[Ffmpeg-devel-irc] ffmpeg.log.20150911

[burek]

[00:34:02 CEST] <t4nk035> Hey guys
[00:34:42 CEST] <t4nk035> can I ask how you compile ffmpeg 2.8 for use with quicksync? I'm using ubuntu
[01:12:24 CEST] <ericwooley> I got a webm file off of youtube, through this site http://youtube.mygeodata.eu/, and ffprobe says the encoder is "Google". Any idea how I can encode something the same way. (http://pastebin.com/uHpaDuuw For the full output from ffprobe)
[01:15:16 CEST] <llogan> ericwooley: what do you mean by "the same way"?
[01:16:03 CEST] <ericwooley>   Metadata:
[01:16:03 CEST] <ericwooley>     encoder         : Lavf56.25.101
[01:16:07 CEST] <ericwooley> I tried running the command `ffmpeg -i project-fi.webm -vf scale=320:-1 small/project-fi.webm` and it came out with that
[01:16:30 CEST] <DHE> make it look like a google encode?
[01:16:33 CEST] <Max-P> ericwooley: That's just metadata. It doesn't really have any use. But https://www.ffmpeg.org/ffmpeg-all.html#Metadata-1
[01:16:47 CEST] <ericwooley> ah ok thanks.
[01:17:03 CEST] <ericwooley> Well, I have a second webm video and I need it to match that one
[01:17:13 CEST] <ericwooley> I am trying to figure out which parts of the video matter
[01:17:24 CEST] <ericwooley> (sorry, I am kind of new to this)
[01:17:53 CEST] <Max-P> ericwooley: Stream #0:0: Video: vp8, yuv420p, 640x360, SAR 1:1 DAR 16:9, 24 fps, 24 tbr, 1k tbn, 1k tbc (default)
[01:17:58 CEST] <Max-P> ericwooley: Stream #0:1: Audio: vorbis, 44100 Hz, stereo, fltp (default)
[01:18:09 CEST] <Max-P> ^ These are the settings that matters if you want the same quality settings
[01:19:02 CEST] <Max-P> Tells you the pixel format, resolution, aspect ratios, framerate and such
[01:19:49 CEST] <ericwooley> Max-P: Alright, thanks. So I should be paying attention to individual stream settings. Thanks
[01:20:08 CEST] <llogan> why are you trying to do whatever it is you're doing?
[01:20:15 CEST] <ericwooley> http://ericwooley.github.io/VideoAppendTesting/
[01:20:23 CEST] <ericwooley> I am trying to append videos
[01:20:29 CEST] <ericwooley> some work, and some don't
[01:20:38 CEST] <ericwooley> They are all webm
[01:21:03 CEST] <ericwooley> It works perfectly depending on the video you play first
[01:21:20 CEST] <ericwooley> playing video 4 (the project-fi) video first makes it work as expected
[01:21:29 CEST] <ericwooley> strange things happen if you add other things first
[01:21:56 CEST] <llogan> how are you appending them?
[01:21:59 CEST] <ericwooley> I think it's got to be a codec issue, or malformed timestamp data
[01:22:22 CEST] <ericwooley> I am using an ajax request to get the data as a buffer, and appending the buffer to the video src buffer
[01:23:03 CEST] <Max-P> ericwooley: I don't think you can do that unless all the videos use the same codecs
[01:23:05 CEST] <ericwooley> https://github.com/ericwooley/VideoAppendTesting/blob/master/src/scripts/libs/media-streamer.js line 40
[01:23:53 CEST] <ericwooley> Thats the weird part, setting video 4 as the source first works great from then on
[01:24:08 CEST] <ericwooley> no matter what you append
[01:24:12 CEST] <ericwooley> but not the other way around
[01:24:26 CEST] <Max-P> Not for me. I can only append 2 and 3 together, appending any other videos just resets the whole thing
[01:25:49 CEST] <ericwooley> hmm
[01:26:00 CEST] <ericwooley> that is more annoying
[01:26:15 CEST] <ericwooley> if you click the src: * buttons, it changes the source
[01:26:51 CEST] <ericwooley> Max-P: if you just click the "video [number]" buttons it should append
[01:27:02 CEST] <ericwooley> what version of chrome are you using?
[01:27:25 CEST] <Max-P> 45.0.2454.85
[01:29:18 CEST] <ericwooley> Max-P: strange, for me, playing video 4, then appending another video transitions to the next video. playing 2 and 3 back to back causes the video to stop playing at the transition between them
[01:29:20 CEST] <Max-P> Humm sometimes it works
[01:29:31 CEST] <Max-P> I think your bug might be somewhere else
[01:29:48 CEST] <ericwooley> Yea, this is pretty experimental
[01:29:57 CEST] <Max-P> Just got a JS error
[01:30:02 CEST] <ericwooley> what was the error?
[01:30:27 CEST] <Max-P> Uncaught TypeError: Cannot read property 'updating' of undefined
[01:31:02 CEST] <Max-P> Yeah I can append all videos in whatever order so long as I append the first video I pick with itself
[01:31:11 CEST] <Max-P> There's some weird resetting errors happening
[01:31:12 CEST] <ericwooley> ah, that can happen, sorry. I haven't gotten this down yet, more of a proof of concept
[01:31:34 CEST] <ericwooley> What OS?
[01:31:40 CEST] <Max-P> You should probably as in an HTML5 channel, that's got nothing to do with ffmpeg it looks like
[01:32:14 CEST] <ericwooley> k, thanks
[01:32:16 CEST] <ericwooley> will do
[01:32:19 CEST] <Max-P> Linux 4.1.6-1-lqx x86_64
[01:32:23 CEST] <ericwooley> appreciate the help
[01:32:43 CEST] <Max-P> I know webdev a bit but haven't messed with videos other than just playing them
[01:33:00 CEST] <Max-P> I'll try in firefox
[01:33:03 CEST] <ericwooley> yea, I am a web dev professionally
[01:33:07 CEST] <ericwooley> doesn't work in firefox
[01:33:24 CEST] <ericwooley> brb
[01:45:37 CEST] <DX099> exit
[01:49:04 CEST] <chungy> who's in charge of the /topic? It should say ffmpeg 2.8 :D
[01:55:28 CEST] <chungy> michaelni: ping ^
[04:21:38 CEST] <fred1807> how can I get the direct url of this stream, so I can play it on vlc media player?  http://www.worldsurfleague.com/events/2015/mct/1281/hurley-pro-at-trestles/live
[04:30:07 CEST] <votz> fred1807: Have you tried tracing the network requests made by your browser?
[04:31:51 CEST] <fred1807> yes... on chrome , dev tools, when I hit play,  .mp4 files start show up... the list is infinite...
[04:32:17 CEST] <fred1807> votz: this is one fo them http://nlds300.cdnllnwnl.neulion.com/nlds/wsl/t107637/as/live/t107637_hd_1600/20150910/19/5334.mp4?nltid=wsl&nltdt=0
[04:32:41 CEST] <votz> fred1807: I noticed the same. However, no video (besides a pre-roll ad) plays in the video player for me.
[04:33:05 CEST] <votz> Instead, the player displays 'THANKS FOR WATCHING'.
[04:33:21 CEST] <fred1807> votz: you have to got back in the timelne
[04:33:46 CEST] <fred1807> the live event already ended for today
[04:35:52 CEST] <votz> fred1807: I see. Once I seek backwards, the video is served in 2s pieces.
[04:36:14 CEST] <fred1807> votz: can we get the head of the snake?
[04:36:24 CEST] <votz> fred1807: Not sure. Poking around now.
[04:37:25 CEST] <votz> I also Googled for NeuLion Adaptive Streaming Server, which is the daemon serving the two second clips
[04:37:28 CEST] <votz> https://www.reddit.com/r/cordcutters/comments/1lyi8u/lets_try_and_get_this_nfl_streaming_thing_vlc/
[04:38:44 CEST] <votz> https://trac.videolan.org/vlc/ticket/9632
[04:39:48 CEST] <votz> http://nlds300.cdnak.neulion.com/nlds/wsl/t107637/as/live/t107637_hd_4500.m3u8
[04:40:06 CEST] <votz> Opening that in VLC plays the 'THANKS FOR WATCHING' loop
[04:41:22 CEST] <Kuukunen> it's called DASH, and if you were asking for a "direct url" I doubt it exists
[04:41:43 CEST] <Kuukunen> (at least I assume it's dash)
[04:42:50 CEST] <votz> Kuukunen: It's HLS, not DASH
[04:43:03 CEST] <Kuukunen> also yea, HLS
[04:43:15 CEST] <fred1807> votz: wow you are fast... many usefull information... How do you find the m3u8?   It may work when it is live and running
[04:43:16 CEST] <Kuukunen> I just saw quickly talk about random pieces :P
[04:44:20 CEST] <votz> Kuukunen: Yep. They're similar
[04:45:17 CEST] <fred1807> do you believe it will only work with a cookie?  my last goal was to get this live show on raspberry, using omxplayer
[04:48:46 CEST] <fred1807> my connection droped
[04:48:51 CEST] <fred1807> votz: http://livestreamer.readthedocs.org/en/latest/
[04:48:55 CEST] <Kuukunen> fred1807: the ticket implies it's fixed in newer vlc versions
[04:49:27 CEST] <Kuukunen> fred1807: also, why aren't you asking this on the vlc channels? :P
[04:50:22 CEST] <fred1807> people here knows more about video, cmd line way of life
[04:50:51 CEST] <votz> fred1807: I don't know how to request the original, long video
[04:51:08 CEST] <Kuukunen> because it most likely isn't requestable
[04:51:15 CEST] <fred1807> votz: who can I find the m3u8 for future events on this same web page?
[04:51:30 CEST] <votz> But you can change the URL and get 10s snippets
[04:51:43 CEST] <votz> And you could trivially write a script that requests snippets 10s at a time and then concat them all together
[04:51:50 CEST] <fred1807> maybe when it is live, that m3u8 should work
[04:51:58 CEST] <votz> A pain? Yes. But a workable solution.
[04:52:05 CEST] <fred1807> yes..
[04:52:11 CEST] <Kuukunen> votz: yea, also a bit harder to make it watchable live
[04:52:23 CEST] <votz> There's almost certainly a way to request the whole thing. We just haven't figured it out yet.
[04:53:05 CEST] <fred1807> I think this is actually it http://livestreamer.readthedocs.org/en/latest/
[04:53:42 CEST] <Kuukunen> votz: why would they expose the whole video if the whole point is to stream it in pieces?
[04:53:48 CEST] <fred1807> livestreamer hlsvarient://ekonk...akamaihd../live/master.m3u8 --player "omxplayer --timeout 20" --fifo       <---- omxplayer is the hardware accelerated playr for raspberry, may work with other cmd line players
[04:55:09 CEST] <votz> The timestamps start at 20150910130932
[05:00:13 CEST] <votz> Not to hijack the discussion, but with a raw h264 stream, how can the appropriate metadata be extracted to feed into x264 to generate a video with identical SPS and PPS values? The two videos need to be concatenated together, which requires identical SPS and PPS values.
[06:19:38 CEST] <JanNeeds32bit> hi guys, I need help building 32-bit version of FFmpeg from source using Msys2 under Windows. Right now it always builds 64-bit version!!
[06:20:43 CEST] <JanNeeds32bit> After running ./config script, the generated config.mak files always has LIBTARGET=i386:x86-64
[06:21:45 CEST] <JanNeeds32bit> searching online, I only see outdated steps to compile under Windows. And since Msys2 and Mingw-64 came into the scene the instructions no longer work.
[06:22:33 CEST] <JanNeeds32bit> Please help, my gut feeling tells me it's simple silly problem (a flag needs to be set for 32-bit somewhere).
[06:25:03 CEST] <Max-P> JanNeeds32bit: I think you need to install 32bit msys to get 32bit builds
[06:26:38 CEST] <Max-P> I don't know anything about Windows, but since it appears to use a bunch of unix-like tools I can only assume it also comes with the same restrictions, that is, have the whole set of 32bit packages installed as well
[06:28:57 CEST] <Max-P> Yeah, 32bit is 50.8MB, 64bit is 56.3MB there's no way the 64bit version has gcc-multilib plus all the 32bit libs that comes with it
[06:30:33 CEST] <JanNeeds32bit> I will try that, thank you very much!!
[06:34:53 CEST] <JanNeeds32bit> @Max-P, by the way, www.mingw.org vs mingw-w64.org, appear to be two different groups, is that the case? Is the old www.mingw.org dead project and no longer maintained?
[06:35:29 CEST] <JanNeeds32bit> I do recall compiling FFmpeg and other open source without problem years ago with old mingw.org.
[06:41:45 CEST] <Max-P> I have no idea. It would logically be the same, but it's possible that one ships gcc-multilib and then you can pass -m32 to the compiler
[06:42:36 CEST] <Max-P> But all I know is that on Linux we usually make 32bit chroots for that as cross-compiling is always a pain to set up
[06:43:41 CEST] <JanNeeds32bit> -m32 flag didn't work in Msys2-64bit...I'm not installing Msys-32bit....
[06:58:46 CEST] <JanNeeds32bit> Max-P, big thank you! FFmpeg 32-bit appears to be compiling fine as I confirmed LIBTARGET=i386.  I gotta get some sleep. Night.
[08:27:01 CEST] <joules> Hi, not sure how "-reserve_index_space" is suppose to be applied (-f matroska). Any tips?
[08:29:05 CEST] <joules> currently I'm doing "0" and wait it to finish encoding with errors with a suggested index size and basically do a copy only transcode with -reserve_index_space {value} - seems non-optimal.
[12:01:21 CEST] <t4nk146> Hey guys, trying to compile FFMPEG 2.8 with --enable -libmfx gives me the error: libmfx not found using pkg-config, but when I pkg-config --list-all| grep mfx it is found.
[12:03:30 CEST] <iive> t4nk146: look in config.log and see what the exact error is
[12:05:09 CEST] <t4nk146> can I pastebin this?
[12:06:08 CEST] <t4nk146> http://pastebin.com/CHaN3tpF
[12:09:17 CEST] <iive> what gcc is that?
[12:09:59 CEST] <iive> hum no
[12:10:06 CEST] <iive> it's not gcc fault.
[12:10:34 CEST] <t4nk146> Urghh I haven't linked it have i?
[12:10:43 CEST] <iive> i see that the test adds "-c" options and that one skips linking, but the test provides the library to link...
[12:11:00 CEST] <t4nk146> I need to use libtool and I'm feeling overwhelmed
[12:11:26 CEST] <iive> just wait until i parse everything :)
[12:11:42 CEST] <iive> developers are free to help :)
[12:12:33 CEST] <t4nk146> Cheers, I've had quicksync hardware for ages, and now that QS encoding is now a thing in Ubuntu.. I can finally drop my dependance on Windows.
[12:13:08 CEST] <iive> the second gcc line tries to do the linking
[12:13:30 CEST] <iive> it's given same libmfx.a
[12:13:49 CEST] <iive> ffconf.K5w0Sowm.c:(.text+0x1): undefined reference to `MFXInit'
[12:15:12 CEST] <t4nk146> someone online said the git was a little broken and that I'd need to rebuild with automake?
[12:15:21 CEST] <t4nk146> autotools, sorry
[12:15:42 CEST] <iive> ffmpeg doesn't use autotools
[12:15:56 CEST] <iive> do you compile mfx on your own?
[12:16:03 CEST] <t4nk146> yeah
[12:16:24 CEST] <t4nk146> I also installed the media sdk, and all the kernel updates and patches
[12:17:04 CEST] <t4nk146> I used this https://git.videolan.org/?p=mfx_dispatch.git
[12:19:37 CEST] <iive> the library might be placed in a wrong place in the command line...
[12:21:40 CEST] <t4nk146> how would I find it?
[12:21:47 CEST] <t4nk146> I shall have a look
[12:23:31 CEST] <iive> if i remember correctly, the algorithm for library names is, that gcc picks the functions it needs and forgets for the rest. So the program should be first, then the local libraries and system libraries last.
[12:24:02 CEST] <t4nk146> yep, I compiled it into the folder that is was downloaded in, libmfx.a is it right?
[12:24:03 CEST] <iive> in the paste i see mxf.a program.o  system.
[12:25:49 CEST] <t4nk146> Hm, is there a way for me to find it?
[12:26:09 CEST] <t4nk146> I basically compiled it and then went on to try and build FFMPEG
[12:33:23 CEST] <iive> ok...
[12:33:53 CEST] <iive> t4nk146: i'd ask you to fill a bugreport in trac and to post the same pastebin there.
[12:34:16 CEST] <iive> i think i found the problem.
[12:34:23 CEST] <t4nk146> OOh
[12:35:33 CEST] <iive> the function that does the test is using a hack, to move the libraries in place, but it checks only for *.so , since you are using .a  it remains in wrong place.
[12:35:57 CEST] <iive> so, you can workaround our problem if you compile dynamic mfx
[12:36:13 CEST] <t4nk146> Ah, how would I do that?
[12:37:02 CEST] <iive> check mfx configure options. there should be something like --enable-dynamic libs or something...
[12:43:18 CEST] <t4nk146> I hate to keep being a bother, but where would I find this options?
[12:44:07 CEST] <iive> t4nk146: wait
[12:44:19 CEST] <iive> t4nk146: i think i might have a fix, if you want to test it
[12:44:34 CEST] <t4nk146> yes please
[12:47:33 CEST] <iive> http://pastebin.com/ain9hqDx
[12:50:22 CEST] <t4nk146> That's a script aint it?
[12:50:56 CEST] <t4nk146> Sorry, I'm a real noob
[12:51:46 CEST] <kapusta> Hi there =)
[12:53:47 CEST] <kapusta> im a total noob when it comes down to ffmpeg and I can't seem to find any info on this. Anyone knows whether ffmpeg is cappable of splitting TGA files into seperate key and fill files?
[12:54:58 CEST] <kapusta> I know it can encode videos from key and fill TGA sequences but I'm looking to split them to Alpha and RGB basically and google doesn't really help much
[12:58:40 CEST] <t4nk146> https://software.intel.com/en-us/forums/intel-media-sdk/topic/560169 I feel a littl better that it's not just me with the problem.
[13:00:29 CEST] <iive> t4nk146: it's a patch
[13:00:44 CEST] <t4nk146> how do I utilise it?
[13:01:00 CEST] <iive> save it to a file e.g. mfx.patch
[13:01:17 CEST] <iive> then go into ffmpeg source directory `cd ffmpeg-2.8/`
[13:01:43 CEST] <iive> and do `patch -p1 --dry-run < mfx.patch`
[13:01:59 CEST] <iive> if it doesn't complain for errors, remove the --dry-run
[13:04:53 CEST] <t4nk146> Nope
[13:05:02 CEST] <t4nk146> problem still persists :'(
[13:06:31 CEST] <iive> you did run it without the --dry-run ?
[13:06:46 CEST] <t4nk146> yeah
[13:07:29 CEST] <t4nk146> Hunk #1 succeeded at 915 (offset -8 lines)
[13:07:58 CEST] <iive> then, i'm out of ideas
[13:08:18 CEST] <t4nk146> me too
[13:08:25 CEST] <t4nk146> Thanks anyway
[13:08:29 CEST] <iive> can you give pastebin of config.log portion?
[13:08:45 CEST] <iive> you did run configure again, didn't you?
[13:09:45 CEST] <t4nk146> yeah ./configure
[13:13:31 CEST] <iive> ok, i want to see if the gcc gets the same options.
[13:13:57 CEST] <t4nk146> http://pastebin.com/1rtBxpDC
[13:14:05 CEST] <t4nk146> different errors I think
[13:15:57 CEST] <iive> it's the same...
[13:18:33 CEST] <iive> open the configure file with text editor and try to find the "check_ld" function . it should be around 922 line (+-8)
[13:18:44 CEST] <iive> see if *.a are present in it.
[13:29:30 CEST] <t4nk146> live 922
[13:29:31 CEST] <t4nk146> libs=$($ldflags_filter $libs)
[13:32:38 CEST] <t4nk146> both *.so
[13:33:19 CEST] <t4nk146> what would I need to add, if I delete .so and replace it with .a it will cause lots of problems right?
[13:36:40 CEST] <alo> AVStream::pts': was declared deprecated
[13:36:45 CEST] <alo> someone can help me?
[13:46:06 CEST] <oreWizard> PICTURE_FORMAT
[13:49:28 CEST] <oreWizard> avcodec_encode_video deprecated?
[13:52:46 CEST] <iive> t4nk146: look at my patch
[13:53:02 CEST] <iive> it adds the |*.a there
[13:53:08 CEST] <BtbN> oreWizard, yes, avcodec_encode_video is deprecated, if that's a question.
[13:53:18 CEST] <iive> in patch, lines that start with space are not changed.
[13:53:20 CEST] <t4nk146> it has
[13:53:31 CEST] <t4nk146> but still the same error
[13:53:52 CEST] <iive> well, somebody smarter should have to look at the issue
[13:54:02 CEST] <t4nk146> I have uninstalled and re-installed libmfx, and I'm paying closer attention
[13:54:38 CEST] <t4nk146> "/usr/local/lib/libmfx.a" is where the file is stored
[13:55:41 CEST] <iive> you need to get libmfx.so ...
[13:58:42 CEST] <t4nk146> I can't find any reference to an .so
[13:59:34 CEST] <t4nk146> https://bbs.archlinux.org/viewtopic.php?id=198935 I can't decipher much of this
[13:59:43 CEST] <t4nk146> but same issue with the .a file
[14:07:28 CEST] <iive> t4nk146:  would you run configure again, this time with bash debugging enabled. start it like this:
[14:07:39 CEST] <iive> bash -x ./configure
[14:08:12 CEST] <iive> and add your options too. you might want to redirect the output to a file.
[14:14:23 CEST] <t4nk146> http://pastebin.com/g14k8VY8
[14:21:35 CEST] <iive> t4nk146: hum... the executed commands are different than these in your last config.log pastebin.
[14:22:04 CEST] <iive> are you sure you haven't given my the old one, by mistake?
[14:22:20 CEST] <t4nk146> No
[14:38:27 CEST] <iive> t4nk146: upload new config.log file, and make sure it is new.
[14:42:02 CEST] <iive> t4nk146: i think i have something else for you to try
[14:50:31 CEST] <iive> t4nk146: i moved only the *.a, but I missed to move the -L that hints where the *.a file is
[14:51:55 CEST] <t4nk146> Ahh, what should I do?
[14:52:12 CEST] <iive> +    flags=$(filter_out '-l*|*.so|-L*|*.a' $@)
[14:52:13 CEST] <iive> +    libs=$(filter '-l*|*.so|-L*|*.a' $@)
[14:52:37 CEST] <iive> these lines should look like that. (without the + at the start)
[14:58:03 CEST] <t4nk146> Nope
[14:58:09 CEST] <t4nk146> still the same issue
[14:58:31 CEST] <t4nk146> Live, could I not give you some remote desktop?
[15:06:00 CEST] <iive> i'd better not...
[15:08:22 CEST] <t4nk146> Are you sure? This is a fresh install, and I am not worried about losing anything
[15:21:20 CEST] <iive> i'm out of ideas.
[15:22:56 CEST] <t4nk146> I'm going to give arch a go as they have a binary for libmfx, and I just want to see whether QS on Linux is worth bothering with
[16:00:57 CEST] <oreWizard> url_fclose deprecated
[16:03:59 CEST] <c_14> oreWizard: https://git.videolan.org/?p=ffmpeg.git;a=blob;f=doc/APIchanges;h=958035ce777d8d5af4d03a7c34fff7f58ec08a96;hb=HEAD#l1723
[16:19:08 CEST] <fred1807> how can I get the m3u8 of this live stream?  http://www.worldsurfleague.com/events/2015/mct/1281/hurley-pro-at-trestles/live
[17:04:37 CEST] <fred1807> votz: Event is live now! How do I find the playlist file m3u8?  http://www.worldsurfleague.com/events/2015/mct/1281/hurley-pro-at-trestles/live
[17:06:38 CEST] <votz> fred1807: http://nlds299.cdnak.neulion.com/nlds/wsl/t107638/as/live/t107638_hd_4500.m3u8
[17:08:50 CEST] <fred1807> votz: hmm does is works for you?  I only see 1 frame...
[17:09:15 CEST] <votz> fred1807: works for me in VLC, though it hiccups between requested video segments
[17:09:44 CEST] <votz> likely my ancient VLC doesn't prefetch the next segment, thus when one segment ends, it requests the next, hence the hiccup
[17:11:40 CEST] <fred1807> maybe this will work? ? http://docs.livestreamer.io/cli.html#playing-built-in-streaming-protocols-directly
[17:12:14 CEST] <votz> fred1807: only one way to find out. looks promising
[17:12:49 CEST] <fred1807> how do you get eh m3u8?
[17:13:01 CEST] <votz> though if it works for me in VLC (with hiccups), I wonder why it doesn't work for you
[17:14:17 CEST] <votz> Get a .mp4 from the website
[17:14:18 CEST] <votz> http://nlds299.cdnak.neulion.com/nlds/wsl/t107638/as/live/t107638_hd_1600/20150911/15/1352.mp4?nltid=wsl&nltdt=0
[17:14:42 CEST] <votz> truncate the date formatting portion of the, i.e. '20150911/15/1352.mp4'
[17:14:51 CEST] <votz> and add .m3u8
[17:14:55 CEST] <votz> so http://nlds299.cdnak.neulion.com/nlds/wsl/t107638/as/live/t107638_hd_1600/20150911/15/1352.mp4?nltid=wsl&nltdt=0
[17:15:03 CEST] <votz> -> http://nlds299.cdnak.neulion.com/nlds/wsl/t107638/as/live/t107638_hd_1600.m3u8
[17:16:08 CEST] <fred1807> hmm, nice!
[17:21:14 CEST] <fred1807> votz: do you have a raspberry?  omxplayer with --live option plays that .m3u8 perfect!
[17:21:24 CEST] <votz> fred1807: I do not
[17:39:37 CEST] <satiender> hi can we compile the ecoding code of ffmpeg outside from ffmpeg directory
[17:39:48 CEST] <satiender> I shared a link of that code
[17:39:52 CEST] <satiender> https://www.ffmpeg.org/doxygen/2.1/doc_2examples_2decoding_encoding_8c-example.html
[17:40:10 CEST] <satiender> please give some idea about use the ffmpeg API
[17:50:39 CEST] <satiender> can we use life=s=640x480:mold=10:r=100:ratio=0.1:death_color=blue:life_color=#00ff00,boxblur=2:2 for add effects in my video ??
[17:56:25 CEST] <durandal_170> satiender: that is video source
[17:57:25 CEST] <satiender> yes
[17:58:13 CEST] <satiender> durandal_170 : i want that kind of effect in  my video how I can use that
[17:58:20 CEST] <satiender> ??
[18:00:25 CEST] <durandal_170> You can overlay it with your video
[18:01:36 CEST] <satiender> yes
[18:02:05 CEST] <satiender> can I modify something in above command
[18:02:19 CEST] <satiender> or use as its with overlay
[18:02:47 CEST] <satiender> durandal_170: ??
[18:03:00 CEST] <durandal_170> yes, you can, depends what you need
[18:12:57 CEST] <fred1807> votz: Do you know how to find the parent multi-bitrate m3u8 ?
[18:13:14 CEST] <votz> fred1807: Not off hand, not.
[18:13:15 CEST] <votz> *no
[18:18:42 CEST] <satiender> durandal_170 : I get errors
[18:19:02 CEST] <satiender> you already specify a input
[18:19:08 CEST] <satiender> something like that
[18:20:43 CEST] <durandal_170> satiender: because its source your changes are incorrect
[18:29:38 CEST] <jasom> so, I have a program stream that I've been having some difficulties with demuxing, so I just tried a -map 0 -c copy -f vob to see what would happen, and I got "Could not write header for output file #0 (incorrect codec parameters ?): Operation not permitted"
[19:10:34 CEST] <ericwooley> I am trying to use ffmpeg to caputre a device, and spit out the last 30 seconds of video every 30 seconds. This is what I have so far
[19:10:35 CEST] <ericwooley> ffmpeg -f avfoundation -i "0" -vf scale=640:360  -c:v libvpx-vp9 -r 24 test.webm
[19:11:18 CEST] <ericwooley> any ideas to make ffmpeg split out a chunk of x amount of seconds?
[19:11:27 CEST] <c_14> segment muxer
[19:11:35 CEST] <ericwooley> I will google that, thanks
[19:11:46 CEST] <c_14> https://ffmpeg.org/ffmpeg-formats.html#segment_002c-stream_005fsegment_002c-ssegment
[19:16:31 CEST] <ericwooley> do you know of a good tutorial on how I would be able to use that? I tried
[19:16:33 CEST] <ericwooley> ffmpeg -f segment -segment_times 20 -f avfoundation -i "0" -vf scale=640:360  -c:v libvpx-vp9 -r 24 test.webm
[19:16:53 CEST] <ericwooley> c_14: I got this error: Option segment_times not found.
[19:16:54 CEST] <c_14> Where you put options matter
[19:17:03 CEST] <c_14> Those belong after -i "0" and before test.webm
[19:17:31 CEST] <ericwooley> ok, and I can use two -f
[19:17:34 CEST] <ericwooley> ?
[19:18:03 CEST] <c_14> ffmpeg -f avfoundation -i "0" -vf scale=640:360 -c:v libvpx-vp9 -r 24 -segment_times 20 -f segment test.webm
[19:18:07 CEST] <c_14> eh
[19:18:12 CEST] <c_14> test%d.webm
[19:18:25 CEST] <c_14> And you'll probably want to use segment_wrap or something
[19:18:28 CEST] <ericwooley> rad, I see what you mean. Thanks
[19:21:15 CEST] <ericwooley> c_14: I simplified to this `ffmpeg -f avfoundation -i "0" -vf scale=640:360 -segment_times 20 -f segment test.webm` and got `Could not write header for output file #0 (incorrect codec parameters ?): Invalid argument`
[19:21:51 CEST] <c_14> I think you need the %d
[19:22:24 CEST] <c_14> Otherwise it can't find the segment to use
[19:25:17 CEST] <ericwooley> perfect thanks, seems to have worked. I have one more question about libvp8, it keeps saying
[19:25:20 CEST] <ericwooley> [avfoundation @ 0x7fb14880da00] Selected pixel format (yuv420p) is not supported by the input device.
[19:25:21 CEST] <ericwooley> [avfoundation @ 0x7fb14880da00] Supported pixel formats:
[19:25:21 CEST] <ericwooley> [avfoundation @ 0x7fb14880da00]   uyvy422
[19:25:22 CEST] <ericwooley> [avfoundation @ 0x7fb14880da00]   yuyv422
[19:25:24 CEST] <ericwooley> [avfoundation @ 0x7fb14880da00]   nv12
[19:25:26 CEST] <ericwooley> [avfoundation @ 0x7fb14880da00]   0rgb
[19:25:28 CEST] <ericwooley> [avfoundation @ 0x7fb14880da00]   bgr0
[19:25:30 CEST] <ericwooley> [avfoundation @ 0x7fb14880da00] Overriding selected pixel format to use uyvy422 instead.
[19:25:34 CEST] <ericwooley> is there anything I can do about that?
[19:26:05 CEST] <c_14> You could explicitly chose an input pixel format
[19:26:10 CEST] <c_14> *choose
[19:26:22 CEST] <c_14> not that it really matters
[19:26:30 CEST] <ericwooley> yea, I have been lookign for a while, and I am not sure how to do that
[19:26:53 CEST] <ericwooley> sorry, I am overwhelmed with options on this
[19:27:07 CEST] <ericwooley> or would that not be a libvpx option, but just a regular ffmpeg?
[19:27:11 CEST] <c_14> -pix_fmt yuyv422p -i "0" for example
[19:27:18 CEST] <c_14> just a regular ffmpeg option
[19:31:02 CEST] <ericwooley> k, do you know of any guides or anything on option placement. I think I am just becoming more confused
[19:31:13 CEST] <ericwooley> c_14: that worked great btw
[19:31:31 CEST] <DX099> hi everyone, I'm a complete noob in programming and is trying to figure out a way to obtain ffmpeg version so that I can do some #define based on it. My end goal is to write a patch for a project to be able to switch dynamically to the some newly renamed variables
[19:31:57 CEST] <c_14> ericwooley: you can read the manpages and look at the examples, but you basically just put the option in front of the thing you want it to modify
[19:32:17 CEST] <c_14> DX099: avutil.h get_version() or something
[19:32:18 CEST] <ericwooley> ah, that one statement actually makes a ton of sense
[19:32:34 CEST] <ericwooley> thank you, ill read the man more closely
[19:33:33 CEST] <c_14> DX099: actually, all the libraries have a [library_name]_version(void); function in their header
[19:34:29 CEST] <DX099> http://hastebin.com/isusuhewun.xml
[19:34:54 CEST] <DX099> c_14: this is what I've done: http://hastebin.com/isusuhewun.xml
[19:35:01 CEST] <DX099> but it fails...
[19:36:25 CEST] <c_14> DX099: can you upload to somewhere that isn't hastebin? I'm only getting a blank (blue) page. (something like sprunge or ix or pastie or pastebin or something)
[19:37:46 CEST] <DX099> c_14: hah, true that hastebin needs its JS
[19:38:23 CEST] <c_14> I tried enabling it, but it still didn't like me.
[19:40:35 CEST] <DX099> c_14: http://pastie.org/10412959
[19:44:34 CEST] <c_14> DX099: http://pb.c-14.de/t/kng.wPJHkL
[19:53:31 CEST] <feliwir> hey i asked yesterday what do about the Unknown OS error
[19:55:16 CEST] <DX099>  c_14, thanks it worked! Now to identify the version before a bunch of variables were prefixed with "AV_" string...
[19:57:30 CEST] <feliwir> this one: Unknown OS 'msys_nt-10.0'.
[19:57:31 CEST] <DelphiWorld> hi all
[19:57:43 CEST] <DelphiWorld> maybe a stupid unreasonable question for someone, but...
[19:57:46 CEST] <feliwir> i run configure from msys shell
[19:57:48 CEST] <DelphiWorld> should i use aac-lc or aac-he?
[19:58:15 CEST] <c_14> DelphiWorld: depends on the bitrate. High bitrate is lc and low bitrate is he (afaik)
[19:58:33 CEST] <DelphiWorld> c_14: i wan high good quality
[19:59:56 CEST] <DelphiWorld> c_14: is google stupid enough converting musics into aac from mp3?
[20:00:36 CEST] <c_14> DelphiWorld: actually you probably want AAC Main
[20:00:36 CEST] <DHE> for youtube?
[20:00:48 CEST] <feliwir> noone wants to answer my noob question? :(
[20:00:57 CEST] <DelphiWorld> c_14: main?
[20:01:03 CEST] <c_14> feliwir: msys2/mingw64 ?
[20:01:09 CEST] <DelphiWorld> aac-lc?
[20:01:14 CEST] <feliwir> c_14, yes
[20:01:34 CEST] <feliwir> msys2 shell
[20:01:36 CEST] <c_14> DelphiWorld: If you require a low audio bitrate, such as d 32kbs/channel, then AAC-HE would be worth considering if your player or device can support AAC-HE decoding. Anything higher may benefit more from AAC-LC due to less processing. If in doubt use AAC-LC. All players supporting AAC-HE also support AAC-LC.
[20:02:11 CEST] <DelphiWorld> c_14: LOL i know that ;)
[20:02:17 CEST] <DelphiWorld> i'm not talking about compatibility
[20:02:26 CEST] <DelphiWorld> but talking about mather of quality
[20:02:34 CEST] <c_14> DelphiWorld: it mentions that as well
[20:02:38 CEST] <c_14> Low bitrates HE otherwise LC
[20:02:45 CEST] <DelphiWorld> ok, so lc
[20:02:55 CEST] <DelphiWorld> and yay libfdk_aac is my frien
[20:02:57 CEST] <DelphiWorld> and yay libfdk_aac is my friend
[20:03:48 CEST] <c_14> feliwir: I've heard there were some issues. You can either try modifying configure to treat mingw64 as mingw32 or modify the msys batch script so it exports some variable differently.
[20:04:03 CEST] <feliwir> c_14, why is there an OS check at all?
[20:05:03 CEST] <feliwir> it does nothing but causing trouble
[20:05:46 CEST] <DX099> feliwir: because ffmpeg is cross-platform AFAIK
[20:06:05 CEST] <c_14> feliwir: because some things aren't supported on some platform.s Because headers are in nonstandard locations
[20:06:16 CEST] <feliwir> c_14, ok
[20:06:29 CEST] <feliwir> is there an option to force an OS?
[20:06:44 CEST] <DX099> c_14: by the way what was that 'extern "C"' thing?
[20:06:55 CEST] <c_14> feliwir: --target-os iirc
[20:07:01 CEST] <c_14> DX099: how to fix your problem
[20:07:06 CEST] <llogan> feliwir: possibly related to the unstable mingw-w64 trunk talk i saw in ffmepg-devel
[20:07:21 CEST] <DX099> c_14: yes but from a programming standpoint, what does it do?
[20:07:28 CEST] <c_14> The ffmpeg libraries are C code, if you want to link to them from C++ you need some magic.
[20:07:37 CEST] <DX099> ahhh
[20:07:44 CEST] <DX099> thanks for explaining
[20:07:57 CEST] <llogan> rcombs: looks like feliwir has the same issue as you with 'msys_nt-10.0'. hwo did you resolve it?
[20:08:26 CEST] Action: llogan knows nothing of msys2 'n junk
[20:08:39 CEST] <rcombs> llogan: I just disabled the relevant stuff
[20:09:01 CEST] <rcombs> didn't actually need it
[20:09:09 CEST] <rcombs> as for the platform name, I tweaked configure
[20:09:22 CEST] <rcombs> made it treat `msys64*` the same as `msys32*`
[20:09:42 CEST] <feliwir> ohm
[20:09:53 CEST] <feliwir> can you send me that configure file perhaps?
[20:11:10 CEST] <rcombs> https://gist.github.com/244bbcb81e7721bc8d0d
[20:11:29 CEST] <rcombs> might be worth actually applying in the repo
[20:11:29 CEST] <feliwir> thanks
[20:14:25 CEST] <feliwir> would make things easier atleast
[21:19:08 CEST] <tlhiv_laptop> all examples that i have seen regarding crossfading two videos seem extremely complicated ... is there a simple way to concatenate two MP4 videos (containing audio) by crossfading from the first video to the second one?
[21:27:48 CEST] <Canar> http://superuser.com/questions/778762/crossfade-between-2-videos-using-ffmpeg
[21:27:59 CEST] <Canar> seems lengthy, but one-line and pretty straightforward
[21:31:46 CEST] <tlhiv_laptop> i've read this ... i didn't understand "at 4th second during 1 second" ... i'm not sure what that means
[21:32:56 CEST] <tlhiv_laptop> and i'm wondering if re-encoding is required for performing this crossfade
[21:33:17 CEST] <c_14> yes
[21:33:19 CEST] <c_14> reencoding is required
[21:33:59 CEST] <tlhiv_laptop> :-/
[21:36:16 CEST] <tlhiv_laptop> i am creating a video from a presentation that i am making, and each "slide" of the presentation has been transformed into its own audio/video MP4 ... i will ultimately need to concatenate about 60 or so MP4's which is easy without crossfading but seems quite a bit more complicated when requiring crossfading
[21:36:44 CEST] <c_14> yeah, ffmpeg isn't good at crossfading
[21:37:03 CEST] <tlhiv_laptop> i would rather not have to re-encode all of these FFMPEG created mp4's
[21:37:19 CEST] <tlhiv_laptop> c_14: by "isn't good" do you mean quality or ease of use?
[21:37:26 CEST] <c_14> ease of use
[21:39:56 CEST] <tlhiv_laptop> in some sense, i presume it could be done by never crossfading more than two at a time because i could always crossfade 1+2 and get an intermediate video "a", crossfade a+3, get "b", crossfade b+4, and get "a", etc.
[21:41:01 CEST] <tlhiv_laptop> since re-encoding is required, that would be no worse than doing them all at once
[21:43:59 CEST] <BtbN> crossfading just doesn't realy fit into the filter pipeline ffmpeg has.
[21:44:03 CEST] <tlhiv_laptop> perhaps that's not the best plan since video 1 would be re-encoded 59 times :-(
[21:44:17 CEST] <BtbN> Just use something lossless untill you're done.
[22:04:02 CEST] <Laurenceb__> hi
[22:04:11 CEST] <Laurenceb__> I have a problem extracting frames from a video
[22:04:17 CEST] <Laurenceb__> all my frames have zero size
[22:04:20 CEST] <Laurenceb__> can anyone help?
[22:05:35 CEST] <Laurenceb__> http://pastie.org/10413225
[22:10:03 CEST] <klaxa> your ffmpeg version is ancient
[22:10:21 CEST] <klaxa> it's from SVN, wow, wouldn't have thought i'd see that again
[22:10:30 CEST] <klaxa> update your ffmpeg and try again
[22:10:36 CEST] <BtbN> wasn't there recently one from 2004, which still mentioned CVS?
[22:18:01 CEST] <feliwir> somewhere here has a clue about the rust progrmaming language? I try to compile ffmpeg from it like this: http://pastebin.com/n891eC6g
[22:18:21 CEST] <feliwir> I get this compile error though: libavdevice/libavdevice.a: 2: libavdevice/libavdevice.a: Syntax error: newline unexpected
[22:47:16 CEST] <BtbN> Why oO
[23:00:02 CEST] <feliwir> BtbN, why?
[23:40:02 CEST] <tlhiv_laptop> BtbN: what is a lossless video format?
[23:50:44 CEST] <c_14> tlhiv_laptop: ffv1, ffvhuff, huffyuv, libx264 has a lossless mode
[00:00:00 CEST] --- Sat Sep 12 2015