[Ffmpeg-devel-irc] ffmpeg-devel.log.20120831

[burek]

[00:08] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * r469f8b1dcb 10ffmpeg/ffmpeg_opt.c: 
[00:08] <CIA-56> ffmpeg: ffmpeg: Fix assert failure
[00:08] <CIA-56> ffmpeg: Fixes Ticket1674
[00:08] <CIA-56> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[00:19] <brimestone> thanks saste, now i just need to know the standard color values of Rec709
[00:45] <CIA-56> ffmpeg: 03Stephan Hilb 07master * r6eac554659 10ffmpeg/ (9 files in 4 dirs): 
[00:45] <CIA-56> ffmpeg: Add CPiA video decoder
[00:45] <CIA-56> ffmpeg: The cpia video decoder is intended to be used with the v4l2 demuxer.
[00:45] <CIA-56> ffmpeg: There are some small changes to the v4l2 demuxer to support the
[00:45] <CIA-56> ffmpeg: variable frame length of the format.
[00:45] <CIA-56> ffmpeg: Fixes ticket #1537
[00:45] <CIA-56> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[00:47] <brimestone> ok, i got to the point where lutyuv needs a little help.. 
[00:48] <brimestone> i got something interesting result with -vf lutrgb='y=gammaval(2):u=gammaval(2):v=gammaval(2)'
[00:49] <brimestone> so aside from "gammaval" what else can i use as a param? lummaval?? and how can i string them up? :v=gammaval(2)lummaval(2)'
[04:21] <ubitux> anyone familiar with ebu r128?
[04:22] <ubitux> afaiu, gating blocks are 400ms of samples, with 75% overlap between them
[04:22] <ubitux> this is used to compute the "integrated" loudness
[04:22] <ubitux> but the momentary & short term loudness seem to be non-gated
[04:22] <ubitux> first one is 0.4s and second one 3s
[04:23] <ubitux> but the refresh for the live meter in the case of short term is at least 10Hz
[04:24] <ubitux> how come you can give results for the last 3s 10x per second if not "overlapping" by the blocks?
[04:39] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * re7b4337777 10ffmpeg/libavdevice/v4l2.c: 
[04:39] <CIA-56> ffmpeg: v4l2: fix compile on systems that lack V4L2_PIX_FMT_CPIA1
[04:39] <CIA-56> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[07:09] <brimestone> has anyone here dealt with Arri Alexa LUT footage?
[09:12] <cypher497> fmpeg
[11:15] <CIA-56> ffmpeg: 03Carl Eugen Hoyos 07master * ra5a90fa577 10ffmpeg/libavcodec/tiffenc.c: Simplify setting colour space for mono* when encoding tiff.
[11:15] <CIA-56> ffmpeg: 03Carl Eugen Hoyos 07master * r5eb532a930 10ffmpeg/libavcodec/tiffenc.c: Cosmetics: Reorder pix_fmts in tiffenc.c.
[11:15] <CIA-56> ffmpeg: 03Carl Eugen Hoyos 07master * rf3683349ae 10ffmpeg/libavcodec/ (tiff.h tiffenc.c): Write TIFF Tag ExtraSamples when encoding pix_fmts that contain transparency information.
[11:40] <panicking_mike> Hi all
[11:41] <panicking_mike> I'm working on audio streaming and evaluating 
[11:41] <panicking_mike> opensource software
[11:41] <panicking_mike> I'm trying to stream a 24bit wav file
[11:41] <panicking_mike> using ffserver
[11:41] <panicking_mike> and receive it using ffplay
[11:41] <panicking_mike> but I stuck in some errors
[11:41] <panicking_mike> I have already written to the mailing list
[11:41] <panicking_mike> but maybe it was a silly question
[11:42] <panicking_mike> Does anyone has the clue here?
[13:08] <CIA-56> ffmpeg: 03Anton Khirnov 07master * re2785fa728 10ffmpeg/cmdutils.c: cmdutils: make -codecs print lossy/lossless flags.
[13:08] <CIA-56> ffmpeg: 03Anton Khirnov 07master * ra2318326f1 10ffmpeg/ (4 files in 2 dirs): lavc: add lossy/lossless codec properties.
[13:08] <CIA-56> ffmpeg: 03Anton Khirnov 07master * rbbcedade00 10ffmpeg/ (Changelog avconv.c avconv.h avconv_opt.c doc/avconv.texi): avconv: make the -passlogfile option per-stream.
[13:08] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * rbbefd27e52 10ffmpeg/libavcodec/ (utvideo.h utvideoenc.c): 
[13:08] <CIA-56> ffmpeg: utvideoenc: Avoid writing into the input picture
[13:08] <CIA-56> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[13:08] <CIA-56> ffmpeg: 03Kostya Shishkov 07master * r04fc5c6bde 10ffmpeg/ (5 files in 3 dirs): g723_1: add comfort noise generation
[13:08] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * r040405b59e 10ffmpeg/libavcodec/ (utvideo.h utvideoenc.c): 
[13:08] <CIA-56> ffmpeg: utvideoenc: Switch to dsputils' median prediction
[13:08] <CIA-56> ffmpeg: Also, align the mangled RGB planes, which is required for the
[13:08] <CIA-56> ffmpeg: SIMD versions of dsputils' median predict.
[13:08] <CIA-56> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[13:08] <CIA-56> ffmpeg: 03Diego Biurrun 07master * rcdaec0b240 10ffmpeg/libavfilter/x86/ (yadif.c yadif_template.c): avfilter: x86: Use more precise compile template names
[13:08] <CIA-56> ffmpeg: 03Anton Khirnov 07master * r11d957fbd8 10ffmpeg/ (6 files): 
[13:08] <CIA-56> ffmpeg: avtools: remove the distinction between func_arg and func2_arg.
[13:08] <CIA-56> ffmpeg: func2_arg is the same as func_arg, except it has one additional
[13:08] <CIA-56> ffmpeg: parameter. Change all func_arg callbacks to take that parameter (and
[13:08] <CIA-56> ffmpeg: ignore it).
[13:08] <CIA-56> ffmpeg: 03Diego Biurrun 07master * r50cd43f2cd 10ffmpeg/configure: configure: Add more fine-grained SSE CPU capabilities flags
[13:08] <CIA-56> ffmpeg: 03Diego Biurrun 07master * rec36aa6944 10ffmpeg/libavcodec/x86/ (mpegaudiodec.c mpegvideoenc.c rv40dsp_init.c): 
[13:08] <CIA-56> ffmpeg: x86: Fix linking with some or all of yasm, mmx, optimizations disabled
[13:08] <CIA-56> ffmpeg: Some optimized template functions reference optimized symbols, so they
[13:39] <CIA-56> ffmpeg: 03Martin Storsjö 07master * r372de27df7 10ffmpeg/tools/pktdumper.c: 
[13:39] <CIA-56> ffmpeg: pktdumper: Use sizeof(variable) instead of the direct buffer length
[13:39] <CIA-56> ffmpeg: Also change the snprintf size to use the full buffer, since
[13:39] <CIA-56> ffmpeg: snprintf always null-terminates the buffer.
[13:39] <CIA-56> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[13:39] <CIA-56> ffmpeg: 03Mans Rullgard 07master * rd96d6ba618 10ffmpeg/configure: 
[13:39] <CIA-56> ffmpeg: configure: handle --disable-asm before check_deps
[13:39] <CIA-56> ffmpeg: This is necessary to avoid spuriously enabling _external or _inline
[13:39] <CIA-56> ffmpeg: variants of arch extensions when they should be disabled.
[13:39] <CIA-56> ffmpeg: Signed-off-by: Mans Rullgard <mans at mansr.com>
[13:39] <CIA-56> ffmpeg: 03Alberto Delmás 07master * rede3d6400d 10ffmpeg/libavcodec/ (mss1.c mss12.c mss2.c): 
[13:39] <CIA-56> ffmpeg: MSS1 and MSS2: set final pixel format after common stuff has been initialised
[13:39] <CIA-56> ffmpeg: This way it won't interfere with WMV9 initialisation inside MSS2 decoder and
[13:39] <CIA-56> ffmpeg: avplay will play it fine.
[13:39] <CIA-56> ffmpeg: Signed-off-by: Kostya Shishkov <kostya.shishkov at gmail.com>
[13:39] <CIA-56> ffmpeg: 03Martin Storsjö 07master * r21411a4102 10ffmpeg/tools/pktdumper.c: 
[13:39] <CIA-56> ffmpeg: pktdumper: Use a custom define instead of PATH_MAX for buffers
[13:39] <CIA-56> ffmpeg: PATH_MAX is not necessarily available on all systems, e.g. it's
[13:39] <CIA-56> ffmpeg: normally not available on MSVC, and is not guaranteed to defined
[13:39] <CIA-56> ffmpeg: on a POSIX system either.
[13:39] <CIA-56> ffmpeg: Signed-off-by: Martin Storsjö <martin at martin.st>
[13:39] <CIA-56> ffmpeg: 03Martin Storsjö 07master * rbcc44873d9 10ffmpeg/tools/pktdumper.c: 
[13:39] <CIA-56> ffmpeg: pktdumper: Use av_strlcpy instead of strncpy
[13:39] <CIA-56> ffmpeg: This takes care of null-terminating the buffer if it is too small,
[13:39] <CIA-56> ffmpeg: which wasn't handled properly before.
[13:39] <CIA-56> ffmpeg: 03Alberto Delmás 07master * ree769c6a7c 10ffmpeg/ (20 files in 4 dirs): 
[13:39] <CIA-56> ffmpeg: MSS2 decoder
[13:39] <CIA-56> ffmpeg: Signed-off-by: Kostya Shishkov <kostya.shishkov at gmail.com>
[13:39] <CIA-56> ffmpeg: 03Michael Niedermayer 07master * rc617bed34f 10ffmpeg/: (log message trimmed)
[13:39] <CIA-56> ffmpeg: Merge remote-tracking branch 'qatar/master'
[13:39] <CIA-56> ffmpeg: * qatar/master:
[13:39] <CIA-56> ffmpeg:  MSS1 and MSS2: set final pixel format after common stuff has been initialised
[13:39] <CIA-56> ffmpeg:  MSS2 decoder
[13:39] <CIA-56> ffmpeg:  configure: handle --disable-asm before check_deps
[13:39] <CIA-56> ffmpeg:  x86: Split inline and external assembly #ifdefs
[13:39] <CIA-56> (1 lines omitted)
[14:31] <Compn> mss2 decoder
[14:31] <Compn> is that part of vlc bounty ?
[14:31] Action: Compn looks at j-b
[14:33] <Compn> commit mixes cosmetics and code changes
[14:35] <Compn> not too many tho
[14:41] <Compn> new mss2 files dont have copyright info
[15:02] <ubitux> is Wolfram Gloger on IRC?
[15:22] <j-b> Compn: yes, it is.
[15:31] <Compn> cool
[15:40] <CIA-53> ffmpeg: 03Carl Eugen Hoyos 07master * r52be5428c0 10ffmpeg/ (10 files in 3 dirs): Add some missing _EXTERNAL suffixes to yasm source files.
[15:50] <CIA-53> ffmpeg: 03Michael Niedermayer 07master * rd7587ba1c6 10ffmpeg/libavformat/flvdec.c: 
[15:50] <CIA-53> ffmpeg: flvdec: remove duplicate return
[15:50] <CIA-53> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[15:50] <CIA-53> ffmpeg: 03Michael Niedermayer 07master * r0b7d12e840 10ffmpeg/libavfilter/vf_mp.c: 
[15:50] <CIA-53> ffmpeg: vf_mp: fix duplicate return
[15:50] <CIA-53> ffmpeg: Signed-off-by: Michael Niedermayer <michaelni at gmx.at>
[16:05] <CIA-53> ffmpeg: 03Stefano Sabatini 07master * re64b941dbc 10ffmpeg/ (5 files in 3 dirs): 
[16:05] <CIA-53> ffmpeg: lavfi/mp: remove smartblur filter
[16:05] <CIA-53> ffmpeg: It was natively integrated into libavfilter.
[16:05] <CIA-53> ffmpeg: 03Stefano Sabatini 07master * r6c0107822d 10ffmpeg/libavfilter/libmpcodecs/vf_decimate.c: 
[16:05] <CIA-53> ffmpeg: lavfi/mp=decimate: fix off-by-one logic in diff_C() x loop
[16:05] <CIA-53> ffmpeg: Set x offset values in the range 0-7, rather than in the range 8-1.
[16:05] <CIA-53> ffmpeg: The y loop is changed accordingly, to avoid confusion.
[16:05] <CIA-53> ffmpeg: This also fixes output difference with the new pending native decimate
[16:05] <CIA-53> ffmpeg: filter.
[16:05] <CIA-53> ffmpeg: 03Stefano Sabatini 07master * r9dd3d6c72a 10ffmpeg/ (7 files in 3 dirs): 
[16:05] <CIA-53> ffmpeg: lavfi: port decimate libmpcodecs filter
[16:05] <CIA-53> ffmpeg: This filter is based on the MPlayer decimate filter by Rich Felker.
[16:10] <saste> michaelni: what's the difference between smartblur and sab?
[16:15] <saste> oh well, they are different things, never mind
[16:29] <Compn> smart adaptive blur :P
[16:29] Action: Compn has never used such filters
[16:29] <Compn> i stick to hqdn3d and uspp/fspp and yadif :P
[16:30] <Compn> decimate is good too
[17:31] <j-b> 17:29 < etix> VDD attendees: you are expected to be in front of the hotel at 6pm :)
[17:50] <CIA-53> ffmpeg: 03Stefano Sabatini 07master * rfa838a0c0f 10ffmpeg/configure: configure: add decimate filter dependency on avcodec
[18:23] CTCP re:  from CIA-53 (CIA-53!cia at cia.vc) to #ffmpeg-devel
[18:23] <CIA-53> ffmpeg: 03Jemy Tran 07master * r54ca7e39da 10ffmpeg/ (doc/filters.texi libavfilter/version.h libavfilter/vf_hue.c): 
[18:23] <CIA-53> ffmpeg: lavfi/hue: add process_command callback
[18:23] <CIA-53> ffmpeg: This allows dynamic reconfiguration of the filter.
[18:23] <CIA-53> ffmpeg: The callback uses some code that was in the init function. Hence this code
[18:23] <CIA-53> ffmpeg: has been moved in its own function.
[18:23] <CIA-53> ffmpeg: Signed-off-by: Stefano Sabatini <stefasab at gmail.com>
[18:43] <brimestone> hey guys.. support i have a LUT file and i want to use that as an argument for -vf lutyuv or lutrgb? how can i achieve that?
[18:46] <saste> brimestone: start by reading the docs and figuring out how to use lut
[18:46] <saste> Compn: what do you use decimate for?
[18:47] <Compn> i recorded some video , but i wanted to make it a timelapse
[18:48] <Compn> so decimate took off a lot of frames, and then i could drop more frames later. ensuring that i dropped duplicate frames instead of good frames later on
[18:49] <saste> what i dislike of decimate are that the prameters are far from be intuitive
[18:49] <saste> you need to tweak a lot to get the wanted result
[18:50] <saste> *what i dislike of decimate is that its parameters are far from being intuitive
[18:52] <Compn> hmm
[18:52] <Compn> i dont remember how i used it
[18:52] <Compn> wonder if i still have the script
[18:58] <brimestone> saste: i did.. but documentation is a little bit gray on how to apply look up table
[18:58] <saste> brimestone: it applies a *expression"
[18:59] <saste> I suppose you have a map of input->output values
[18:59] <saste> if it can't be reduced to an expression you are out of luck
[18:59] <saste> and patches are welcome
[18:59] <saste> (since this is the dev channel)
[19:01] <brimestone> guess i need to learn expression
[19:04] <Compn> saste : he needs an example command line
[19:04] <Compn> cant figure out what an 'expression' is
[19:04] <Compn> and i'm a bit confused myself or i'd help
[19:05] <saste> brimestone: there are plenty examples in the lut manual
[19:05] <Compn> brimestone : you have to bribe us with money and chocolate :P
[19:05] <saste> also check the "expression evaluation" in the manual
[19:05] <saste> expression = mathematical expression
[19:06] <brimestone> where to i slide my credit card? hahaha
[19:10] <Compn> into paypal
[19:10] <Compn> or flattr
[19:10] <Compn> or amazon payments
[19:11] <brimestone> sounds resonable.. 
[19:11] <Compn> if only
[19:11] <brimestone> how much we talking?
[19:11] <brimestone> any takers?
[19:11] <brimestone> here is my need
[19:11] <Compn> you might want to try ffmpeg-user mailing list
[19:11] <Compn> as this isnt a development question really
[19:12] <brimestone> i look into that mailing list including the ffmbc and nothing
[19:14] <Compn> i mean
[19:14] <Compn> post there , and ask there for help
[19:15] <brimestone> ohhh
[19:15] <Compn> :)
[19:15] <Compn> thats where people go to help other people with ffmpeg
[19:15] <brimestone> who knows better than the developer them self :)
[19:15] <Compn> there is also an ffmpeg forum if you like forums/boards better than mailing lists
[19:15] <Compn> yes, but developers dont like answering questions 
[19:16] <brimestone> ouch.. even for money?
[19:16] <Compn> answering users questions all day would take time out from developing :)
[19:16] <Compn> maybe for money, i cant speak for everyone
[19:17] <brimestone> understood. but not a single sample outthere about anything lut related
[19:18] <Compn> so uh, are you a video professional ?
[19:18] <Compn> or work in the broadcast, film, photography field ?
[19:18] <brimestone> yes
[19:19] <Compn> neat
[19:19] <brimestone> i have a current issue where we need to "automate" application of lut profile to camera files
[19:19] <brimestone> we can do it via DaVinci Resolve now, but we need something automated
[19:20] <Compn> if i had any idea of what that all was i'd  be able to google help :P
[19:20] <brimestone> tried that.. all i get involves "manual" workflow
[19:21] <Compn> you tried outputting it to images, using imagemagick to convert , and then whatever program back to camera file ?
[19:21] <brimestone> yes.. 
[19:22] <Compn> not working huh?
[19:22] <brimestone> the main issue is the "dynamic" nature of a color profile from the field..  different Directory of Photography will have their "beloved" Lut file.. 
[19:22] <saste> brimestone: i have no idea what is a DaVinci resolve, but I suppose it is a file containing the in->out map, am I right?
[19:22] <Compn> ah
[19:23] <brimestone> so somekind of lookup table to map yuv rgb values base on a file (icc profile) type stuff
[19:24] <brimestone> most of the time, they just use the Standard unmodified Rec709 LUT
[19:28] <saste> brimestone: can you show how it looks such a file?
[19:29] <saste> reading from a file the map should be feasible, but you may need to export the lut file to some format which could be read by ffmpeg
[19:29] <saste> also keep in mind that currently lut works only with 8-bits per component values
[19:29] <brimestone> sure
[19:30] <brimestone> standby creting sreeshots
[19:34] <brimestone> http://tedsimbajon.com/color.php
[19:34] <brimestone> 1st image is a ProRes 444 12Bit with massive Dynamic Range on contrast.. 
[19:34] <brimestone> the second image is the result of the LUT
[19:36] <saste> brimestone: I mean the file used to apply the LUT
[19:37] <brimestone> oh yes.. wait
[19:37] <saste> i mean: is it in a readable format? in that case may be easy to export it
[19:37] <saste> (human readable)
[19:37] <brimestone> yes.. XML
[19:38] <brimestone> http://tedsimbajon.com/profile_test.xml
[19:41] <saste> oh i see...
[19:41] <saste> brimestone; which is the name of the software?
[19:41] <brimestone> which software? the one who made the XML?
[19:41] <saste> yep
[19:43] <brimestone> i believe they used this http://pomfort.com/livegrade/
[19:44] <saste> brimestone: basically that file specify to apply a sequence of filters
[19:45] <brimestone> i think its the same filter throughout the shoot day.. 
[19:45] <saste> parameters are specified in the file, and the filters are implementation dependent
[19:45] <saste> that is you can't use that profile outside of that application
[19:45] <saste> it is not as simple as a simple LUT map, but a sequence of filters
[19:46] <saste> note that you may achieve the same effects using ffmpeg filters (you can concatenate them)
[19:46] <saste> what you can do is automatically convert such a profile to a sequence of ffmpeg filter
[19:46] <brimestone> yes. :) thats what im trying to achieve
[19:46] <saste> much less to a lut table
[19:47] <saste> fuck i meant
[19:47] <saste> *what you *can't* do is to automatically convert such a profile to a sequence of ffmpeg filters
[19:48] <brimestone> can't
[19:55] <saste> brimestone: what you can do is trying to achieve the same effect with ffmpeg filters
[19:56] <saste> in case a specific filter is not implemented you may file a feature request or fund a developer to add it
[19:56] <saste> but the tricky part is to map a sequence of filters in a software to an equivalent sequence of filters in another
[19:57] <brimestone> maybe i could test out the filters
[20:23] <brimestone> saste.. thank you very much.. it looks like i got something working with "-vf mp=eq2=1.2:1.5:0.0:3.2"
[23:17] <ubitux> michaelni: should we add a libav fate instance built against ffmpeg libraries?
[23:23] <saste> INCOMPATIBLE_FORK_ABI
[23:23] <saste> uhm.. it's a bit confusing
[23:24] <saste> what about FORK_ABI_COMPATIBILITY?
[23:26] <ubitux> FORK_COMPAT_IS_A_PAIN
[23:27] <ubitux> --enable-pain
[23:42] <michaelni> ubitux, the idea with such a fate instance sounds interresting
[00:00] --- Sat Sep  1 2012