[Ffmpeg-devel-irc] ffmpeg-devel.log.20140917

[burek]

[00:07] <cone-798> ffmpeg.git 03Lou Logan 07master:9c9aec33c0cf: doc/filters: add geq gradient examples
[02:44] <cone-798> ffmpeg.git 03Michael Niedermayer 07master:3c020b695959: avformat/avienc: Dont search for startcodes in inserted empty avi frames
[03:22] <cone-798> ffmpeg.git 03Michael Niedermayer 07release/2.4:cc73b4f574b6: tools/crypto_bench: fix build when AV_READ_TIME is unavailable
[04:52] <cone-798> ffmpeg.git 03Michael Niedermayer 07master:9ffa705edd77: ffplay: try to fix build failure on apple
[09:44] <cone-798> ffmpeg.git 03James Almer 07master:d1c49bcae9b7: avfilter: remove obsolete FF_API_ACONVERT_FILTER cruft
[09:56] <cone-798> ffmpeg.git 03Paul B Mahol 07master:f1a8943ddd15: doc/filters: remove aconvert
[10:25] <abaudoux> Hello
[10:29] <abaudoux> I run into an error when using the hls format
[10:29] <abaudoux> I think it may be a bug
[10:30] <abaudoux> Here is the command:http://pastebin.com/JmUXNhWz
[11:04] <ubitux> abaudoux: please provide the file(s) in the ticket
[11:04] <ubitux> so it can be reproduced
[11:12] <abaudoux> Its quite big, can I upload it somewhere?
[11:15] <ubitux> see http://ffmpeg.org/bugreports.html
[11:15] <ubitux> "Submitting Sample Media"
[11:15] <abaudoux> thx
[11:22] <abaudoux> uploaded on ftp as ticket_3957.ts
[11:39] <abaudoux> Regarding #3957, I am in fact capturing a live udp stream, to generate HLS. But ffmpeg always stops after a few seconds and complains always with this same error.
[11:41] <cone-798> ffmpeg.git 03James Almer 07master:0456d169c469: x86/me_cmp: port mmxext and sse2 sad functions to yasm
[13:12] <ubitux> eerr wait
[13:12] <ubitux> michaelni: you sure about the me_cmp stuff
[13:12] <ubitux> im sorry i hadn't had a look
[13:13] <ubitux> but typically the sad8 is not height variable
[13:13] <ubitux> why it was originally
[13:13] <ubitux> the h arg is ignored
[13:15] <ubitux> (that's one of the reason it was reimplemented differently in pixelutils)
[13:23] <michaelni> ubitux, i missed that the 8x8 functions lost h :(
[13:23] <ubitux> he probably forked it from the pixelutils
[13:23] <ubitux> where the width are fixed
[13:26] <ubitux> the number of xmm reg looks off in sad16_y2
[13:26] <ubitux> well i should probably review it a bit
[13:29] <ubitux> mmh actually no
[13:29] <ubitux> it's in the mmx scope so shouldn't matter
[14:10] <cone-798> ffmpeg.git 03Michael Niedermayer 07master:85f2c0124d89: avcodec/x86/me_cmp: fix sad8xh
[14:52] <ubitux> oh there is also a sad for alpha?
[15:05] <Daemon404> ubitux, oh, you mean the arch?
[15:05] <ubitux> yes
[15:05] <Daemon404> i was confused for a sec thinking "why is alpha channel SAD special"
[16:05] <J_Darnley> Which moon is out of alignment today?
[16:05] <J_Darnley> (Why is ffmpeg just recieveing black frames?  What did I change to cause this?)
[16:33] <someone-noone> How can I use ffmpeg to capture UTF8 window title with gdigrab device?
[17:24] <J_Darnley> Ha!  What an idiot I am.  I was trying to draw lines of zero width
[18:02] Action: kepstin-laptop waits for the fate tests to finish running on his 266mhz k6.
[18:28] <J_Darnley> how long has it been running?
[18:59] <kepstin-laptop> J_Darnley: started it about 4 hours ago
[20:08] <cone-913> ffmpeg.git 03Diego Biurrun 07master:2dd09ebf5597: cmdutils: Print a more sensible message in show_filters() w/o libavfilter
[20:08] <cone-913> ffmpeg.git 03Michael Niedermayer 07master:9ac08d93b2aa: Merge commit '2dd09ebf5597fd5e4a573bee9cccf21ae821f286'
[20:21] <cone-913> ffmpeg.git 03James Almer 07master:fffc9a077d39: avfilter: remove obsolete FF_API_FILL_FRAME cruft
[20:26] <llogan> whay was aconvert removed from doc/filters? f1a8943
[20:31] <jamrial> because the filter was also removed
[20:31] <jamrial> check d1c49bca
[20:37] <llogan> jamrial: i see. i missed that.
[20:38] <cone-913> ffmpeg.git 03Katerina Barone-Adesi 07master:44caf99ecae7: fate: Add tests for RealAudio 1.0 (14.4) and RealAudio 2.0 (with 28.8)
[20:38] <cone-913> ffmpeg.git 03Michael Niedermayer 07master:a4cefc25ed83: Merge commit '44caf99ecae7bc1b907fab849ecaa72dd340ba2e'
[20:38] <cone-913> ffmpeg.git 03Michael Niedermayer 07master:43633c5388d3: fate: disable fate-ra4-288, the code uses floats and does not produce the same results on different platforms
[20:49] <cone-913> ffmpeg.git 03Katerina Barone-Adesi 07master:c5560e72d0bb: apetag: Fix APE tag size check
[20:49] <cone-913> ffmpeg.git 03Michael Niedermayer 07master:8c71de762b84: Merge commit 'c5560e72d0bb69f8a1ac9536570398f84388f396'
[21:20] <ubitux> wtf is the guy doing with this libffmpeg.a :D
[21:34] <ubitux> < ChocolateArmpits> Ok, made a few tests myself, the speed increase is by about 4%-5%. [...]
[21:34] <ubitux> < koda> ok, hardly noticeable :)
[21:34] <ubitux> ???
[21:45] <iive> ubitux: yeh, that is a lot for overall speed increase.
[22:16] <ubitux> http://www.hydrogenaud.io/forums/index.php?showtopic=106911
[22:25] <iive> ubitux: these test seems way too perfectly grouped to be real...
[22:26] <ubitux> i like the confidence interval of faac
[22:26] <ubitux> :D
[22:29] <J_Darnley> Good.  More evidence that lame -b 320 is idiotic.
[22:30] <microchip_> J_Darnley: pfff! they want lossless now because they clearly hear the difference ;)
[22:31] Action: J_Darnley uses flac
[22:34] Action: ubitux as well
[22:34] Action: microchip_ pities...
[22:40] <J_Darnley> I'm pleased by the conclusions further down the page.
[22:41] <J_Darnley> Some of which can be summed up with "stop fucking with the settings and it will give a good result"
[23:30] Action: kepstin-laptop continues to wait for his k6 box to finish running fate.
[23:58] <cone-913> ffmpeg.git 03Michael Niedermayer 07master:41d82b85ab0e: avcodec/x86/vp9lpf: Always include x86util.asm
[00:00] --- Thu Sep 18 2014