[Ffmpeg-devel-irc] ffmpeg-devel.log.20120611

burek burek021 at gmail.com
Tue Jun 12 02:05:03 CEST 2012


[02:21] <CIA-119> ffmpeg: 03James Zern 07master * rcb7f062286 10ffmpeg/libavcodec/libvpxenc.c: 
[02:21] <CIA-119> ffmpeg: libvpxenc: add support for forced key frames
[02:21] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[03:04] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r0478d5f9fe 10ffmpeg/ffmpeg.c: 
[03:04] <CIA-119> ffmpeg: ffmpeg: allow overriding framerate for stream copy
[03:04] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[07:42] <CIA-119> ffmpeg: 03Clément BSsch 07master * r9846a9c701 10ffmpeg/ (doc/ffprobe.texi doc/general.texi libavformat/riff.c): riff: add ISMP/timecode tag.
[07:42] <CIA-119> ffmpeg: 03Clément BSsch 07master * r9fb2e234d0 10ffmpeg/ (4 files in 3 dirs): movenc: add timecode track support.
[07:47] <CIA-119> ffmpeg: 03Clément BSsch 07master * r0d96ec19eb 10ffmpeg/libavformat/ (isom.h mov.c): mov: parse tref atom.
[07:50] <CIA-119> ffmpeg: 03Clément BSsch 07master * r20e46aafbd 10ffmpeg/libavformat/mov.c: mov: export orphan tmcd track metadata to global format metadata.
[08:17] <CIA-119> ffmpeg: 03Clément BSsch 07master * r1ec23d9c02 10ffmpeg/libavformat/mov.c: (log message trimmed)
[08:17] <CIA-119> ffmpeg: mov: copy timecode metadata from tmcd track to the related video stream.
[08:17] <CIA-119> ffmpeg: Apple softwares seem not to add a tref for the timecode (the next commit
[08:17] <CIA-119> ffmpeg: fixes this issue), but at least FFmpeg does.
[08:17] <CIA-119> ffmpeg: This can be used to generate a sample that demonstrates the feature:
[08:17] <CIA-119> ffmpeg:  ./ffmpeg -f lavfi -i testsrc \
[08:18] <CIA-119> ffmpeg:  -f lavfi -i mptestsrc \
[10:58] <CIA-119> ffmpeg: 03Carl Eugen Hoyos 07master * r37f1a7d51d 10ffmpeg/libavcodec/pngenc.c: 
[10:58] <CIA-119> ffmpeg: Make png-in-mov QuickTime-compatible for more colour-spaces.
[10:58] <CIA-119> ffmpeg: Reported and tested by ami_stuff.
[10:58] <CIA-119> ffmpeg: See ticket #1191.
[12:20] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r66531c75d3 10ffmpeg/libavutil/random_seed.c: 
[12:20] <CIA-119> ffmpeg: random_seed: rewrite generic code.
[12:20] <CIA-119> ffmpeg: The new code is faster and reuses the previous state in case of
[12:20] <CIA-119> ffmpeg: multiple calls. For testing the iterations are reduced and several
[12:20] <CIA-119> ffmpeg: entropy sources are disabled.
[12:20] <CIA-119> ffmpeg: Its based on SHA-1
[12:20] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[14:06] <shroomM> I'm having a problem compiling ffmpeg for windows via mingw under ubuntu 12.04
[14:06] <shroomM> here's the gist of the error: https://gist.github.com/48276ab942789d851006
[14:07] <shroomM> this worked on thursday
[14:11] Action: itrufeng how to use ffmpeg convert wma to mp3.HELP~
[14:13] Action: itrufeng Help~
[14:16] <JEEB> shroomM, a quick way to make it build is to remove the attribute_align_arg from line 67 in w32pthreads.h, but I'm not sure if that'll make it work 100% correctly. It seems to be missing the declaration for that attribute.
[14:16] <merbanan> +#include "libavutil/internal.h"
[14:18] <Compn> itrufeng : ffmpeg -i file.wma out.mp3
[14:18] <Compn> shroomM : missing pthreads
[14:19] <Compn> did you enable pthreads ?
[14:19] <shroomM> hmm
[14:19] <itrufeng> Compn: you are right,but I want to use c++ to do.
[14:19] <Compn> itrufeng : ask on libav-user mailing list for c++ application help
[14:20] <Compn> not many people in here use c++ , and this channel is for ffmpeg program development, not 3rd party programs
[14:20] <shroomM> Compn enable them where ?
[14:20] <Compn> shroomM : in your ./configure line
[14:20] <JEEB> Compn, that happens when win32threads are used and when avconv is built
[14:20] <JEEB> argh
[14:20] <JEEB> ffmpeg I mean
[14:20] <itrufeng> Compn: thanks
[14:21] <JEEB> probably doesn't happen with pthreads though
[14:21] <shroomM> I didn't specify anything
[14:21] <shroomM> lemme try with pthreads
[14:21] <JEEB> and w32threads should be the default for win32/64
[14:21] <Compn> yes
[14:21] <Compn> wonder what mingw enviro shroomM is using too
[14:22] <Compn> most people switched to mingw64 or whatever updated fork of mingw, as it was easier to compile...
[14:22] <shroomM> umm
[14:22] <JEEB> I think my 32bit setup is still mingw32
[14:22] <shroomM> based on a blog
[14:22] <shroomM> sec
[14:22] <JEEB> and that breaks, too
[14:22] <shroomM> http://ffmpeg.zeranoe.com/blog/
[14:22] <shroomM> this one
[14:22] <JEEB> I would guess that mingw64 breaks as well
[14:23] <shroomM> I'm making with --disable-w32threads now
[14:23] <itrufeng> Error
[14:23] <itrufeng> ffmpeg-devel roster authentication failed.
[14:23] <itrufeng> why
[14:24] <Compn> itrufeng : your email client sent the wrong auth code
[14:24] <Compn> copy and paste it instead 
[14:30] <shroomM> not using w32threads works
[14:31] <shroomM> I'm guessing the performance should be on similiar with pthreads and w32threads ?
[14:38] <Compn> yes
[14:38] <Compn> i think theres some bug about pthreads only supporting 16 threads, while w32threads can do any..
[14:38] <Compn> but shouldnt matter for most people
[14:39] <shroomM> ah, ok, thanks
[14:40] <shroomM> do I need to specify -threads when decoding the input to enable multithreaded decoding?
[14:40] <shroomM> or does ffmpeg figure out
[14:40] <shroomM> the number of threads to use
[14:40] <shroomM> but I'm getting off topic for the devel chan, sorry
[17:32] <michaelni> av500, about that beaglebord LAN-USB thing ...
[17:32] <michaelni> which distro/kernel would you recommand ?
[17:39] <av500> michaelni: angstrom
[17:41] <av500> the only drawback is that recent angstrom is all systemd infested
[17:42] <av500> but if you have a hdcp server in your setup, booting until you can ssh into the machine should be ok
[17:42] <av500> and I guess you dont need much more
[17:42] <av500> michaelni: which BB do you have?
[17:42] <av500> revC or XM?
[17:43] <av500> michaelni: for a start, please use of these images: http://circuitco.com/support/index.php?title=BeagleBoard
[17:43] <michaelni> not XM
[17:43] <av500> ok, so you need some usb2eth dongle
[17:44] <Daemon404> michaelni, "Note that this returns the size of a compact representation as generated by avpicture_layout(), which can be smaller than the size required for e.g. avpicture_fill()." <-- from doc for avpicture_get_size
[17:44] <Daemon404> yet in av_picture_get_size code:
[17:44] <Daemon404> return avpicture_fill(&dummy_pict, NULL, pix_fmt, width, height);
[17:44] <Daemon404> ^ it's returning this
[17:44] <Daemon404> and causing avpicture_layout to fail if the dest buffer size is not larger than the padded size
[17:44] <michaelni> av500, thilo send me 2 usb-lan adapters but neither works with the ubuntu thats installed
[17:44] <Daemon404> which seems Wrong^TM
[17:51] <Daemon404> michaelni, asking you cause the proper fix has potential api breakage
[17:57] <michaelni> Daemon404, has any of these functions changed in behavior ?
[17:57] <Daemon404> michaelni, im not sure
[17:57] <Daemon404> but teh docs dont match teh behavior
[17:57] <Daemon404> and teh behavior is broken
[17:57] <Daemon404> avpicture_layout's purpose is to copy to a contiguous buffer
[17:57] <Daemon404> but it fails if the buffer size is less than teh padded size
[17:57] <Daemon404> and it shouldnt care about the padded size.
[18:00] <Daemon404> i dont really want to switch to manually memcpy-ing each line...
[18:08] <michaelni> Daemon404, there are several kinds of padding, for example 1bpp is padded to byte boundary
[18:08] <michaelni> also the palette position is aligned to 4
[18:08] <Daemon404> michaelni, my problem is for yuv420p my buffer size is obviously w*h*3/2
[18:09] <Daemon404> and it's deciding it needs more than that (it doesnt) and failing
[18:09] <Daemon404> unelss im missing something
[18:09] <michaelni> well, when either w or h is odd w*h*3/2 is not correct anymore
[18:09] <Daemon404> h is odd
[18:10] <michaelni> think of 1 pixel it needs 3 bytes YUV420 noz 3/2
[18:10] <michaelni> noT
[18:10] <Daemon404> oh indeed
[18:10] <Daemon404> /facedesk
[18:10] <Daemon404> actually
[18:10] <Daemon404> why the heck didnt i just call get_size in my own code
[18:11] <Daemon404> i think i need a hot coffee injection
[18:11] <michaelni> av500, the board i got from thilo seem rev B4 the page with the images talks about C*
[18:13] <michaelni> so i dunno which to download ...
[18:31] <av500> michaelni: B4 is ancient
[18:31] <av500> and they had usb issues AFAIK
[18:38] <michaelni> av500, so which board that is in stock somewhere in europe should i get ?
[18:39] <av500> sec, I am checking for the usb issue
[18:45] <av500> michaelni: 
[18:45] <av500> michaelni: your B4 does not even have a full size usb port, right?
[18:46] <av500> so you would need ethernet via the usb OTG connector which I would not recommend
[18:47] <av500> http://www.watterott.com/de/BeagleBoard-xM
[18:47] <michaelni> av500, yes ...
[18:47] <av500> oop, 0 in stock
[18:47] <av500> http://www.watterott.com/de/BeagleBoard should work a a simple fate client too
[18:48] <av500> the usb port is high speed only but that should be OK with a modern usb2eth
[18:50] <michaelni> av500, what about http://www.watterott.com/de/PandaBoard ? it seems in stock and more powerfull but i have no clue about these arm boards ...
[19:10] <av500> michaelni: sure
[19:10] <av500> if its in stock, get this one
[19:10] <av500> it's a dual core so you can write mt codecs optimizations for ARM :)
[19:17] <j-b> av500: http://www.videolan.org/videolan/events/vdd12/
[20:22] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * rd9a9518fba 10ffmpeg/libavformat/flvenc.c: 
[20:22] <CIA-119> ffmpeg: flvenc: use av_assert instead of assert
[20:22] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[20:22] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r709b54eae9 10ffmpeg/libavcodec/rv20enc.c: 
[20:22] <CIA-119> ffmpeg: rv20enc: use av_assert
[20:22] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[21:00] <CIA-119> ffmpeg: 03Paul B Mahol 07master * r751dcd9243 10ffmpeg/libavcodec/Makefile: 
[21:00] <CIA-119> ffmpeg: Remove eightsvx_raw decoder line
[21:00] <CIA-119> ffmpeg: Decoder have been renamed to pcm_s8_planar long ago.
[21:00] <CIA-119> ffmpeg: Signed-off-by: Paul B Mahol <onemda at gmail.com>
[21:13] <CIA-119> ffmpeg: 03Nedeljko Babic 07master * r3827a86eac 10ffmpeg/ (24 files in 5 dirs): (log message trimmed)
[21:13] <CIA-119> ffmpeg: Optimization of AMR NB and WB decoders for MIPS
[21:13] <CIA-119> ffmpeg: AMR NB and WB decoders are optimized for MIPS architecture.
[21:13] <CIA-119> ffmpeg: Appropriate Makefiles are changed accordingly.
[21:13] <CIA-119> ffmpeg: Cnfigure script is changed in order to support optimizations.
[21:13] <CIA-119> ffmpeg:  Optimizations are enabled by default when compiling is done for
[21:13] <CIA-119> ffmpeg:  mips architecture.
[21:55] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r7925f923a5 10ffmpeg/libavcodec/4xm.c: 
[21:55] <CIA-119> ffmpeg: 4xm: use av_assert()
[21:55] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[21:55] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r65efc66d8e 10ffmpeg/libavcodec/pngenc.c: 
[21:55] <CIA-119> ffmpeg: pngenc: use av_assert
[21:55] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[21:59] <CIA-119> ffmpeg: 03Samuel Pitoiset 07master * rb4c92e94bb 10ffmpeg/doc/protocols.texi: 
[21:59] <CIA-119> ffmpeg: doc: Add documentation for the newly added rtmp_* options
[21:59] <CIA-119> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[21:59] <CIA-119> ffmpeg: 03Samuel Pitoiset 07master * r8ee3e1874e 10ffmpeg/ (doc/protocols.texi libavformat/rtmpproto.c): 
[21:59] <CIA-119> ffmpeg: rtmp: support connection parameters
[21:59] <CIA-119> ffmpeg: Allow using connection parameters in order to append arbitrary
[21:59] <CIA-119> ffmpeg: AMF data like "B:1 S:authMe O:1 NN:code:1.23 NS:flag:ok O:0" to the
[21:59] <CIA-119> ffmpeg: Connect message. You can pass these parameters through the -rtmp_conn
[21:59] <CIA-119> ffmpeg: option.
[21:59] <CIA-119> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[21:59] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r94d5650266 10ffmpeg/: 
[21:59] <CIA-119> ffmpeg: Merge remote-tracking branch 'qatar/master'
[21:59] <CIA-119> ffmpeg: * qatar/master:
[21:59] <CIA-119> ffmpeg:  rtmp: Do not send extension for flv files
[21:59] <CIA-119> ffmpeg: rtmp: Do not send extension for flv files
[21:59] <CIA-119> ffmpeg: This fixes bugzilla bug #304.
[21:59] <CIA-119> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[23:31] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * rf91b635b52 10ffmpeg/libavformat/rtmpproto.c: 
[23:31] <CIA-119> ffmpeg: rtmpproto: replace strtok_r by av_strtok its unavailable in mingw
[23:31] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[23:31] <CIA-119> ffmpeg: 03Michael Niedermayer 07master * r9ec5e956a2 10ffmpeg/ffmpeg.c: 
[23:31] <CIA-119> ffmpeg: ffmpeg: disable threading on mingw, it doesnt work due to dependance on internal code.
[23:31] <CIA-119> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[23:33] <Daemon404> wtf?
[23:33] <Daemon404> disable threads on win32?
[23:41] <michaelni> Daemon404, no
[23:42] <michaelni> only the multithreaded demuxing in ffmpeg.c
[23:42] <Daemon404> ah.
[23:42] <michaelni> it never worked on win32
[23:42] <michaelni> see our fate
[23:42] <michaelni> mingw is red :(
[23:42] <Daemon404> this reminds me
[23:42] <Daemon404> i set up a native windows mingw fate instance
[23:42] <Daemon404> ill register it with fate tonight
[23:43] <michaelni> great, thanks!
[00:00] --- Tue Jun 12 2012


More information about the Ffmpeg-devel-irc mailing list