[Ffmpeg-devel-irc] ffmpeg.log.20120828

[burek]

[00:05] <JEEB> Iszak, that's the recommended way, but if you have time to use you could try to play around the APIs Intel gives
[00:05] <Iszak> I did a google around for the API's they seem to be lacking information.
[00:08] <Iszak> JEEB: well that goes any chance for me, the SDK is built for Windows, I'm on Mac.
[00:09] <JEEB> there's a linux API as well, dunno about macs
[00:10] <Iszak> no I mean the SDK is built for windows, the api is cross platform.
[00:11] <JEEB> uhh
[00:12] <JEEB> http://cgit.freedesktop.org/libva/
[00:16] <Iszak> JEEB: yeah.. nothing about it.
[00:19] <JEEB> Iszak, they even have basic apps there as documentation of how to use the API :P
[00:20] <JEEB> http://cgit.freedesktop.org/libva/tree/test/encode/avcenc.c
[00:20] <Iszak> JEEB: hang on what? is that repo for libavcodec or intels?
[00:21] <JEEB> that's the intel libva repo
[00:21] <JEEB> libavcodec is part of ffmpeg/libav and that's where video decoders and encoders go (as opposed to container-level stuff that goes to libavformat)
[00:22] <Iszak> so someones added the examples into the libav repo?
[00:22] <JEEB> ...
[00:23] <JEEB> re-read my sentences
[00:24] <Iszak> not following.
[00:24] <JEEB> ffmpeg and libav are the two projects that provide libavcodec and friends, libva is intel's thing
[00:25] <Iszak> Oh right, gotcha.
[00:25] <Iszak> so basically look at libva's example, port it to libavcodec?
[00:25] <JEEB> that'd be one way of looking at it, yes
[00:26] <Iszak> I don't know C++ though, I'm high level programming, python, ruby, php.
[00:26] <JEEB> you could take a look at one of the video encoders that use separate encoding libraries over at the libavcodec folder
[00:26] <JEEB> Iszak, it's alright, I was barely making patches to x264 last year :P
[00:26] <JEEB> and now I coded a simple'ish video encoder over the summe
[00:26] <JEEB> *summer
[00:26] <JEEB> also it's C
[00:26] <JEEB> not C++
[00:26] <Iszak> ouch.
[00:27] <Iszak> way over my head I think.
[00:27] <JEEB> although C++ is somewhat tolerated at ffmpeg, like the libutvideo thing in libavcodec
[00:28] <JEEB> Iszak, dunno, with some effort you can grind in rather well
[00:29] <JEEB> looking at either libutvideoenc.cpp or libx264.c should give you some sense of how to make whatever you want usable in libavcodec
[00:30] <Iszak> I think it's probably best to leave it to the experts, knowing my luck I'd cause a memory leak.
[00:31] <JEEB> valgrind's for that
[00:31] <Iszak> Still, I have no knowledge of how video encoding/decoding works.
[00:32] <JEEB> well, you only need to know about encoding in this case
[00:32] <Iszak> oh? how come?
[00:33] <JEEB> because you'd be wriiting an encoder and you'd just limit your encoder to certain type of colorspace (most probably 4:2:0 YCbCr, planar), and you'd be getting a buffer with a picture in that format
[00:33] <Iszak> See, this is why it's not a good idea. I know nothing about colorspace, etc.
[00:33] <JEEB> and then you just do what you need with that buffer, and return a packet (the first thing an encoder does usually is to set the packet unless it's already given to you)
[00:34] <JEEB> http://www.xiph.org/video/vid1.shtml
[00:34] <JEEB> ^ primer to digital video and audio
[00:34] <Iszak> bookmarked it, along with that other link and will watch it when I get time. A bit late to start hacking up something at 11:30 PM
[00:35] <JEEB> also planar means that you basically have all of the things in their own memory plane (like with planar R,G,B you'd have one plane with all the R, second with all the G and third with all the B)
[00:35] <JEEB> non-planar would be just one plane with everything one after another
[00:35] <JEEB> (RGB is usually non-planar)
[00:36] <Iszak> is this how the video is stored, or how it's done in programming? converting all the rgb values to separate arrays?
[00:36] <JEEB> depends on the format on how it's stored
[00:36] <JEEB> the encoder I did had everything planar inside
[00:36] <JEEB> so I had non-planar RGB which I then planarized
[00:37] <Iszak> surely there's a utility to planar everything?
[00:37] <JEEB> what is "everything"?
[00:38] <Iszak> well it doesn't make much sense to have to write (the same?) code to convert non-planar into planar all the time, surely.
[00:38] <JEEB> dunno, didn't know of such a function so I just wrote a function myself
[00:40] <JEEB> anyways, each encoder has an infobox like this http://git.videolan.org/?p=ffmpeg.git;a=blob;f=libavcodec/utvideoenc.c;h=e497333d06bd9feb499a14ba128af148df979533;hb=HEAD#l610
[00:41] <JEEB> which contains the name, the type, the id, the functions to be called for initalization of the encoder, the actual encoding and for closing it, and then the pixel formats it can take in
[00:41] <JEEB> and then the "human readable" name
[00:44] <JEEB> something even simpler is the v410 encoder I guess, as a basic example of how an encoder works
[00:50] <Iszak> JEEB: I appreciate the enthusiasm to get me involved, but I'm unlikely to develop an encoder.
[00:51] <JEEB> whatever, you seemed to be interested in trying to get something working with that thing :P
[00:51] <JEEB> or well, interested in the feature set generally
[00:53] <Iszak> JEEB: I would give it a try if I knew C/C++ and an understanding of digital video.
[00:53] <JEEB> heh
[00:53] <Iszak> But alas, I do not - and I probably never will write C/C++ as it's just too low level for me. I'm not used to GC and static typing.
[00:54] <JEEB> I kind of feel you're underestimating yourself if you've coded already. I mean, if _I_ can code up an encoder, I'm pretty sure you can too, esp. when it's just about feeding some third-party function the raw data in a way it likes it
[00:54] <JEEB> and then taking the output bitstream and stuffing it into a packet
[00:54] <Iszak> JEEB: are you doing a BCS degree?
[00:54] <JEEB> bachelor computer science?
[00:55] <Iszak> yes, or any computer-related degree?
[00:55] <JEEB> aye
[00:55] <JEEB> that said, I've only coded java, php, haskell and python at the university
[00:55] <Iszak> I thought as much, when you mentioned summer holidays.
[00:55] <Iszak> only? ha. more than most people touch on.
[00:56] <Iszak> I've dabbled in Java, PHP, Ruby, Python, ActionScript 3, JavaScript, C#, as you can see, none of which are low level.
[00:56] <JEEB> C# is close enough to java and friends, and other C-likes
[00:57] <JEEB> http://c.learncodethehardway.org/book/
[00:57] <JEEB> quickly going through this should get you started with C
[00:57] <Iszak> oh yeah, C# and Java are quite similar, just the API/Libs. AS3 and JS are similar too.
[00:59] <Iszak> JEEB: most of it I can understand, it's just things like *var and &var which confuses me the most.
[00:59] <JEEB> *var = value of pointer var
[01:00] <JEEB> as in, the value of var gets interpreted as a memory address and the thing that it's leading to is used
[01:00] <JEEB> &var = address of var
[01:00] <JEEB> also * is often used to define that something is a pointer in function definitions
[01:01] <JEEB> int herpderp(uint8_t *src) {}
[01:01] <JEEB> function herpderp that takes in a pointer to a memory area of 8 bit values
[01:02] <Iszak> and it returns an integer?
[01:02] <JEEB> yup
[01:02] <JEEB> anyways, the book covers pointers pretty well as well
[01:03] <Iszak> yeah, I saw.
[01:03] <Iszak> Learn C The Hard Way, heh.
[01:03] <JEEB> it's a surprisingly good guide
[01:04] <Iszak> You're a man of good references.
[01:04] <Iszak> JEEB: what about unit tests?
[01:05] <JEEB> ffmpeg uses the fate system to keep sure encoders and decoders output what they're supposed to
[01:05] <JEEB> so you basically have test files and then you have various decoders and encoders run over them with various settings
[01:06] <JEEB> http://fate.ffmpeg.org/ <- and then ffmpeg has its own set of machines running the tests on various architectures
[01:09] <Iszak> well, maybe if I get some spare time I'll look into all this, it's a bit late now.
[01:09] <JEEB> sure
[01:10] <Iszak> book marked that learncodethehardway, seems good.
[01:30] <hypnocat> so, here's what i had to do to get ffplay to work with jack:
[01:30] <hypnocat> first, i had to start jackd and start jamin (an eq app)
[01:30] <hypnocat> then, i put this in to my ~/.asoundrc:  http://pastie.org/4599911
[01:31] <hypnocat> and when i started ffplay, the sound output went to via jack to jamin (which is a jack app), and through jamin to my speakers
[01:31] <hypnocat> for some reason, i couldn't get this to work without first outputting to jamin
[01:32] <hypnocat> i tried putting this in to my ~/.asoundrc instead:  http://pastie.org/4600283
[01:33] <hypnocat> but that didn't work
[01:33] <hypnocat> ffplay would show up in patchage, but it wasn't connected to my system output ports, and when i manually connected it, no sound would come out
[01:34] <hypnocat> i also tried this:  http://pastie.org/4600290
[01:34] <hypnocat> that made ffplay connect to my system output ports, but the sound was all weird, sped-up and distorted
[01:35] <hypnocat> so i think the connection to jamin was necessary for this to work.. though once ffplay was running and the connection was made, then i could use patchage to disconnect ffplay from jamin and connect it directly to my system output ports and it worked fine
[01:36] <ubitux> isn't ffplay using SDL audio output?
[01:36] <hypnocat> someone here said earlier that it does
[01:37] <saste> ubitux: SDL -> alsa -> jack
[01:38] <ubitux> ok :)
[04:32] <FelipeS> what's the easiest container to segment?
[04:43] <arch_is_awesome> Hey, I was wondering where I could get more presets for ffmpeg libvpx?
[04:44] <arch_is_awesome> There have to be more than the ones in the ffmpeg git repository... http://wiki.oz9aec.net/index.php/High_quality_screen_capture_with_Ffmpeg shows the usage of lossless_ultrafast
[04:51] <JEEBsv> that's for libx264, and nowadays you would select a lossless rate control mode and just use -preset ultrafast (lossless rate control modes are -crf 0 and -q 0 with libx264)
[04:51] <JEEBsv> (ffmpeg now has a way to use the libx264's internal presets, so separate preset files for presets pretty much became unneeded)
[04:54] <arch_is_awesome> Oh, so I can run lossless_ultrafast (example) without a file?
[04:54] <arch_is_awesome> JEEBsv: It does not seem to be working :P
[04:56] <JEEBsv> why don't you actually read what I wrote?
[04:56] <arch_is_awesome> I _did_
[04:57] <JEEBsv> ok, just that there is no such preset s lossless_ultrafast
[04:57] <arch_is_awesome> Ah, but there used to be?
[04:57] <JEEBsv> as an ffpreset file for libx264, maybe. Which either set crf to zero or q to zero, and set the settings for -preset ultrafast
[04:58] <arch_is_awesome> How do I get the options for -preset?
[04:58] <arch_is_awesome> *available options
[04:59] <JEEBsv> http://mewiki.project357.com/wiki/X264_Settings#preset
[04:59] <JEEBsv> I don't think other things than libx264 have internal presets
[04:59] <arch_is_awesome> Is there anything like that for vp8?
[04:59] <JEEBsv> nope
[05:00] <JEEBsv> one day when xvp8 gets finished, maybe
[05:00] <JEEBsv> but not now
[05:00] <JEEBsv> only ffpreset files defined by someone
[05:00] <arch_is_awesome> So, how would I get libvpx to encode quickly and with good quality?
[05:00] <arch_is_awesome> (I don't care about size)
[05:00] <JEEBsv> set the quants low enough and try to find fast settings?
[05:01] <JEEBsv> the last I checked libvpx wasn't really on par with libx264 with regards to speed tho
[05:01] <arch_is_awesome> Ah, okay. Thank you!
[05:01] <JEEBsv> so I would most probably use it, or something like ffvhuff for the capture
[05:01] <JEEBsv> -qp 0/-crf 0 for libx264 for lossless mode
[05:01] <JEEBsv> ugh, q not qp most probably
[05:02] <JEEBsv> (x264's command line encoder uses --qp)
[05:02] <arch_is_awesome> Maybe capturing in raw would work?
[05:02] <arch_is_awesome> And then encoding later?
[05:02] <JEEBsv> lol, only if your hard drives are quick enough
[05:02] <JEEBsv> better to use something fast lossless
[05:02] <arch_is_awesome> I've got an ssd
[05:02] <JEEBsv> ffvhuff or maybe ultrafast libx264 lossless
[05:02] <JEEBsv> both should be quick
[05:02] <arch_is_awesome> Okay, thanks
[05:02] <JEEBsv> and ffvhuff is one of the most simplest formats around
[05:03] <JEEBsv> (basically prediction + huffman)
[05:03] <arch_is_awesome> Thanks.
[05:04] <JEEBsv> raw video generally tends to be a drag, even if it possibly would use less power to encode
[05:04] <arch_is_awesome> Okay, GTG now. Bye
[05:04] <arch_is_awesome> Thank you again :)
[05:04] <JEEBsv> I did meet one girl who insisted on raw RGB capture, and in the end ended up buying some SAS or whatever thingy to have the bandwidth
[05:05] <arch_is_awesome> I'll take x264 with lossless and then convert it to vp8 later.
[05:05] <JEEBsv> well, whatever is quick enough for your needs
[06:14] <redxiizzle> this doesn't compile in Mingw: http://git.videolan.org/?p=ffmpeg.git;a=commitdiff;h=78cca74222157e3b0f1ad9ce9a14a31c525f69ea
[06:15] <redxiizzle> CoreFoundation.h is missing (doesn't exist), two other source files mention it, device.c "#ifdef DARWIN_DVD_IOCTL" or path.c "#ifdef CONFIG_MACOSX_BUNDLE" but no ifdef for h264 vda
[06:16] <redxiizzle> libavcodec/vda_h264_dec.c:28:43: fatal error: CoreFoundation/CoreFoundation.h: No such file or directory
[06:16] <redxiizzle> compilation terminated.
[06:17] <redxiizzle> vda_h264.c also includes some CoreFoundation files
[06:23] <pingUone> how do you drop the idv3 tags and framerate info into a textfile?
[06:26] <pingUone> or at least direct the output of ffmpeg -i to stdout?
[06:52] <hotfm> hi
[10:29] <dv_> hi
[10:30] <dv_> when I call ffmpeg -i input.mp3 -map_metadata 0:g:0 -ar 48000 -ab 192000 output.mp3 , the produced mp3 contains the input.mp3 metadata, as it should
[10:30] <dv_> however, I noticed that album art originally contained as jpeg is now present as png
[10:31] <dv_> is there a way to prevent this conversion?
[10:31] <dv_> I just want ffmpeg to copy metadata without transforming it at all
[11:54] <tiborfabian> Hi everyone. Can i use ffmpeg also to normalize the audio track of video files?
[11:55] <saste> tiborfabian: you need a two-pass process
[11:56] <saste> check the volume and volumedetect filters
[11:56] <saste> volumedetect is pretty new, so you need a pretty recent version of ffmpeg
[12:58] <zap0> what other 'detect' filters are there?
[12:59] <ubitux> libavfilter/af_silencedetect.c  libavfilter/vf_blackdetect.c  libavfilter/vf_edgedetect.c
[12:59] <ubitux> libavfilter/af_volumedetect.c   libavfilter/vf_cropdetect.c
[15:13] <Rodeo> hi!
[15:13] <Rodeo> trying to play with the recently-committed H.264 VDA decoder
[15:14] <Rodeo> seems to be compiled successfully, I think: http://pastebin.com/qNGsGjUb
[15:14] <Rodeo> h264_vda in Enabled decoders
[15:14] <Rodeo> but neither ffplay nor ffmpeg seem to use it
[15:14] <Rodeo> the video I'm trying to decode and/or re-encode is an M2TS from Blu-ray
[15:16] <rocktop> I am not able to convert avi to mp4 with ffmpeg and h264 can some help me to get the right comman line ?
[15:16] <rocktop> command *
[15:16] <Rodeo> I've added logging and it seems that libavcodec/h264.c decode_frame() is being used, rather than libavcodec/vda_h264_dec.c vdadec_decode()
[15:18] <rocktop> anyidea ?
[15:19] <Rodeo> well, what command-line are you currently using?
[15:22] <rocktop> Rodeo:  ffmpeg -i eoxp.avi -acodec aac -b:a 128k -vcodec libx264  -b:v 1200k -flags +aic+mv4 -strict experimental eoxp.mp4
[15:23] <Rodeo> and that's not working, how?
[15:27] <rocktop> Rodeo: http://pastebin.com/1xHc6CNd
[15:28] <Rodeo> x264 [error]: malloc of size 107520 failed
[15:28] <Rodeo> doesn't look like something you're doing wrong (keeping in mind, I'm not very familiar with using libx264 via ffmpeg)
[15:33] <alyawn> Is there a precedent for exposing additional optional SEI message types in the h264 decoder included in libavcodec?
[15:34] <rocktop> Rodeo: I am using  libx264 just for embidding the video in flash player , what codec I should to use to get mp4 work under flashplayer ?
[15:35] <Rodeo> rocktop: my point is, AFAIK, you're doing the right thing
[15:35] <Rodeo> for some reason, x264 fails to allocate memory for something
[15:38] <rocktop> Rodeo:yes exactly I have this problem
[15:38] <rocktop> with some videos
[15:40] <Rodeo> could be a bug somewhere
[15:41] <rocktop> Rodeo: I wouldl ike to test something else could you please give me a command line to convert avi to mp4 and get it work in flashplayer ?
[15:42] <Rodeo> I'm not 100% sure what other encoders would work with Flash
[15:43] <Rodeo> but let's try something: does it work if you use, say, -vcodec mpeg4?
[15:43] <Rodeo> i.e. does the encoding fail like it does for x264, or does it complete successfully?
[15:43] <rocktop> Rodeo: doesn't work
[15:44] <Rodeo> pastebin?
[15:44] <JEEB> btw, what are -flags +aic+mv4 ?
[15:45] <JEEB> I've never seen those used and they aren't needed for flash compatibility
[15:45] <rocktop> Rodeo: mpeg4 codec doesn't wok with flash
[15:45] <Rodeo> yes, but that's not the point
[15:46] <Rodeo> does encoding work or does it fail like it does if you encode with x264?
[15:46] <rocktop> Rodeo: not fail
[15:46] <Rodeo> hmm, OK
[15:47] <Rodeo> what you could try is encoding to a intermediate format like DNxHD or ProRes, and then re-encode that to H.264 with x264
[15:47] <Rodeo> maybe than x264 encoding won't fail
[15:48] <JEEB> shouldn't matter
[15:48] <JEEB> the encoder would have to malloc as much
[15:49] <rocktop> this video work with flash http://h264-demo.code-shop.com/demo/apache/trailer2.mp4 I need to encoding to same this codec or similare one
[15:50] <rocktop> JEEB : I will try with mpeg4 without -flags +aic+mv4
[15:51] <JEEB> uhh, well if mpeg4 works you don't have to test it without it :P
[15:51] <JEEB> also
[15:51] <JEEB> I have no idea how your libx264 is compiled
[15:51] <JEEB> --extra-cflags='-O2 -g -pipe -Wall -Wp,-D_FORTIFY_SOURCE=2 -fexceptions -fstack-protector --param=ssp-buffer-size=4 -m64 -mtune=generic -fPIC'
[15:51] <JEEB> some of these might be stopping you from mallocing
[15:51] <JEEB> but first of all
[15:51] <JEEB> tell me how much ram you have :P
[15:51] <JEEB> usable, that is
[15:52] <rocktop> JEEB: 8GB
[15:56] <JEEB> rocktop, if without those flags it still fails, then I'm just going to have to say that either your system is derped, or that your ffmpeg build is broken
[15:56] <JEEB> ffmpeg/libx264
[15:58] <rocktop> JEEB : with mpeg4 its not going to fails
[15:58] <JEEB> yes, but it's a separate code path
[15:58] <rocktop> JEEB: I have this error : x264 [error]: malloc of size 118272 failed
[15:59] <JEEB> the libx264 or ffmpeg you have compiled there is broken, or suddenly your system cannot malloc more memory
[15:59] <JEEB> I see some weird cflags in there
[15:59] <JEEB> that have been manually added
[15:59] <JEEB> so yeah
[15:59] <JEEB> :P
[15:59] <rocktop> jeeb: ok what I should to do ?
[16:00] <JEEB> try compiling libx264 and ffmpeg yourself?
[16:00] <JEEB> shouldn't be too hard as you only need libx264 as a dependency
[16:00] <JEEB> you need yasm, a compiler and git
[16:02] <rocktop> JEEB: I just installed them from pkg
[16:03] <rocktop> JEEB: you advice to to compile them from source ?
[16:03] <JEEB> (install/compile current yasm) -> git clone x264 -> ./configure --enable-static --prefix=/home/username/ownapps -> make -> make install -> git clone ffmpeg -> PKG_CONFIG_PATH=/home/username/ownapps/lib/pkgconfig ./configure --enable-gpl --enable-libx264 --prefix=/home/username/ownapps
[16:03] <JEEB> this is all you need
[16:03] <JEEB> yes, my advice would be to compile  from source without any weird additions to settings
[16:04] <rocktop> ok I will install them from the source now
[16:04] <JEEB> just don't use any weird settings and use current sources :P
[16:59] <rocktop> JEEB: I have installed them as you provide me I have this error : http://pastebin.com/WPcs4v9B
[16:59] <JEEB> drop those flags
[17:00] <JEEB> also if you want to use libfaac, use -acodec libfaac
[17:00] <JEEB> not aac
[17:00] <JEEB> and remove the -strict experimental
[17:02] <JEEB> also I see you didn't use up-to-date x264
[17:03] <JEEB> also it seems like your input audio trac is borked
[17:03] <JEEB> [mp3 @ 0x322c1c0] Header missing
[17:04] <JEEB> also I have no idea which version of ffmpeg you've built
[17:11] <rocktop> JEEB: I get the version from git clone git://git.videolan.org/ffmpeg.git ffmpeg
[17:12] <JEEB> I'm just not seeing the hash within the commits from the last two hours or so
[17:12] <rocktop> JEEB: I do not understand what you mean by "also I see you didn't use up-to-date x264"
[17:12] <JEEB> core 118
[17:12] <JEEB> the current libx264 is version 122
[17:13] <JEEB> git://git.videolan.org/x264.git
[17:15] <rocktop> JEEB: I get it from here git clone git://git.videolan.org/x264.git
[17:15] <JEEB> then why is it not up to date?
[17:15] <JEEB> or wait
[17:16] <JEEB> inb4 your current libx264 got linked instead
[17:16] <JEEB> even with the PKG_CONFIG_PATH set
[17:16] <JEEB> --extra-cflags="-I/where/you/installed/x264/include" --extra-ldflags="-L/where/you/installed/x264/lib"
[17:16] <JEEB> ^ for ffmpeg
[17:17] <JEEB> also the current ffmpeg is hash-wise 416d2f7
[17:17] <rocktop> JEEB: in command line ?
[17:17] <JEEB> try git pulling
[17:17] <JEEB> in configure
[17:17] <JEEB> git pull -> make clean -> re-run configure with those two added -> make
[17:18] <rocktop> right now I converting an other video I don't see any warning or error at this time
[17:18] <JEEB> yes, I'm pretty sure that was just a broken audio stream
[17:18] <JEEB> but you should still update your libx264 from 118 to 122
[17:20] <rocktop> JEEB: how can I check which version of libx264 I have installed ?
[17:20] <JEEB> well, the one your ffmpeg is linked against is pretty damn obvious
[17:20] <JEEB> x264 outputs its version and settings string when you start encoding
[17:20] <JEEB> and the pastebin you posted has it still say 118
[17:23] <rocktop> JEEB:  264 - core 118 - H.264/MPEG-4 AVC codec
[17:23] <JEEB> yes
[17:23] <JEEB> that means your ffmpeg is still linked against version 118 :P
[17:24] <rocktop> JEEB: I get this version from git it should point to newest version
[17:24] <JEEB> well, that's not what got linked :P
[17:24] <JEEB> as I said
[17:25] <JEEB> git pull your ffmpeg code base, make clean, configure with those two settings I added, make once more
[17:25] <rocktop> JEEB: ok
[17:26] <Rodeo> shouldn't he first make sure that the latest x264 is actually installed?
[17:26] <JEEB> well he is saying he did install it :P so I'm believing him, the 118 is most probably his packaged libx264
[17:26] <JEEB> which is going in because a specific -I and -L aren't set
[17:26] <Rodeo> OK
[17:27] <JEEB> if it will still be 118 after he specifically sets -I and -L I will tell him to remove his packaged libx264
[17:34] <rocktop> JEEB: ./configure --enable-gpl --enable-libfaac --enable-nonfree --enable-libx264 --prefix=/usr/local/src/ownapps --extra-cflags="-I/usr/local/src/x264/include" --extra-ldflags="-L/usr/local/src/x264/lib"
[17:35] <JEEB> yeah
[17:35] <JEEB> now see what actually gets linked
[17:46] <rocktop> JEEB: 264 - core 118
[17:46] <rocktop> same version
[17:46] <JEEB> ok...
[17:46] <JEEB> remove your packaged libx264, it's being used
[17:48] <rocktop> from system ?
[17:48] <JEEB> or before that, tell me what you get in your x264 git repo when you look at x264.h
[17:48] <JEEB> it should have a define in the beginning
[17:49] <JEEB> http://git.videolan.org/gitweb.cgi?p=x264.git;a=blob;f=x264.h;h=b6c258eccba24b5414ac1f78c1e651511f157eff;hb=HEAD#l44
[17:49] <JEEB> #define X264_BUILD 125 is current
[17:49] <JEEB> (not 122 as I was saying)
[17:50] <rocktop> #define X264_BUILD 125
[17:50] <JEEB> yup
[17:50] <JEEB> so the x264 you've built is not getting linked
[17:50] <JEEB> did you build x264 with --enable-static btw?
[17:51] <rocktop> JEEB: this is what I used ./configure --enable-static --prefix=/usr/local/src/ownapps
[17:51] <JEEB> ok
[17:51] <JEEB> that should work
[17:51] <JEEB> and you make installed, right?
[17:51] <rocktop> exact
[17:51] <JEEB> ok
[17:51] <JEEB> wait
[17:52] <JEEB> then why are you
[17:52] <JEEB> -I/usr/local/src/x264/include <-
[17:52] <JEEB> -L/usr/local/src/x264/lib <-
[17:52] <JEEB> why the x264?
[17:52] <JEEB> if the prefix is /usr/local/src/
[17:52] <JEEB> it should be /usr/local/src/include and /usr/local/src/lib
[17:53] <rocktop> JEEB: oh Sorry I will built it once again
[17:54] <JEEB> make clean ffmpeg, re-configure and re-make
[17:54] <rocktop> sould be /usr/local/src/ownapps/include not /usr/local/src/include ?
[17:54] <JEEB> depends on the prefix you installed x264 into
[17:55] <JEEB> <prefix>/include and <prefix>/lib
[18:02] <rocktop> JEEB: 264 - core 125 r2208 d9d2288
[18:03] <JEEB> there we go
[18:03] <JEEB> r2208 is current revision of x264
[18:03] <rocktop> JEEB: checking ...
[18:12] <rocktop> JEEB: Thank you I have last question , what is codec used to convert this   http://pastebin.com/eXgYmCKU I get it work in flash but my converted video won't to work since I used same codec
[18:12] <rocktop> my video here http://pastebin.com/696sbr8t
[18:13] <rocktop> ffmpeg -i test.avi -acodec libfaac -b:a 128k -vcodec libx264  -b:v 1200k  test.mp4
[18:13] <JEEB> those should both work on flash, you only need to put the index atom to front after encoding to have it load straight up
[18:14] <rocktop> JEEB: how ?
[18:15] <JEEB> either the qtfaststart app that is in the ffmpeg source, or by this python script https://github.com/danielgtaylor/qtfaststart
[18:18] <rocktop> JEEB: I get it under ffmpeg/tools
[18:19] <rocktop> JEEB: I do :  make qt-faststart && cp qt-faststart /usr/local/bin
[18:22] <rocktop> and I recnvert the vide but still same problem
[18:22] <JEEB> what is this exact problem?
[18:22] <JEEB> an mp4 file with H.264 video and aac audio should work just fine
[18:23] <JEEB> flash by itself has no special limitations
[18:25] <rocktop> JEEB: the exact problem is I have 2 video one external work and my convert one doesn't work since they have same codec as I show you here mine: http://pastebin.com/696sbr8t ext: http://pastebin.com/eXgYmCKU
[18:26] <JEEB> could you please say more than "doesn't work"
[18:26] <JEEB> that doesn't say anything to me
[18:26] <JEEB> I mean, the only weird thing I see in "your file" is that it gives a warning
[18:26] <JEEB> [mov,mp4,m4a,3gp,3g2,mj2 @ 0x2a21240] multiple edit list entries, a/v desync might occur, patch welcome
[18:27] <JEEB> but really, I'm not sure if that affects whether or not flash likes it
[18:27] <JEEB> I've tested flash enough to know that it generally is pretty damn lenient on what exactly you stick into it
[18:27] <JEEB> and generally that "your file" looks just fine
[18:28] <rocktop> JEEB: THank you very much I will check with other video
[18:28] <JEEB> that said, as I've noted already, I have no effing idea what exactly you're doing and what you're trying -- all I know is that you're trying to make something that should work in flash. You should already have that.
[18:31] <rocktop> JEEB: THank you
[18:33] <Mariol> hello, I'm trying to watermark using ffmpeg but I have a problem. The command and the output are both here http://pastebin.com/BN6tRba3
[18:33] <Mariol> may you please help me?
[18:39] <Mariol> I tried searching the web about watermarking with ffmpeg but it seems there are not a lot of official informations about that
[18:40] <zap0> officially from whom?
[18:40] <Mariol> from the ffmpeg website or ffmpeg dev team
[18:40] <ubitux> Mariol: this is not ffmpeg
[18:41] <Mariol> ubitux isn't this channel the official support irc chat?
[18:41] <zap0> Mariol, what lead you to believe your command line arguments are right?
[18:42] <ubitux> i mean what you are using isn't ffmpeg
[18:42] <ubitux> use an up-to-date ffmpeg, and check the examples at https://ffmpeg.org/ffmpeg.html#overlay-1
[18:42] <ubitux> especially the filter_complex example
[18:43] <Mariol> thank you very much
[18:44] <ubitux> feel free to re-pastebin after that if you have trouble or if it doesn't suit your needs
[18:45] <ubitux> here are static build for quick testing
[19:41] <raven> do you have any experience with watchfolder automation with existing scripts?
[21:05] <raven> are there any running tools for doing watchfolder transcoding automation?
[21:06] <Neuhmaan> Hi !
[21:06] <Neuhmaan> I want to record Screen and Sound. Screencast work, but I can't get audio :/
[21:42] <cbsrobot> raven: like with inotify ?
[21:50] <raven> cbsrobot ?
[21:50] <cbsrobot> raven: I havent heard of any existing tool
[21:51] <cbsrobot> but I think you can do it with inotify
[21:53] <raven> oh thats a good hint
[21:53] <raven> ill look for a way using it tnx :)
[23:31] <Lns> Is it true that ffmpeg2theora can only encode video -OR- audio with a single instance? I'm trying to encode video+audio and send it to my icecast server but only video is being sent. I did online research and a couple of places tell me I need to first use 'ffmpeg' and then pipe it to 'ffmpeg2theora'. Is this true?
[23:36] <Lns> It sure doesn't seem that way anyway since there's both --no-video and --no-audio parms
[00:00] --- Wed Aug 29 2012