[Ffmpeg-devel-irc] ffmpeg-devel.log.20130114

[burek]

[00:25] <cone-117> ffmpeg.git 03Michael Niedermayer 07master:2d8d63a9a07c: ffplay: fix sws_flags
[00:25] <cone-117> ffmpeg.git 03Michael Niedermayer 07master:c13e4e288c84: h264: fix () placement
[03:05] <cone-117> ffmpeg.git 03Michael Niedermayer 07master:55d052866964: swr/dither: fix division by 0
[03:05] <cone-117> ffmpeg.git 03Michael Niedermayer 07master:aaa7d2fafcc3: h264: don t leave stale pointers in delayed_pic in flush_changes.
[05:28] <cone-117> ffmpeg.git 03Michael Niedermayer 07master:48a7981e6f93: doc/examples: fix assignments in if()
[05:28] <cone-117> ffmpeg.git 03Michael Niedermayer 07master:98fed59427ce: aacenc: Fix assignments in if()
[05:28] <cone-117> ffmpeg.git 03Michael Niedermayer 07master:68a25c64cda1: libfaac: Fix assignments in if()
[05:28] <cone-117> ffmpeg.git 03Michael Niedermayer 07master:9302ad1ac89d: libfdk-aacenc: Fix assignments in if()
[05:28] <cone-117> ffmpeg.git 03Michael Niedermayer 07master:871b6ec01d27: libmp3lame: Fix assignments in if()
[05:28] <cone-117> ffmpeg.git 03Michael Niedermayer 07master:d6180aa29741: libopencore-amr: Fix assignments in if()
[05:28] <cone-117> ffmpeg.git 03Michael Niedermayer 07master:3b8d66d5317d: libspeexenc: Fix assignments in if()
[05:28] <cone-117> ffmpeg.git 03Michael Niedermayer 07master:0ccb31dcad5a: libvo-aacenc: Fix assignments in if()
[05:28] <cone-117> ffmpeg.git 03Michael Niedermayer 07master:bdd71abe5f34: libvorbisenc: Fix assignments in if()
[05:28] <cone-117> ffmpeg.git 03Michael Niedermayer 07master:795d2dc23b16: nellymoserenc: Fix assignments in if()
[05:28] <cone-117> ffmpeg.git 03Michael Niedermayer 07master:e2704381e5f1: ra144enc: Fix assignments in if()
[05:28] <cone-117> ffmpeg.git 03Michael Niedermayer 07master:37be1d802f2e: svq1enc: Fix assignments in if()
[05:28] <cone-117> ffmpeg.git 03Michael Niedermayer 07master:659546b42d65: tiff: Fix assignments in if()
[05:28] <cone-117> ffmpeg.git 03Michael Niedermayer 07master:6f88d2d786b8: vsrc_testsrc: Fix assignments in if()
[05:28] <cone-117> ffmpeg.git 03Michael Niedermayer 07master:1ac5a8d7e334: lavf/mux: Fix assignments in if()
[05:28] <cone-117> ffmpeg.git 03Michael Niedermayer 07master:a601eb9543ec: rtmpproto: Fix assignments in if()
[11:09] <ludde> anyone knows any .ts analysis tools to see if they're valid with respect to timing / muxing?
[12:01] <kierank> ludde: nothing that doesn't cost less than $10k
[12:10] <ludde> ok
[12:11] <cbsrobot> kierank: maybe dnxhd-tr uses it's own quant table too
[12:11] <cbsrobot> there is some discussion on ffmbc, but bcoudurier is just saying it is not implemented.
[12:51] <durandal_1707> there is bunch of spam for av_restrict on same fate machines
[13:34] <cone-774> ffmpeg.git 03Rémi Denis-Courmont 07master:44e065d56c87: vdpau: Add context and common helpers for hwaccel support
[13:34] <cone-774> ffmpeg.git 03Michael Niedermayer 07master:8ee7b3881bfe: Merge commit '44e065d56c87d6a9d0effccec5f31517f72924ec'
[13:54] <cone-774> ffmpeg.git 03Rémi Denis-Courmont 07master:aaf9d0e8f396: vdpau: Add MPEG-1/2 decoding via hwaccel infrastructure
[13:54] <cone-774> ffmpeg.git 03Rémi Denis-Courmont 07master:200e8ac92007: vdpau: Add MPEG-4 decoding via hwaccel infrastructure
[13:54] <cone-774> ffmpeg.git 03Michael Niedermayer 07master:41bdef1d3958: Merge commit '200e8ac92007bc2fe30da05d3bd00ab620842a6b'
[14:05] <cone-774> ffmpeg.git 03Rémi Denis-Courmont 07master:51b56a071602: vdpau: Add H.264 decoding via hwaccel infrastructure
[14:05] <cone-774> ffmpeg.git 03Rémi Denis-Courmont 07master:775253278914: vdpau: Add VC-1 decoding via hwaccel infrastructure
[14:05] <cone-774> ffmpeg.git 03Anton Khirnov 07master:246da0b13551: v4l2: avoid pointless indirection.
[14:05] <cone-774> ffmpeg.git 03Michael Niedermayer 07master:53c2f401f726: Merge commit '246da0b13551b1f80f067e4f258e5bd691f5ab33'
[14:20] <Compn> funman : so are you reviewing that h264 dxva2 deocder patch ?
[14:21] <funman> Compn: i only had a quick look, i saw nothing strange
[14:23] <cone-774> ffmpeg.git 03Anton Khirnov 07master:838b849e70f1: v4l2: set the average framerate instead of codec timebase.
[14:23] <cone-774> ffmpeg.git 03Anton Khirnov 07master:49dc82eef776: v4l2: do not assert on a value received from outside of Libav
[14:23] <cone-774> ffmpeg.git 03Anton Khirnov 07master:90cfc084e3e6: avpacket: free side data in av_free_packet().
[14:24] <cone-774> ffmpeg.git 03Michael Niedermayer 07master:a646ac8ef5c5: Merge commit '90cfc084e3e6d37ab88fc96a95f0401d8e8b4cd1'
[14:25] <Compn> funman : ok thanks
[14:25] <Compn> michaelni : any chance you could review/commit the h264 dxva2 decoder for ffmpeg patch ? 
[14:26] <Compn> or say whats up on the list, rejected, need more info, etc
[14:27] <funman> Compn: except missing copyright of vlc authors though
[14:30] <Compn> Subject: [FFmpeg-devel] [PATCH] libavcodec:add h264_dxva2 decoder correct the copyright line
[14:30] <michaelni> Compn, iam happy to leave review to some other volunteer if there is one
[14:31] <Compn> michaelni : i've been asking few people ... :)
[14:31] <Compn> maybe i need to ask some vlc devs
[14:34] <nevcairiel> review would be much easier if the inline patch wouldn't be mangled to death :p
[14:35] <Compn> theres an attached patch as well ...
[14:36] <Compn> nevcairiel : are you going to review it? 
[14:38] <Compn> i forgot , i think i asked you 
[14:38] Action: Compn losing brain cells
[14:38] <michaelni> Compn, if all else fails you have to review it :)
[14:38] <Compn> :)
[14:39] <Compn> well there was some small things that could be fixed
[14:40] <Compn> but i didnt want to bug patch author with those until i knew if the patch was correct / right design
[14:40] <nevcairiel> i wonder if that person finds my comments in the middle of his patch, or if i should write a big text blob at the end :p
[14:40] <nevcairiel> i think the whole design is terrible, but apparently noone wants to hear that
[14:41] <nevcairiel> features above anything, i guess
[14:41] <Compn> the hwaccel design
[14:41] <Compn> ?
[14:41] <nevcairiel> its not a hwaccel, thats the problem :P
[14:41] <Compn> wasnt his first patch for the hwaccel ?
[14:42] <Compn> before someone said to use the other api...
[14:42] <divVerent> what does the new API do better/different?
[14:43] <nevcairiel> there is no new api
[14:43] <nevcairiel> its just a wrapper around the hwaccel so stupid applications can use it
[14:43] <divVerent> the "hwaccel infrastructure"
[14:43] <divVerent> so the old way that e.g. mplayer uses stays supported?
[14:44] <Compn> >code copied from vlc, blame mplayer 
[14:44] <Compn> :P
[14:44] <divVerent> that means?
[14:44] <Compn> no idea, i'm not the person who decides what api is good or bad
[14:44] <Compn> i dont have that power
[14:44] <Compn> nor expertise
[14:45] Action: michaelni wants fate tests for hwaccel
[14:45] <Compn> michaelni : what about nevcairiel 's and divVerent 's comments about some hwaccel mess and wrappers that mplayer uses ?
[14:46] <divVerent> mplayer used "some hwaccel" for H.264, VC-1 and the other two for ages now
[14:46] <divVerent> for VDPAU support
[14:46] <divVerent> I am mainly wondering, what
[14:46] <divVerent> s the direction we're going, as apparently a new "hwaccel" API is being made
[14:46] <nevcairiel> vdpau is a even worse hack in avcodec =p
[14:47] <divVerent> well, both are in avcodec
[14:47] <Compn> did you see that new vdpau that was committed ? i thought it was there already ...
[14:47] <Compn> 'new' ... 2008 copyright from nvidia
[14:47] <divVerent> yes, that's why I am asking
[14:47] <nevcairiel> there is a AVHWAccel infrastructure, which should be the way to go forward, and vdpau was finally converted to that after years and years
[14:47] <divVerent> thing is, we're already USING "some api" to decode H.264 through VDPAU with ffmpeg's help
[14:47] <divVerent> so the old API is going to be removed?
[14:48] <nevcairiel> not without a major bump, but personally, i hope so
[14:48] <divVerent> what's better with the new one?
[14:48] <Compn> whats the difference between them ?
[14:48] <divVerent> also:
[14:48] <divVerent> 14:40:54     nevcairiel | i think the whole design is terrible, but apparently noone wants to hear that
[14:48] <nevcairiel> the integration with the decoders is much cleaner
[14:48] <divVerent> so the old API#s design is MORE terrible?
[14:48] <nevcairiel> divVerent: that was about the dxva patch currently on review
[14:49] <cone-774> ffmpeg.git 03Luca Barbato 07master:d9cf5f516974: lavc: check for overflow in init_get_bits
[14:49] <cone-774> ffmpeg.git 03Luca Barbato 07master:e28ac6e5e27e: lavc: introduce the convenience function init_get_bits8
[14:49] <cone-774> ffmpeg.git 03Daniel Kang 07master:67360ccd51f3: yadif: x86: fix build for compilers without aligned stack
[14:49] <cone-774> ffmpeg.git 03Anton Khirnov 07master:8fd4d1f9b925: cscd: return meaningful error codes.
[14:49] <cone-774> ffmpeg.git 03Anton Khirnov 07master:9e6764900a07: msrle: return meaningful error codes.
[14:49] <cone-774> ffmpeg.git 03Anton Khirnov 07master:6bf70127d1d4: tmv: return meaningful error codes.
[14:49] <cone-774> ffmpeg.git 03Anton Khirnov 07master:a1c525f7eb07: pcx: return meaningful error codes.
[14:49] <cone-774> ffmpeg.git 03Michael Niedermayer 07master:329675cfd71f: Merge commit 'a1c525f7eb0783d31ba7a653865b6cbd3dc880de'
[14:49] <divVerent> how would code using the new API even look?
[14:50] <wm4> what new api
[14:50] <divVerent> I know that what mplayer does is "somehow related to CODEC_CAP_HWACCEL_VDPAU"
[14:50] <wm4> the old way is selecting specific decoders
[14:50] <wm4> for vdpau
[14:50] <wm4> the new way, you just use the correct decoder, and request a hardware pixel format
[14:50] <nevcairiel> oh my the decoder uses a static data buffer
[14:50] <nevcairiel> that cant go well <.<
[14:51] <wm4> wat
[14:51] <Compn> wm4 : oh, no more mplayer -vc ffh264vdpau,ffvc1vdpau,ffmpeg2vdpau ?
[14:51] <Compn> ehe
[14:51] <Compn> hacks upon hacks
[14:51] <nevcairiel> it should really behave more like any other decoder if it wants to be one
[14:52] <wm4> Compn: no, that already works in mpv (mpv -hwdec vdpau -vo vdpau file.mkv)
[14:52] <divVerent> and is there documentation on how to change the code to the new API?
[14:52] <wm4> divVerent: there never is
[14:52] <divVerent> as I suppoose choosing the pixel format isn't everything
[14:52] <divVerent> well, USUALLY there is at least documentation in the headers
[14:52] <divVerent> but that's not existing for this either
[14:53] <Compn> i've seen various reports about 10bit not working with vdpau
[14:53] <Compn> 10bit h264*
[14:53] <wm4> Compn: mpv will fallback to software decoding in this case
[14:53] <nevcairiel> obviously, the hardware only does 8-bit 
[14:53] <nevcairiel> and only 4:2:0
[14:53] <nevcairiel> and only main/high profile
[14:53] <nevcairiel> and i guess baseline too =P
[14:53] <Compn> nevcairiel : how would a calling application figure out 10bit before sending it to a decoder ?
[14:53] <divVerent> or is the only difference between the APIs how the decoding is set up?
[14:53] <divVerent> I see the new one also uses AV_PIX_FMT_VDPAU
[14:54] <wm4> divVerent: the new one introduced that pixel formats; before that each codec had its own pixel format because mplayer needs that
[14:54] <divVerent> ah, I see
[14:54] <wm4> divVerent: as the decoder is created in the VO instead of the decoder wrapper
[15:04] <durandal_1707> what happened to ubitux?
[15:06] <mateo`> durandal_1707: he's on vacation
[15:06] <durandal_1707> where?
[15:08] <pigoz> south pole
[15:09] <cone-774> ffmpeg.git 03Anton Khirnov 07master:3344f5cb747b: nuv: return meaningful error codes.
[15:09] <cone-774> ffmpeg.git 03Anton Khirnov 07master:62d9655217f9: qpeg: return a meaningful error code.
[15:09] <cone-774> ffmpeg.git 03Anton Khirnov 07master:01cbc6f6adf5: targa: return meaningful error codes.
[15:09] <cone-774> ffmpeg.git 03Anton Khirnov 07master:14cf33e957fb: lcldec: return meaningful error codes.
[15:09] <cone-774> ffmpeg.git 03Anton Khirnov 07master:089b3d6815a8: interplayvideo: return meaningful error codes.
[15:09] <cone-774> ffmpeg.git 03Anton Khirnov 07master:3d973e461be6: interplayvideo: remove a static variable.
[15:09] <cone-774> ffmpeg.git 03Anton Khirnov 07master:048ffb9bb26f: gifdec: return meaningful error codes.
[15:09] <cone-774> ffmpeg.git 03Anton Khirnov 07master:8f1782945530: qtrle: return a meaningful error code.
[15:09] <cone-774> ffmpeg.git 03Anton Khirnov 07master:688b132b881d: qdrw: return meaningful error codes.
[15:09] <cone-774> ffmpeg.git 03Michael Niedermayer 07master:8943925d1f68: Merge commit '688b132b881d423877e38dc82f17e23a604be676'
[15:09] <mateo`> durandal_1707: japan
[15:11] <michaelni> j-b, i just got "remote: mail: /tmp/mail.RsXXXXRzM96x: No space left on device" 
[15:13] <durandal_1707> OOO
[15:13] <michaelni> seems the push worked fine though otherwise
[15:27] <ubitux> hey, japan is connected to internet!
[15:27] <ubitux> @_@
[15:28] <michaelni> ubitux, :))))))
[15:28] <JEEB> ubitux, yes -- if you are talking about public wifis they are indeed scarce as hell
[15:28] <JEEB> it's easier to get into an internet cafe than find a publically usable wifi spot
[15:28] <JEEB> even in Starbucks!
[15:29] <ubitux> :)
[15:29] Action: JEEB still rememembers him wardriving through Tokyo station's Yaesu exit
[15:30] <JEEB> Starbucks -> "Sorry, we don't have wifi. But I think the McDonalds next to us has." -> McDonalds -> "Sorry, but the WiFi is only for Yahoo! BB users." -> ...
[15:31] <JEEB> I think the widest free wifi in Tokyo is the Tokyo Metro one
[15:31] <mateo`> ubitux: \o/
[15:31] <JEEB> there were some WEP networks in there, I just wish my eeepc had more than 2h of battery power :V
[15:32] <JEEB> (@Yaesu Exit)
[15:32] <cone-774> ffmpeg.git 03Anton Khirnov 07master:80cf2ebc59ba: qdrw: cosmetics, reformat
[15:32] <cone-774> ffmpeg.git 03Anton Khirnov 07master:e4ca055b3ce3: rpza: return a meaningful error code.
[15:32] <cone-774> ffmpeg.git 03Anton Khirnov 07master:74a9a624c5c4: vcr1: return a meaningful error code.
[15:32] <cone-774> ffmpeg.git 03Michael Niedermayer 07master:4adc8a29a056: Merge commit '74a9a624c5c4d50760d8d57458eba57366f6cb26'
[15:33] <mateo`> is dnxhd-tr standardized ?
[15:40] <cone-774> ffmpeg.git 03Anton Khirnov 07master:6b3a1229bb30: vcr1: remove disabled encoder stub
[15:40] <cone-774> ffmpeg.git 03Anton Khirnov 07master:a9d970a019df: flicvideo: return meaningful error codes.
[15:40] <cone-774> ffmpeg.git 03Anton Khirnov 07master:1a31dff9370b: loco: return meaningful error codes.
[15:40] <cone-774> ffmpeg.git 03Michael Niedermayer 07master:f9d5bdd0f652: Merge commit '1a31dff9370b4732c91df5cb1ca4b39f2cb3050e'
[15:51] <cone-774> ffmpeg.git 03Anton Khirnov 07master:b627c3010b9b: loco: cosmetics, reformat
[15:51] <cone-774> ffmpeg.git 03Anton Khirnov 07master:ac1e93f555aa: tscc: return meaningful error codes.
[15:51] <cone-774> ffmpeg.git 03Anton Khirnov 07master:a4a26f518856: tscc: remove some pointless comments and empty lines.
[15:51] <cone-774> ffmpeg.git 03Anton Khirnov 07master:df9036830b15: truemotion2: return meaningful error codes.
[15:51] <cone-774> ffmpeg.git 03Michael Niedermayer 07master:92b238715929: Merge commit 'df9036830b15997a3e9c3f2c632ed98d64f9deef'
[16:21] <cone-774> ffmpeg.git 03Anton Khirnov 07master:60a42ef44cc0: truemotion2: cosmetics, reformat
[16:21] <cone-774> ffmpeg.git 03Michael Niedermayer 07master:59d568031008: h264: Fix assignments in if()
[16:21] <cone-774> ffmpeg.git 03Michael Niedermayer 07master:1d7ffd06e41e: lavc: Fix assignments in if() when calling ff_af_queue_add
[16:21] <cone-774> ffmpeg.git 03Michael Niedermayer 07master:30aeada3d8e9: svq1enc: Fix assignments in if()
[16:21] <cone-774> ffmpeg.git 03Michael Niedermayer 07master:d641ee94b57a: lavf: Fix assignments in if()
[16:22] <cone-774> ffmpeg.git 03Michael Niedermayer 07master:6dc8505417c9: rtmpproto: Fix assignments in if()
[16:22] <cone-774> ffmpeg.git 03Michael Niedermayer 07master:57d59e82cbdd: Merge remote-tracking branch 'qatar/master'
[16:29] <cone-774> ffmpeg.git 03Carl Eugen Hoyos 07master:c29c7c1470f9: The c99-to-c89 binaries are now hosted on videolan.org.
[16:36] <cone-774> ffmpeg.git 03Michael Niedermayer 07master:ac73d3a12a33: init_get_bits8: check byte_size against being positive
[16:36] <cone-774> ffmpeg.git 03Michael Niedermayer 07master:153fad14e5a2: init_get_bits8: zero pointers & struct on error
[16:36] <cone-774> ffmpeg.git 03Michael Niedermayer 07master:7980cca05c7c: init_get_bits: fix off by 1 error
[17:10] <cone-774> ffmpeg.git 03Michael Niedermayer 07master:922332e60df3: nuv: remove unused variable
[17:14] Action: Compn remembers paying 100Y for 10 minutes internet
[17:14] <Compn> stupid hotel rates!
[17:15] <Compn> that was before internet cafes made it big
[17:26] <Jumpyshoes> michaelni: yadif should be fixed
[17:59] <mateo`> sorry to ask again, is dnxhd-tr standardized ?
[18:13] <durandal_1707> there is no dnxhd spec?
[18:14] <merbanan> smpte vc-3
[18:15] <merbanan> but lots of stuff is not specified
[18:26] <kierank> mateo`: no, cid 1244 is not specified in vc-3
[18:31] <mateo`> kierank: i was looking for an update of the SMPTE document :(
[18:31] <kierank> i don't think there is a revision
[19:14] <michaelni> Jumpyshoes, thanks!
[22:11] <Skyler_> if someone wanted to pay for an LXF demuxer/decoder(?), who should I point them to?
[22:12] <Skyler_> (is that format related to MXF?)
[22:12] <Skyler_> bcoudurier: ^
[22:12] <kierank> it has been done
[22:12] <Skyler_> oh, already supported?
[22:12] <kierank> yes
[22:12] <kierank> ifb wrote a patch for newer variants
[22:13] <kierank> but it didn't get accepted into libav
[22:13] <kierank> because at the time they were rewriting the whole audio side
[22:13] <Skyler_> okay, will poke them with that info
[22:14] <kierank> http://lists.libav.org/pipermail/libav-devel/2011-December/017642.html
[22:14] <kierank> 21:12:33 <"Skyler_> (is that format related to MXF?) --> only in that it is a broadcast codec
[22:14] <kierank> container*
[22:15] <Skyler_> I can't find it on the changelog on the website
[22:15] <Skyler_> is it not in a release yet?
[22:16] <beastd> should be in an ffmpeg release if i am not mistaken
[22:17] <kierank> Skyler_: do they need vbi support?
[22:17] <Skyler_> I don't know, they'll poke me if they do I suppose
[22:17] <kierank> because there's no chance in hell that'll work any time soon in lavc
[22:45] <Compn>  Demuxer for Leitch/Harris' VR native stream format (LXF)
[22:46] <Compn> is that the same LXF format Skyler_ ? 
[22:46] <Compn> it sounds like an old format ...
[22:57] <Skyler_> Yes it's that
[22:58] <pettter> is there any UPnP/DLNA server currently using FFmpeg for input/output/transcoding?
[23:03] <cone-902> ffmpeg.git 03Michael Niedermayer 07master:30981a966f17: lavc: split snow and dirac DWTs
[23:17] <Compn> pettter : oh, i remember a project doing that using ffmpeg
[23:17] <Compn> but i dont remember now
[23:18] <Compn> either xbmc or ps3mediaserver
[23:18] <Compn> probably xbmc...
[23:18] <Compn> XBMC also have a built-in UPnP-server (since the 24th of January 2007)
[23:18] <Compn> PS3 Media Server is a DLNA compliant Upnp Media Server for the PS3, written in Java
[23:18] <Compn> ah both of them :)
[23:19] <pettter> Compn: right
[23:19] <Compn> so whats the question ?
[23:20] <pettter> Compn: basically, I'm running raspbmc (XBMC on a raspberry pi) and am considering various ways to get transcoding and streaming to work
[23:20] <pettter> running xbmc on the media server _as well_ feel excessive...
[23:20] <pettter> currently setting up minidlna, we'll see how it goes
[23:24] <Compn> well , you expect the raspberry pi to transcode ?
[23:25] <Compn> i dont think it has much cpu
[23:27] <kierank> it has a hardwaren encoder
[23:33] <cone-902> ffmpeg.git 03Michael Niedermayer 07master:9cb3c1a4d9db: x86/dirac: fix asm on win64
[23:45] <cone-902> ffmpeg.git 03Michael Niedermayer 07master:274e48d8ac1e: x86/Makefile: move dirac_dwt to right type
[23:53] <Compn> kierank : still have to decode it , if its not h264 already
[23:53] <kierank> it has mpeg-2/vc-1/h264 decoder
[23:53] <Compn> not those then
[23:54] <gnafu> Well, you have to pay for MPEG-2 and VC-1, so I could see situations where someone might want to reencode even those on another machine ;D.
[23:55] <gnafu> You could reencode all non-H.264 content to H.264 that the Pi can play out of the box.
[23:56] <Compn> they can have my money when they pry it from my cold dead hands
[23:59] <gnafu> Hehe.
[00:00] --- Tue Jan 15 2013