log

age author description
Tue, 26 Jan 2010 17:17:48 +0000 lorenm optimize h264_loop_filter_strength_mmx2 libavcodec
Tue, 26 Jan 2010 15:34:21 +0000 michael remove unneeded check libavcodec
Tue, 26 Jan 2010 14:57:53 +0000 michael Use left_mb_xy from fill_caches instead of recalculating it. libavcodec
Tue, 26 Jan 2010 14:55:19 +0000 michael Simplify left_xy content for the loop filter, this also makes it closer to libavcodec
Tue, 26 Jan 2010 13:39:26 +0000 michael Simplify loop filter a little by using top/left_type. libavcodec
Tue, 26 Jan 2010 13:38:18 +0000 michael Set top & left types for deblock in fill_caches(). libavcodec
Tue, 26 Jan 2010 13:30:33 +0000 michael 100l typo fix, mixed up +-1 forcing the loop filter skip to never skip. libavcodec
Tue, 26 Jan 2010 13:28:55 +0000 michael cosmetic libavcodec
Tue, 26 Jan 2010 13:27:22 +0000 michael Fix qp_thres loop filter check for MBAFF. libavcodec
Tue, 26 Jan 2010 10:35:36 +0000 michael Optimize mb neighbor initialization for MBAFF in fill_caches(). libavcodec
Tue, 26 Jan 2010 10:30:24 +0000 cehoyos Metasoft MJPEG codec has Cb and Cr swapped, fixes issue 1611. libavcodec
Mon, 25 Jan 2010 13:26:10 +0000 michael Fix heap overflow due to lack of nb_components check. libavcodec
Mon, 25 Jan 2010 10:46:32 +0000 cehoyos Release buffer when decoding is finished. libavcodec
Mon, 25 Jan 2010 02:44:34 +0000 michael Optimize decode_cabac_field_decoding_flag(). libavcodec
Mon, 25 Jan 2010 01:56:13 +0000 astrange 100l, correct wrong H.264+adaptive MBAFF decoding libavcodec
Mon, 25 Jan 2010 00:30:44 +0000 astrange H.264: Use 64-/128-bit write-combining macros for copies libavcodec
Sun, 24 Jan 2010 23:55:33 +0000 jbr Only check frame size if the header is valid. libavcodec
Sun, 24 Jan 2010 22:47:50 +0000 cehoyos Fix decoding ATI VCR2 codec, sample in libavcodec
Sun, 24 Jan 2010 20:54:09 +0000 michael Branchless setting of MB_TYPE_8x8DCT. libavcodec
Sun, 24 Jan 2010 20:52:49 +0000 michael Remove cruft. libavcodec
Sun, 24 Jan 2010 19:19:09 +0000 michael Mark index as unsigend so gcc knows it doesnt need a sign extension on x86_64. libavcodec
Sun, 24 Jan 2010 18:42:22 +0000 michael a[b-1] -> (a-1)[b]. libavcodec
Sun, 24 Jan 2010 18:23:46 +0000 michael Optimize suffix_length computation, 1 cpu cycle speedup. libavcodec
Sun, 24 Jan 2010 18:18:08 +0000 michael Simplify suffix_length computation, same speed. libavcodec
Sun, 24 Jan 2010 18:17:01 +0000 michael Optimize level_code computation, 6cpu cycles speedup. libavcodec
Sun, 24 Jan 2010 18:07:29 +0000 reimar Fix crash in MLP decoder due to integer overflow. libavcodec
Sun, 24 Jan 2010 18:05:02 +0000 michael 1 cpu cycle faster suffix_length calculation. libavcodec
Sun, 24 Jan 2010 17:27:00 +0000 fenrir VC-1/WMV3 DXVA2 implementation libavcodec
Sun, 24 Jan 2010 17:02:44 +0000 michael Remove is_complex from loop_filter() its useless there in its current form. libavcodec
Sun, 24 Jan 2010 16:40:58 +0000 fenrir Added a few missing consts in dxva_h264. libavcodec
Sun, 24 Jan 2010 16:37:12 +0000 fenrir Added a missing const to ff_h264_get_slice_type(). libavcodec
Sun, 24 Jan 2010 16:22:09 +0000 fenrir Added missing const to get_bits_count(). libavcodec
Sun, 24 Jan 2010 13:43:26 +0000 michael Prefer cbp over cbp_table. libavcodec
Sun, 24 Jan 2010 13:20:17 +0000 michael Remove all uses of slice_type* from the loop filter, also remove its libavcodec
Sat, 23 Jan 2010 22:11:46 +0000 michael Remove unneeded reset of non_zero_count_cache for deblock. libavcodec
Sat, 23 Jan 2010 21:57:36 +0000 michael Remove useless things from the deblock side of fill_caches(). libavcodec
Sat, 23 Jan 2010 21:45:12 +0000 michael make mv_cache init 64bit where possible. libavcodec
Sat, 23 Jan 2010 18:17:57 +0000 fenrir Moved reusable functions from dxva2_h264.c to dxva2.c libavcodec
Sat, 23 Jan 2010 18:05:30 +0000 michael Move +52 from the loop filter to the alpha/beta offsets in the context. libavcodec
Sat, 23 Jan 2010 18:04:09 +0000 reimar Release buffer when the codec is closed. libavcodec
Sat, 23 Jan 2010 18:01:38 +0000 reimar Use qstride = 0 and allocate only a single line for qscale_table since libavcodec
Sat, 23 Jan 2010 18:01:28 +0000 fenrir Added ff_dxva2_ prefix to get_surface(_index) and commit_buffer functions. libavcodec
Sat, 23 Jan 2010 18:00:51 +0000 fenrir Reindent the content of one if(){} in ff_dxva2_common_end_frame. libavcodec
Sat, 23 Jan 2010 17:59:08 +0000 fenrir Moved reusable code from dxva2_h264.c:end_frame to ff_dxva2_common_end_frame. libavcodec
Sat, 23 Jan 2010 15:28:34 +0000 michael Set edges based on cbp and mv partitioning, not just skiped MBs. libavcodec
Sat, 23 Jan 2010 15:22:48 +0000 reimar Use AVPALETTE_COUNT instead of hard-coding 256. libavcodec
Sat, 23 Jan 2010 15:19:34 +0000 reimar Remove redundant initialization of the palette, it is part of the context libavcodec
Sat, 23 Jan 2010 14:50:56 +0000 michael Optimize filter_mb_mbaff_edge*() libavcodec
Sat, 23 Jan 2010 13:54:02 +0000 michael Optmize 8x8dct check used to skip some borders in the loop filter. libavcodec
Sat, 23 Jan 2010 13:49:09 +0000 reimar Use av_freep intead of av_free to free pointers in vp56 context. libavcodec
Sat, 23 Jan 2010 10:25:22 +0000 reimar Zero the frame data on allocation for VB codec, e.g. the FATE sample seems to libavcodec
Sat, 23 Jan 2010 08:39:50 +0000 daniel Export wma_critical_freqs as ff_wma_critical_freqs libavcodec
Fri, 22 Jan 2010 23:40:18 +0000 reimar The SSSE3 version of dct_quantize in mpegvideo_mmx_template.c needs libavcodec
Fri, 22 Jan 2010 23:07:58 +0000 alexc Implement an sse version of scalarproduct_float(). libavcodec
Fri, 22 Jan 2010 19:07:44 +0000 kostya Check for chunk boundaries when decoding VB codec data libavcodec
Fri, 22 Jan 2010 18:28:41 +0000 kostya Zero palette in case not all entries are initialized later libavcodec
Fri, 22 Jan 2010 03:25:11 +0000 mru Move array specifiers outside DECLARE_ALIGNED() invocations libavcodec
Fri, 22 Jan 2010 01:59:17 +0000 michael Gcc idiocy fixes related to filter_mb_edge*. libavcodec
Thu, 21 Jan 2010 21:01:26 +0000 michael Move dquant check into qscale overflow check. libavcodec
Thu, 21 Jan 2010 16:50:31 +0000 michael Make calculation of mask_edge free of branches, faster of course but probably libavcodec