[Ffmpeg-devel-irc] ffmpeg-devel.log.20170501

burek burek021 at gmail.com
Tue May 2 03:05:03 EEST 2017

[01:06:10 CEST] <cone-186> ffmpeg 03James Almer 07master:095147ae0650: avformat/matroskadec: export Content Light Level metadata
[01:06:10 CEST] <cone-186> ffmpeg 03James Almer 07master:37cc1c1e9137: avformat/matroskaenc: add support for writing Content Light Level elements
[04:18:56 CEST] <cone-350> ffmpeg 03James Almer 07master:54a4c9b4e9a1: avcodec/options: factorize avcodec_copy_context() cleanup code
[04:18:56 CEST] <cone-350> ffmpeg 03James Almer 07master:cac8de2da5c4: avcodec/options: do a more thorough clean up in avcodec_copy_context()
[11:44:54 CEST] <cone-274> ffmpeg 03Martin Vignali 07master:37f4d22075c3: libavcodec/exr : fix piz uncompress on big endian
[11:44:54 CEST] <cone-274> ffmpeg 03Martin Vignali 07master:5ad18f279af1: fate/exr : add tests for piz uncompress
[11:44:54 CEST] <cone-274> ffmpeg 03Martin Vignali 07master:89812e423de6: fate/exr : add test for negative float value
[14:48:02 CEST] <durandal_1707> michaelni: thoughts on bigger fft sizes in our implementation?
[14:49:48 CEST] <michaelni> we should support te sizes we need
[14:50:10 CEST] <atomnuker> you need bigger than 1048576?
[14:50:36 CEST] <atomnuker> that's 21 seconds worth of samples at 48000
[15:18:22 CEST] <durandal_1707> atomnuker: max is 131072 and i need bigger because nb taps are fft size / 4
[15:19:45 CEST] <durandal_1707> so supporting up to 7 seconds is enough for common impulse responses
[17:43:13 CEST] <durandal_1707> michaelni, atomnuker: could i use already available fft sizes for huge number of taps?
[18:13:08 CEST] <durandal_1707> apparently this is called partitioned fft
[18:13:46 CEST] <cone-881> ffmpeg 03Micah Galizia 07master:28b24670741e: libavformat/http: Ignore expired cookies
[18:13:46 CEST] <cone-881> ffmpeg 03Michael Niedermayer 07master:63b8d4146d78: avcodec/bmp: Use ff_set_dimensions()
[18:13:46 CEST] <cone-881> ffmpeg 03Michael Niedermayer 07master:b706ddbae3f4: doc/developer: Add terse documentation of assumed C implementation defined behavior
[20:03:02 CEST] <cone-881> ffmpeg 03Michael Niedermayer 07master:2f00300b779e: avcodec/vp3: Check remaining bits in unpack_dct_coeffs()
[20:03:03 CEST] <cone-881> ffmpeg 03Michael Niedermayer 07master:b29feec9829c: avcodec/indeo2: Check remaining bits in ir2_decode_plane()
[23:00:07 CEST] <nevcairiel> jamrial: i'll look over the hevc patches tomorrow
[23:00:26 CEST] <durandal_1707> michaelni: know how partitioned convolution works?
[23:04:49 CEST] <michaelni> id guess thats FFT based convolution with blocks smaller than the whole signal
[23:13:07 CEST] <durandal_1707> yes
[23:45:44 CEST] <cone-651> ffmpeg 03Carl Eugen Hoyos 07master:a88b0b0ba7b4: lavc/mips/hevc_idct_msa: Add missing const qualifier.
[23:46:59 CEST] <cone-651> ffmpeg 03Carl Eugen Hoyos 07master:f4c133c70874: lavc/mips/iirfilter_mips: Include config.h.
[00:00:00 CEST] --- Tue May  2 2017

More information about the Ffmpeg-devel-irc mailing list