[Ffmpeg-devel-irc] ffmpeg.log.20150407

[burek]

[00:00:07 CEST] <DragonsLord> May I ask you why?
[00:00:31 CEST] <c_14> The protocol is the same thing as doing cat video01.avi video02.avi video03.avi > video0x.avi
[00:01:56 CEST] <DragonsLord> So something like this > ffmpeg -f concat -i mylist.txt -c copy output <
[00:02:11 CEST] <c_14> yep
[00:06:32 CEST] <pyBlob> is there a way of outputting audio data in pcm_s16le format preceded by a length field?
[00:07:40 CEST] <pyBlob> using "-f wav" generates a wav header that would be needed to parse, and "-f s16le" doesn't store a length field
[00:12:14 CEST] <seasc> The new PKG_CONFIG_PATH and LD_LIBRARY_PATH made it possible to build ffmpeg with --enable-libx265, HOWEVER, see: http://pastebin.com/PXQRF85c
[00:13:16 CEST] <pyBlob> and why does the created wav file start with "-f wav" start with "RIFF\xc3"+<4byte-length> instead of "RIFF"+<4byte-length>?
[00:13:57 CEST] <c_14> seasc: did you export LD_LIBRARY_PATH to the correct dir?
[00:16:34 CEST] <seasc> c_14, i have LD_LIBBARY_PATH to the dir with the 'custom' codecs, x265 however is installed in /usr/local/lib, because of ccmake installation... LD_LIB_P is set to $CHROOT/usr/lib, not /usr/local/lib (but that is added as -L/usr/local/lib as --extra-ldflags)
[00:17:10 CEST] <c_14> export LD_LIBRARY_PATH=customlibthingy:/usr/local/lib
[00:17:14 CEST] <seasc> can i set multipla paths just like for PKG_CONFIG?
[00:17:17 CEST] <c_14> yep
[00:17:22 CEST] <seasc> ok
[00:18:13 CEST] <seasc> isnt it the same after all? i mean, is there a difference between --extra-ldflags=-L/path1 and export LD_LIB_P=LD_LIB_P:path1 ?
[00:18:39 CEST] <c_14> The --extra-ldflags is only useful during compliation, LD_LIBRARY_PATH is used when executing the resulting binary
[00:18:52 CEST] <c_14> They're "completely" unrelated
[00:18:54 CEST] <seasc> Ohhh
[00:21:56 CEST] <c_14> (this also means that you only need to set the LD_LIBRARY_PATH when you run the resulting binary, not when compiling)
[00:22:00 CEST] <seasc> ok, last compile run for 'tdoay'... moved all --extra-ldflags to LD_LIBRARY_PATH, lets see.
[00:22:19 CEST] <seasc> err uh?
[00:22:40 CEST] <c_14> The linker uses the ldflags to link the object files.
[00:22:45 CEST] <seasc> Cant i make the binary to use those paths by its own default?
[00:22:45 CEST] <c_14> That's what the --extra-ldflags is fore
[00:22:56 CEST] <c_14> No
[00:23:10 CEST] <seasc> ah wait, but the LD_LIB_PATH does so?
[00:23:14 CEST] <c_14> The binary on execution looks in the paths listed by LD_LIBRARY_PATH (and some defaults)
[00:23:20 CEST] <seasc> or are both paths to be set before exe the bin?
[00:23:46 CEST] <c_14> You only need the ldflags when compiling and you only need the LD_LIBRARY_PATH when you execute/run the resulting binary[ies]
[00:24:52 CEST] <seasc> and if LD_LIBRARY_PATH if set upon compiling, i have to set that to run the result, anyway? No way to 'avoid' that?
[00:25:01 CEST] <smo_> if i have a Video: mpeg4 (Advanced Simple Profile) (XVID / 0x44495658) can i only use the mpeg4 decoder to decode it ?
[00:25:33 CEST] <c_14> seasc: setting it on compilation is a noop, it only takes effect when running the result
[00:25:43 CEST] <seasc> smo_, ffmpeg -i video.mp4 output.mkv
[00:26:16 CEST] <c_14> smo_: most likely yes
[00:26:18 CEST] <smo_> i know seasc thnaks :) just want to know if i can force another decoder :p ?
[00:26:28 CEST] <seasc> smo_, meaning, ffmpeg should have mp4 capabilities included, so... for 'most' videos, yes
[00:28:09 CEST] <c_14> smo_: if there is another decoder for it, yes
[00:28:11 CEST] <smo_> still my chromium boring, since if i make a live transoding of this file type with copy for audio and video the video is totally normal, very fluid but if i play it with the libffmpegsumo.so it has artefacts :/
[00:28:15 CEST] <c_14> ffmpeg -c codec -i video
[00:29:04 CEST] <smo_> i see but i have to find how chromium do to choose the codec for decoding the video
[00:29:32 CEST] <smo_> which other codec can i try for this file type c_14 ?
[00:29:39 CEST] <smo_>  mpeg4 (Advanced Simple Profile) (XVID...
[00:30:41 CEST] <c_14> I don't know of another decoder within ffmpeg that handles mpeg4. Maybe libxvid if you have that compiled in.
[00:33:01 CEST] <seasc> c_14, YEAH, though i dont like the required workaround... it works! :) thanks alot for your patience and explanations!
[00:33:25 CEST] <smo_> seems i hvae this problem only with files where i have the message "Video uses a non-standard and wasteful way to store B-frames no way to find workaround for it ?
[00:33:54 CEST] <c_14> reencode the video?
[00:34:15 CEST] <smo_> (without reencoding, that s the problem)
[00:34:59 CEST] <smo_> arf, don t know why peoples still encode their videos with avi and mpeg4... :)
[00:37:01 CEST] <seasc> because they just copy the streams to another container, because its faster than the actual encoding
[00:43:31 CEST] <smo_> np i ll force reencoding only for avi files ... :) thanks
[00:49:51 CEST] <adrian_1908> hello, what command would you recommend for overlaying a PNG input sequence with transparency onto a plain white background? I want to do it on-the-fly, because of harddrive space restrictions.
[00:50:20 CEST] <adrian_1908> (the output will be x264 video)
[00:51:25 CEST] <c_14> ffmpeg -i thing%03d.png -filter_complex 'color=white[w];[0][w]overlay[v]' -map '[v]' out.mkv
[00:51:45 CEST] <c_14> https://ffmpeg.org/ffmpeg-filters.html for docs on each of the options for the filters
[00:52:27 CEST] <adrian_1908> c_14: thank you, that looks promising.
[00:54:26 CEST] <llogan> adrian_1908: it may look better with overlay=format=rgb
[00:55:33 CEST] <adrian_1908> llogan: in place of 'color=white[...', correct?
[01:06:15 CEST] <llogan> adrian_1908: append =format=rbg to your overlay filter
[01:06:29 CEST] <adrian_1908> llogan: yeah, i noticed, sorry.
[02:11:14 CEST] <Prelude2004c> hey guys.. looking for a developer to help me port NVENC to work with AMD cards or build a process that will handle the work of multiple sessions and only send 1 or 2 sessions to NVidia cards ( overcome session limit )  .. Anyone interested ? If so please PM me
[05:25:36 CEST] <YaMoonSun> Anyone know where I can find a video tutorial of creating your own ffmpeg build? I need to compile with libfdk_aac
[06:52:19 CEST] <relaxed> YaMoonSun: did you see https://trac.ffmpeg.org/wiki/CompilationGuide ?
[08:01:00 CEST] <ScottSteiner> Is it possible to combine two videos (including audio) to have them playing side by side without resizing the input videos?
[08:44:49 CEST] <anshul_mahe> how to make ffmpeg listen for rtsp stream, just not I need to make ffplay listen on my rtsp stream
[08:45:21 CEST] <anshul_mahe> I want fffmpeg to listen and ffplay can play the live stream any time it wants
[08:48:47 CEST] <anshul_mahe> here is my command http://pastebin.com/kG71BKQQ and logs
[08:49:12 CEST] <anshul_mahe> I am trying to set timeout value, but ffmpeg has ignored it
[10:14:15 CEST] <dude42> Good day, I have two video files, both contain a video steam encoded with h264. However one play smoothly when streaming from a mediatomb dlna server using vlc on my phone, the other is choppy. The weird thing is, the choppy video has a lower bitrate than the smooth video according to ffprobe. Why could this be ?
[10:16:35 CEST] <BtbN> Broken video, broken device, bad hardware decoder, bad player software, and every possible combination of those.
[10:18:53 CEST] <dude42> BtbN, if I stream from the same dlna server using a faster machine like a laptop, the video is no longer choppy
[10:19:35 CEST] <dude42> I just feel that if the higher bitrate video could stream smoothly on my phone with the h264 encoding, then surely the lower bitrate video should also be smooth ?
[10:20:05 CEST] <BtbN> sounds like your dlna server does some transcoding, and is too slow to handle it.
[10:21:25 CEST] <Keshl> Agreed. While on the topic, by the way, I've had cases where lower bitrates played locally are choppier than higher bitrates and vice-versa. Bitrate honestly doesn't have as much baring on how smoothly video plays as most people think. -É-.
[10:22:11 CEST] <dude42> Thank you, I just streamed from my laptop and the lower bitrate video was not choppy, so it can't be the server
[10:23:17 CEST] <Keshl> It sounded like you said that you switched which system was hosting the server to a faster one, and that fixed it. If that wasn't the case, then re-read what I said after BtbN. oÉo.
[10:23:22 CEST] <dude42> What does encoder: Lavf56.25.101 mean ?
[10:24:07 CEST] <BtbN> That this file was muxed by libavformat 56.25.101
[10:25:34 CEST] <dude42> Keshl, the machine hosting the dlna server never changed, and the lower bitrate video streams smoothly on my laptop
[10:25:39 CEST] <dude42> Thank BtbN
[10:25:58 CEST] <dude42> If the bitrate does not play a role and the same codec was used, what does play a role ?
[10:26:08 CEST] <Keshl> The phase of the moon, pretty much.
[10:26:43 CEST] <Keshl> Seriously, there's so many variables in play that it's next to impossible to just go "Okay, if we use /this/ encoder and /that/ container, it'll play just fine on every device ever".
[10:27:52 CEST] <Keshl> If your phone's running Android, the first thing you should try doing is grabbing alternative video players. There's two really common ones almost everyone uses; BSPlayer and.. Something, else, that starts with an A that I forgot the name of. x.x
[10:28:01 CEST] <dude42> would aspect ratios make a difference ?
[10:28:07 CEST] <Keshl> For some reason, the ones that come with phones are absolutely horrible.
[10:28:20 CEST] <dude42> I am using mediahouse and it plays through vlc on android
[10:28:36 CEST] <Keshl> Seriously, go grab BSPlayer. <É<
[10:28:47 CEST] <dude42> OK I'll give it a shot
[10:28:50 CEST] <dude42> Thanks
[10:29:02 CEST] <Keshl> Aspect ratio could make a difference, but that depends vastly on everything BtbN said earlier. oÉo.
[10:29:23 CEST] <dude42> The only reason I'm hunting this is because the video that looks a LOT better runs smoothly
[10:29:32 CEST] <dude42> I'll try the other players
[10:29:53 CEST] <Keshl> Oh thank goodness. x.x
[10:30:11 CEST] <Keshl> Sorry, I assumed you were one of those "Why doesn't my video play perfectly everywhere? RAWR!" guys. x.x
[10:30:50 CEST] <dude42> I am :D but I'm not unwilling to eliminate the player just yet
[10:30:58 CEST] <Keshl> ... x.x
[10:31:05 CEST] <Keshl> I'll cut to the chase -- It can't.
[10:31:42 CEST] <Keshl> Minor differences in hardware mean that certain encoders, paramaters in those encoders, and /general video content/ will play differently. You /literally can't make a video play perfectly everywhere/.
[10:32:14 CEST] <Keshl> Different players do this, too, making specific players work better on specific hardware, further narrowing what'll actually play well.
[10:32:50 CEST] <Keshl> If you're streaming, it gets even more stupid-complicated, because, yep, different hardware can be tuned for different protocols, which narrows it even more.
[10:33:41 CEST] <dude42> Ok, but one video works with exactly the same setup and codec, so I'm thinking it's the video somehow.
[10:33:57 CEST] <Keshl> It can be. I said "general video content will play differently". <É<
[10:34:21 CEST] <Keshl> See, videos aren't actually moving. Nothing is.
[10:34:55 CEST] <dude42> Can bsplayer actually stream dlna ?
[10:34:59 CEST] <Keshl> Videos are just made up of individual, non-moving pictures. They're shown for such brief periods of time, though, that it fools people into thinking they're moving. Generally, it's as low as 24 images per second, though it can go as high as 60 for most devices.
[10:35:02 CEST] <Keshl> (Not sure.)
[10:35:48 CEST] <Keshl> Now, if you do the math, for a ten second video at 24 frames per second, judging on typical pictures you take with your phone's camera, you might wonder why they don't turn out to be around 240 megs for a ten second video.
[10:36:09 CEST] <Keshl> Even if you scale it down, five-minute videos should logically be a few gigs in size.
[10:36:55 CEST] <Keshl> The reason they're not is because the file, be it an AVI, MP4, or what have you, is constructed /knowing it'll be a video/. Videos generally aren't random images -- They have patterns from one frame to the next.
[10:37:23 CEST] <Keshl> So, the frames in a video just contain the /differences/, not the entire image.
[10:37:47 CEST] <dude42> is that what the codecs do ?
[10:37:48 CEST] <Keshl> That's the crux of all video encoding ever -- Find patterns from frame to frame, and somehow record them.
[10:37:52 CEST] <dude42> generate the differences ?
[10:37:56 CEST] <Keshl> No, encoders do that.
[10:38:13 CEST] <dude42> well the encoder is part of codec
[10:38:18 CEST] <Keshl> Codecs specify what patterns are allowed to be detected, in a nutshell. (This is _very_ over-simplified, but it works.)
[10:39:47 CEST] <Keshl> In the olden days, no hardware was made to decode video. At all. It was all software, all the time.
[10:40:13 CEST] <Keshl> Time went on and, eventually, people realized that people watch so many videos that it'd be useful to make specific areas of CPU's dedicated to decoding them.
[10:40:33 CEST] <Keshl> That said, these areas only work on specific codecs, and /only if you use the correct paramaters that they expect/.
[10:41:08 CEST] <dude42> yaaaaay I fixed it
[10:41:13 CEST] <dude42> my OCD is satisfied
[10:41:14 CEST] <Keshl> To make matters worse, they don't all follow the same ones. Even two chips from the same manufactorer might expect different parameters.
[10:41:38 CEST] <dude42> don't ask me why but ffmpeg -i file.mkv out.mkv
[10:41:44 CEST] <dude42> re encoded it smaller
[10:41:48 CEST] <dude42> and now its smooth
[10:41:58 CEST] <Keshl> Shiny. -É-.
[10:42:04 CEST] <dude42> maybe the original just had an older version of the codec
[10:42:14 CEST] <Keshl> It can't "have" a codec. xwx.
[10:42:22 CEST] <Keshl> Video files just contain parts of frames. Nothing else.
[10:42:46 CEST] <Keshl> A codec is a standard used to define how to encode and decode video files. Something like that, anyway.
[10:42:58 CEST] <dude42> I meant maybe an old version of a codec was used to encode the original
[10:43:14 CEST] <dude42> I also stripped all metadata by doing map_metadata -1
[10:43:21 CEST] <dude42> not sure if that would make a difference
[10:43:30 CEST] <Keshl> Doubtful. That said, gut feeling.. Was your original video recorded live? Like, did you record you playing a game?
[10:44:45 CEST] <dude42> I'm not sure, I got it from someone else
[10:45:06 CEST] <Keshl> Well, if you ever get a video like that, just doing what you did will probably make it play way saner.
[10:45:28 CEST] <Keshl> When people do live recordings, it's usually made to be encoded really, /really/ fast so there's a lot of CPU time left over for the thing they're recording.
[10:46:06 CEST] <Keshl> Simply going over it again and giving it more time to work'll fix a lot of issues.
[10:46:37 CEST] <dude42> oh yes the audio also changed from aac to vorbis in the process, not sure if that would make a difference either
[10:46:45 CEST] <Keshl> It could.
[10:47:09 CEST] <dude42> Anyhoo, thanks :D This was fun
[10:47:15 CEST] <Keshl> Shiny. -É-.
[10:48:02 CEST] <Keshl> ... Suddenly, a thought strikes me. There's certain applications where a less general-purpose codec might be useful..
[10:48:23 CEST] <Keshl> For example, first-person games and gopro cameras. The motion will almost always expand from the center or contract towards it.
[10:49:41 CEST] <dude42>   Stream #0:0 -> #0:0 (h264 (native) -> h264 (libx264))
[10:49:42 CEST] <dude42>   Stream #0:1 -> #0:1 (aac (native) -> vorbis (libvorbis))
[10:49:55 CEST] <dude42> not sure what h264 native means
[10:51:24 CEST] <Keshl> Me either. xwx.
[11:01:51 CEST] <relaxed> dude42: I think it refers to the decoder. In both cases it's using ffmpeg's native decoders.
[11:13:00 CEST] <dude42> thanks relaxed
[11:49:14 CEST] <anshul_> how to send feed to ffserver, what format of video is required by feed, please link some document  to me
[11:49:44 CEST] <anshul_> just now I am using config file present in doc/ffserver.com in ffmpeg source
[11:52:45 CEST] <relaxed> anshul_: did you see https://trac.ffmpeg.org/wiki/StreamingGuide ?
[12:03:59 CEST] <anshul_> thanks relaxed
[12:10:27 CEST] <anshul_> i am not able to feed ffserver, I have started fffserver with default config file
[12:10:52 CEST] <anshul_> now when I feed I get http://pastebin.com/3dQWYtDR
[12:12:56 CEST] <anshul_> error is MPEG1/2 does not support 3/1 fps
[12:17:34 CEST] <anshul_> if i write VideoFrameRate 1/15 there then also I get 15/1 not supported if I write 15/1 then also I get same error
[12:21:42 CEST] <anshul_> I commented out most of stream related option in config file it worked
[12:26:34 CEST] <Mavrik> why are you using MPEG-1 as output.
[12:26:42 CEST] <Mavrik> and WMV?!
[12:36:30 CEST] <anshul_> Mavrik: I just want to test server, they are byb default in that format
[12:36:38 CEST] <anshul_> by*
[13:07:38 CEST] <anshul_> how I have made this config for rtsp http://pastebin.com/8SZESsyY
[13:07:50 CEST] <anshul_> how to play it using ffplay
[13:10:22 CEST] <Keshl> The more I watch this place, the more I realize why people always sound rude and the more I feel sorry about all the silly questions I asked two years ago. For the fifth time or so, I'm sorry, guys. x.x...
[13:11:51 CEST] <anshul_> either I get connection refused or ffplay does not play the stream at all,
[13:25:25 CEST] <anshul_> in ffserver config file it is stated that port number is 5454, but where  they have actually stated it
[14:03:11 CEST] <anshul_> should I use Format in ffserver in conf as RTSP or rtp.
[14:03:52 CEST] <anshul_> I want to play rtsp in my vlc, but some blogs are saying to use rtp in format for rtsp
[14:51:25 CEST] <anshul_> I debuged the code, and I found that rtsp does not work with feed
[15:57:56 CEST] Action: Mavrik hugs Keshl.
[16:08:13 CEST] <ffmpegusr> hi, is the windows precompiled ffmpeg come with hardware accelerationf or h264 decoding ?
[16:08:53 CEST] <ffmpegusr> (how can i check it ?)
[16:11:30 CEST] <seasc> I'd like to stream my desktop to a friend, what do i need to change on the commandline? :: http://pastebin.com/xVFdY6gz
[16:11:47 CEST] <seasc> ffmpegusr, just type ffmpeg in the command line, and check its output (--enable-XY ..)
[16:12:16 CEST] <ffmpegusr> seasc: yeah, which enable am i lookign for ?
[16:12:18 CEST] <ffmpegusr> ...
[16:13:53 CEST] <seasc> all there is --enable-XY is enabled, if its there, its enabled.
[16:14:18 CEST] <seasc> otherwise, please rephrase your question (idk about windows specific, on linux each distro has its own build - if any at all)
[16:14:51 CEST] <seasc> but just entering the executable, will return a list of all enabled stuff it has
[16:15:21 CEST] <ffmpegusr> anyone else ?
[16:20:24 CEST] <Mavrik> fflogger, you're looking for dxva
[16:20:30 CEST] <Mavrik> eh.
[17:22:06 CEST] <Anoia> hi all
[17:23:14 CEST] <Anoia> is the Zeranoe build of FFmpeg based off this FFmpeg?
[17:23:19 CEST] <c_14> yes
[17:23:29 CEST] <Anoia> it looks liek it's just a windows build
[17:23:37 CEST] <Anoia> ok, thanks
[17:24:08 CEST] <Anoia> I'm tryign to rationalise our usage which seems to include both FFmpeg (from Zeranoe) and ffdshow
[17:30:48 CEST] <c_14> It is just a windows build.
[17:36:31 CEST] <Anoia> thanks, I see it's linked form the download page now
[17:52:52 CEST] <seasc> Trying to include more codecs, ffmpeg fails to reckognize x264, though x264 doesnt provide any *pc or *h files ? --> http://ur1.ca/k4ijq
[17:55:30 CEST] <c_14> What's your configure line for x264?
[17:56:09 CEST] <seasc> c_14, "--prefix=/home/sea/ffmpeg_chroot/usr" "--disable-static"
[17:56:25 CEST] <c_14> Add --enable-shared
[17:56:29 CEST] <c_14> Or else you won't build libraries
[17:56:42 CEST] <seasc> c_14, ok, thank you
[18:20:15 CEST] <seasc> xvid, its --enable-libxvid, but i only have xvid.h files, do i have to rename them?
[18:20:30 CEST] <seasc> +a -s
[18:20:42 CEST] <seasc> them=it
[18:21:24 CEST] <seasc> nah.. doesnt help
[18:22:02 CEST] <c_14> You only need the .h file for xvid
[18:22:09 CEST] <c_14> (and the .so)
[18:27:35 CEST] <seasc> hahaa, figured, xvid created $CHROOT/include, instead of $PREFIX/include... :doh: changed that and voila :D next codec
[19:05:07 CEST] <luc4> Hello! Im trying to stream a video file using ffmpeg. I came up with this: ffmpeg -re -i big_buck_bunny_1080p_h264.mov -an -f rtp rtp://192.168.1.73:1234. It seems to work locally using 127.0.0.1 but not when trying to play remotely. Any idea if the command is wrong?
[19:06:08 CEST] <c_14> If it works locally, but not remotely, it's probably firewall.
[19:07:01 CEST] <luc4> c_14: firewalls are down
[19:07:30 CEST] <luc4> c_14: can you confirm that rtp://192.168.1.73:1234 is the address of the interface that will publish the stream?
[19:07:57 CEST] <c_14> No, that's where it's going.
[19:08:39 CEST] <c_14> afaik
[19:09:00 CEST] <luc4> c_14: and that was the problem :-) thanks
[19:09:12 CEST] <luc4> c_14: I have to say I see many many errors.
[20:06:25 CEST] <justinX> if I want to compress my filmclip making sure it wastes as few bytes as possible on the minute in the middle where nothing happens, should I run ffmpeg three times like   ffmpeg -i in.avi -pass 1 out.mp4    ffmpeg -i in.avi -pass 2 out.mp4    ffmpeg -i in.avi -pass 3 out.mp4   ?
[20:08:45 CEST] <justinX> also are there other parameter I should use to tell it to use variable bitrate for video?
[20:10:42 CEST] <llogan> just use -crf. no need for multiple passes.
[20:10:45 CEST] <lordkrondor> i'd use -crf with a -maxrate:v
[20:10:47 CEST] <lordkrondor> yeah
[20:12:35 CEST] <maxsilver> Is it possible to use lavfi with a UDP stream? For instance, in a command like 'ffmpeg -f lavfi -i "movie=test.mpg[out0+subcc]" -map s output.srt' -- is it possible to have the source be a UDP stream instead of a local file?
[20:14:40 CEST] <maxsilver> Using the normal method doesn't seem to work. 'ffmpeg -i "udp://239.1.1.1:1234"' -- and I'm sure I'm doing something wrong -- but I'm struggling to understand the docs around libavfilter
[20:18:56 CEST] <justinX> ok.
[20:22:32 CEST] <justinX> lordkrondor: hwo do I use maxrate? is v a number there? (or a "v" ?)
[20:22:36 CEST] <justinX> *how
[20:23:59 CEST] <lordkrondor> well using -maxrate:v is just to set a maximum bit-rate for video
[20:24:12 CEST] <lordkrondor> you don't have to use if if you don't have any limitations
[20:24:14 CEST] <justinX> oh
[20:24:25 CEST] <maxsilver> llogan - http://pastebin.com/Gwt4PMVS
[20:24:49 CEST] <llogan> we already know all of that info.
[20:25:10 CEST] <llogan> please re-read the message from fflogger
[20:25:41 CEST] <maxsilver> I'm sorry, I don't follow you. There is no console output because I don't know the command to run on the console.
[20:30:14 CEST] <maxsilver> If it's helpful, here's the complete console output of a successful command - http://pastebin.com/76V4An8G
[20:30:52 CEST] <justinX> is   "ffmpeg -i SDC18234.AVI -crf out.mp4" wrong way on writing it?   it answers "At least one output file must be specified" :-(
[20:32:40 CEST] <ChocolateArmpits> justinX: you need to specify the crf number
[20:33:05 CEST] <ChocolateArmpits> now ffmpeg takes "out.mp4" as a value for crf, which then leaves the output not existing and problems follow
[20:33:24 CEST] <ChocolateArmpits> obviously "out.mp4" is not a correct value for crf either way
[20:34:24 CEST] <ChocolateArmpits> justinX: read this for CRF https://trac.ffmpeg.org/wiki/Encode/H.264#crf
[20:39:50 CEST] <llogan> maxsilver: why not just do: ffmpeg -i input -map s out.srt
[20:40:04 CEST] <llogan> why are you using the movie source filter?
[20:43:41 CEST] <maxsilver> llogan: I thought it was required, since I'm trying to pull closed captioning out of user data in video frames. I'm probably 100% wrong on that -- it just seemed suggested from the docs at https://www.ffmpeg.org/ffmpeg-devices.html
[20:44:48 CEST] <maxsilver> (under section 3.11 and 3.11.2). I'll try your command right now.
[20:49:54 CEST] <seasc> v4l-utils to be included, but it fails to find libjpeg.h, eventhough i've set the proper --libdir=$DIR_LIB :: http://ur1.ca/k4jww
[20:51:57 CEST] <c_14> can you pastebin ffmpeg's config.log as well as the output of `ls $DIR_LIB'
[20:52:44 CEST] <seasc> c_14, sry, its the v4l build, not ffmpeg yet. and v4l's config.log is pasted
[20:54:30 CEST] <c_14> Can you still pastebin the configuration line and output of `ls $DIR_LIB' when compiling libv4l2
[20:56:03 CEST] <seasc> configuration line for v4l is on line 446, or isnt that what you are looking for atm? dir content:: -> http://fpaste.org/208135/
[20:56:23 CEST] <seasc> looking/asking
[20:57:21 CEST] <luc4> Hello! Im trying to stream h264 via rtp from my machine to a virtual machine locally. It seems to work but the result is all messed up. I suppose it is losing packets& is this possible? But how is it possible to stream h264 via rtp if it is unreliable?
[20:57:55 CEST] <c_14> seasc: did you set the CFLAGS to include the path containng jpeglib.h ?
[20:58:04 CEST] <seasc> c_14, sry, accidently used dir_li instead of dir_inc... .. this is the inc: http://fpaste.org/208136/
[21:00:29 CEST] <seasc> c_14, line 289 of http://ur1.ca/k4jww says its installed to $CHROOT/usr/include // but http://fpaste.org/208136/ doesnt show that file
[21:00:40 CEST] <maxsilver> llogan : That command, as written, results in http://fpaste.org/208137/84331871/
[21:01:32 CEST] <seasc> c_14, no its there, 1st of 4th column
[21:01:33 CEST] <c_14> seasc: yes it does. line 4, 3rd column
[21:01:43 CEST] <seasc> :)
[21:02:46 CEST] <maxsilver> and this is my attempt to fix the "stream map s matches no streams" issue - http://fpaste.org/208140/84333361/
[21:03:51 CEST] <seasc> and on line 446 of http://ur1.ca/k4jww i tell the v4l-configure even twice to use --includedir=$DIR_INC, must been a typo
[21:04:29 CEST] <c_14> seasc: I'm pretty sure that option just tells it where to put things, not where to look for them.
[21:06:04 CEST] <seasc> c_14, i pretty much belive on this :) so back to the LD_LIBRARY_PATH? which already is.... very long.... /home/sea/ffmpeg_chroot/usr/lib:/home/sea/ffmpeg_chroot/usr/lib:/home/sea/ffmpeg_chroot/usr/lib64:/usr/local/lib:/usr/local/lib64:/usr/lib64:/usr/lib:/home/sea/ffmpeg_chroot/usr/lib:/home/sea/ffmpeg_chroot/usr/lib64:/home/sea/ffmpeg_chroot/usr/include:/home/sea/ffmpeg_chroot/include:/usr/local/lib:/usr/local/lib64:/usr/lib64
[21:06:04 CEST] <seasc> /home/sea/ffmpeg_chroot/usr/lib:/home/sea/ffmpeg_chroot/usr/lib:/home/sea/ffmpeg_chroot/usr/lib64:/usr/local/lib:/usr/local/lib64:/usr/lib64:/usr/lib:/home/sea/ffmpeg_chroot/usr/lib:/home/sea/ffmpeg_chroot/usr/lib64:/home/sea/ffmpeg_chroot/usr/include:/home/sea/ffmpeg_chroot/include:/usr/local/lib:/usr/local/lib64:/usr/lib64
[21:06:28 CEST] <seasc> hmm... ran the script too often, added itself to the string.. sry
[21:06:50 CEST] <c_14> Nah, in this case you want the CFLAGS="-I$INC_DIR" and LDFLAGS="-L$LIB_DIR"
[21:07:02 CEST] <seasc> why do i want this in this case?
[21:07:36 CEST] <c_14> Because you're compiling/linking and not executing.
[21:08:51 CEST] <seasc> ok, but i really do not want this.... i get a short log now than i did before.. fails earlier.. sec
[21:10:08 CEST] <seasc> @ c_14 http://fpaste.org/208143/
[21:10:36 CEST] <seasc> i did export those two suggest values
[21:13:27 CEST] <seasc> c_14, ahh v4l has no CFLAGS, but CXXFLAGS.. is that the same?  not really isnt it?
[21:13:53 CEST] <c_14> you can usually set CXXFLAGS=CFLAGS
[21:14:50 CEST] <seasc> well either way, the log is short, as the 208143'ish
[21:15:18 CEST] <c_14> yeah, because it decided that g++ couldn't create executables
[21:15:20 CEST] <c_14> for some reason
[21:16:01 CEST] <seasc> fantastic :( so i have a g++ that cant create executeables, and a header that is not found.
[21:16:16 CEST] <llogan> maxsilver: i don't know anything about cc. Anshul Maheshwari may answer if you ask on ffmpeg-user mailing list
[21:16:23 CEST] <seasc> why dont they look in the includedirs and libdirs for libs and include files as well?
[21:17:11 CEST] <c_14> AFAIK, those are just directives that tell it where to place the files on install.
[21:18:13 CEST] <maxsilver> llogan: Ok, I'll try that. Thank you for your time and assistance today.
[21:19:03 CEST] <seasc> c_14, must be that way, otherwise it would find the libjpeg.h which is in that location.
[21:20:24 CEST] <seasc> its not a PKG_CONFIG_PATH candidate either, as it doesnt ask for *pc file. the path is already provided in LD_LIBRARY_PATH, is there any other way to tell ./configure where to look for required header files?
[21:20:51 CEST] <c_14> CFLAGS and/or CXXFLAGS
[21:20:59 CEST] <seasc> They obviously didnt work :(
[21:21:10 CEST] <c_14> maybe with --extra-cxxflags=
[21:21:28 CEST] <seasc> didnt see that for v4l configure
[21:21:53 CEST] <seasc> just as the variable= with that -i$DIR_INC?
[21:22:03 CEST] <c_14> -I
[21:22:31 CEST] <seasc> command not found
[21:22:49 CEST] <c_14> hmm?
[21:22:53 CEST] <c_14> Did you add that to the configure line?
[21:23:02 CEST] <seasc> lol
[21:23:44 CEST] <seasc> well i just placed the option above the ./configure, so it was the command... but confgure doesnt know that option either
[21:24:34 CEST] <c_14> mhm, drat. have you tried exporting only CXXFLAGS?
[21:26:50 CEST] <seasc> just tried... export CFLAGS="-I$INC_DIR" and we're back to where we were... libjpeg.h not found...
[21:27:02 CEST] <c_14> CXXFLAGS, not CFLAGS
[21:27:06 CEST] <seasc> then it fails
[21:27:14 CEST] <c_14> eeeeh
[21:27:37 CEST] <seasc> i was mistaken, didnt see CFLAGS, and we already figured with CXXFLAGS it thinkgs gcc canot create binaries
[21:29:01 CEST] <c_14> oh
[21:29:06 CEST] <c_14> $INC_DIR is empty
[21:29:21 CEST] <c_14> eh, no
[21:29:23 CEST] <seasc> crappy typo.. thanks :D
[21:29:51 CEST] <c_14> but yeah, one of the things was expanding to an empty string
[21:31:14 CEST] <seasc> gngn ngngng its further now... but... failing on the (almost) same.... http://fpaste.org/208165/
[21:32:06 CEST] <c_14> Did you export both CXXFLAGS and LDFLAGS?
[21:32:25 CEST] <seasc> err no LDFLAGS... did disable them to test
[21:32:54 CEST] <seasc> lol, back to the short one
[21:33:12 CEST] <c_14> what did you export LDFLAGS as?
[21:33:59 CEST] <seasc> export LDFLAGS="-L$LIB_DIR" // http://fpaste.org/208135/
[21:34:44 CEST] <c_14> Eh, do you maybe want $DIR_LIB instead of $LIB_DIR ?
[21:35:11 CEST] <c_14> What you can also do is, echo $LDFLAGS
[21:35:19 CEST] <c_14> to make sure that the variable is expanding correctly
[21:36:11 CEST] <seasc> compiling now... was blind for the last typo, as i just uncommented it
[21:37:05 CEST] <seasc> c_14, ffmpeg compiling now thank you very much
[22:41:33 CEST] <DragonsLord> Hello everybody ... I'm experincing a problem: option -vf fps=fps=25 is not working ... the resulting file is still 50 fps :(
[22:41:53 CEST] <DragonsLord> Is someone so kind to help me?
[22:44:52 CEST] <DragonsLord> You are right, c_14 ... indeed
[22:45:03 CEST] <DragonsLord> here is the command line > http://pastebin.com/yq789NXV <
[22:46:09 CEST] <c_14> You can't have more than one filtergraph
[22:46:18 CEST] <c_14> Put the fps=fps=25 before pp=lb
[22:46:26 CEST] <c_14> and get rid of the other -vf
[22:48:17 CEST] <DragonsLord> Thank you very much, as always, c_14
[22:48:22 CEST] <DragonsLord> I am trying ...
[22:48:59 CEST] <DragonsLord> .. but the console output is already saying "25 fps" so it should work
[22:49:15 CEST] <DragonsLord> Do you think I could experience some A/V desync with this filter?
[22:49:51 CEST] <c_14> from 50 to 25, probably not
[22:50:55 CEST] <DragonsLord> could you please elaborate on this a little more, just to understand ...
[22:52:31 CEST] <c_14> The fps filter shouldn't destroy A/V desync at all, especially not when it's such an easy conversion.
[22:52:57 CEST] <DragonsLord> thank you very much for this explanation
[22:53:23 CEST] <DragonsLord> c_14, are you a FFmpeg developer?
[22:53:46 CEST] <c_14> Not really.
[22:53:53 CEST] <DragonsLord> -_-
[22:54:04 CEST] <DragonsLord> Y/N Not really is not exaustive XD
[22:54:13 CEST] <c_14> I have some minor patches in tree, but nothing major.
[22:54:45 CEST] <DragonsLord> you have a very large knowledge in FFmpeg
[22:54:49 CEST] <DragonsLord> as far as I know
[22:54:54 CEST] <DragonsLord> congratulations
[22:54:57 CEST] <DragonsLord> :)
[22:54:59 CEST] <DragonsLord> compliments
[22:55:13 CEST] <DragonsLord> I don't know which is the beast word to say this in English
[22:55:31 CEST] <DragonsLord> but I hope you can understand the basic concept
[22:55:31 CEST] <DragonsLord> ;9
[23:23:36 CEST] <seegee__> Hey guys, quick question. How can I take a screenshot from directly in the middle of a video?
[23:24:23 CEST] <DragonsLord> seegee__, I usually use VLC
[23:25:35 CEST] <seegee__> DragonsLord: I'm looking for a command line tool that I can automate with bash. I have around 8000 videos that I need screenshots for atm
[23:26:38 CEST] <c_14> seegee__: you want the exact middle?
[23:27:05 CEST] <c_14> ffprobe to get the length. divide by two `ffmpeg -ss [time] -i video -frames:v 1 out.png'
[23:27:59 CEST] <Prelude2004c> hey guys.. can anyone tell me why i get these errors ? ( http://pastebin.com/xqDemsiA ) ?
[23:28:15 CEST] <seegee__> c_14: Will that keep the aspect ratio correct?
[23:28:27 CEST] <Prelude2004c> cpu usage is like nothing
[23:28:32 CEST] <c_14> seegee__: yep. should at least
[23:29:05 CEST] <seegee__> Alright, thanks. Is taking screenshots like that a timley / resource intensive process?
[23:29:28 CEST] <c_14> should be relatively fast. limited mainly by the speed of your io device
[23:30:08 CEST] <seegee__> Ok, but It's just insanley intesnive? (I won't have any problems starting 2000 threads at once with bash?)
[23:31:06 CEST] <c_14> Depends on your cpu/io. you might want to add -threads 1 as an input/output option to make it easier to start numerous instances
[23:32:11 CEST] <Prelude2004c> c_14 is that an answer to my question?
[23:32:52 CEST] <c_14> Nah, was seegee__'s question. As to your question; no clue. Looks like a corrupt source to me.
[23:34:09 CEST] <DragonsLord> Prelude2004c, I am quite ignorant ... but I guess a problem at the source video
[23:34:31 CEST] <DragonsLord> Prelude2004c, are you experiencing this problem for all of your video?
[23:39:43 CEST] <Prelude2004c> its weird.. when i run one stream it runs fine
[23:39:58 CEST] <Prelude2004c> when i try to add mutli outputs it gives me those errors.. sometimes works and sometimes it doesnt
[23:41:26 CEST] <DragonsLord> <Prelude2004c> sometimes works and sometimes it doesnt <<< with the same source video?
[23:49:11 CEST] <debianuser> Prelude2004c: some "udp" errors there, not enough network bandwidth or dropped packets maybe?
[23:53:13 CEST] <Prelude2004c> no, we have enough capacity
[23:53:15 CEST] <Prelude2004c> here is another odd thing
[23:53:24 CEST] <Prelude2004c> i can only get like 100fps total
[23:53:32 CEST] <Prelude2004c> the board is 24 cores
[23:53:40 CEST] <Prelude2004c> running at 2.2Ghz
[23:53:48 CEST] <Prelude2004c> the CPU usage is 100% of one core approx
[23:53:51 CEST] <Prelude2004c> not even all cores
[23:53:59 CEST] <Prelude2004c> and i can't seem to pash the 100fps threshhold
[23:54:05 CEST] <Prelude2004c> anyone know why that might be
[23:54:22 CEST] <__jack__> feeling: more thread
[23:55:28 CEST] <Prelude2004c> more threads ?  am saying threads 0
[23:57:37 CEST] <__jack__> hum ?
[00:00:00 CEST] --- Wed Apr  8 2015