[Ffmpeg-devel-irc] ffmpeg.log.20141015

[burek]

[00:03] <c_14> Would that value make sense?
[00:05] <JaviSRK> trying to see where I can find it.
[00:05] <c_14> The timebase?
[00:06] <JaviSRK> yea
[00:07] <c_14> It's the tbc one, not tb
[00:08] <c_14> > Stream #0:0(und): Video: h264 (Constrained Baseline) (avc1 / 0x31637661), yuv420p, 480x360 [SAR 1:1 DAR 4:3], 554 kb/s, 29.97 fps, 29.97 tbr, 30k tbn, 59.94 tbc (default) <- for this video it would be 59.94
[00:10] <JaviSRK> I see.  My tbc is 60.
[00:10] <JaviSRK> so based on the two values i posted above
[00:11] <JaviSRK> First value would be 1457.916667
[00:11] <c_14> Hmm, that value still doesn't really make sense.
[00:19] <c_14> I have a strong feeling it's off by a factor of 10...
[00:19] <c_14> What's the output framerate?
[00:24] <jellosea> hi, im trying to build ffmpeg on ubuntu. however it cant seem to find libass although ive installed it via apt-get. i checked via apt to see where it was installed (/usr/lib and /usr/include) and included this in the configure via extra-cflags and extra-ldflags but it still cant seem to find libass and the configure fails. i was following this guide: https://trac.ffmpeg.org/wiki/CompilationGuide/Ubuntu
[00:24] <jellosea> any help would be appreciated
[00:25] <c_14> Did you install libass-dev ?
[00:25] <jellosea> yes
[00:25] <jellosea> http://pastie.org/9648291
[00:26] <c_14> What does pkg-config --exists --print-errors libass output?
[00:26] <JaviSRK>     Stream #0:0[0x1e0]: Video: mpeg1video, yuv420p(tv), 640x480 [SAR 178:163 DAR 712:489], 104857 kb/s, 60 fps, 60 tbr, 90k tbn, 60 tbc
[00:28] <Baked_Cake> im tempted to compile ffmpeg in a linux vm
[00:29] <c_14> JaviSRK: you're running a suitably new version of FFmpeg?
[00:29] <JaviSRK> let me check
[00:30] <Baked_Cake> or at least compile for windows with the latest libfdk_aac
[00:30] <JaviSRK> N-62720-gb217dc9
[00:32] <JaviSRK> i suppose I should get the latest.  Let me see what happens.
[00:33] <c_14> If getting the newest doesn't help, send a mail with the question to the ffmpeg-user mailing list asking about what units the time is in.
[00:33] <c_14> Baked_Cake: Or just don't use AAC? :P
[00:49] <JaviSRK> Thanks for the help.
[03:35] <jellosea> hi i cant seem to get cbr video with vp8 using ffmpeg. ive tried all the supposed settings (setting minbitrate, maxbitrate and bitrate to the same value and qcomp to 0 but it never gives me true CBR, the video bitrate varies. how can i fix this?
[03:42] <jellosea> here is my command http://pastie.org/9648549
[04:23] <yoyowinwin> c_14: http://ffmpeg.gusari.org/viewtopic.php?f=11&t=1737
[04:23] <yoyowinwin> c_14: thank you for your attention
[05:06] <hungnv> which options of ffmpeg will effect video duration? When encode video, the output video is shorter than the original one
[08:32] <rule_2> did the new version of ffmpeg fix the v4l2 buffer errors when recording?
[08:33] <rule_2> i've been using cvlc to capture v4l2 video recently but it doesn't let you set the framerate of input video :(
[08:47] <hungnv> any idea to keep original duration of video after transcoding?
[08:58] <Mavrik> hungnv, original video duration is always kept by default
[08:59] <Mavrik> unless you explicitly tell ffmpeg to start later or truncate it
[08:59] <hungnv> Mavrik, but when transcoding video, I got video duration changed, without any option telling ffmpeg change the duration
[10:12] <heftig> the live-ness of an RTMP stream is nothing that can be discovered (at least until seeking is attempted)?
[10:46] <c_14> yoyowinwin: What are you using to detect the 1-2G memory leak?
[11:05] <yoyowinwin> c_14: top
[11:07] <yoyowinwin> it only happened  when i use rtmp stream as a input
[11:11] <c_14> Under VIRT or RES ?
[11:15] <c_14> Also, how long did you have valgrind run?
[11:21] <c_14> In either case, since I don't know very much about why this could be happening, you should probably send an email to ffmpeg-user containing: your complete command line, the ffmpeg console output up to the point where it has already leaked the 1-2G, the valgrind output up to the point where it has leaked 1-2G, the line from top showing how much memory the process is currently using at that time (ie right
[11:21] <c_14> before you stop process execution, you also might want to get it with ps with -o and the info top outputs so you can get the column headers as well). That should be it afaik. Place all that as plain text into the email or as plain text attachements (don't link to outside sources). Ah yes, also the ffserver output, a valgrind of the ffserver process and the ps/top output for that as well. That should hopefully
[11:21] <c_14> cover all the bases.
[11:25] <yoyowinwin> thank you very very much
[11:26] <yoyowinwin> both
[11:26] <yoyowinwin> both  VIRT or RES
[11:26] <yoyowinwin> 	both VIRT and RES
[11:28] <yoyowinwin> the valgrind had run for about one hour
[11:30] <ffffffff11111111> hi, how to set the max/avg bitrate for "libxvid"? TIA
[11:30] <ffffffff11111111> "-c:v libxvid -b:v 1000k" results in "2524.1 kbps (308.1 kbyte/s)" file
[11:30] <ffffffff11111111> "-c:v libxvid -b:v 100k" - the same result
[11:30] <ffffffff11111111> "-c:v libxvid -b 1000k" results in "Please use -b:a or -b:v, -b is ambiguous" and the bitrate is ignored again
[13:33] <termos> is the FFT functionality bundled with ffmpeg something that I can use to do the fft and inverse fft of an AVFrame?
[13:34] <termos> I'd like to use it in some noise reduction stuff that does work in the frequency domain and then transfers back to signal domain
[15:09] <chuckleplant> Hi there
[15:10] <chuckleplant> Regarding the following issue : https://trac.ffmpeg.org/ticket/1783 , that conflicts with Android NDK's system headers
[15:11] <chuckleplant> There is a fix that consists on changing ffmpeg's time.h to avtime.h and the corresponding includes throughout the code... Afaik that violates the LGPL license right? Anyone here knows about this on Android platforms?
[15:18] <BtbN> chuckleplant, just don't put libavutil in your include path.
[15:19] <BtbN> It's not supposed to be on there anyway.
[15:22] <iive> i thought that the convention is to always #include <libavutil/time.h> aka having the pathname.
[15:23] <BtbN> yep
[15:27] <chuckleplant> Thanks! I'm trying that now
[15:34] <chuckleplant> BtbN, I was already including like that, from what I see it's ffmpeg that includes time.h directly. In avformat.h line 253. So I can't control which one is being included from outside ffmpeg.. can I?
[15:46] <iive> chuckleplant: are you sure that that time.h is not there to include the system one?
[16:04] <chuckleplant> iive, I don't think so, there is a time.h file inside the ffmpeg folder hierarchy, within the libavutil folder
[16:07] <iive> chuckleplant: avformat.h and time.h are in different directories.
[16:17] <iive> also the system includes are first and after time.h there is stdio
[16:46] <chuckleplant> Sorry I took so long iive, time.h was definitely being included somewhere in ffmpeg. Causing the issue that was closed as an Android problem. I will downgrade my android NDK version to tackle this without violating LGPL. Thanks anyways!!
[16:47] <Proto> Âàó
[16:47] <Proto> åñòü êòî?
[16:47] <Proto> somebody here?
[16:49] <Proto> can you help me with ffmpeg?I don't understand how to install it,for web server,php,on UBUNTU
[16:53] <Proto> i know how to run it from terminal,it's works great,but it not works from php
[19:05] <relaxed> Proto: That's more of a question for #php
[21:01] <dagerik> how can I draw border on seleced region?
[21:01] <dagerik> im using ffmpeg and capturing x11screen
[21:01] <dagerik> I would like to see which rectangle is being recorded
[21:02] <llogan> -show_region 1
[21:02] <llogan> http://ffmpeg.org/ffmpeg-devices.html#x11grab
[21:03] <dagerik> llogan: ah thanks
[21:12] <dagerik> https://dvikan.no/files/vim-omnicomplete-php.webm
[21:20] <Paul___> Hello there people of the ffmpeg world...
[21:20] <Paul___> I've have a rather simple question...
[21:21] <Paul___> I've been trying to convert an audio file into a 8 bit unsigned pcm plain file (just the samples one after the other)
[21:22] <Paul___> According to my needs this file must be sampled at 8Khz
[21:22] <Paul___> so the command I've come with is like so: http://pastebin.com/eid8TxVP
[21:23] <Paul___> I only need a "small" piece of the file to test some algorithims, that's why I have the -ss and -t part
[21:24] <ChocolateArmpits> You should put -t after input
[21:24] <c_14> Doesn't matter.
[21:24] <Paul___> The wierd thing is that the output file seems to be sampled at 16 Khz
[21:28] <Paul___> yeah, the -t is just to limit the size of the file....
[21:29] <Paul___> what I really find confusing is that the sample rate is really doubled by ffmpeg...
[21:30] <Paul___> I was wondering if it's something that I got wrong
[21:31] <Paul___> or maybe it has to do with this: http://en.wikipedia.org/wiki/Nyquist_frequency
[21:32] <Paul___> meaning that ffmpeg -ar argument actually takes the target frequency and doubles it to make sure there is no aliasing...
[21:33] <Baked_Cake> 16khz might be the lowest it can go
[21:34] <ChocolateArmpits> Did you try using an audio filter to specify sample rate through that ?
[21:55] <Baked_Cake> maybe try some variations or leave it out and see what it does
[21:56] <ChocolateArmpits> Or try out SoX
[23:15] <zzz> How can I do lossless encoding with libx265 through ffmpeg? -qp 0 doesn't seem to work.
[23:22] <zzz> Nevermind, it's -x265-params lossless=1.
[23:24] <Matador> Hows x265 coming along ? :)
[23:27] <llogan> Matador: i don't know, but you could probably browse the commit log in hg (mercurial)
[23:28] <Matador> I just meant in general :)
[23:28] <Matador> right now doing live flash streaming w/ ffmpeg + x264 on Win*
[23:29] <Matador> just friendly talk today llogan :)  How are you doin ? :)
[23:30] <llogan> good. took a long lunch break.
[00:00] --- Thu Oct 16 2014