| Mon, 18 Jan 2010 21:22:09 +0000 |
michael |
Rather call filter_mb_mbaff_edge*v() more often than do extra calculations
libavcodec
|
| Mon, 18 Jan 2010 20:13:53 +0000 |
michael |
Use h->slice_num where possible.
libavcodec
|
| Mon, 18 Jan 2010 19:45:56 +0000 |
michael |
Enable filter_mb_fast for CAVLC P slices.
libavcodec
|
| Mon, 18 Jan 2010 16:29:16 +0000 |
michael |
PAFF CABAC P slices seem to work as well, so enable them for ff_h264_filter_mb_fast() too.
libavcodec
|
| Mon, 18 Jan 2010 16:16:22 +0000 |
michael |
Reenable filter_mb_fast for I slices and progressive CABAC P slices.
libavcodec
|
| Mon, 18 Jan 2010 13:09:53 +0000 |
michael |
Move CAVLC 8x8 DCT special case from ff_h264_filter_mb() to fill_caches
libavcodec
|
| Mon, 18 Jan 2010 05:15:31 +0000 |
michael |
Perform reference remapping at fill_cache() time instead of in the
libavcodec
|
| Mon, 18 Jan 2010 00:20:44 +0000 |
michael |
Move the qp check to skip the loop filter up.
libavcodec
|
| Sun, 17 Jan 2010 23:44:23 +0000 |
michael |
Reorganize how values are stored in h->non_zero_count.
libavcodec
|
| Sun, 17 Jan 2010 20:35:55 +0000 |
michael |
Rearchitecturing the stiched up goose part 1
libavcodec
|
| Sat, 16 Jan 2010 17:41:40 +0000 |
michael |
Comment for() ; out
libavcodec
|
| Sat, 16 Jan 2010 17:27:17 +0000 |
michael |
Mark a few functions as noinline, this makes ff_h264_filter_mb() a bit smaller
libavcodec
|
| Sat, 16 Jan 2010 16:21:12 +0000 |
michael |
Apply last 2 optimizations to similar code i forgot.
libavcodec
|
| Sat, 16 Jan 2010 16:14:32 +0000 |
michael |
Another microopt, 4 cpu cycles for avoidance of FFABS().
libavcodec
|
| Sat, 16 Jan 2010 15:19:05 +0000 |
michael |
Minor (2 cpu cycles) optimization ||->|.
libavcodec
|
| Sat, 16 Jan 2010 11:55:35 +0000 |
michael |
Avoid wasting 4 cpu cycles per MB in redundantly calculating qp_thresh.
libavcodec
|
| Tue, 12 Jan 2010 06:01:55 +0000 |
michael |
Split h264 loop filter off h264.c.
libavcodec
|
