Yaowu Xu 737179f275 Reduced bias in picking loop filter level ...

The bias in picklpf intended to bias toward less greedy in getting
best frame level psnr while maximize overall quality for a clip.
This commit reduced the bias for frames using 8x8 transform to
achieve better compression overall.

The change improve compression by ~.15% consistently on most of the
HD clips tested.

http://www.corp.google.com/~yaowu/no_crawl/LPFBias_FixSkip.html

Change-Id: Ic30932d2b8eaebd52339b0195f569edc48eed7bc

2012-02-17 16:44:08 -08:00

..

arm

Move neon only arm functions under arm/neon.

2011-09-20 10:51:06 -07:00

generic

Simplification of experimental code base.

2012-02-14 09:03:56 +00:00

ppc

remove one set of 16x16 variance funcations

2011-06-09 11:23:05 -07:00

x86

Simplification of experimental code base.

2012-02-14 09:03:56 +00:00

asm_enc_offsets.c

use GCC inline magic

2011-06-08 16:19:37 -04:00

bitstream.c

Reworked context conversion between 8x8 and 4x4

2012-02-16 15:00:10 -08:00

bitstream.h

Add runtime CPU detection support for ARM.

2010-10-25 09:23:29 -04:00

block.h

Improved coding using 8x8 transform

2012-02-10 14:23:27 +00:00

boolhuff.c

Moved vp8_encode_bool into boolhuff.h

2011-07-19 09:17:25 -04:00

boolhuff.h

Moved vp8_encode_bool into boolhuff.h

2011-07-19 09:17:25 -04:00

dct.c

optmized rounding for transforms

2012-02-16 07:03:57 -08:00

dct.h

Add 8x8 transform to experimental branch

2011-07-20 09:49:22 -07:00

defaultcoefcounts.h

Merge remote branch 'internal/upstream' into HEAD

2011-08-25 08:36:19 -04:00

encodeframe.c

Fix overflows in dual prediction mode selection.

2012-02-15 15:57:49 -08:00

encodeintra.c

Improved coding using 8x8 transform

2012-02-10 14:23:27 +00:00

encodeintra.h

Fix more warnings.

2011-12-22 16:33:06 -08:00

encodemb.c

Reworked context conversion between 8x8 and 4x4

2012-02-16 15:00:10 -08:00

encodemb.h

a bit code clean-up

2012-02-16 15:15:06 -08:00

encodemv.c

Use full-pixel MV in mvsadcost calculation

2011-04-01 16:41:58 -04:00

encodemv.h

Use full-pixel MV in mvsadcost calculation

2011-04-01 16:41:58 -04:00

find_rotation.c

vpn common -> implicit segmentation

2012-01-24 11:20:13 -08:00

firstpass.c

Rate control on static scenes plus Y2dc delta Q fix.

2012-01-17 17:42:46 +00:00

firstpass.h

Move firstpass motion map to stats packet

2010-10-21 14:04:20 -04:00

lookahead.c

Simplification of experimental code base.

2012-02-14 09:03:56 +00:00

lookahead.h

Copy less when active map is in use

2011-08-19 10:29:41 -04:00

mbgraph.c

Reference frame prediction:

2012-02-03 12:44:45 +00:00

mbgraph.h

Further segment feature extensions.

2011-10-24 15:52:18 +01:00

mcomp.c

made vp8_mode_context adaptive

2011-12-07 11:01:59 -08:00

mcomp.h

Revised coding using adaptive mode context to depend on frame type

2011-12-09 12:13:42 -08:00

modecosts.c

Enabled adaptive UV intra coding for inter frames

2011-12-09 14:44:13 -08:00

modecosts.h

Use WebM in copyright notice for consistency

2010-09-09 10:01:21 -04:00

onyx_if.c

moved scaling from dequantization to inverse transform for T8x8

2012-02-16 07:03:55 -08:00

onyx_int.h

Fix overflows in dual prediction mode selection.

2012-02-15 15:57:49 -08:00

pickinter.c

Merge dualpred (compound prediction) experiment.

2012-02-09 16:29:18 -08:00

pickinter.h

Move intra block mode selection to pickinter.c

2011-06-08 11:44:57 -04:00

picklpf.c

Reduced bias in picking loop filter level

2012-02-17 16:44:08 -08:00

psnr.c

Merge remote branch 'internal/upstream' into HEAD

2011-03-23 00:27:52 -04:00

psnr.h

Increase static linkage, remove unused functions

2011-03-17 20:53:47 -04:00

quantize.c

moved scaling from dequantization to inverse transform for T8x8

2012-02-16 07:03:55 -08:00

quantize.h

fix more merge issues

2011-07-21 09:01:53 -07:00

ratectrl.c

Improved coding using 8x8 transform

2012-02-10 14:23:27 +00:00

ratectrl.h

Removal of Qindex LUTS.

2011-11-22 08:42:33 +00:00

rdopt.c

Reworked context conversion between 8x8 and 4x4

2012-02-16 15:00:10 -08:00

rdopt.h

Fix more warnings.

2011-12-22 16:33:06 -08:00

sad_c.c

Preload reference area to an intermediate buffer in sub-pixel motion search

2011-07-22 09:28:06 -04:00

satd_c.c

Further segment feature extensions.

2011-10-24 15:52:18 +01:00

segmentation.c

Removal of SEGFEATURES placeholder comments

2012-02-09 17:25:05 +00:00

segmentation.h

Merge CONFIGURE_SEGMENTATION experiment.

2011-11-15 16:15:23 +00:00

ssim.c

Reclassify optimized ssim calculations as SSE2.

2011-08-22 12:36:28 -07:00

temporal_filter.c

Removal of threading code.

2012-02-10 16:23:59 +00:00

temporal_filter.h

x86 sse2 temporal_filter_apply

2011-01-06 14:00:30 -05:00

tokenize.c

Reworked context conversion between 8x8 and 4x4

2012-02-16 15:00:10 -08:00

tokenize.h

fix more merge issues

2011-07-21 09:01:53 -07:00

treewriter.c

Use WebM in copyright notice for consistency

2010-09-09 10:01:21 -04:00

treewriter.h

Fix relative include paths

2011-02-10 15:09:44 -05:00

variance_c.c

remove redundant functions

2011-06-06 16:44:05 -07:00

variance.h

Further segment feature extensions.

2011-10-24 15:52:18 +01:00