mmhh that doesn't look like a subtitle issue
[01:07] <iwanttobefreak> with XVID and ffmpeg I can put subtitles
[01:08] <iwanttobefreak> I have mkvmerge in spanish
[01:10] <iwanttobefreak> http://pastebin.com/sDFRthd0
[01:11] <iwanttobefreak> http://pastebin.com/czrY3iGp
[01:11] <iwanttobefreak> Or how can I convert h264 to xvid?
[01:29] <f0x> Hi there, is there anyone with experience in piping output from RTMPDump (livestream) -> FFMpeg using pipe:0 ?
[01:51] <f0x> the saddest part is when you've been trying so many different things on server/client side that eventually it works and you just dont know why.
[01:51] <f0x> :)
[01:53] <klaxa> rtmpdump <whatever options> | ffmpeg -i - <whatever optoins> out-file ?
[01:53] <klaxa> that doesn't work?
[01:53] <f0x> i had an issue with reading from pipe:0
[01:53] <f0x> just kept on saying invalid
[01:53] <klaxa> argss
[01:54] <klaxa> sorry!
[01:54] <klaxa> stupid 1337sp34k in my brain
[01:54] <f0x> Its sorted now thought
[01:54] <f0x> -t
[01:54] <f0x> also, it was ffmpeg to output to a RTMP server
[01:55] <f0x> so it was probably my mistake somewhere not escaping correctly or a flag I had missed somewhere
[01:56] <foonix> klaxa: np ;)
[01:56] <iwanttobefreak> klaxa, finally I change audio codec to ogg and it works!
[01:56] <iwanttobefreak> avconv -i 3x01.mp4 -vcodec copy -acodec libvorbis  3x01.mkv
[01:57] <iwanttobefreak> thanks!
[01:57] <iwanttobefreak> bye !
[02:18] <grepper>  would this be better than using "-vcodec copy -f null" to quickly get duration?  Assuming it does decode the stream ... "ffmpeg -i out.m2v -vcodec rawvideo -an -y -f rawvideo /dev/null"
[02:19] <klaxa> lol dunno do some tests maybe?
[02:22] <grepper> seems to work on a few, not sure about vbr files yet
[02:26] <grepper> hm, biggest issue right now seems to be avconv misidentifying framerate of m2v  files and getting duration wrong.  ffmpeg does it correctly.  Unless there is something messed up about the m2v files I'm creating.
[02:26] Action: grepper starting to wish he didn't decide to support both
[02:26] <llogan> drop support for avconv. done.
[02:28] <llogan> why not just "ffmpeg -i input" or use ffprobe to get duration?
[02:44] <grepper> llogan: well, I need something that is more frame accurate
[02:45] <grepper> those just guess based on reading a certain number of bytes of the file afaik
[09:13] <Soho> hi all
[09:13] <Soho> when I try to use vlc to connect ffserver via WiFi interface
[09:14] <Soho> ffserver says "invalid stream index" every time
[09:14] <Soho> does anyone have idea about this issue?
[09:14] <Soho> but when I try to connect via ethernet interafce,
[09:14] <Soho> it is ok
[09:15] <Soho> is it cause by wifi environment?
[09:15] <Soho> any idea?
[11:07] <divVerent> I need a bit help in order to debug a playback problem
[11:07] <divVerent> (on the N900)
[11:07] <divVerent> I made these files using ffmpeg: https://gist.github.com/5194949
[11:07] <divVerent> all by the same command lines
[11:07] <divVerent> but, the top three don't play on the N900, the others do
[11:08] <divVerent> anyone have any ideas what ELSE could it be that makes the player accept/reject files?
[11:08] <divVerent> the N900 is a bit "special" in that it claims to only be able to decode Level 3.0, which is slightly below the display resolution... so I did use a larger resolution, which works for SOME videos (even 30fps) but not others (even 23.98fps ones where the Level 3.0 MB rate is not exceeded)
[11:09] <divVerent> my next attempt today in the evening would be rerunning the broken ones, with the H.264 level 3.0 VBV restrictions applied, but I highly doubt the files I have exceed them
[11:09] <divVerent> given that a bandwidth of 10Mbps is allowed, and the average is WAY below that
[11:12] <divVerent> in parallel, I will try to figure out if there is any way on the N900 to find out WHY a file is rejected
[11:14] <relaxed> It could be a specific libx264 setting triping it up. Compare the SEI from the streams.
[11:14] <divVerent> SEI is?
[11:14] <JEEB> strings welp.file |grep "x264"
[11:16] <relaxed> for example, my phone won't playback streams with weightp enabled.
[11:19] <divVerent> ah, that's identical for them
[11:19] <divVerent> they did come out of the same command line after all, just different inputs
[11:19] <divVerent> actually, they are not... WTF, good hint
[11:20] <divVerent> ok, I apparently had used two different --presets...
[11:20] <JEEB> but I'd be surprised in general that the N900 could decode such high-res clips
[11:20] <JEEB> it was as if it was decoding in software the last I tested
[11:20] <JEEB> my underclassman was trying to get rtsp streaming working for it :)
[11:20] <divVerent> BUT, two of the files that do play have 100% identical settings to the three that do not play
[11:21] <divVerent> jeeb: it worked fine for me
[11:21] <divVerent> but at high battery use
[11:21] <divVerent> I didn't overclock, even
[11:21] <JEEB> so yeah, I remembered correctly
[11:21] <JEEB> it decodes with software :s
[11:21] <divVerent> it SUPPOSEDLY has a DSP, but I have not checked if it decodes in HW or SW
[11:22] <divVerent> it is a bit too fast to eb able to decode this in SW
[11:22] <divVerent> *too slow
[11:22] <divVerent> the settings only differ in analyse=, me=, subme=, mixed_ref=, trellis=, chroma_qp_offset=, rc_lookahead=, mbtree=
[11:22] <divVerent> me=, subme=, chroma_qp_offset=, rc_lookahead= can sure be ruled out, I think
[11:22] <JEEB> yeah
[11:22] <JEEB> mixed refs and subme left
[11:22] <JEEB> mbtree and rc_lookahead are encoder-side only things too
[11:22] <divVerent> mixed ref and trellis, you mean?
[11:23] <JEEB> yes
[11:23] <JEEB> d'oh
[11:23] <JEEB> was looking at the wrong point of the screen when blindly writing :P
[11:23] <JEEB> brainderp
[11:24] <divVerent> I GUESS one is the default settings... and the other is one of the faster profiles
[11:24] <divVerent> I just forgot which one I had set :P
[11:24] <JEEB> anyways, you should be using --tune fastdecode with that thing anyways
[11:25] <JEEB> profile baseline --level 30 (remember to set refs too) --tune fastdecode
[11:25] <JEEB> libx264 doesn't auto-set refs to level
[11:25] <JEEB> :s
[11:25] <JEEB> x264cli does
[11:25] <divVerent> ah, good idea
[11:25] <divVerent> also, I know I am exceeding level limits a bit
[11:25] <JEEB> no idea why they left that feature out of the library
[11:25] <divVerent> (about 10% more MB/sec)
[11:26] <divVerent> but I have no way to avoid that without downscaling below native screen res
[11:26] <JEEB> just stop expecting full res video playback :P
[11:26] <JEEB> srsly
[11:26] <divVerent> but anyway, now checking why two of the files DID play with the same settings...
[11:26] <divVerent> I hate extra scaling
[11:26] <JEEB> then enjoy your shit
[11:26] <divVerent> it looks better than the blur
[11:26] <divVerent> but anyway, I had also now fixed my scripts to avoid upscaling, ever
[11:27] <divVerent> so I now can really ensure there's only ONE scale at any time
[11:27] <divVerent> the two videos at same settings that DID play are the 648x480 and 652x480 ones
[11:27] <divVerent> ok... so it IS either trellis or mixed_refs
[11:27] <JEEB> my first phone did hw decoding, but could only do baseline, level 3.0 and had a problem with weightp. My newer one actually has a proper qualcomm chipset that has a high profile, level 4.1 compatible ASIC
[11:28] <relaxed> JEEB: was it the Atrix?
[11:28] <tbo> is autodetecting of hw decoding on in ffmpeg 1.2?
[11:28] <divVerent> fastdecode turns off deblock, weightb only
[11:28] <divVerent> weightp and cabac aren't in baseline anyway
[11:28] <JEEB> no, weightp was afaik
[11:28] <divVerent> weightb also doesn't matter
[11:28] <JEEB> at least it had an effect
[11:28] <divVerent> because there are no B frames
[11:29] <divVerent> x264's help says that --profile baseline implies --weightp 0
[11:29] <JEEB> oh wait, it had problem with deblock
[11:29] <JEEB> yeah, another brainderp
[11:29] <divVerent> deblock makes sense
[11:29] <divVerent> it's the one thing taking most CPU
[11:29] <JEEB> well, this was HW decoding
[11:29] <JEEB> the ASIC was failing at life
[11:29] <JEEB> :V
[11:29] <JEEB> relaxed, nah -- samsung galaxy spica (i5700) and Huawei U8800
[11:29] <divVerent> right... and I don't WANT to disable deblock, as I then could just as well use MPEG4 ASP again+
[11:29] <JEEB> I can only afford cheap'o phones :V
[11:30] <divVerent> but good hint, will test if it was mixed_ref or trellis
[11:30] <JEEB> divVerent, you'd probably still get a better compression ratio with x264 :P
[11:30] <divVerent> hehe
[11:30] <relaxed> my nexus 4 should arrive today :)
[11:30] <JEEB> because MPEG-4 Visual is a bad format, and the encoders aren't that great either
[11:30] <divVerent> I actually don't care much for the ratio, I just want it to look as good as possible :P
[11:30] <JEEB> yeah, have fun
[11:30] <JEEB> with that thing
[11:30] <divVerent> hehe, sure
[11:30] <divVerent> that's the only thing it is for
[11:30] <JEEB> it's just that I've already enjoyed my share of N900 and video encoding :)
[11:30] <divVerent> just wondering which one "in theory" is more evil to CPU, trellis or mixed_refs
[11:31] <divVerent> haha, I see
[11:31] <divVerent> mixed-refs SOUNDS like something that may kill CPU cache, but else it looks harmless
[11:31] <JEEB> the PSP on the other hand was simple and nice http://x264.fushizen.eu/random/psp_encoding.txt
[11:31] <JEEB> :V
[11:31] <divVerent> and trellis is something I have not any idea about
[11:33] <divVerent> I mean, how does trellis affect the DEcoder
[11:34] <divVerent> ok, so I will actually try TWO things... turning off mixed_ref, and if nothing else works, deblock
[11:41] <durandal_1707> michaelni: kick spammer
[11:53] <durandal_1707> michaelni: lol, but i think flood was just temporal and would not happen again (hope)
[11:54] <michaelni> hopefully :)
[12:05] <dmonjo> how can i take video form my webcam and stream to vp8?
[12:12] <dmonjo> i am getting Unknown encoder 'libvpx'
[12:12] <dmonjo> libvpx-dev installed
[12:12] <dmonjo> ffmpeg version N-50910-ga95a387
[12:12] <dmonjo> what am i missing
[12:13] <JEEB> dmonjo, ffmpeg -codecs |grep "vpx"
[12:14] <JEEB> if you don't get anything with that, then you didn't enable libvpx encoding
[12:14] <JEEB> :)
[12:14] <JEEB> when compiling ffmpeg, that is
[12:14] <dmonjo> i get an output
[12:14] <dmonjo> that shows vpx
[12:14] <JEEB> pastebin pl0x
[12:15] <relaxed> we get signal
[12:15] <burek> dmonjo, read the topic
[12:15] <burek> or the channel on-join notice, something.. :)
[12:15] <dmonjo> noted
[12:16] <dmonjo> http://pastie.org/6625060
[12:16] <JEEB> uhh yeah
[12:16] <JEEB> that pretty much means that you don't have libvpx enabled
[12:17] <JEEB> time to recompile :)
[12:17] <dmonjo> ok one more question please
[12:17] <JEEB> also you probably will want to get libvorbis-dev too
[12:17] <JEEB> because the internal vorbis encoder is horrible
[12:17] <JEEB> it's the epitome of why NIH should only be done when absolutely needed
[12:17] <dmonjo> i want to replace the drop.avi by the output of my webcam /dev/video0
[12:17] <dmonjo> how can i do that?
[12:17] <JEEB> depends on the input module used
[12:18] <JEEB> see the documentation related video4linux
[12:18] <JEEB> and stuff like that
[12:18] <burek> :roll:
[12:21] <dmonjo> ok great
[12:21] <dmonjo>  -f video4linux2 -r 25 -s 320x240 -i /dev/video0 -codec:v libvpx  -b:v 500k -qmin 10 -qmax 42 -maxrate 500k -bufsize 1000k -threads 4 -vf scale=-1:480 -codec:a libvorbis -b:a 128k output.webm
[12:21] <dmonjo> works fine
[12:21] <dmonjo> now i want to replace output.webm by an http:// address
[12:22] <dmonjo> can that be done?
[12:22] <burek> sure, just use the text editor
[12:22] <JEEB> dmonjo, ffmpeg itself can't serve http
[12:22] <JEEB> so nope
[12:22] <dmonjo> text editor?
[12:22] <JEEB> you can use ffserver, but good luck and have fun with that
[12:22] <dmonjo> JEEB: tried that
[12:22] <dmonjo> very bogus
[12:22] <JEEB> more like very much without documentation or help :P
[12:23] <dmonjo> so how can i transfer that? :/
[12:23] <JEEB> if you need something simple that can do HTTP I recommend VLC
[12:23] <JEEB> of course you would have to get support from them, naturally
[12:25] <burek> dmonjo, what exactly do you want to achieve with "replacing output.webm with http:// address"
[12:26] <JEEB> he wants to serve it over http?
[12:26] <dmonjo> yes
[12:26] <Bor0> how can I extend verbosity of "h264: no picture" ?
[12:26] <dmonjo> i want to send itto icecast
[12:26] <dmonjo> vp8 not h264.
[12:26] <burek> etc
[12:26] <dmonjo> i use the HLS
[12:27] <burek> get yourself a web server and use segment format http://ffmpeg.org/ffmpeg-formats.html#segment_002c-stream_005fsegment_002c-ssegment
[12:27] <dmonjo> h264 transcoindg works fine
[12:27] <dmonjo> but dunno about vp8
[12:28] <burek> Bor0?
[12:28] <Bor0> hi?
[12:29] <Bor0> I'm using libffplay, which is basically using the ffplay codebase. however, when I try to play a bit larger mp4 files I get "h264: no picture" and player hangs, but it works for small video files
[12:29] <Bor0> logs say it crashes at queue_picture(), at "wait until we have space to put a new picture"
[12:30] <dmonjo> burek: the idea ok i treasncode and segment
[12:30] <dmonjo> but how can i send to http?
[12:30] <burek> you dont
[12:31] <dmonjo> ?
[12:31] <burek> Bor0, are you using the latest version of that library
[12:31] <dmonjo> burek: how can i serve the vp8 fiel on the webserveR?
[12:32] <burek> i dunno, i dont use vp8
[12:32] <Bor0> no, it's a project I found on github. they are using a bit older version of the library, that is FFmpeg @ 8d07742
[12:32] <burek> Bor0, then ask people who own that project?
[12:33] <Bor0> I can't reach them :)
[12:33] <Bor0> I compared libffplay.c to ffplay.c, the only difference is SDL -> pthreads
[12:33] <Bor0> for queue_picture()
[12:34] <burek> we don't offer a support for outdated versions of ffmpeg libraries/tools because your bug has probably been fixed already
[12:34] <burek> so, why waste time on that
[12:34] <Bor0> do you have information if this kind of bug was present in previous ffmpeg versions?
[12:34] <burek> im not a bug tracker :)
[12:34] <Bor0> :)
[12:43] <dmonjo> burek: ok
[12:43] <dmonjo> burek: is it possible to transcode ogv to vp8?
[12:43] <burek> sure it is
[12:43] <dmonjo> burek now the ogv is a live stream
[12:44] <dmonjo> and on the webserver
[12:44] <dmonjo> i wantthe transcoding to happen live
[12:44] <burek> i'll just quote the answer i got here some time ago
[12:44] <dmonjo> so for ewxmaple now when a user opens the stream he sees the live stream not plays the stream from the time it started streaming
[12:44] <burek> "ffmpeg is an encoder, not a streamer"
[12:45] <dmonjo> yea but i mean when you encode it and try to play it it, everytime you open it it plays from the very start?
[12:45] <dmonjo> or plays from where it is actualyl being encoding in that point in time
[12:46] <burek> i never streamed ogv live, so im not sure if it's even possible
[13:08] <dmonjo> how can i convert ogv to vp8 plz
[13:12] <spaam> use libvpx
[13:13] <spaam> ffmpeg -i super.ogv -c:v libvpx superDuperMegaFile.webm   dont forget att some other options!
[13:13] <dmonjo>  avconv -i http://67.227.206.166:8000/event1.ogv  -codec:v libvpx  -b:v 500k -qmin 10 -qmax 42 -maxrate 500k -bufsize 1000k -threads 4 -vf scale=-1:480 -codec:a libvorbis -b:a 128k output.webm
[13:13] <dmonjo> working fine
[13:13] <dmonjo> thans
[13:14] <dmonjo> but the question is does webm support indexing in this case?
[13:14] <dmonjo> like for example you start transcoding at time 0
[13:14] <dmonjo> but a client views the webm generates at time 10
[13:14] <dmonjo> he should see time 10 instead of time 0
[13:15] <dmonjo> so is this indexing happening with the transcoding?
[14:28] <dmonjo> quick question
[14:28] <dmonjo> -c:v libx264 -c:a libfdk_aac -b:v 64k -flags -global_header -map 0:0 -map 0:1 -hls_time 5 -hls_wrap 64000 -hls_list_size 999 wso35.m3u8
[14:28] <dmonjo> is the output considered h264/mpeg4?
[14:28] <dmonjo> is mpeg-ts same as mpeg 4 ?
[14:43] <dmonjo> how can i convert ogv to h264?
[14:43] <dmonjo> mpeg4
[14:50] <DeXa> @dmonjo try ffmpeg -i <input> -q 0 filename.mp4
[14:50] <dmonjo> Demon_Fox: the input is a stream file
[14:50] <dmonjo> i need to create segments of m3u8 pointing to mp4
[14:52] <burek> dmonjo, this is not avconv support channel
[14:53] <dmonjo> huh?
[14:57] <dmonjo> burek: the question is can i create m3u8 with mp4 inside it?
[14:57] <dmonjo> from an ogv
[14:57] <dmonjo> i managed to get hls streams with m3u8
[14:57] <dmonjo> dunno if i can get mp4
[14:57] <dmonjo> is it feasible?
[14:57] <burek> try and see
[15:06] <dmonjo> burek: you mean try with  -segment_format mpeg4 ?
[17:20] <trose> Is there a way to encode high quality AAC's using ffmpeg that is GPL compliant? if not, are there alternative tools?
[18:05] <exco> Is ffmpeg the right tool to take a video stream (mjpeg/h.264 over rtp) from a webcam, transcode + repackage in a different container (flv) and stream that to an online streaming service?
[18:05] <ubitux> yes
[18:06] <Fjorgynn> :)
[18:07] <exco> thought so - but ran into some problems
[18:07] <Demon_Fox> ubitux, I think the visitor wanted some code for it
[18:08] <kaffeebohne> Hi, i still have got a problem with avisynth and ffmpeg. :( http://paste42.de/47a319e6116b2f22d1fc8e68c6f970d5/5151/ (ffmpeg stops with an "input/output error")
[18:08] <exco> e.g. how do I handle dropped rtp packets - so the stream doesn't die?
[18:10] <LithosLaptop> thats weird, I would think it shouldn't die
[18:10] <LithosLaptop> rtp allows packets to get dropped
[18:19] <exco> how does my ffmpeg cmd look? $FFMPEG  -re  -i "$INURL" -s "$INRES" -r 25 -vcodec libx264 -x264opts keyint=1 -r "$FPS" -preset ultrafast -crf 26 -threads 0 -s "$OUTRES" -f flv "$URL"
[18:41] <gp> How do I make sense of the idet filter output?  Example output here: http://pastebin.com/HN8RxXsh
[18:42] <gp> I.e. I am trying to automatically deinterlace when required... how would I use this output to determine if deinterlace is a good idea?
[18:44] <sacarasc> exco: It looks out of date.
[18:44] <exco> sacarasc, tips for improvement?
[18:44] <exco> *hints
[18:44] <sacarasc> And having 2 -s things in the output is silly.
[18:46] <sacarasc> -c:v is the way to choose a video codec now. -vf scale=123x456 to scale it. -threads 0 has been the default for libx264 for a while.
[18:47] <Demon_Fox> Is ffmpegs aac codecs up to par yet with the other aac codecs?
[18:47] <Demon_Fox> Like nero
[18:47] <Demon_Fox> and maybe the faac one?
[18:48] <exco> sacarasc,  how do I go about rtp errors? http://pastebin.com/V8TpyLxY
[18:49] <LithosLaptop> Demon_Fox: only fdk_aac
[18:49] <Demon_Fox> LithosLaptop, Is it good though?
[18:49] <LithosLaptop> yes
[18:49] <Demon_Fox> and if so, how would I set it to the optimal settings, unless it already does that?
[18:52] <LithosLaptop> there aren't lots of settings to change
[18:52] <LithosLaptop> http://ffmpeg.org/trac/ffmpeg/wiki/AACEncodingGuide
[18:53] <LithosLaptop> thrtr is a 'afterburner' setting that gives higher quality. I don't know if it is enabled by default
[18:55] <LithosLaptop> *there
[19:04] <asdfqwer> Is there somewhere that has a list of compatible codec_types and codec_ids? For instance codec_id: CODEC_ID_TEXT only seems to work with codec_type: AVMEDIA_TYPE_SUBTITLE.
[19:18] <llogan> LithosLaptop: afterburner should be default
[19:57] <eric___> i am trying to transcode a video from Metacafe site... I see following behaviour. While transcoding normal videos ffmpeg transcode fine but when a person is standing and giving speech and lot of people are taking his photographs and  when the flash shot comes here comes the quality distraction
[19:57] <eric___> quality degrades during the photo shots alot
[19:58] <eric___> any obvious reason behind it?
[20:00] <eric___> example video http://www.metacafe.com/watch/9979920/anne_hathaway_practiced_oscar_speech_to_be_likable/
[20:02] <LithosLaptop> transcoding to what and at what bitrate?
[20:04] <eric___> mp4 transcoding and at 500kbps
[20:07] <LithosLaptop> The bitrate from the source seems more than 500Kbps and you are transcoding from a more efficient codec(H.264) to a less efficient codec (MPEG4)
[20:08] <LithosLaptop> So you will need a higher bitrate than the source video to get about the same quality
[20:11] <LithosLaptop> try not to transcode the audio. use 'copy' as the audio codec
[20:15] <eric___> it works fine till the photo shoots comes in picture
[20:16] <eric___> as soon as people start taking pictures vudei quality degrades
[20:16] <eric___> video quality
[20:48] <eric___> @<LithosLaptop> any suggestions?
[20:49] <llogan> eric___: how can we make any suggestions without seeing your command and the console output as requested?
[20:50] <eric___> @llogan: did you see the example video
[20:50] <eric___> http://www.metacafe.com/watch/9979920/anne_hathaway_practiced_oscar_speech_to_be_likable/
[20:51] <funyun> hi. does anyone know what multiplexing means? and if it's possible to do with ffmpeg?
[20:54] <LithosLaptop> isn't multiplexing just muxing?
[20:59] <funyun> LithosLaptop: i'm not sure. i encoded a video with ffmpeg. and i'm trying to create a burnable iso image but it tells me the .mpg has to be re-multiplexed
[21:00] <saste> funyun: multiplexing == muxing
[21:00] <saste> interleave packets from different streams in the same multimedia stream
[21:00] <saste> it's what muxers are supposed to do
[21:01] <funyun> saste: is there anything i can do with ffmpeg so that i don't get this error?
[21:19] <divVerent> jeeb: fun n900 fact: it does hw decode h:264 but copies to screen in SW
[21:19] <divVerent> other fun fact: the default background contains the color key
[21:20] <divVerent> ffmpeg can decode fullscreen at realime only with loopfilter disabled
[21:21] <divVerent> kinda like on eeepc 701 (celeron 630 mhz)
[21:21] <Fjorgynn> celeron?
[21:22] <divVerent> the first eeepc has a celeron 900
[21:22] <divVerent> clocked at 630 to save battery
[21:22] <funyun> saste: http://pastebin.com/BX6ZqCPL
[21:22] <divVerent> the atoms are the later models
[21:24] <saste> funyun: dunno, you should ask dvdauthor people
[21:27] <llogan> funyun: -crf is not a mpeg2video option, and i think target ntsc-dvd will set -ab anyway. also you're missing the console output
[21:27] <funyun> saste: any idea where i can find them?
[21:28] <funyun> llogan: you mean ffmpeg's full output?
[21:28] <llogan> yes. you've been here enough to know to include the console output
[21:29] <llogan> although i doubt i'll see anything obvious, it is always good in include.
[21:29] <saste> funyun: dvdauthor ML? website?
[21:29] <funyun> llogan: http://pastebin.com/jJNm3EPs
[21:30] <funyun> saste: what's ML?
[21:30] <saste> mailing list
[21:30] <funyun> oh i see
[21:33] <funyun> llogan: so it's possible that because i added "-ab 384k" that's messing it up? i feel like this problem has to do with the audio..
[21:33] <llogan> i don't know, but 384k for 6 channels seems low. your source audio is already ac3. you probably don't need to re-encode.
[21:34] <llogan> try adding -c:a copy after -target ntsc-dvd
[21:36] <funyun> llogan: does that copy the audio? if so, it will make the final output too large for a dvd.
[21:39] <llogan> yes, it does copy the audio.
[21:39] Action: llogan goes afk for an hour
[22:36] <DeXa> anyone knows how does ffmpeg deal with VDA for machines with no hardware support?
[22:36] <DeXa> https://github.com/mxcl/homebrew/pull/18583
[23:19] <stiell> Some frames are getting lost (counted as "dup" with -vsync 1) when capturing from v4l2. Is there any way to avoid this?
[23:28] <kangstarr> hi everyone
[23:28] <kangstarr> brb...relog
[23:31] <kangstarr> Are there plans for ffmpeg to support Rec. 2020 colorspace?
[23:35] <kangstarr> hmm, anyone at keyboard here?
[23:48] <kangstarr> rec. 2020? anyone?
[23:49] <ubitux> i answered you on the other channel
[23:49] <ubitux> open an issue if you want it
[00:00] --- Wed Mar 20 2013