[Ffmpeg-devel-irc] ffmpeg.log.20121202

[burek]

[00:01] <joules> hi, i have a 560x480 (svideo), is that ok to encode with x264 considering its lowres?
[00:01] <JEEB> uhh, that's not even lowres
[00:01] <JEEB> lol
[00:02] <joules> :/
[00:02] <JEEB> anyways, yes -- x264 is usable just fine
[00:02] <cbreak> x264 can encode stuff that is a multiple of 16 width/height
[00:02] <joules> yeh I haven't seen x264 spit out an "Eww thats not worthy!" error.
[00:02] <cbreak> (otherwise it'll add padding or so)
[00:03] <dericed> haha
[00:03] <dericed> please file an enhancement request for "Eww thats not worthy!"
[00:03] <JEEB> cbreak, it used to be more efficient with mod16 (it would warn if you would encode non-mod16), but that was removed a couple of years ago
[00:04] <joules> or "Use mpeg1 old man!"
[00:04] <JEEB> no
[00:04] <JEEB> that won't happen
[00:32] <cronix> mhm
[00:32] <cronix> windows ffmpeg crashes when trying to sent a live-stream to ffserver running on linux host
[01:02] <llogan> cronix: are you using a recent build?
[01:07] <cronix> i tryed several builds now
[01:07] <cronix> my server is running ffmpeg 0.9
[01:08] <cronix> ive tryed random client prebuild windows packages dating back to somtime in june 2011
[01:08] <cronix> and even latest crashes on me
[01:09] <cronix> it crashes right after or shortly before it connects to remote ffserver
[01:09] <cronix> at least i can tell in the ffserver log that something touched the feed1.ffm at the same time the windows ffmpeg crashed on me
[01:10] <cronix> other variant with ffserver 0.7.12 was a permission denied in recent windows builds when trying to access feed1.ffm
[01:10] <cronix> or crash on older version
[01:11] <cronix> cmdline on windows ffmpeg: ffmpeg -f vfwcap -i 0 http://ip:port/feed1.ffm
[01:11] <cronix> capturing live feed from webcam
[01:12] <cronix> the webcam can be accessed by ffmpeg, it works when i write it out to a mpeg file
[01:12] <cronix> i can play back the mpeg file fine
[01:45] <llogan> cronix: crashes are important to the developers. consider making a bug report.
[01:46] <llogan> does it crash if you use a file as an input instead of the webcam?
[02:27] <TheCycoONE> what's the audio equivalent to sws_scale, if I want to change e.g. the frequency
[02:35] <TheCycoONE> or an example
[02:45] <TheCycoONE> ok, looks like there is swresample and avresample
[02:49] <ubitux> an example for swr will be added soon
[02:49] <ubitux> the doxy contains some though
[02:50] <ubitux> see http://ffmpeg.org/pipermail/ffmpeg-devel/2012-December/135098.html
[02:57] <TheCycoONE> thanks ubitux, I'll see if I can sort it out from the doxygen and this.
[02:58] <ubitux> you can also look at various code using it
[02:58] <ubitux> such as libavfilter/af_aresample
[02:59] <ubitux> or libswresample/swresample-test.c
[02:59] <ubitux> ah, and ffplay.c :)
[03:01] <TheCycoONE> spiffy.
[03:02] <TheCycoONE> I try to avoid reading ffplay.c because of the LGPL - wouldn't want to end up copying it.
[03:03] <TheCycoONE> I think I have enough to go on though
[04:16] <Emmanuel_Chanel> Hello, all!
[04:17] <Emmanuel_Chanel> Is the current git version of ffmpeg ok now?
[04:24] <ubitux> what do you mean by "now"?
[04:24] <ubitux> what wasn't ok?
[04:30] <Emmanuel_Chanel> ubitux: that uses too much resource to do other things...
[04:30] <ubitux> use -threads 1 maybe?
[04:32] <Emmanuel_Chanel> Certainly, it would solve. But formerly, -threads 0 (in my case, 6 cores of AMD's) didn't use too much.
[04:49] <Emmanuel_Chanel> ubitux: Oh... I forgot that it uses too much memory... :(
[07:58] <phunyguy_t430s> just making my rounds.  Any word on when decoding of DTS-HD Master Audio will be fully supported?
[07:58] <sacarasc> When you code it, it will work.
[08:06] <phunyguy_t430s> yeah yeah
[08:06] <phunyguy_t430s> If only I knew how
[08:07] <phunyguy_t430s> I remember hearing about it being a potential project, just wasnt sure if anyone tackled it yet or is tackling it.
[08:08] <phunyguy_t430s> also, with a specific movie and TrueHD, I am getting a few Lossless Check Failed warnings.
[08:08] <phunyguy_t430s> seems like copy protection.
[08:08] <phunyguy_t430s> as well as some End of stream indicated.s/s warnings
[08:25] <phunyguy_t430s> oh, also, what about encoding flac in anything higher than 16 bit?
[08:32] <joecool> phunyguy_t430s: it's done, just not useful in most applications
[08:34] <phunyguy_t430s> joecool, to which question? flac?
[08:35] <joecool> phunyguy_t430s: yes
[08:35] <phunyguy_t430s> not sure what you mean by that
[08:36] <phunyguy_t430s> joecool, can you elaborate?
[08:38] <joecool> phunyguy_t430s: this article explains why http://people.xiph.org/~xiphmont/demo/neil-young.html <
[08:40] <phunyguy_t430s> joecool, that article talks about music.  This is to preserve the quality of movie audio tracks in 5.1 or 7.1 - I'm kind of a nerd like that, however I hate closed codecs like TrueHD, etc.
[08:41] <phunyguy_t430s> But honestly I would be happy with DTS-HD/MA working before 24 bit flac.  For now I can ouput as raw pcm_s24le, and pipe to the flac command
[09:05] <phunyguy_t430s> nevermind, read more into it.
[09:06] <phunyguy_t430s> perhaps they are right.  meh.  Still, would love to see DTS-HD/MA decoded, even if it gets converted to 16 bit in FLAC.
[11:20] <joules> hi, I have a some original source video @ 680x576 has anyone heard of that resolution.
[11:20] <joules> ?
[11:22] <joules> sar is 1.180
[11:22] <vtwo> http://en.wikipedia.org/wiki/576i ?
[11:23] <joules> yeh
[11:24] <joules> more interested in the x resolution and why :/
[11:27] <joules> all i get is thermal imaging/infrared cameras.
[11:35] <vtwo> i have a problem with ffserver and ffmpeg. here's both cmd's + console output and my ffserver.conf, would anyone be so kind and help please? -> http://pastebin.com/KNyVn5mX
[11:36] <vtwo> i get "av_interleaved_write_frame(): Connection reset by peer"
[11:52] <joules> perhaps its 672x576, yet the analog stream was digitised @680
[11:52] <joules> or whatever.
[11:55] <joules> crops pretty close to where its starts to bleed.
[11:55] <joules> hello i'm talking to myself.
[11:57] <joules> oh and a sar of 7:6
[13:25] <joelio> Are there any ruby bindings available for current stable (ubuntu 12.04
[13:25] <joelio> the ones I've tried fail to compile as they look old
[13:26] Action: joelio finds https://github.com/streamio/streamio-ffmpeg -will look
[13:33] <t4nk152> Hello
[13:34] <t4nk152> I've a problem compiling ffmpeg under Centos 6.3
[13:34] <t4nk152> I followed this procedure http://pastebin.com/mgu0YUKC
[13:34] <t4nk152> but making ffmpeg i recevied this error
[13:34] <t4nk152> .. just a moment, it's compiling right now
[13:35] <t4nk152> shared object; recompile with -fPIC /usr/lib/gcc/x86_64-redhat-linux/4.4.6/../../../libx264.a: could not read symbols: Bad value collect2: ld returned 1 exit status make: *** [libavcodec/libavcodec.so.54] Error 1
[13:37] <t4nk152> http://pastebin.com/QV7i6dwx
[13:37] <t4nk152> I pasted pastebin  for detail
[13:37] <t4nk152> someone can help me ?
[13:38] <t4nk152> I tried to remove flag "--enable-libx264" and make operation it's ok .... but I think that this flag is quite important ...
[13:39] <JEEB> if you remove --enable-libx264 then libx264 will not be enabled, if you need H.264 encoding that's a must
[13:39] <JEEB> also that "procedure" seems awfully complicated
[13:43] <vtwo> wouldnt he have to compile x264 with --disable-cli ?
[13:44] <vtwo> dunno if that makes a difference
[13:46] <JEEB> ah
[13:46] <JEEB> "can not be used when making a shared object; recompile with -fPIC"
[13:46] <JEEB> ^ this is the point
[13:47] <JEEB> mixing static and shared, you need to enable PIC
[13:48] <JEEB> t4nk152, recompile x264 with --enable-pic
[13:49] <vtwo> JEEB any idea about solving my ffserver problem i pasted way up there?
[13:49] <JEEB> I have absolutely no idea about ffserver
[13:49] <vtwo> k :(
[15:19] <t4nk152> it doesn't work ....
[15:19] <t4nk152> Iìve recompiled x264 with --enable-pic flag
[15:19] <t4nk152> but making ffmpeg I've same result
[15:20] <JEEB> with another library, right?
[15:20] <JEEB> why are you even mixing shared and static even? why are you even using shared somewhere?
[15:22] <t4nk152> no the same
[15:22] <t4nk152> libx264a
[15:23] <t4nk152> I don't know beacuse, this my first time that I tried to compile ffmpeg
[15:23] <t4nk152> Now I'm triyng to recompile ffmpeg without --enable-pic
[15:24] <JEEB> poor soul :P Having no idea about anything
[15:24] <JEEB> also did you make sure that ffmpeg's compilation picks up the newly PIC-built x264?
[15:35] <t4nk152> how can I sure of that ?
[15:58] <smj> when I'm capturing video with x11grab, audio seems to fall behind over time...
[15:59] <smj> should I figure out what's causing it and fix that, or should I use some audio syncing thing?
[16:00] <smj> documentation mentions this 'asyncts' filter but I don't see it listed when running 'ffmpeg -filters'
[16:02] <smj> if it has to be enabled in configuration when compiling, is it not the optimal solution?
[16:16] <ubitux> smj: check if playing with -async helps
[16:17] <smj> ok
[16:18] <ubitux> if you need more control, look at the aresample filter
[16:21] <ubitux> seems the documentation is pretty weak on this mmh.
[16:21] <JEEB> smj, do you get frame drops or anything else like that?
[16:22] <JEEB> aka are you sure that your video encoding is going fast enough
[16:23] <ubitux> smj: if you have a recent git, look at the compensation options in doc/ffmpeg-resampler.texi
[16:23] <ubitux> min_comp, min_hard_comp, comp_duration, ... etc
[16:23] <ubitux> then you can pass them to af resample
[16:23] <ubitux> something like -af resample=min_comp=...:comp_duration=... etc
[16:24] <ubitux> -af aresample sorry
[16:24] <ubitux> note that -async option insert the filter with the appropriate settings, but you have less control
[17:59] <smj> and setting -itsoffset 100 for video doesn't do anything
[18:28] <smj> -itsoffset for audio works
[20:21] <tekk> hi guys, just building a small tool for myself to demux mkv's on my NAS. As its embedded, I wanna use libavcodec/format rather than calling ffmpeg --options directly& is this possible using the API's available? any tips?
[20:32] <Mavrik> tekk, it's possible, APIs are available and documented in ffmpeg source code
[20:32] <Mavrik> tekk, and I recommend you use ffmpeg anyway if you're not really familiar with libav APIs and C coding
[20:32] <tekk> i'm a C++/Objective-C dev, and like a bit of a challenge :) so will read docs and give it a go
[20:32] <tekk> thanks
[21:48] <smj> do I have to set -itsoffset for video after -vsync?
[22:00] <brx_> should i be using fmpeg or mencoder to concatenate multiple mp4 files?
[22:00] <brx_> ffmpeg*
[22:06] <smj> if you have a recent-enough version http://stackoverflow.ccom/a/11175851
[22:07] <smj> also, you should read the small print
[22:07] <brx_> ok thanks smj
[22:09] <smj> try it, but I'm not sure if it works, I don't trust mp4
[22:10] <brx_> there needs intermediary with ffmpeg
[22:10] <brx_> thing is i have compiled ffmpeg for arm, i am bundling with an app im writing
[22:11] <brx_> i only wanna bundle either ffmpeg or mencoder, not both
[22:11] <brx_> overhead is low
[22:16] <brx_> smj, http://www.molecularbear.com/blog/?p=799
[22:16] <brx_> that worked
[22:17] <brx_> weird though, both vids are 1.5 seconds long...in the end result file the first file takes more duration
[22:18] <brx_> its like its trimmed the end of the second file
[22:18] <brx_> not too bothered though tbh lol
[22:25] <witheld> Okay no one in libav will help so I guess I'm using ffmpeg....I cannot figure out how to get video4linux input
[22:25] <witheld> It just breaks in various ways
[22:42] <klaxa> ffmpeg -f video4linux -i /dev/video0 ?
[22:44] <klaxa> witheld, if you wouldn't withhold (lol) your error messages we could help
[22:44] <klaxa> pastebin everything and we might be able to help
[00:00] --- Mon Dec  3 2012