[Ffmpeg-devel-irc] ffmpeg-devel.log.20130426

burek burek021 at gmail.com
Sat Apr 27 02:05:02 CEST 2013


[00:21] <ubitux> is it relevant for a windowing function to start at index 0.5 (instead of 0) with inc=1 and using a win size of N (instead of N-1) for the normalize?
[00:22] <ubitux> basically:
[00:22] <ubitux>          for (i = 0; i < win_size; i++)
[00:22] <ubitux> -            showspectrum->window_func_lut[i] = .5f * (1 - cos(2*M_PI*i / (win_size-1)));
[00:22] <ubitux> +            showspectrum->window_func_lut[i] = .5f * (1 - cos(2*M_PI*(i+.5) / (win_size)));
[03:19] <cone-660> ffmpeg.git 03highgod0401 07master:f31247cf47dc: lavu/opencl:add opencl set param function
[03:19] <cone-660> ffmpeg.git 03highgod0401 07master:cd134963e8fd: lavfi/unsharp: add opencl unsharp filter
[03:51] <cone-660> ffmpeg.git 03Michael Niedermayer 07master:05f4c0506198: vc1dec: remove interlaced warning
[11:02] <cone-198> ffmpeg.git 03Luca Barbato 07master:c2cb01d418dd: lavf: introduce AVFMT_TS_NEGATIVE
[11:02] <cone-198> ffmpeg.git 03Michael Niedermayer 07master:c4bf74022562: Merge commit 'c2cb01d418dd18e1cf997c038d37378d773121be'
[11:16] <cone-198> ffmpeg.git 03Martin Storsjö 07master:74685f6783e7: x86: Rename dsputil_rnd_template.c to rnd_template.c
[11:16] <cone-198> ffmpeg.git 03Michael Niedermayer 07master:721ffc691ab4: Merge remote-tracking branch 'qatar/master'
[11:20] <ubitux> oh shit a user using --enable-random
[11:26] <ubitux> durandal_1707: afade doc are somehow broken
[11:26] <ubitux> afade example is using "ss" instead of "st"
[11:26] <ubitux> also, AVOption description are mentioning "expression" while they aren't
[11:27] <ubitux> durandal_1707: and btw, why INT32 and not INT64?
[11:27] <ubitux> (for durations)
[11:40] <durandal_1707> ubitux: and what descriptions should be instead?
[11:42] <ubitux> "expression" is just used for AVExpr, use "set starting sample for fading", "set duration of the fade" etc.
[11:56] <durandal_1707> anyone get this: gmake: *** No rule to make target ibavcodec/x86/dsputil_mmx.o'.  Stop. ?
[11:56] <nevcairiel> you are missing a l there
[11:57] <durandal_1707> i selected all line, but xterm copy eat most of it
[11:57] <durandal_1707> gmake: *** No rule to make target `libavcodec/x86/dsputil_rnd_template.c', needed by `libavcodec/x86/dsputil_mmx.o'.  Stop.
[11:57] <nevcairiel> did you do a make clean?
[11:58] <ubitux> cheatcode: rm libavcodec/x86/dsputil_mmx.d
[11:58] <durandal_1707> too late
[11:58] <ubitux> (or even sed to nicely remove the exact dep)
[12:30] <cone-198> ffmpeg.git 03Paul B Mahol 07master:1a530f9ed897: lavfi: support timeline for colorbalance and colorchannelmixer
[12:30] <cone-198> ffmpeg.git 03Paul B Mahol 07master:09250e344591: lavfi/curves: add support for more rgb formats
[12:30] <cone-198> ffmpeg.git 03Paul B Mahol 07master:e0a636fadff0: doc/filters: fix afade 2nd example
[12:30] <cone-198> ffmpeg.git 03Paul B Mahol 07master:42d61fd9680c: lavfi/afade: fix AVOptions descriptions
[12:30] <cone-198> ffmpeg.git 03Paul B Mahol 07master:6a8d05061f9f: lavfi/stereo3d: avoid division when changing aspect ratio
[12:35] <ubitux> thx
[13:19] <durandal_1707> michaelni: the lfe_fir error that address sanitizer is complaining is correct, but table used by dcaenc and one in dcadata.h are different eg. positions of values differes, but values itself are same
[13:35] <durandal_1707> and it looks like lfe is not tested at all in dca decoder
[13:37] <durandal_1707> lol, there is FIXME
[13:43] <cone-198> ffmpeg.git 03Paul B Mahol 07master:30579b7ef013: dcaenc: silence address sanitizer warning
[13:49] <durandal_1707> libdca use same lfe_fir_64 table as dcaenc
[13:53] <durandal_1707> libdca have same FIXME there...
[13:53] <durandal_1707> so why we use different lfe_fir_64 table?
[14:03] <durandal_1707> hmm, it was changed for simd, so looks like i need to find out how to undo reorder....
[14:22] <ubitux> wth is the unification guy talking about
[14:22] <durandal_1707> ?
[14:22] <ubitux> https://lists.ffmpeg.org/pipermail/ffmpeg-devel/2013-April/142942.html
[14:25] <durandal_1707> someone inventing another api for multimedia
[14:25] <ubitux> "It's not like Unix, every function every dime has a specification"
[14:25] <ubitux> parse error
[14:28] <durandal_1707> is our audio example still buggy?
[14:58] <cone-198> ffmpeg.git 03Michael Niedermayer 07master:060f14f49310: avformat/mux: change avoid_negative_ts code to support later addition of AVStreams
[14:58] <cone-198> ffmpeg.git 03Michael Niedermayer 07master:c329195a2712: avformat/mux: use a local variable for offset
[14:58] <cone-198> ffmpeg.git 03Michael Niedermayer 07master:647f306b461a: avformat/mux: assert that timestamps are positive after offsetting them
[14:58] <cone-198> ffmpeg.git 03Michael Niedermayer 07master:e5891a23372d: avformat/movenc: mov/mp4 should support negative timestamps through edit lists
[14:58] <cone-198> ffmpeg.git 03Michael Niedermayer 07master:eba33396f007: avformat/mux: set avoid_negative_ts based on flags if not overridden by user or muxer
[15:19] <ubitux> durandal_1707: i took the liberty to notice luca about the 256 colors support
[15:19] <ubitux> but well, i'll get ignored again i guess
[15:20] <durandal_1707> what 256 color support?
[15:20] <durandal_1707> ubitux: you better spam them with bug fixes
[15:20] <ubitux> i do already, but there are too much
[15:20] <ubitux> durandal_1707: 1842b2756 ’ http://lists.libav.org/pipermail/libav-devel/2013-April/046009.html
[15:21] <durandal_1707> that is most useless feature i ever did
[15:22] <ubitux> :D
[15:23] <durandal_1707> you can have true color in kde konsole
[15:52] <cone-198> ffmpeg.git 03Michael Niedermayer 07master:ba323d67fa55: avformat/write_packet: drop disabled code
[15:52] <cone-198> ffmpeg.git 03Michael Niedermayer 07master:ee4c01bef65d: vc1dec: factor lut clean code
[15:52] <cone-198> ffmpeg.git 03Michael Niedermayer 07master:d78329700df2: vc1dec: make sure next_use_ic is set correctly
[17:33] <durandal_1707> michaelni: why at encode_init logs do not display name of encoder?
[17:34] <durandal_1707> happens here when encoding float to flac
[20:13] <ubitux> hey beastd 
[20:13] <beastd> hi
[20:13] <ubitux> when are you going to write us some good news again?
[20:16] <beastd> yeah right. need to start cooking up a news article
[20:16] <ubitux> :)
[20:18] <beastd> i have been kind of sick for some weeks now. not feeling to well yet  :(
[20:18] <ubitux> :(
[20:48] <cone-198> ffmpeg.git 03Michael Niedermayer 07master:74b1c32a8027: avutil/log: leave prefix printing state on empty input.
[21:40] <cone-198> ffmpeg.git 03Clément BSsch 07master:039c2b7828da: lavfi/hqdn3d: remove timeline flag.
[23:49] <cone-308> ffmpeg.git 03Michael Niedermayer 07master:a057552212fe: avutil/log: dont count&print repeats for empty strings
[23:53] <funman> buddhist proverb: if an empty string is displayed, can it be seen?
[23:56] <ubitux> you have to trust your ears
[00:00] --- Sat Apr 27 2013


More information about the Ffmpeg-devel-irc mailing list