[Ffmpeg-devel-irc] ffmpeg-devel.log.20160812

[burek]

[01:14:02 CEST] <cone-286> ffmpeg 03Moritz Barsnick 07master:e8b355a0288c: libavformat/http: add support for headers option in listen mode
[11:08:26 CEST] <ubitux> > Anniversary edition
[11:08:28 CEST] <ubitux> wat
[11:16:38 CEST] <Loriker> yes, I thought the same... :)
[11:37:33 CEST] <jsebechlebsky> Hello, is Michael N. here?
[11:38:10 CEST] <durandal_1707> Yes, his nick is michaelni 
[11:38:34 CEST] <michaelni> jsebechlebsky, yes
[11:41:35 CEST] <jsebechlebsky> michaelni: I wanted to ask you about the failed fate test for my fifo muxer patch. Do you think it was caused by that uint8_t used as boolean option (which caused the following field was modified if the structure was packed), or there is some another issue and I should set up mips toolchain to find out?
[11:41:59 CEST] <michaelni> jsebechlebsky, i think it was the uint8
[11:43:01 CEST] <michaelni> the test actually worked but 3 lines where missing in the output the printing code was under a check of that variable IIRC
[11:43:14 CEST] <jsebechlebsky> ok, so I'll just fix that and resend the patch. Thanks!
[13:02:55 CEST] <thardin> unicode supports colors now?
[16:19:53 CEST] <durandal_1707> ubitux: are signalstats changes ok?
[16:24:03 CEST] <ubitux> i don't like much the large amount of duplication instead of using (big) macro, but well... 
[16:24:12 CEST] <ubitux> don't you redefine SET_META macro btw?
[16:25:04 CEST] <ubitux> pretty sure compute_bit_depth() can be simplified
[16:26:10 CEST] <ubitux> maybe using av_popcount?
[16:27:03 CEST] <ubitux> btw, the doc says the range is [0-16]
[16:27:18 CEST] <ubitux> is it really? not [0-15]?
[16:27:34 CEST] <ubitux> or maybe 1-16 dunno
[16:29:39 CEST] <ubitux> i feel like you would want to use av_popcount(x)+1 and assume [1-16] as range
[16:39:15 CEST] <cone-094> ffmpeg 03Aman Gupta 07master:b5f30fe8fea5: avcodec/utils: only warn when passed invalid lowres value
[16:47:00 CEST] <durandal_1707> ubitux: what's bitdepth of image full of 0?
[16:47:33 CEST] <ubitux> how can you code any information (even full of zero) on a 0 bitdepth?
[16:48:42 CEST] <durandal_1707> it must have at least one bit change, no?
[16:49:41 CEST] <ubitux> av_popcount(x)+1 is indeed incorrect, you might want av_popcount(x|1)
[16:53:52 CEST] <ubitux> or maybe you just want av_popcount(x) with [0-16], but i feel like a picture full of 0 needs is a picture with as much information as one full of 1s
[16:58:14 CEST] <ubitux> that second patch is useful though
[16:58:45 CEST] <ubitux> i wonder how it detect upscaled 8-bit footages 
[16:59:08 CEST] <ubitux> i guess the dithering causes an actual large bitdepth
[16:59:53 CEST] <nevcairiel> upscaling and dithering would definiteily increase the bitdepth
[21:46:21 CEST] <CFS-MP3b> Is any mentor for GSoC here by any chance? If yes... I wonder if the org has already decided who's going to the GSoC submit in october?
[21:59:33 CEST] <Compn> CFS-MP3b : the gsoc summit ? 
[21:59:50 CEST] <Compn> durandal_1707 was a mentor, dunno if hes working on it now
[21:59:52 CEST] <Compn> michaelni is a mentor
[22:06:31 CEST] <michaelni> CFS-MP3b, i think thilo and rostislav ( atomnuker ) or carl will probably go to the summit, why do you ask ?
[23:18:45 CEST] <llogan> michaelni, Compn: I'll be gone Aug 15-22. No internet or mobile/cell connection, so i won't be able to clear the riff-raff from the ML queue then.
[00:00:00 CEST] --- Sat Aug 13 2016