[Ffmpeg-devel-irc] ffmpeg-devel.log.20141204

[burek]

[00:44] <cone-642> ffmpeg.git 03Benoit Fouet 07master:08aec6f68e6a: libavcodec/pngdec: support 'previous' dispose operation for APNG.
[00:44] <cone-642> ffmpeg.git 03Benoit Fouet 07master:6e9b060e4f0c: avcodec/pngdec: fix mem leak in init()
[00:44] <cone-642> ffmpeg.git 03Benoit Fouet 07master:1d69ab9d1c3d: avcodec/pngdec: fix indentation in handle_row()
[01:17] <cone-642> ffmpeg.git 03Benoit Fouet 07fatal: ambiguous argument 'refs/heads/release/2.5': unknown revision or path not in the working tree.
[01:17] <cone-642> Use '--' to separate paths from revisions
[01:17] <cone-642> refs/heads/release/2.5:HEAD: avcodec/pngdec: fix indentation in handle_row()
[01:38] <cone-642> ffmpeg.git 03Michael Niedermayer 07master:f478bdabf2af: Changelog: add 2.5 entry
[01:38] <cone-642> ffmpeg.git 03Michael Niedermayer 07master:0f16ab18db80: doc/APIchanges: add 2.5 cut marker
[01:58] <cone-642> ffmpeg.git 03Carl Eugen Hoyos 07master:963aa1daf880: Fix standalone compilation of the iec61883 input device.
[01:58] <cone-642> ffmpeg.git 03Michael Niedermayer 07master:8d87c0b3ad76: Merge remote-tracking branch 'cehoyos/master'
[01:58] <cone-642> ffmpeg.git 03Michael Niedermayer 07master:5fe026f2bda6: avcodec/aacdec: reduce noisiness of missing channel elements
[02:05] <cone-642> ffmpeg.git 03Carl Eugen Hoyos 07release/2.5:36e7385d0eb7: Fix standalone compilation of the iec61883 input device. (cherry picked from commit 963aa1daf88066d43c7495803b870800c5faaa4d)
[02:05] <cone-642> ffmpeg.git 03Michael Niedermayer 07release/2.5:a588e1e5609c: avcodec/aacdec: reduce noisiness of missing channel elements
[02:17] <cone-642> ffmpeg.git 03Michael Niedermayer 07release/2.5:d9bdf7d9ae6c: Changelog: release is 2.5 not "next"
[02:31] <cone-642> ffmpeg.git 03Michael Niedermayer 07fatal: ambiguous argument 'refs/tags/n2.5': unknown revision or path not in the working tree.
[02:31] <cone-642> Use '--' to separate paths from revisions
[02:31] <cone-642> refs/tags/n2.5:HEAD: avcodec/aacdec: reduce noisiness of missing channel elements
[05:39] <Zeranoe> I'm getting a link time error with the new DeckLink patches http://paste.ubuntu.com/9362028/
[05:41] <Zeranoe> Looks like 5a57f389f45e9d6663746d7f060cccee7db04f16 introduced SysFreeString
[05:43] <Zeranoe> Looks like liboleaut32 or libole32 includes SysFreeString. Perhaps configure should be updated to include the lib
[05:53] <Zeranoe> liboleaut32 seemed to fix it
[05:54] <michaelni> Zeranoe, can you send a patch (to ffmpeg-deve ML) ? 
[05:54] Action: michaelni falls asleep
[06:09] <cone-760> ffmpeg.git 03Timothy Gu 07release/2.5:bb8614cb7b5c: Changelog/RELEASE_NOTES: Add APNG decoder
[06:09] <cone-760> ffmpeg.git 03Timothy Gu 07release/2.5:192d46e6d1e5: Update RELEASE_NOTES
[06:36] <arwa> by working with uspp you mean porting it to ffmpeg, right?
[08:13] <j-b> 'morning
[12:09] <cone-887> ffmpeg.git 03Timothy Gu 07master:af14ae803739: RELEASE: Bump
[12:09] <cone-887> ffmpeg.git 03Rong Yan 07master:36cd2bcbc343: avcodec/ppc/vc1dsp_altivec: add POWER LE support to vc1_inv_trans_8x4_altivec()
[12:44] <ubitux> saste: aah i like that a lot :)
[12:45] <ubitux> will look later
[12:49] <anshul_mahe> In our documentation there is something wrong about decklink device
[12:50] <anshul_mahe> we say that decklink only give uyvy422 but its not true, 
[12:51] <anshul_mahe> I use decklink for 10bityuv
[12:51] <anshul_mahe> though using ffmpeg we can only get that uyvy4222
[12:58] <saste> ubitux, :-)
[13:01] <kierank> anshul_mahe: you don't use 10-bit yuv
[13:02] <kierank> you use v210
[13:08] <anshul_mahe> kierank: I am actualy using PIX_FMT_YUV422P10 so i hope its 10 bit yuv, i dont know what is v210 may be some other pixel format
[13:08] <kierank> then it won't work
[13:08] <anshul_mahe> to decklink device I specify bmdFormat10BitYUV
[13:08] <anshul_mahe> kierank: why 
[13:08] <kierank> because blackmagic wants v210
[13:08] <kierank> read the docs
[13:11] <anshul_mahe> I am using bmd device as input
[13:13] <anshul_mahe> According to spec only I specify pix format to bmdFormat10BitYUV. I dont know which document you are talking about but I am referering to sdk api specification
[13:14] <anshul_mahe> its also working with that pixel format, if you want I can share my code
[13:14] <kierank> 10bitYUV is v210 in the docs
[13:14] <kierank> believe me
[13:14] <kierank> it definitely won't work with yuv422p10
[13:16] <anshul_mahe> ok, I will get back to it after studing it again
[13:18] <anshul_mahe> I am specifying video_codec as AV_CODEC_ID_V210
[13:18] <anshul_mahe> and pixel format as PIX_FMT_YUV422P10
[13:19] <kierank> yes then that will work
[13:22] <anshul_mahe> so my point is still valid that "In docs:we say that decklink only give uyvy422 but its not true"
[13:23] <kierank> well it's true in ffmpeg
[14:47] <cone-887> ffmpeg.git 03Rong Yan 07master:d23e8832486f: avcodec/ppc/vp8dsp_altivec.c: POWER LE support put_vp8_epel_h_altivec_core() put_vp8_epel_v_altivec_core() put_vp8_pixels16_altivec() add marcos GET_PIXHL() GET_OUTER() LOAD_HL()
[14:47] <cone-887> ffmpeg.git 03Michael Niedermayer 07master:a4f21a8dd9d8: MAINTAINERS: add 2.5
[15:41] <kierank> Does anyone require a shell on an avx2 machine?
[15:44] <Compn> kierank : ask on ml
[15:44] <kierank> most of the connectivity is setup at work now
[15:44] <kierank> so it's possible
[16:00] <ubitux> kierank: i provide such a machine already but it would be nice if i can shut it down and use that machine at some point
[16:01] <ubitux> it's a always on laptop on my home connection so... :p
[16:01] <kierank> I have 3, I need to figure out which one to let people use
[16:06] <kierank> probably the lowest power one
[16:39] <funman> aaaS / avx2 as a Service ?
[17:09] <cone-887> ffmpeg.git 03Benoit Fouet 07master:0f6931f4b60c: avcodec/pngdec: use memcpy instead of byte loops for P frames.
[18:03] <Daemon404> am i correct in thinking the only way to stream copy from mpegts to mp4 using the API is to manually invoke the h264 parser?
[18:03] <Daemon404> maybe nevcairiel?
[18:03] <kierank> yes
[18:03] <Daemon404> fuck
[18:03] <Daemon404> am i deldued to think tehre is a good example of how to use that API
[18:07] <kierank> don't you mean h264 bsf
[18:08] <Daemon404> ?
[18:08] <Daemon404> no?
[18:08] <kierank> never mind then
[18:08] <kierank> i thought you wanted to convert between annexb and mp4
[18:08] <Daemon404> the problem i have is that extradata is 0x0
[18:08] <Daemon404> so i cant mux to mp4
[18:08] <Daemon404> because it needs the global header
[18:09] <kierank> oh
[18:09] <Daemon404> mayeb lavf will do it for me
[18:09] Action: Daemon404 checks
[18:19] <rcombs> works for me with ffmpeg.c, so something's doing something right
[18:24] <nevcairiel> lavf with the parser should extract an extradata and put it into the codec context
[18:24] <nevcairiel> but does the mp4 muxer accept annexb input? does it change it around by itself then?
[18:25] <cone-887> ffmpeg.git 03Christophe Gisquet 07master:291747581e3d: imc: make issue non-fatal
[18:25] <cone-887> ffmpeg.git 03Michael Niedermayer 07master:cbf09545f250: avcodec/imc: only check remaining bits in case bits will be read
[18:29] <Daemon404> nevcairiel, well... im getting... closer... i think
[18:30] <Daemon404> [mp4 @ 0x281f5c0] Tag avc1/0x31637661 incompatible with output codec id '28' ([33][0][0][0])
[18:30] <Daemon404> ^ wtfing at this right now
[18:36] Action: Daemon404 is stumped...
[18:42] <Daemon404> ok i had to reset the codec tag to 0
[18:42] <Daemon404> fun.
[18:47] <Daemon404> nevcairiel, [mp4 @ 0x28375c0] Codec for stream 0 does not use global headers but container format requires global headers
[18:47] <Daemon404> are you sure lavf does it?
[18:48] <nevcairiel> yes
[18:48] <Daemon404> what am i missing here then
[18:48] <nevcairiel> its probably checking something else though
[18:48] <nevcairiel> or it just doesnt like annexb format
[18:48] <Daemon404> extradata is not filled with teh global header.
[18:49] <Daemon404> full stop.
[18:49] <Daemon404> i dont know how to fill it, other than calling parser_parse2
[18:49] <nevcairiel> parsing doesnt fill it
[18:49] <Daemon404> then what the hell does
[18:49] <nevcairiel> lavf fills it using the split method of the parser
[18:49] <nevcairiel> during find_stream_info etc
[18:49] <Daemon404> yes im not using find_stream_info
[18:50] <Daemon404> since i know the layout and types of streams
[18:50] <nevcairiel> well there you go
[18:50] <Daemon404> have i mentioned lately that find_stream_info is satan
[18:52] <nevcairiel> be that as it may, you need a function like it to gather such information
[18:53] <Daemon404> heh
[18:53] <Daemon404> trying to find where it actually does stuff
[18:55] <Daemon404> ugh
[18:55] <Daemon404> parse_packet is fucking huge
[18:56] <Daemon404> and parse_queue and stuff
[18:56] <Daemon404> this api makes me wanna kill myself...
[18:58] <Daemon404> looks insanely hard to use...
[19:00] <wm4> it was probably made for ffmpeg.c
[19:00] <wm4> trollol
[19:00] <nevcairiel> all it does is read packets and call the parser split function until the split returns a positive return
[19:01] <Daemon404> ther is no "Split" function
[19:01] <nevcairiel> sure there is
[19:01] <Daemon404> do tell
[19:01] <Daemon404> i onyl see _change and _parse2
[19:02] <nevcairiel> there is no wrapper function, lavf just calls parser->split directly
[19:02] <Daemon404> o...
[19:02] <Daemon404> i cant find where it actually sets the extradata
[19:02] <Daemon404> at all
[19:03] <nevcairiel> i can't cure blindness yet, sorry :)
[19:03] <Daemon404> i grepped for extradata
[19:03] <Daemon404> i dont see it
[19:03] <nevcairiel> lavf utils.c, 3243 downwards
[19:03] <nevcairiel> not sure if recent version
[19:04] <Daemon404> oh there it is
[19:06] <Daemon404> nevcairiel, i assume i have to call split on every single packet i demux/remux?
[19:06] <nevcairiel> until its done
[19:06] <Daemon404> yes
[19:06] <Daemon404> im not sure which packets NOT to pass to the demuxer?
[19:07] <Daemon404> er
[19:07] <Daemon404> mxuer
[19:07] <Daemon404> e.g. the extradata/global header i obviously would not pass to the muxer during remux, i think... or...
[19:07] <Daemon404> ffs
[19:08] <cone-887> ffmpeg.git 03Michael Niedermayer 07master:bde27e1e617d: doc/examples/transcoding: check encoder before using it
[19:10] <Daemon404> oh derp
[19:10] <Daemon404> i an just set teh codec flag
[21:43] <cone-887> ffmpeg.git 03Steve Jiekak 07master:6e9ac02af8b6: add av_dv_codec_profile2 : uses framerate to select best matching profile. default on first matching profile
[21:43] <cone-887> ffmpeg.git 03Steve Jiekak 07master:a32c3c2392ec: replaced av_dv_codec_profile by av_dv_codec_profile2 in encoder and dv muxers
[23:39] <cone-887> ffmpeg.git 03Mark Reid 07master:4f69477870dd: libavformat/mxfdec.c refactor resolving MultiDescriptor and remove essence group hack
[00:00] --- Fri Dec  5 2014