[Ffmpeg-devel-irc] ffmpeg-devel.log.20150916

[burek]

[00:03:19 CEST] <jamrial> Gramner: seems to work, so lgtm
[03:57:44 CEST] <cone-437> ffmpeg 03Michael Niedermayer 07master:3496a20bb925: avformat/flvdec: Change packet loop to return EAGAIN instead of looping until a valid packet is foud
[03:57:44 CEST] <cone-437> ffmpeg 03Michael Niedermayer 07master:e01ad950bb1b: avformat/rtmpproto: Fix 2nd packet size
[03:57:44 CEST] <cone-437> ffmpeg 03Michael Niedermayer 07master:cbbd906be615: avformat/flvdec: Check that sizes match and resync if not
[05:03:46 CEST] <jamrial> https://gcc.gnu.org/git/?p=gcc.git;a=commitdiff;h=1bca8cbd0c68366f07277f98ce6963e10c2aa617 this change apparently made the warnings during compilation of ffmpeg go from ~90 to more than 500
[05:50:19 CEST] <dioo> Hi, is avcodec_decode_video2() supposed to set picture->format if it doesn't fail?
[05:51:50 CEST] <dioo> Because to my untrained eye, it looks like the if (HAVE_THREADS &) (line 2095 in latest git) branch doesn't set it.
[05:52:05 CEST] <dioo> And it makes unpaper fail
[05:52:46 CEST] <peloverde> anyone have recommendations for an msys variant for developing ffmpeg?
[05:53:06 CEST] <dioo> Ohh, wait. Does it just spawn another thread to do the decode which will go through the else?
[05:58:27 CEST] <dioo> Nope, looks like the else branch is never hit
[05:58:46 CEST] <jamrial> peloverde: msys2?
[06:01:00 CEST] <jamrial> it's actively maintained, and their package manager has gcc using mingw-w64 instead of mingw32 like msys' does (assuming you care about that and don't use your own gcc)
[06:23:22 CEST] <peloverde> jamrial: thanks
[07:32:16 CEST] <lilibox_> hi
[07:33:08 CEST] <lilibox_> i have got special wish, i would like to get ffmpeg revision 17733 from unknown reason from svn repo is libswscale folder missing :( 
[07:43:59 CEST] <peloverde> lilibox_: back then swscale lived in mplayer svn
[07:45:49 CEST] <peloverde> You should be able to search git for "Originally committed as revision 17733 to svn://svn.ffmpeg.org/ffmpeg/trunk"
[07:45:56 CEST] <peloverde> and that will have sws integrated
[07:47:26 CEST] <lilibox_> thank you for fast replay
[07:47:44 CEST] <lilibox_> this svn co -r 17733 svn://svn.ffmpeg.org/ffmpeg/trunk got me empty ...scale folder 
[07:47:55 CEST] <lilibox_> soo, i will try harvest it from mplayer
[07:48:03 CEST] <lilibox_> thank you for advice
[09:03:10 CEST] <fritsch> nevcairiel: ping
[09:03:30 CEST] <fritsch> nevcairiel: is it true that a GTX960 can already decode and render via dxva main10 hevc content?
[09:08:43 CEST] <JEEB> fritsch: yes
[09:12:20 CEST] <fritsch> JEEB: but using a copy back method?
[09:12:32 CEST] <fritsch> e.g. only the decoder works but rendering afterwards is an issue?
[09:12:40 CEST] <fritsch> or does this also work now on the dxva chain?
[09:14:03 CEST] <nevcairiel> fritsch: yes its true, both copy back and native dxva, although the renderer converts it to 8-bit rgb at some point
[09:14:40 CEST] <fritsch> okay - so basically on linux a lot infrastructure work is needed yet?
[09:15:10 CEST] <fritsch> or as first solution, the decoder already converts to 8 bit for us?
[09:17:15 CEST] <fritsch> nevcairiel: and is there other hardware already, that is able to do the decoding?
[09:17:25 CEST] <nevcairiel> personally i wouldnt trust it to do a properly dithered conversion to 8-bit
[09:17:58 CEST] <nevcairiel> there is also the GTX950 of course, and on windows there is also Intels hybrid decoder which can do 10-bit,  but partly on CPU
[09:19:21 CEST] <fritsch> cool, thx - that clears it up
[09:19:39 CEST] <fritsch> but having a 10 -> 8 conversation at one point would be an intermediate solution
[09:19:54 CEST] <fritsch> btw. the 8 bit conversion that is done in the render - did you add dithering here?
[09:20:35 CEST] <nevcairiel> i dont control the renderer
[09:20:51 CEST] <nevcairiel> i just provide a decoder
[09:22:18 CEST] <fritsch> okay - but if I was the render I could add a shader to add some dithering?
[09:22:59 CEST] <nevcairiel> if you keep the bitdepth high enough until that point, sure
[09:23:22 CEST] <fritsch> may I cite this conversation?
[09:23:30 CEST] <fritsch> cause a lot of noob knowledge is out there
[09:23:38 CEST] <fritsch> and people demanding things that are just not given
[09:26:33 CEST] <fritsch> nevcairiel: ah, that channel is publically logged - so thanks again for that information
[09:29:22 CEST] <nevcairiel> If you think that will make the noobs listen, which they probably wont =p
[10:16:40 CEST] <kierank> durandal_1707: are you at vdd?
[10:21:55 CEST] <durandal_1707> why would I, its full of ego tripping :-P
[10:24:23 CEST] <durandal_1707> maybe next year when its closer to my city
[11:10:00 CEST] <cone-583> ffmpeg 03Vittorio Giovara 07master:10de408738d2: lavf: Update to the new crypto API
[11:10:00 CEST] <cone-583> ffmpeg 03Hendrik Leppkes 07master:182550e42c02: Merge commit '10de408738d28ab17aa5c1fdccd809b0637c12d5'
[11:12:28 CEST] <cone-583> ffmpeg 03Vittorio Giovara 07master:18adc1fd3bb7: avconv_filter: Use full named options
[11:12:29 CEST] <cone-583> ffmpeg 03Hendrik Leppkes 07master:352eb4b5992e: Merge commit '18adc1fd3bb739a240c59f87ec6837fdfa89d32e'
[11:12:40 CEST] <cone-583> ffmpeg 03Vittorio Giovara 07master:5b142a4d334b: lavc: Keep idct integer option
[11:12:41 CEST] <cone-583> ffmpeg 03Hendrik Leppkes 07master:a600e7145f9e: Merge commit '5b142a4d334bfa8ee5ec5f364c26c98ddb425fac'
[11:18:23 CEST] <cone-583> ffmpeg 03Vittorio Giovara 07master:e17db9ac4964: fate: Move mv0 option from global to mpv private
[11:18:24 CEST] <cone-583> ffmpeg 03Hendrik Leppkes 07master:25a7c2aed6f4: Merge commit 'e17db9ac496420f8824190ff5059a8a3646cbb7e'
[11:18:36 CEST] <cone-583> ffmpeg 03Vittorio Giovara 07master:2c6014c62934: fate: Change global error option to private error_rate option
[11:18:37 CEST] <cone-583> ffmpeg 03Hendrik Leppkes 07master:03d47f2ff1b5: Merge commit '2c6014c629346ebfeb25fc715581a8df9175f71f'
[11:19:28 CEST] <cone-583> ffmpeg 03Vittorio Giovara 07master:dc54c78c4d56: buffersrc: Improve initialization log message
[11:19:29 CEST] <cone-583> ffmpeg 03Hendrik Leppkes 07master:1eb2c0de9eee: Merge commit 'dc54c78c4d56cdf5549f35ccff2fd66e7ffa0af2'
[11:19:47 CEST] <cone-583> ffmpeg 03Vittorio Giovara 07master:c7247eb7fe2b: lavr: Remove unreachable code
[11:19:48 CEST] <cone-583> ffmpeg 03Hendrik Leppkes 07master:c2972a280024: Merge commit 'c7247eb7fe2b95803983e8f7c4f0e904f050ca7b'
[11:21:10 CEST] <cone-583> ffmpeg 03Luca Barbato 07master:285fe5b098c8: avfilter: Support both syntaxes for the scale filter
[11:21:11 CEST] <cone-583> ffmpeg 03Hendrik Leppkes 07master:643a1560024f: Merge commit '285fe5b098c8fbefaed4aacdb839b760a5b7f20d'
[11:23:29 CEST] <cone-583> ffmpeg 03Vittorio Giovara 07master:525f58977c93: mpegvideo: Move macros to more appropriate headers
[11:23:30 CEST] <cone-583> ffmpeg 03Hendrik Leppkes 07master:b3ff8ef6ec23: Merge commit '525f58977c93e189fda49a5c4928feaf4d89fac6'
[11:24:00 CEST] <cone-583> ffmpeg 03Vittorio Giovara 07master:5d14cf199990: mpegvideo: Make sure mpegutils.h is included where needed
[11:24:01 CEST] <cone-583> ffmpeg 03Hendrik Leppkes 07master:7b865c222e43: Merge commit '5d14cf199990cd378904a2618b5c72c4b02290f6'
[11:28:56 CEST] <ubitux> nevcairiel :)
[11:29:11 CEST] <cone-583> ffmpeg 03Vittorio Giovara 07master:e60a6e7545dd: mpegvideo: Drop mpegvideo.h where not needed
[11:29:12 CEST] <cone-583> ffmpeg 03Hendrik Leppkes 07master:52ce8b7b3177: Merge commit 'e60a6e7545dd6f5b25e3a65de9c6fdcc6e2e9d6b'
[11:29:21 CEST] <nevcairiel> i sure hope vaapi doesnt break in that one, i couldnt test that =p
[11:29:52 CEST] <ubitux> i'm smiling for 643a1560
[11:30:03 CEST] <ubitux> anyway, thx
[11:30:22 CEST] <BtbN> nevcairiel, break in what way?
[11:30:30 CEST] <nevcairiel> compiling
[11:30:30 CEST] <nevcairiel> :D
[11:31:32 CEST] <cone-583> ffmpeg 03Luca Barbato 07master:6536da7371b4: doc: Add travis badge
[11:31:33 CEST] <cone-583> ffmpeg 03Hendrik Leppkes 07master:5de28d341f24: Merge commit '6536da7371b4847a2824ed46b98829f127f24dad'
[11:31:42 CEST] <BtbN> I'll test-compile it.
[11:31:53 CEST] <cone-583> ffmpeg 03Luca Barbato 07master:41bcc3d15204: jpeg2000: Split codeblock decoding from the main tile decoding
[11:31:54 CEST] <cone-583> ffmpeg 03Hendrik Leppkes 07master:5a754f92d1ab: Merge commit '41bcc3d15204f290400ba02e4e8f87fc07bcc00e'
[11:34:40 CEST] <BtbN> nevcairiel, breaks compilation.
[11:34:58 CEST] <nevcairiel> want to fix it? :)
[11:35:10 CEST] <BtbN> https://bpaste.net/show/626f60886814
[11:35:31 CEST] <BtbN> I guess some header is missing?
[11:36:30 CEST] <nevcairiel> yeah, add an include to internal.h
[11:36:37 CEST] <BtbN> Which commit is the one that potentialy caused this?
[11:37:00 CEST] <nevcairiel> http://git.videolan.org/?p=ffmpeg.git;a=commitdiff;h=52ce8b7b31773f5dd749184445189ecd0cbca24d
[11:37:10 CEST] <nevcairiel> or the origin http://git.videolan.org/?p=ffmpeg.git;a=commit;h=e60a6e7545dd6f5b25e3a65de9c6fdcc6e2e9d6b
[11:37:29 CEST] <nevcairiel> the error seems like you want to include interna.h in vaapi_internal.h
[11:37:44 CEST] <nevcairiel> internal*
[11:38:17 CEST] <BtbN> yes, was aparently included from some other header before.
[11:39:47 CEST] <BtbN> Should i just push the fix?
[11:40:02 CEST] <nevcairiel> if it works agian now, then go ahead
[11:40:25 CEST] <cone-583> ffmpeg 03Michael Niedermayer 07master:fee2b97d563d: avcodec/libxvid: add missing include
[11:41:32 CEST] <cone-583> ffmpeg 03Timo Rothenpieler 07master:dad38bff81bb: avcodec/vaapi_internal: Add missing include
[11:42:24 CEST] <durandal_1707> michaelni: why you beat me? WHY?
[11:43:15 CEST] <michaelni> sorry
[11:50:25 CEST] <cone-583> ffmpeg 03Paul B Mahol 07master:03f34e0d3c1c: doc/filters: ladspa: add two more examples, this time using SWH plugins
[11:50:26 CEST] <cone-583> ffmpeg 03Paul B Mahol 07master:dc1050a3e855: avfilter/af_ladspa: process all channels for nb_handles > 1
[12:14:16 CEST] <cone-583> ffmpeg 03Luca Barbato 07master:29b00f880faa: jpeg2000: Templatize the frame writer
[12:14:17 CEST] <cone-583> ffmpeg 03Hendrik Leppkes 07master:64f72bb61f0f: Merge commit '29b00f880faa404aa1d0d6820310c510c5996479'
[12:26:53 CEST] <durandal_1707> FFmpeg and libav metric? What is this about?
[12:27:02 CEST] <JEEB> random numbers
[13:27:04 CEST] <durandal_1707> BBB: what is ffmpeg and libav metric?
[13:27:14 CEST] <BBB> ?
[13:27:34 CEST] <BBB> sorry, missing context
[13:37:31 CEST] <durandal_1707> your talk at vdd
[13:56:48 CEST] <cone-583> ffmpeg 03Henrik Gramner 07master:985e7d8cc109: checkasm: v210: s/Libav/FFmpeg/
[13:56:49 CEST] <cone-583> ffmpeg 03Henrik Gramner 07master:6115966ad347: checkasm: v210: Fix array overwrite
[13:59:13 CEST] <cone-583> ffmpeg 03Luca Barbato 07master:1b709f23fb5f: jpeg2000: Refactor decode_packets
[13:59:14 CEST] <cone-583> ffmpeg 03Luca Barbato 07master:7fb93eae43cc: jpeg2000: Factor out band initialization
[13:59:15 CEST] <cone-583> ffmpeg 03Hendrik Leppkes 07master:8f80f3d074d8: Merge commit '1b709f23fb5f505c834d4c855703225795def01d'
[13:59:16 CEST] <cone-583> ffmpeg 03Hendrik Leppkes 07master:054ce2d6ab15: Merge commit '7fb93eae43cc89bd7c1c275344e1cdbadab7155b'
[14:02:07 CEST] <cone-583> ffmpeg 03Luca Barbato 07master:95a41311ac3a: jpeg2000: Factor out band stepsize initialization
[14:02:08 CEST] <cone-583> ffmpeg 03Hendrik Leppkes 07master:2fc552b629d7: Merge commit '95a41311ac3a44773cc4dc407408aca35b1f8e26'
[14:03:35 CEST] <cone-583> ffmpeg 03Carl Eugen Hoyos 07master:f17ca460093b: lavf/riff: Support decoding Ambisonic wav files.
[14:07:00 CEST] <cone-583> ffmpeg 03Luca Barbato 07master:570fcaf3326a: jpeg2000: Factor out prec init
[14:07:01 CEST] <cone-583> ffmpeg 03Hendrik Leppkes 07master:b2a72e8ab98d: Merge commit '570fcaf3326aef9313b10863e2f6e6ae664d9dae'
[14:09:40 CEST] <cone-583> ffmpeg 03Martin Storsjö 07master:3c525b8b4770: rtpdec: Increase the max size of the jitter buffer to 500 packets
[14:09:41 CEST] <cone-583> ffmpeg 03Hendrik Leppkes 07master:7cecab2f886b: Merge commit '3c525b8b4770c1ac5f466a12c5523802bd5d40eb'
[14:10:59 CEST] <cone-583> ffmpeg 03Eloi BAIL 07master:ff7f6ea9db2a: rtpdec: add a trace when jitter buffer is full
[14:11:00 CEST] <cone-583> ffmpeg 03Hendrik Leppkes 07master:3d481f39a272: Merge commit 'ff7f6ea9db2a77d74f7e68a716f53ba1f3f85017'
[14:11:14 CEST] <cone-583> ffmpeg 03Eloi BAIL 07master:d555bd69e792: rtpdec: inform jitter buffer size
[14:11:15 CEST] <cone-583> ffmpeg 03Hendrik Leppkes 07master:85c85d453d79: Merge commit 'd555bd69e792e42d94bd909b30d9bfef01db4e25'
[14:12:00 CEST] <nevcairiel> fuck all that refactoring, hours of time trying not to break j2k :(
[14:12:08 CEST] <cone-583> ffmpeg 03Eloi BAIL 07master:1ec611a10228: rtsp: warning when max_delay reached
[14:12:09 CEST] <cone-583> ffmpeg 03Hendrik Leppkes 07master:78af3ffaa934: Merge commit '1ec611a10228945d2ec8a9cf6c5531dee6b7ee56'
[14:13:00 CEST] <nevcairiel> doesnt help that fate coverage of j2k seems to be minimal at best
[14:18:50 CEST] <ubitux> it's not refactoring since it's only used in one place
[14:19:28 CEST] <nevcairiel> technically in the end some loop content was put into a function so the overall binary size may be smaller
[14:19:38 CEST] <nevcairiel> who k nows
[14:19:49 CEST] <nevcairiel> still a lot of work to merge :p
[14:20:16 CEST] <nevcairiel> some people just dont like long functions, i guess
[14:35:56 CEST] <cone-583> ffmpeg 03Michael Niedermayer 07master:a860adb49c4c: tests/checkasm/vp9dsp: Use snprintf() for safetey
[14:47:22 CEST] <durandal_1707> BBB: your talk at VDD about metric..
[14:48:16 CEST] <durandal_1707> to what branch push fix for 2.8 release?
[14:48:55 CEST] <ubitux> release/2.8
[14:50:40 CEST] <BBB> durandal_1707: its a whole bunch of metrics
[14:52:11 CEST] <durandal_1707> would you share slides after its over?
[14:52:57 CEST] <cone-583> ffmpeg 03Paul B Mahol 07release/2.8:fb0d41932d4e: avfilter/af_ladspa: process all channels for nb_handles > 1
[14:55:07 CEST] <BBB> durandal_1707: ok
[14:59:33 CEST] <BBB> j-b: can I rename the vp9 talk to more accurately reflect talk content?
[15:07:20 CEST] <andrelec1> hello
[15:07:49 CEST] <andrelec1> i need some help for starting my school project ...
[15:08:46 CEST] <andrelec1> when i try compile my projet i have ffmpeg/swscale.h missing !
[15:09:06 CEST] <j-b> BBB: sure
[15:11:27 CEST] <ubitux> andrelec1: wrong channel, see topic
[15:18:42 CEST] <andrelec1> oupsss, sorry ;)
[15:26:01 CEST] <BBB> I wonder why intel decide to add packuswd in sse4 rather than in sse2
[16:08:34 CEST] <durandal_1707> ubitux: the sine example is buggy
[16:10:58 CEST] <ubitux> durandal_1707: ./ffplay -v repeat -f lavfi -i "sine=1000:samples_per_frame='st(0,mod(n,5)); print(1602-not(not(eq(ld(0),1)+eq(ld(0),3))))'"
[16:11:03 CEST] <ubitux> i just tried this
[16:11:12 CEST] <ubitux> and it looks ok, unless i'm mistaken
[16:11:17 CEST] <ubitux> what's wrong?
[16:11:49 CEST] <ubitux> (you can add -nostat too)
[16:11:59 CEST] <durandal_1707> I need to escape ;
[16:12:14 CEST] <ubitux> you shouldn't
[16:12:53 CEST] <ubitux> you have to wrap everything into ""
[16:13:37 CEST] <ubitux> the example is the exact string you have to send to the filter
[16:17:30 CEST] <Gramner> j-b: is there any vdd hotel room list available?
[16:23:02 CEST] <j-b> Gramner: privately, yes. Why?
[16:23:57 CEST] <Gramner> was just wondering who I'm sharing with
[16:24:06 CEST] <j-b> someone you hate
[16:24:22 CEST] <Gramner> so I'm alone then, got it!
[16:24:34 CEST] <Gramner> :D
[16:35:50 CEST] <durandal_1707> in what room is Carl?
[16:37:26 CEST] <j-b> durandal_1707: yours
[16:39:02 CEST] <j-b> durandal_1707: carl did not request a room
[16:41:09 CEST] <durandal_1707> heh
[16:42:34 CEST] <lilibox_> svn://svn.ffmpeg.org/ffmpeg/trunk<@peloverde> You should be able to search git for "Originally committed as revision 17733 to svn://svn.ffmpeg.org/ffmpeg/trunk"
[18:26:21 CEST] <BBB> who is good at macros here
[18:26:40 CEST] <BBB> lets say I have this code: http://pastebin.com/Xwm9Te1X
[18:27:05 CEST] <BBB> you see all these occurences of the number 10, and Id like to have two versions of this file, one for 10bpp and one for 12bpp
[18:27:24 CEST] <BBB> so I change all 10 to BPC_TEMPLATE, rename the file to blabla_tempalte and include that template in this file
[18:27:42 CEST] <BBB> and then it fails to compile because in some cases it replaces BPC_TEMPLATE with 10, but in others it doesnt
[18:27:43 CEST] <BBB> why?
[18:29:03 CEST] <atomnuker> do you really need a template for this?
[18:29:40 CEST] <BBB> itll grow a lot larger over time, so yes
[18:29:49 CEST] <atomnuker> I woud just add a int n_bits argument, make the function an inline and then go about my business
[18:30:10 CEST] <atomnuker> the performance should be the same and any sane compiler undestands how to optimize it out
[18:30:25 CEST] Action: durandal_1707 looks at swscale
[18:30:28 CEST] <BBB> Im not interested in performance
[18:30:34 CEST] <BBB> Im interested in not duplicating the source code twice
[18:30:43 CEST] <BBB> (tautology?)
[18:32:28 CEST] <nevcairiel> you want to template the init code?
[18:32:36 CEST] <atomnuker> I don't think you'd be repeating code if you inlined the function
[18:33:50 CEST] <BBB> I dont think I understand how inlining would solve it
[18:33:59 CEST] <BBB> oh I think I get it
[18:34:07 CEST] <BBB> the init_* is not a function, its a macro
[18:34:13 CEST] <BBB> I cant give it function arguments, that wouldnt work
[18:34:18 CEST] <BBB> brb
[18:35:36 CEST] <atomnuker> it doesn't matter if it's a macro, it'll sill inline and take whatever you give it as an argument
[18:48:39 CEST] <cone-583> ffmpeg 03Rostislav Pehlivanov 07master:5ba811bdf00f: aacenc_tns: fix coefficient compression condition
[18:54:46 CEST] <durandal_1707> atomnuker: so you going to use showspectrum?
[18:58:36 CEST] <atomnuker> well, debian fixed spek's ridiculous libary requirements for now, so haven't done much to avf_showspectrum
[18:58:50 CEST] <atomnuker> though I still plan to
[19:27:17 CEST] <feepk> hello guys, sorry for the off-topic: if you join us on Saturday for the VideoLAN Dev Days dinner, please reply to this form _NOW_: http://goo.gl/forms/4yZx5l8bZg
[19:39:01 CEST] <BBB> atomnuker: its a macro argument thats concatenated
[19:39:30 CEST] <BBB> decl_mc_func(a,b,c,d,e) expands to vp9_##a##_##b##_##c##_##d##_##e## or some variant thereof
[19:39:39 CEST] <BBB> atomnuker: that doesnt work with function arguments, sadly
[20:33:20 CEST] <durandal_1707> BBB: what's error?
[20:55:51 CEST] <BBB> durandal_1707: it works now, n/m
[20:55:57 CEST] <BBB> mustve been some weird thing I did wrong
[20:56:04 CEST] <BBB> also: yay, more vp9 simd
[21:03:58 CEST] <BBB> lets do loopfilter next I guess
[21:17:06 CEST] <kinnu323>  /msg NickServ VERIFY REGISTER kinnu323 kkfecpdozysw
[21:24:00 CEST] <BBB> kinnu323: you just pasted your password in this channel, you probably want to change it if you used it anywhere else
[21:24:16 CEST] <Daemon404> thats not a password
[21:24:23 CEST] <Daemon404> it's a registration token
[21:33:26 CEST] <BBB> j-b deleted my ffmpeg/libav metrics from the wiki :-p
[21:34:05 CEST] <Daemon404> im sure announcing it beforehand brought out the insane people
[21:34:06 CEST] <BBB> I guess we should make that a libav/ffmpeg only thing
[21:34:08 CEST] <Daemon404> the zealots
[21:34:14 CEST] <BBB> who cares
[21:34:20 CEST] <Daemon404> j-b
[21:34:22 CEST] <BBB> the worst thing they could do is kill me and put the place on fire
[21:34:26 CEST] <Daemon404> j-b's pm window*
[21:34:27 CEST] <BBB> thats not so bad
[21:34:42 CEST] <BBB> (assuming j-b has fire insurance)
[21:34:56 CEST] <Daemon404> the building itself is an ad company i think
[21:44:43 CEST] <BBB> jamrial: Im posting updates on github as-we-go btw, Ill send a new patch email after youve reviewed them all
[21:44:47 CEST] <BBB> otherwise it gets confusing
[21:47:45 CEST] <BBB> the interesting thing is that with C implementations for anything except MC (emu_edge, itxfm_add, intra_pred, loop_filter), were already faster than libvpx (which, to my knowledge, has full simd coverage)
[21:47:59 CEST] <Daemon404> wut
[21:48:30 CEST] <BBB> its a fast decoder ;)
[21:55:31 CEST] <llogan> BBB: preparing for your talk, eh?
[21:55:46 CEST] <BBB> the talk only covers 8bit
[21:55:52 CEST] <BBB> so the stuff Im doing now is unrelated to the talk
[21:55:56 CEST] <BBB> (10/12bpp optimizations)
[21:56:07 CEST] <llogan> ah
[21:56:12 CEST] <BBB> I dont think anyone is using 10/12bpp vp9 in the wild yet
[21:56:29 CEST] <BBB> you still need a config option to enable it in libvpx, I dont think its enabled by default, so its not trivial to generate content for it
[21:56:54 CEST] <llogan> looks like some interesting topics on the schedule
[21:57:14 CEST] <BBB> very codec-focussed, I like it
[21:57:15 CEST] <jamrial> is google pushing for vp9 10bit adoption? for example with the blu ray 4k spec it seems that hevc 10bit yuv444 is going to be more popular than 8bit yuv420
[21:57:27 CEST] <BBB> jamrial: ask google :-p
[21:57:34 CEST] <JEEB> jamrial: > 4:4:4
[21:57:37 CEST] <JEEB> that's not gonna happen
[21:57:39 CEST] <JEEB> 10bit is
[21:57:50 CEST] <JEEB> but 4:2:0 is still what 2160p blu-ray is gonna use
[21:57:59 CEST] <BBB> 444 sounds weird, yes; people typically use that only for desktop remoting etc.
[21:58:01 CEST] <JEEB> http://i.imgur.com/FDVvfTY.jpg
[21:58:01 CEST] <jamrial> if you mean blu ray 4k, i thought it was already decided it was yuv444
[21:58:07 CEST] <JEEB> nope
[21:58:13 CEST] <jamrial> guess i read wrong then
[21:58:22 CEST] <JEEB> and as far as I know, that stuff is what went into the specs
[21:59:04 CEST] <BBB> I like how that slide says confidential and is on imgur
[21:59:13 CEST] <JEEB> yup
[22:33:21 CEST] <kierank> BBB: that's marketing
[22:44:04 CEST] <nevcairiel> with the higher DPI of a 4K display, 4:2:0 becomes less of a "problem" anyway
[22:44:25 CEST] <nevcairiel> at least you get FullHD chroma now ;)
[22:46:38 CEST] <JEEB> :)
[22:53:02 CEST] <BBB> jamrial: awh youre not liking my cute 10/12bpp init templating, I thought it was so cute
[22:57:13 CEST] <durandal_1707> going for 300th filter
[22:57:37 CEST] <llogan> what percentage are yours?
[22:58:46 CEST] <durandal_1707> small
[22:59:56 CEST] <durandal_1707> I mean ones I 100% wrote
[23:01:17 CEST] <llogan> what filters do you have planned?
[23:07:54 CEST] <jamrial> BBB: I replied saying i was fine with it if you liked it that much :p
[23:08:00 CEST] <BBB> awh
[23:08:01 CEST] <BBB> ty
[23:16:07 CEST] <durandal_1707> llogan: more audio ones, agate, more sidechain ones, multiband filters, lots of fun stuff
[23:51:30 CEST] <BBB> why do so many people ask us to do their homework nowadays?
[23:51:38 CEST] <BBB> hi, my boss asked me to do this. can you please do it for me?
[23:51:45 CEST] <BBB> (email from roshanta on ML)
[23:53:40 CEST] <durandal_1707> and for free
[23:55:05 CEST] <BBB> and QUICK
[23:55:08 CEST] <BBB> I HAVE A DEADLINE
[00:00:00 CEST] --- Thu Sep 17 2015