[Ffmpeg-devel-irc] ffmpeg.log.20150115

[burek]

[00:31] <n0-pann1es> hi folks
[00:31] <n0-pann1es> is this the right place to learn about video editing?
[00:32] <c_14> Depends, what are you trying to do.
[00:32] <n0-pann1es> I want to cut and paste differnt clips together into one vid
[00:32] <n0-pann1es> but I use windows
[00:33] <c_14> You can do that with ffmpeg.
[00:34] <n0-pann1es> i thought it was justa codec
[00:34] <n0-pann1es> is it a program?
[00:36] <c_14> It's a program, yes.
[00:36] <c_14> https://ffmpeg.org/
[00:36] <n0-pann1es> can I buy it from you?
[00:37] <n0-pann1es> cool thx
[00:43] <AlexRussia> n0-pann1es: yep, may be you can do some kind of donation....
[00:43] <AlexRussia> :D
[00:44] <n0-pann1es> I will send you some money if you comment on my facebook
[00:48] <AlexRussia> n0-pann1es: http://ffmpeg.org/donations.html
[00:48] <n0-pann1es> cool thx, you think they get enough?
[00:49] <AlexRussia> n0-pann1es: IDK, I am not part of them
[00:49] <AlexRussia> just you said buy
[00:49] <AlexRussia> but ffmpeg is open source
[00:49] <AlexRussia> so buy here easy to convert in donation
[00:50] <n0-pann1es> but it does nto work with windows os?
[00:50] <c_14> It works on windows.
[00:51] <c_14> Check under Downloads Windows
[00:51] <n0-pann1es> o cool thx
[00:53] <n0-pann1es> why cant i find it?
[00:58] <robotbrain> hello
[00:58] <robotbrain> I am trying to get an extremely low latency livestream of a screen region
[00:59] <robotbrain> here is the command: http://pastebin.com/Num6NLgq
[00:59] <c_14> Why do you have -tune zerolatency twice?
[00:59] <robotbrain> I am using this as a/v for a wireless VR rig with google cardboard
[01:00] <robotbrain> and that is accidental
[01:00] <c_14> The second -s is redundant
[01:01] <c_14> What's the -strict -2 for there?
[01:03] <voip_> can you guys chek http://pastebin.com/nsRs4B5t  i have many different errors
[01:05] <voip_> Number of bands (13) exceeds limit, Multiple frames in a packet from stream 0B, Prediction is not allowed in AAC-LC, Reference 4 >= 2,error while decoding MB 14 64, concealing 515 DC, 515 AC, 515 MV errors in P frame, Assuming an incorrectly encoded 7.1 channel layout ...........
[01:05] <voip_> llogan, can yo check ?
[01:08] <robotbrain> c_14: I was using aac earlier and it was complaining
[01:08] <c_14> Ye, but you don't need it for opus.
[01:09] <robotbrain> I realize that, had it left in
[01:09] <robotbrain> ive just been tinkering the same cli
[01:09] <n0-pann1es> sorry I found it... I am very idiot
[01:09] <llogan> voip_: i don't know. looks like a shitty input.
[01:09] <c_14> Other than the things I mentioned, I can't find much improvement. You can read through https://trac.ffmpeg.org/wiki/StreamingGuide#Latency if you want
[01:10] <robotbrain> I have been
[01:10] <robotbrain> there is about a one second delay atm
[01:10] <robotbrain> for a/v input in vr that is unacceptable
[01:10] <n0-pann1es> cool so is ther a graphics program with layering that will work with widows and is free?
[01:11] <voip_> thamks. llogan
[01:13] <c_14> robotbrain: what if you remove the audio stream?
[01:13] <llogan> n0-pann1es: sounds like you want an editor. you could just use "movie maker". it's free and basic i think.
[01:14] <robotbrain> c_14: same
[01:15] <n0-pann1es> llogan MM sux
[01:16] <robotbrain> I also get data is not aligned if that helps
[01:16] <c_14> robotbrain: try enabling intra refresh?
[01:17] Action: robotbrain facepalms after realizing he never put that in
[01:18] <robotbrain> c_14: longer delay
[01:18] <c_14> longer?
[01:18] <c_14> That's weird
[01:18] <robotbrain> 1sec vs 5sec
[01:20] Action: robotbrain adds more threads and it cuts time in half
[01:21] <c_14> Wait, more threads decreases the latency?
[01:21] <robotbrain> oh wait
[01:21] <robotbrain> I must have had more before
[01:21] <robotbrain> I didnt have it set
[01:22] <c_14> Unset is automatic
[01:22] <robotbrain> down to ~.5sec
[01:22] <c_14> Ie as many as it wants
[01:22] <robotbrain> by setting it low
[01:23] <c_14> Do you know what -compression_level does?
[01:23] <c_14> Cause I sure don't.
[01:23] <robotbrain> 10 is best quality
[01:23] <robotbrain> 0 is worst
[01:23] <robotbrain> but 10 tends to high latency
[01:23] <robotbrain> its specific to opus
[01:24] <c_14> aaaah, now I see it.
[01:24] Action: c_14 was looking at the compression_level global option
[01:25] <robotbrain> I have to head out
[01:25] <robotbrain> ill be back tommorow
[01:41] <robotbrain> do note I use quassel as a bouncer so I get anything that is said while im gone
[02:06] <pisto> I have tried anything I can think of, but still I cannot get sound and vide  sync when recording with x11grab pulse
[02:06] <pisto> anyone any pointer?
[02:06] <pisto> real life success story?
[02:48] <pisto> http://pastebin.com/K2iAM6wb command + output
[02:49] <pisto> pisto.horse/test.mkv file output
[02:49] <pisto> llogan ^
[02:50] <klaxa> pisto: http://ffmpeg.org/ffmpeg.html#Video-Options see: -framerate
[02:51] <klaxa> try using -framerate 30 instead of -r 30 and see if it still occurs
[02:51] <llogan> or http://ffmpeg.org/ffmpeg-devices.html#Options-8
[02:52] <llogan> hmm...x11grab -framerate default is 30000/1001. i assumed it was 25.
[02:52] <pisto> klaxa, yes, identical problem
[02:53] <llogan> can you use -f alsa instead? (not sure what device you're capturing though)
[02:55] <pisto> alsa_output.pci-0000_00_1b.0.analog-stereo.monitor
[02:56] <pisto> doing -f alsa I get an error
[02:56] <pisto> [alsa @ 0xc62980] cannot open audio device alsa_output.pci-0000_00_1b.0.analog-stereo.monitor (No such file or directory)
[02:56] <pisto> alsa_output.pci-0000_00_1b.0.analog-stereo.monitor: Input/output error
[02:56] <pisto> not sure what I shohuld put there indeed
[02:56] <llogan> http://ffmpeg.org/ffmpeg-devices.html#alsa
[02:56] <llogan> https://trac.ffmpeg.org/wiki/Capture/ALSA
[02:58] <pisto> it seems n sync but it0s capturing my microphne
[02:58] <llogan> when did horse become a TLD?
[02:58] <pisto> this fall
[02:58] <llogan> why did you get a .horse?
[02:59] <pisto> because I can
[02:59] <pisto> and I love raw horse meat
[02:59] <pisto> I want to capture my audio output, not my microphone
[02:59] <llogan> what? no .mule?
[02:59] <pisto> never tried mule meat
[03:00] <llogan> https://trac.ffmpeg.org/wiki/Capture/ALSA#Recordaudiofromanapplication
[03:02] <pisto> this is starting to get scaring
[03:03] <pisto> can this be done without persisting it into a configuration file?
[03:11] <nia> Hi, I'm trying to use ffmpeg to render/convert a swf file and export it as an mp4. I've noticed a few video streams popped up while the process was running and I'd like to be able to include those in my output file. Currently the output is audio only. Here is the output from ffmpeg showing the two video streams that popped up: https://bpaste.net/show/43fbf08a2600
[03:14] <llogan> nia: your command is missing
[03:15] <nia> Ah... I apologize
[03:15] <nia> ffmpeg -i output.swf test.mp4
[03:19] <llogan> experiment with -probesize and/or -analyzeduration
[03:20] <llogan> as input options
[03:20] <llogan> also the 1.2 branch is old
[03:43] <pisto> ok, the alsa loopback card trick worke
[03:43] <pisto> d
[03:44] <pisto> is there a magical streaming protocol that is on-demand and doesn't require particular setup on clients and is dare I ask cross platform?
[05:52] <fahadash> hello
[05:53] <fahadash> Can anyone tell me what type of Sound attributes do FLV files carry ?
[05:53] <fahadash> and the sample rate
[07:04] <fffan> can I use av_frame_get_best_effort_timestamp to get audio's pts
[11:42] <DeadSix27> if i would want to seek to a certain percentage, is there a better way than my slow method of probing a video first, and then calculating the percentage by its length?
[12:34] <maximg> hi, I am using libavcodec+libavformat to encode audio to aac and put it in an mpeg-ts stream. However, mpeg-ts requires aac/adts stream; any hints how to set up my codec/muxer chain for that? (to be clear, this is not about the ffmpeg command line tool)
[12:37] <aspiers> any ideas why screencasting via ffmpeg -f x11grab would result in artefacts when using urxvt/xterm but not gnome-terminal? not sure but I think the problem is that coloured pixels don't always get reset to black when they are supposed to, e.g. the text cursor is supposed to flash but stays solid, and scrolling text leaves remnants behind.
[12:47] <fahadash> Hello
[13:25] <Yengas> I'm trying to merge an audio and video file with ffmpeg libraries in c++. Are there any beginner level tutorials, books about audio/video encoding and how the ffmpeg library works?
[13:56] <ubitux> muxing example in doc/examples maybe
[14:54] <Yengas> yeah i was pretty glad that i found that. i rewrote the same code in cpp to get a feel about the library. i also filtered out the sound from an avi by finding the stream with the audio type and not writing that stream into the output file.
[14:57] <Yengas> but i still dont know how to seek(skip streams) while reading format contexts, but i just took a look at the ffmpeg source code and it seems pretty well documented. i think i will just read the source code until i understand what the hack is really going on under the hood.
[15:15] <tholin_> it looks like the flag avctx->err_recognition & AV_EF_CRCCHECK is used to error out on broken crc but some decoders doesn't check AV_EF_CRCCHECK before returning an error. Is that normal or should AV_EF_CRCCHECK always be respected?
[15:16] <tholin_> the reason I ask is because I'm doing some fuzzing and it's better if crc are disabled for that.
[15:22] <Mavrik> encoders pretty much do whatever they want
[15:28] <tholin_> I'm only talking about decoders. Like libavcodec/ffv1dec.c:592 looks like it error out on bad crc without checking AV_EF_CRCCHECK. Some probes used crc to set probe score but that doesn't look like the case here.
[15:31] <Mavrik> tholin_, sorry, I meant decoders too
[15:32] <Mavrik> all the decoders aren't equally maintained and don't equally conform to those kind of parameters :/
[15:46] <NapoleonWils0n> afternoon all
[15:46] <NapoleonWils0n> im trying to use amerge to convert stereo stream to mono
[15:46] <NapoleonWils0n> any tips
[15:47] <NapoleonWils0n> want to convert stream 0 from stereo to mono
[15:48] <NapoleonWils0n> my mic only outputs to right channel for some reason the left channel is blank
[15:48] <NapoleonWils0n> so need to output to 1 channel
[16:33] <jumpysnake> hi qualcuno parla italiano?
[16:35] <jumpysnake>  i ave this problem http://pastebin.com/Gx7ZHMgk how can I solve
[16:35] <jumpysnake> ?
[16:48] <ShadowCruiser> Hello
[17:04] <fahadash> Where can I find ffmpeg codecs for FLV ?
[17:42] <fahadash> Do I exist ?
[17:43] <justinX> fahadash: no you are just a fiction of my mind
[17:44] <justinX> I imagine in scifi they would call it a "fictoid"
[17:56] <c_14> fahadash: https://git.videolan.org/?p=ffmpeg.git;a=blob;f=libavformat/flv.h;h=db9468f4691d259608160bbba0639ac25fd738c3;hb=HEAD#l89
[17:59] <fahadash> I am actually having trouble loading a FLV file with 0x0 video (hench only audio) in Goldwave, I have to first FFMPEG -VN it and then load the wave file into Goldwave. Goldwave could not recognize the flv format and asks for Bitrate, Frequency etc
[18:07] <c_14> What does ffprobe say about the video?
[19:11] <Tythus> Hi guys I'm looking at doing a mozaic recompile to merge multiple rmtp cmaera streams into one stream I have the rmtp streams for an nginx addon so now what I need to do is recompile them into a single stream and then send them out in another rmtp stream
[20:41] <fahadash> c_14: Duration: 00:00:09.29, start: 0.000000, bitrate: 20 kb/s Stream #0:0: Audio: nellymoser, 8000 Hz, mono, flt
[21:00] <Tythus> Hi guys I'm looking at doing a mozaic recompile to merge multiple rmtp cmaera streams into one stream I have the rmtp streams for an nginx addon so now what I need to do is recompile them into a single stream and then send them out in another rmtp stream
[21:02] <goulard> Is there an ffmpeg call I can make to obtain the original x264 encode settings of a source video?  I would like to obtain keyint, keyint_min, gop_size etc... of a source video for use in a transcode
[21:03] <goulard> avcodec_decode_video2 only really seems to populate the resolution
[21:03] <goulard> of the codec context
[21:10] <jvcleave> goulard - also looking for gop_size - ran across this
[21:10] <jvcleave> http://ffmpeg.org/doxygen/trunk/ffmdec_8c_source.html#l00453
[21:20] <goulard> jvcleave: thanks!
[21:21] <goulard> jvcleave: how did you end up calling into ff_read_header?
[21:22] <jvcleave> goulard: it seems like those may only be used for encoding tho - mine are showing up with values of 12 when MediaInfo is saying 1
[21:22] <goulard> yeah I have same issue
[21:22] <goulard> I know ffmpeg can do it because you can do it at the command line
[21:25] <jvcleave> Im looking to see where the hints are processed
[21:28] <jvcleave> goulard: somewhat like this http://ffmpeg.org/doxygen/trunk/avio_reading_8c-example.html#a23
[21:29] <goulard> jvcleave: cool
[21:30] <goulard> where input file may be raw h264?
[21:30] <goulard> or a TS?
[21:31] <jvcleave> that one is a file (which is also my case)
[22:53] <goulard> jvcleave: do you know how to disable ffmpeg dump to stdout?
[22:54] <goulard> when calling avformat_open_input
[22:55] <jvcleave> goulard: I dont - Im working with an existing project
[22:56] <goulard> I was trying to feed it raw h264 - but it doesnt look like it can detect what it is
[22:56] <goulard> It is able to detect a TS though
[22:57] <goulard> I wonder if there is a special way to initialize the "fmt" argument so I can pass it raw h264
[22:59] <JEEB> yes, set the format to "h264". the documentation or headers should have it noted somewhere
[23:08] <goulard> JEEB: fmt is a struct
[23:17] <JEEB> goulard, av_find_input_format("h264");
[23:17] <JEEB> I just googled and looked through the documentation since I haven't forced the stuff myself :P
[23:17] <goulard> JEEB: you rock
[23:18] <JEEB> I mean I just looked through the functions returning AVInputFormat in avformat.h :P
[23:33] <am11> hello guys
[23:33] <am11> is it possible to statically compile libstdc++ (say gcc/g++ v4.6 and glibc 3.4.9), so to make the binary work on older target (glibcxx 2.x for example)?
[00:00] --- Fri Jan 16 2015