[Ffmpeg-devel-irc] ffmpeg.log.20141122

[burek]

[00:17] <c_14> You shouldn't find a static ffmpeg compiled with libfdk_aac.
[00:18] <c_14> Anybody who offers such a build is going against the libfdk-aac license.
[00:18] <c_14> s/going against/violating
[00:31] <uex1fi> hello
[00:31] <BtbN> c_14, no, he's not. He's violating the GPL. The libfdk_aac license is actualy quite permissive.
[00:32] <uex1fi> i am reading about using mcdeint with yadif
[00:32] <uex1fi> the third parameter in mcdeint is qp, or quantization parameter.  could someone help me to understand what this means?
[00:35] <c_14> BtbN: aah, I just know it's a license issue somewhere. Never really looked at the specifics.
[00:36] <c_14> uex1fi: https://en.wikipedia.org/wiki/Quantization_(image_processing)
[00:37] <uex1fi> "Higher values should result in a smoother motion vector field but less optimal individual vectors."
[00:37] <uex1fi> For one, what's the highest value?  Also, what is a vector.
[00:38] <uex1fi> is it basically that the higher the value, the smoother the video?
[00:38] <uex1fi> and if it is set too high, the video will look too smooth?
[00:39] <BtbN> c_14, The libfdk license is very strange. There are quite a few GPL projects which interpret it in a GPL compatible way.
[00:41] <c_14> uex1fi:   qp                <int>        ..FV.... set qp (from INT_MIN to INT_MAX) (default 1)
[00:42] <uex1fi> c_14: I don't quite understand that
[00:42] <uex1fi> i know the default is 1, but i just don't know the number range allowed
[00:42] <c_14> from INT_MIN to INT_MAX
[00:43] <uex1fi> i don't know that those mean....
[00:43] <c_14> Usually -32767 to 32767
[00:43] <uex1fi> okay, wow
[00:43] <uex1fi> but the examples i see just have it set to maybe 10 or 20
[00:44] <c_14> Just because you can set it from INT_MIN to INT_MAX doesn't mean you have to.
[00:44] <c_14> Or should.
[00:44] <uex1fi> yes
[00:45] <uex1fi> ok, so it's not like a 1-10 range... i'll just have to do some tests to find the right value, then.
[00:45] <uex1fi> as for mcdeint, what's a really good deinterlace setting without going crazy?
[00:46] <uex1fi> i was thinking of doing extra_slow, but i don't know if it's going overboard
[00:52] <uex1fi> or let me put it this way.  in mcdeint, does 'extra_slow' always produce a better result than 'fast'?
[00:53] <c_14> I've never really used mcdeint. You might just have to compare results.
[00:54] <uex1fi> yeah
[00:57] <punto> hi.. is there a way to get ffmpeg to write a .srt subtitle file without the html tags?
[00:57] <punto> my player can't interpret some of them, so they show up on the screen, I just want the text
[01:50] <setuidbit> Hello, I am trying to statically compile ffmpeg on Mac osx 10.9.5.  I have gotten my heuristic to work on linux but now I need it on mac and I am not sure the best way to go about getting a static binary compiled.  Here is my console in-out: http://pastebin.com/z5C5Pcuw  and here is config.log: http://pastebin.com/xEzGX1x0  Thank you in advance!
[01:54] <uex1fi> for yadif, 2 is "Like send_frame, but it skips the spatial interlacing check."
[01:55] <uex1fi> what is this?
[01:55] <uex1fi> why would you skip that check?
[07:15] <cyphase_> when i run ffprobe on an mp3 i encoded from some wavs (and sometimes with a few mp3s thrown in), i get a non-zero 'start'; as in the following string: "Duration: 02:01:02.88, start: 0.025056, bitrate: 64 kb/s". why would it not be 0.000000, what are the effects of that, and how can i fix it?
[07:52] <lookatmeyou> anyone help! I compile using this command: g++ deco.c -o deco -D__STDC_CONSTANT_MACROS -lavdevice -lavformat -lavfilter -lavcodec -lswresample -lswscale -lavutil
[07:53] <lookatmeyou> but got lots of undefined reference error
[07:53] <lookatmeyou> for example, deco.c:(.text+0x6a): undefined reference to `avcodec_find_encoder(AVCodecID)'
[07:54] <lookatmeyou> I have already installed ffmpeg libs in /usr/local/lib
[08:34] <gabriel> Hey, could someone help me with a question I have about image sequences -> video
[08:34] <gabriel> ?
[08:35] <gabriel> I would like to create a video using the mpeg4 video codec with *just* two images as an input. Id like to be able to use two arbitrary images rather than using the image-%03d.png syntax. Is this possible>
[08:58] <c_14> Try using the concat filter.
[09:01] <c_14> lookatmeyou: make sure your PKG_CONFIG_PATH is set correctly and then pas the output of `pkg-config --cflags --libs libavdevice libavformat libavfilter libavcodec libswresample libswscale libavutil' to g++
[09:02] <c_14> *pass
[09:02] <lookatmeyou> thank you, later I use gcc instead of g++ and works
[10:05] <lookatmeyou> Anyone help, when I run a program I got this error
[10:05] <lookatmeyou> me at ubuntu:~/codec$ ./deco
[10:05] <lookatmeyou> H264 Codec not found
[10:05] <lookatmeyou> But I've already installed x264 on my machine
[10:05] <lookatmeyou> And the ffmpeg example using x264 runs OK
[10:05] <lookatmeyou> root at ubuntu:~/ffmpeg-2.4.3/doc/examples# ./decoding_encoding h264
[10:05] <lookatmeyou> Encode video file test.h264
[10:05] <lookatmeyou> http://paste.ubuntu.com/9168165/, this is my codes
[10:16] <lookatmeyou> anyone please help
[10:24] <lookatmeyou> solved, thanks
[11:11] <t4nk081> I have a question to ffmpeg
[11:13] <t4nk081> We do livestreaming over the internet and transmit the stream to rtmp streaming server over udp. If theres no input to udp we wanne show a picture that signals that the stream is currently unavailable. Can it be done with the videofilter complex switch?
[11:25] <viruser> Hi. is there any command line to tell encoder to ignore invalid data from input?
[11:28] <user39564> Hello, I try to convert images to video. This works if the images are named 1.png, 2.png, 3.png ... but I can not make it work with 00001.png, 00002.png, 00003.png. I am forced to change their name. Is it possible to do otherwise? Thank you for your help!
[11:33] <relaxed> user39564: use %05d.png as your input
[11:36] <relaxed> user39564: http://ffmpeg.org/ffmpeg-formats.html#image2-2
[11:37] <user39564> relaxed: thank you, sorry for the loss of time.
[11:38] <relaxed> viruser: pastebin the problem
[11:41] <viruser> relaxed: Here you go http://paste.debian.net/132952/
[11:42] <viruser> background: I have salvaged a video from a scratched cd. the program filled unreadable sectors with zeros. Most players crash when they reach the zero parts. but mplayer can skip it. I want to fix the video
[11:44] <relaxed> does mplayer handle the audio we
[11:44] <relaxed> well* ?
[11:45] <viruser> yes. Like i said, it works fine. it just skips the broken areas and continues to read the unbroken parts.
[11:46] <viruser> I want ffmpeg to create a file without the broken area so other players can handle it as well
[11:46] <viruser> if that's possible
[11:46] <relaxed> and ffmpeg stops encoding at these sections?
[11:47] <viruser> Well, as the dump suggests, it stops at 35kth frame
[11:49] <relaxed> use mplayer's '-vo yuv4mpeg' and '-ao pcm' to dump the video and audio, then feed those to ffmpeg as input
[11:51] <viruser> I'll try that. thanks
[14:34] <lookatmeyou> When I decode a h264 video using avcodec_decode_video2, I got this error http://paste.ubuntu.com/9172231/
[14:34] <lookatmeyou> This is the avcodec_decode_video2 output.
[14:34] <lookatmeyou> What's that mean?
[14:35] <JEEBcz> either your code is wrong or your AVC stream is broken
[14:35] <JEEBcz> as in, I hope you are using lavf to open the actual annex b stream
[14:35] <JEEBcz> and then decoding with lavc
[14:36] <lookatmeyou> http://paste.ubuntu.com/9172489/, this is my codes
[14:37] <lookatmeyou> And th input file Avatar.h264 can use ffplay to play.
[14:37] <JEEBcz> exactly
[14:37] <JEEBcz> you are only using libavcodec
[14:37] <JEEBcz> not libavformat
[14:38] <JEEBcz> you need to first read the raw annex b with libavformat
[14:38] <JEEBcz> and then push its output to libavcodec
[14:38] <lookatmeyou> OK, let me try, thanks
[14:44] <lookatmeyou> Could you please give me some codes to show the process?
[14:46] <JEEBcz> I don't have to, there are already samples in the FFmpeg code base :P
[14:46] <JEEBcz> where you most probably copied your code so far
[14:48] <lookatmeyou> Is it in avio_reading.c?
[14:49] <Popara> Hello guys, i noticed that the -f tee Mux is very buggy. Some streams are unable to be restreamed while some others are loosing synch with the video/audio. This with all the other mux doesnt happen.
[14:49] <Popara> How can i reproduce it to make a bug report. Its not something that is happening all the time but most of the time yes
[14:52] <Wader8> hello
[14:53] <Wader8> how do I stream RTSP with ffmpeg, without pipe option, thank you
[14:53] <Wader8> http://forum.videohelp.com/threads/368465-Stream-from-RTMP-to-Nokia-72-%28RTSP%29-on-home-network?p=2358217
[14:53] <Wader8> the code is in the thread, but it didn't work for me
[15:25] <AlfredWangHanWuD> Hello Guys
[15:26] <AlfredWangHanWuD> I try to use a media player to play a mpg stream..but when fffmpeg finish filling up the feed1.ffm
[15:26] <AlfredWangHanWuD> the player stoped ...
[15:27] <AlfredWangHanWuD> anyone know what is wrong here
[16:25] <Wader8> hello
[16:26] <Wader8> my ffmpeg config does not produce a working RTSP link
[16:27] <Wader8> http://pastebin.com/2JKAPzkH
[16:27] <Wader8> I have already followed the tutorials I found on the web, much of the material is several years old, please assist thanks
[16:43] <Wader8> so i find out i need ffserver for streaming
[16:44] <Wader8> and zeranoe says it can't be compiled for windows and it's outdated
[16:44] <Wader8> worthless
[16:45] <JEEBcz> ffserver is barely used, barely maintained
[16:45] <JEEBcz> if you can't do what you want with just ffmpeg, you're pretty much out of luck
[16:47] <Popara> Wader8
[16:47] <Popara> there is one way, you can output it to HLS
[16:47] <Wader8> no i don't want HLS sorry
[16:47] <Popara> and make a webserver and place the m3u8 file there
[16:47] <Popara> yes just saying
[16:48] <Wader8> because my device only supports 640x360 MP4 Visual
[16:48] <Wader8> and over RTSP
[16:48] <Wader8> and VLC does this 99% except the breaking 1% ... there's a bug that times it out every 60 secs, impractical to watch like that
[16:49] <JEEBcz> yeah, VLC is simpler to use and most probably more maintained
[16:49] <JEEBcz> at least in that part
[16:49] <Wader8> basically im trying to get video over, and video's not necessary, i get audio okay, but the cutoff 60 s is always there
[16:50] <Wader8> there would be no issue if VLC didn't have that bug, or actually RealPlayer on Nokia didn't have that bug
[16:50] <JEEBcz> either fix the issue in VLC or try to find an ffserver user that uses RTSP
[16:50] <JEEBcz> (and then find out if ffserver has the same issue)
[16:50] <Wader8> i was trying to do that but never could get ffmpeg to work
[16:50] <Wader8> to find out if VLC is to blame or not
[16:50] <JEEBcz> well, ffserver being what it is...
[16:51] <Wader8> i've tried TVersity, very confusing program had no idea how to setup, and can't lower bitrate below 0.8 which is unacceptable that won't work over my wlan distance
[16:52] <Wader8> does anybody know anything else, Wowza is not free and will probably run out of a few days right ?
[16:52] <Wader8> or a few weeks
[16:53] <Wader8> VLC is so good in this because I can take directly an external stream, transcode and publish as a stream, it's just exactly the right type of a thing, only that bug, it's working, just loses connectiong every 60 secs I have to clik Play and it's impractical to do that while doing something else
[16:53] <JEEBcz> automate it somehow?
[16:54] <JEEBcz> or debug the issue
[16:54] <JEEBcz> to see which end has it
[16:54] <Wader8> automate what ?
[16:55] <Wader8> don't understand that
[16:55] <Wader8> well im not an expert on VLC so I have to yet find out how to open logs and traces
[19:15] <vrexx> noob question alert: is it possible to convert one audio channel of a stereo stream (dvd ac-3 codec) to a mono stream without re-encoding?
[19:15] <JEEBcz> no
[19:16] <vrexx> ok, thx!
[19:16] <Guiri> I have a curious issue.  I'm trying to remux an NSV stream, and stream copy the VP6 and AAC into an FLV container for RTMP streaming.  It plays fine in ffplay, but when c:v copy are used in ffmpeg, the image produced is upside down.  Horizontal flip.
[19:31] <gabriel_> I would like to create a video using the mpeg4 video codec with *just* two images as an input. Id like to be able to use two arbitrary images rather than using the image-%03d.png syntax. Is this possible?
[19:31] <gabriel_> Also, Hi everyone :)
[19:32] <gabriel_> If you have an answer, Ive posted this question on the pastebin http://pastebin.com/nnxqkytf
[20:37] <raray> is there an #ifdef that will tell me wether i am on libav or ffmpeg?
[20:39] <JEEBcz> https://github.com/FFMS/ffms2/blob/master/configure.ac#L123
[20:41] <raray> lol
[20:41] <raray> how long do you think will the LIBAVCODEC_VERSION_MICRO  trick work?
[20:42] <JEEBcz> probably will never be removed
[20:42] <JEEBcz> the only thing that might break it is libav going to 100 with micro, but I think they won't be able to do that :P
[20:43] <raray> JEEBcz: thx a lot!
[20:46] <raray> Now my little program at least compiles on ubuntu :)
[20:47] <raray> But what I don't get is why we have 2 versions of ffmpeg with libav lagging behind
[21:13] <raray> I want to convert audio data to 32bit float mono. Is there a function ready to use, or do i need to code it myself?
[21:19] <c_14> Set the codec to pcm_f32le and audio channels to 1?
[21:27] <raray> c_14: found it: audio_resample seems to be what i was searching for.
[21:41] <Corin> Hey, how do you make webms?
[21:54] <c_14> ffmpeg -i video webm.webm
[21:55] <c_14> https://trac.ffmpeg.org/wiki/Encode/VP8
[22:46] <cyphase> i currently have an ffmpeg command line that takes multiple audio inputs, concatenates, and encodes to mp3. how can i specify a silence buffer to go between/after each clip? i know i could create a silent audio file and just use it as an input, and that would work well for a static buffer length (especially since the command line is being generated programmatically), but i want to be able to specify the buffer length dynamically. bonus points if i
[22:46] <cyphase> can specify each buffer's length separately. if i don't find a solution, my fallback is to create the silent audio dynamically with ffmpeg -f lavfi -i aevalsrc=0 -t SECS silence.wav
[23:24] <pmart> does dca codec decode the diff part of DTS-HD MA stream for lossless output?
[00:00] --- Sun Nov 23 2014