Commit Graph

9345 Commits

Author SHA1 Message Date
Johann
e10c95dc83 Update vp9_fdct8x8_quant_ssse3 for highbitdepth
Borrow transition functions from fdct.h nee vpx_quantize_b_sse2

BUG=webm:1304

Change-Id: I9c88c3eec3ff8bb461411d98c26c3c236ea28ef1
2016-11-05 01:23:07 +00:00
Marco Paniconi
cca774c7df Merge "vp9: Non-rd pickmode: fix logic in reference masking." 2016-11-03 23:12:05 +00:00
Marco
86b0042f44 vp9-svc: Add decoder control to decode up to x spatial layers.
Change-Id: I85536473b8722424785c84c5b5520960b4e5744a
2016-11-03 11:18:00 -07:00
Marco
da9f762e24 vp9: Non-rd pickmode: fix logic in reference masking.
Add condition that usable_ref_frame > LAST.
This is to avoid potentially skipping all last-nonzero mv modes,
if golden is used as a reference but skipped completely for the
current block.

This has no effect currenty, as we always consider testing golden
mode for each block.

Change-Id: I3182cf44664081935a90ed43aa7b32e710e60e22
2016-11-03 10:32:57 -07:00
Debargha Mukherjee
f93305aa07 Merge "Speed-up recode loop for extreme bitrate diffs" 2016-11-03 17:04:17 +00:00
Paul Wilkins
295cd3b493 Merge "Fixed bug in formatting of debug stats." 2016-11-02 17:10:07 +00:00
paulwilkins
de76d2e315 Fixed bug in formatting of debug stats.
Fixed formatting bug introduced by the fix to BUG=webm:1322
( Iedc4477aef1746aa0a4f84d88a1156296fd3ba87)

Change-Id: I715ee446c0e8584967ab87ba4e355759dd394187
2016-11-02 09:38:18 +00:00
James Zern
1961a92a94 vp9,tile_worker_hook: correctly set jmp target
vp9_init_macroblockd() resets the error_info to cm's global copy; this
needs to be set to the thread-level target to avoid jumping to the
incorrect stack, resulting in hang or crash.
broken since:
1f4a6c8 vp9/tile_worker_hook: add multiple tile decoding
includes v1.5.0, v1.6.0

BUG=629481

Change-Id: Icbf1696b25ba8c479e845fbf227b3c3ca73542f5
2016-11-01 18:45:50 -07:00
Paul Wilkins
84dcfced5b Merge "Change to KF boost calculation." 2016-11-01 09:29:30 +00:00
Paul Wilkins
715c65914b Change to KF boost calculation.
This  change is a step in a larger change to the way boost and interval are
determined for ARF and Key frames.

This patch contains some pluming for the general case but focuses on the
key frame boost calculation. This now relies more heavily on the rate at
which the error score increases between the primary and secondary reference
frame. This seems to be less fragile when dealing with different frame sizes.
For example larger image formats tend in the first pass to see a higher
% of intra coded blocks and the use of this number in calculating the frame
decay factor was leading to much lower boost numbers for 4K, for example,
than the same clip coded at 2K.

This change does give overall gains but they are MUCH larger for the 4K Netflix
set. For the 4K Netflix set the average gain is around 3% with some clips > 20%
whereas for the same set at 2K the average gain is 0.5-1%.

In general for small image formats the boost is most often reduced a little whereas
4K clips the boost is increased. There are some -ve cases such as Akiyo at 352x288
where the reduced boost hurts the metrics, especially for SSIM, even while
the set as a whole improves. This is most notable at very low Q and may be the
subject of a future patch.

Some common code for KF and ARF was separated in this patch for the purposes of
tuning but may later be re-merged if appropriate.

Change-Id: Iaa15ac5a58d2be89181100d95cef6a8dc4b12d0d
2016-10-28 15:35:59 +01:00
Debargha Mukherjee
4f7a59c802 Merge "Force recode if framesize exceeds max allowed size" 2016-10-28 04:21:44 +00:00
Debargha Mukherjee
1cd987d922 Speed-up recode loop for extreme bitrate diffs
Adjusts the q adjustement step depending on how far the
projected and target rates differ.

Change-Id: I498d03523ca233a270512ca3972c372daa4ca2a8
2016-10-27 11:08:44 -07:00
Debargha Mukherjee
54e03017b6 Force recode if framesize exceeds max allowed size
Fixes a case where recode is not triggered based on the value
of maxq passed into the recode loop test function.

BUG=b/32375284

Change-Id: I15ad985d0525c68e0443cfaf842440d2754b2266
2016-10-27 09:52:51 -07:00
Paul Wilkins
de859676dd Changes to KF boost calculation.
Remove double counting of decay. Limit maximum KF boost.

Change-Id: I0fb2344d0f78b5e95bb899dfad12b0ca84034b2c
2016-10-26 17:53:29 +01:00
paulwilkins
ccd6a8e2fa Removal of a couple of two pass adjustments.
Removed a couple of adjustments that no longer move the needle
much but complicate the process of tuning.

Change-Id: Ie320f5cf155e6aac14a4757ea9ada2cd59f27590
2016-10-26 17:52:37 +01:00
Yunqing Wang
c192def8f3 Change 2 motion search counts to be tile data
This patch modified the motion search counts used in:
https://chromium-review.googlesource.com/#/c/305640/

These 2 counts were originally added as thread data, and used to
make decisions in motion search. The tile encoding order can be
inconsistent while using different number of threads, which can
cause bitstream mismatch. Here moved them to tile data to solve
the issue.

BUG=webm:1322

Change-Id: Iedc4477aef1746aa0a4f84d88a1156296fd3ba87
2016-10-25 10:12:41 -07:00
Vignesh Venkatasubramanian
9a032fa262 Merge "vp9_bitstream: Encode tiles in parallel" 2016-10-22 02:23:06 +00:00
Vignesh Venkatasubramanian
5deffa1175 vp9_bitstream: Encode tiles in parallel
Re-use the tile worker threads to pack the bitstream in parallel
on a per-tile basis.  Restricting this to real-time only for now
(further testing is needed to ensure this does not make 2-pass
worse in any case).

BUG=webm:1309

Change-Id: I8a80da7c5089b837d0df79a5c49d5e3022dfc8ec
2016-10-21 17:35:03 -07:00
Marco
ee1b3f34c0 vp9: Nonrd variance partition: increase threshold for using 4x4 avg.
In variance partition low resolutions may use varianace based on
4x4 average for better partitioning.
Increase the threshold for doing this at speed = 8.

Improves speed by ~5%, with little loss, < 1%, on RTC_derf set.

Change-Id: Ib5ec420832ccff887a06cb5e1d2c73199b093941
2016-10-21 11:51:06 -07:00
Marco
a7d116aa67 vp9: Speed=8 real-time: Keep the bias_golden feature on.
Small/no change in metrics on RTC set, speed increase by 2-3%.

Change-Id: Iee997bd7433e8e508216e9267b1c31c5a9aa5121
2016-10-20 17:03:51 -07:00
James Zern
7f31bfeddb Revert "vp9_bitstream: Encode tiles in parallel"
This reverts commit 9e8efa5b18.

this change causes ubsan warnings, failures in
vpxenc_vp9_webm_rt_multithread_tiled

BUG=webm:1309

Change-Id: I020c7be985c771bfff4b3de1afe51cc8edb980da
2016-10-18 22:47:48 -07:00
Marco Paniconi
f6980ca68e Merge "vp9: Non-rd variance partition: add condition for 64x64 split." 2016-10-18 00:03:17 +00:00
Marco
55a2b67368 vp9: Non-rd variance partition: add condition for 64x64 split.
Add stronger condition for splitting 64x64, for low noise content.
This reduces dragging artifact near moving head.

Little/no change in metrics on RTC set.

Change-Id: I39b38cfd20f2ece53ff49c2aaf76ba9f82761be1
2016-10-17 12:54:27 -07:00
Vignesh Venkatasubramanian
9e8efa5b18 vp9_bitstream: Encode tiles in parallel
Re-use the tile worker threads to pack the bitstream in parallel
on a per-tile basis.  Restricting this to real-time only for now
(further testing is needed to ensure this does not make 2-pass
worse in any case).

BUG=webm:1309

Change-Id: Ia2c982da56697756e12f02643f589189b3271d98
2016-10-17 10:42:03 -07:00
Vignesh Venkatasubramanian
769292017b vp9_bitstream: Parameterize interp_filter_selected
Facilitates encoding tiles in parallel.

BUG=webm:1309

Change-Id: I37aa336d47babffc8352188dc767eebdb8a99474
2016-10-12 20:22:03 -07:00
Vignesh Venkatasubramanian
04a6010742 Merge "vp9_bitstream: Parameterize max_mv_magnitude" 2016-10-12 21:52:42 +00:00
Vignesh Venkatasubramanian
d03d1c8cd3 vp9_bitstream: Parameterize max_mv_magnitude
Facilitates encoding tiles in parallel.

BUG=webm:1309

Change-Id: I614a5a492c30b6773c30e7294cd6a6f456e02ab4
2016-10-12 12:50:17 -07:00
Marco
57c6bf291e 1 pass vbr: Allow for lookahead alt-ref in real-time mode.
For 1 pass vbr real-time mode:
Allow for the usage of alt-ref frame when non-zero lag-in-frames is used.
Use non-filtered alt-ref, and select usage based on fast scene/content
analysis/detection within the lag of frames.

Positive gains on ytlive set: overall avgPSNR ~3-4%.
Several clips are up between 5-14%, a few clips are neutral/small change.

Current speed decrease is about ~5-10%.

Use the flag USE_ALTREF_FOR_ONE_PASS to enable this feature
(off by default for now).

Change-Id: I802d2bf3d44f9cf01f6d15c76be9c90192314769
2016-10-11 10:13:17 -07:00
Marco
cdbd89197e vp9: 1 pass vbr: some adjustments to gf interval.
Put limit on gf interval based on lag, and allow
for the adjustment on next gf group also on key frame.

Small/neutral change on ytlive metrics.

Change only affects 1 pass vbr real-time mode.

Change-Id: I339c8f4398848698b6e10fe9482c52ca661b94a5
2016-10-11 08:34:12 -07:00
Vignesh Venkatasubramanian
ed50e7710c write_modes: add MACROBLOCKD as a parameter
This will enable bit stream packing of each tile column in
parallel.

BUG=webm:1309

Change-Id: Ie349d8cc5825326218ffda893a50730b2e68ed34
2016-10-07 10:25:02 -07:00
Alex Converse
3063c37600 Remove vpx_realloc()
It only handles the realloc constraint (preserving low elements) by
serendipity, and we don't actually rely on that behavior anyway.
Meanwhile the calls may do extra copying that gets immediately clobbered
by the callers.

Change-Id: I8dfa89e4a81084b084889c27bd272fdf85184e8d
2016-10-05 10:57:56 -07:00
Angie Chiang
5d635365bb Merge "Move highbd txfm input range check from 2d iht transform to 1d idct/iadst" 2016-10-04 16:57:37 +00:00
Angie Chiang
5b073c695b Move highbd txfm input range check from 2d iht transform to 1d idct/iadst
This change will make the highbd txfm input range check more comprehensive

The 25-bit highbd input range is composed by
12 signal input bits + 7 bits for 2D forward transform amplification + 5 bits for
1D inverse transform amplification + 1 bit for contingency in rounding and quantizing

BUG=https://bugs.chromium.org/p/webm/issues/detail?id=1286
BUG=https://bugs.chromium.org/p/chromium/issues/detail?id=651625

Change-Id: I04c0796edd7653f8d463fba5dc418132986131e7
2016-10-03 17:21:08 -07:00
James Zern
fb020805f9 Merge "Fix warning when building with GCC 5." 2016-10-03 22:42:52 +00:00
James Zern
c6bc7499d9 Merge "cosmetics,*_neon.c: rm redundant return from void fns" 2016-10-03 22:40:42 +00:00
Geza Lore
0dc12b4a1c Fix warning when building with GCC 5.
These caused the following warning with GCC 5:
     warning: logical not is only applied to the left hand side of
     comparison [-Wlogical-not-parentheses]
     assert(!is_compound == (cm->reference_mode == SINGLE_REFERENCE));

Change-Id: If296aabb2311ceb7d903b395c1549ef81c2cbf9b
(cherry picked from commit c6cf7a6111)
2016-10-01 12:23:15 -07:00
Marco Paniconi
0a9f56f146 Merge "vp9: On change_config() only call update_frame_size if needed." 2016-09-30 21:43:33 +00:00
Marco Paniconi
5e908aff34 Merge "vp9 real-time mode: Change loopfilter speed feature at speed 8." 2016-09-30 21:42:05 +00:00
James Zern
db80c23fd4 cosmetics,*_neon.c: rm redundant return from void fns
+ a couple of 'break's after a return

Change-Id: Ia21f12ebcef98244feb923c17b689fc8115da015
2016-09-30 13:09:57 -07:00
James Zern
b6277a47c7 Merge changes from topic '8bit-hbd-idct'
* changes:
  *idct*_neon.c: add missing rtcd include
  idct,msa/neon: exclude idct files from hbd build
  *rtcd_defs.pl: remove empty specialize calls
2016-09-30 19:36:08 +00:00
Linfeng Zhang
8c744fd978 Merge "Unify loopfilter function names" 2016-09-30 15:58:08 +00:00
James Zern
ed62d27c71 *rtcd_defs.pl: remove empty specialize calls
add_proto adds a 'c' specialization

Change-Id: I0ed0c2240d45264b0e0056ce7c8f63f4a00780bc
2016-09-29 20:38:26 -07:00
Yunqing Wang
9afe2cf599 Merge "Fix an issue in vp9_first_pass for non-mulitple of 16 resolutions" 2016-09-30 00:49:06 +00:00
Linfeng Zhang
7f1f35183a Unify loopfilter function names
Rename vpx_lpf_horizontal_edge_8() to vpx_lpf_horizontal_16().
Rename vpx_lpf_horizontal_edge_16() to vpx_lpf_horizontal_16_dual().

Change-Id: I798ca8fbbd657d06d3db2bfb0fb3321168f49e52
2016-09-29 16:25:42 -07:00
Deepa K G
2745f94deb Fix an issue in vp9_first_pass for non-mulitple of 16 resolutions
This patch sets the 16x16 src_diff to zero and ensures correct calculation
of this_error for block sizes smaller than 16x16.

Change-Id: I7b7c02d267433c9f22c8ac9b8d5df2f499175172
2016-09-29 16:19:23 -07:00
Marco
e765435293 vp9: On change_config() only call update_frame_size if needed.
change_config() may be called often in real-time application,
to update bitrate/framerate or qp-max/min.
No need to do update_frame_size() unless frame size has changed.

Change-Id: I23a51deade1e03adc91c468f9ffde3235298770c
2016-09-29 13:03:26 -07:00
Marco
d017548be6 vp9 real-time mode: Change loopfilter speed feature at speed 8.
For real-time mode at speed 8: turn off MINIMAL_LF at speed 8,
for non-screen content mode.

Visually better, avgPSNR/SSIM on rtc set go up by ~4-5%.
Speed decrease of about ~3%.

Change-Id: I8eb69330f02e0ceece1507d43cfc8a049a1d8291
2016-09-29 12:59:01 -07:00
James Zern
450d89034b vp9_detokenize,decode_coefs: fix signed int overflow
when decoding an invalid bitstream with --enable-vp9-highbitdepth

BUG=webm:1297

Change-Id: I401d87033b4293f2ca595bc51678aad9951ecf15
2016-09-28 22:42:03 -07:00
James Zern
e094e151de Merge "vp9: fix compilation for g++ 6.2.x" 2016-09-28 23:36:23 +00:00
Tristan Matthews
32c375447c vp9: fix compilation for g++ 6.2.x
Inline function called from test/dct16x16_test.cc wouldn't build due to:
  invalid operands of types ‘__gnu_cxx::__enable_if<true, double>::__type
  {aka double}’ and ‘int’ to binary ‘operator>>’
  return (abs(ref->row) >> 3) < COMPANDED_MVREF_THRESH &&

this converts the test to abs() < COMPANDED_MVREF_THRESH << 3 which
hides the promotion issue.

Regression from commit de993a847f

BUG=webm:1291

Change-Id: I73b5943d07d5b61b709d299114216a2371a8fd62
2016-09-27 23:17:31 -07:00
Debargha Mukherjee
74f038e6f8 Merge "Fix for compile error with range checking" 2016-09-28 00:05:37 +00:00
Paul Wilkins
b3ebea5e8a Merge "Limit max arf boost and scale motion breakout for image size." 2016-09-27 14:08:29 +00:00
Peter de Rivaz
8db503063f Fix for compile error with range checking
Current version does not build with options:
  --enable-vp9-highbitdepth --enable-coefficient-range-checking

Change-Id: Ic3285f1a3e0d6be88da7f2cd8fa5a631368dd03b
2016-09-27 09:28:44 +01:00
Marco
d9fc28c0a1 vp9: Reduce frame loopfilter-level for 1 pass cbr.
Reduce the filt_guess for 1 pass cbr on inter-frames.
This reduces visual artifact seen in rtc clip (jimred.vga),
and improves metrics on rtc set.

Metrics on rtc set for cbr mode overall positive, most clips are up:
Speed 7 rtc: avgPSNR/SSIM up by: ~2.6/3.9%
Speed 8 rtc: avgPSNR/SSIM up by: ~1.3/2.5%

Change-Id: Ia4eccea1c19d65b583516df28823cd756c49464f
2016-09-26 10:12:43 -07:00
paulwilkins
0421d8e318 Limit max arf boost and scale motion breakout for image size.
Added a cap on the maximum boost for an arf based on interval length.
Fixed bug where by the image size was not accounted for in determining
two of the motion breakout thresholds.

Overall small gains of 0.2-0.4% psnr but on large image format clips with
slow zooms the gain may be as much as 20% or more (e.g. in_to_tree
at 1080P)

Change-Id: Id0a47391203026742daa9c97afac5705fd8c4dfb
2016-09-26 15:38:29 +01:00
Scott LaVarnway
60624aa53a Merge "VP9: token decoder expansion" 2016-09-26 12:06:50 +00:00
James Zern
f8c056a895 Merge "vp9_idct: delete dead TODOs" 2016-09-24 01:47:00 +00:00
Scott LaVarnway
87b689f97a VP9: token decoder expansion
This version is based on Change 267683, but does not
use the macros.

Change-Id: I0619fa618decf8bdeef250584d75d70318b5d9a7
2016-09-23 06:24:20 -07:00
Scott LaVarnway
ada850786c Merge "VP9: pass TileWorkerData instead of MACROBLOCKD and vpx_reader." 2016-09-23 11:59:16 +00:00
Scott LaVarnway
7a34f85955 VP9: pass TileWorkerData instead of MACROBLOCKD and vpx_reader.
Change-Id: I869ef0f113c022143b531c44aefa0f1bb267052d
2016-09-22 13:18:36 -07:00
Angie Chiang
99ef84c65a Merge "Detect invalid highbd iht input" 2016-09-22 01:06:38 +00:00
Angie Chiang
80338b91d3 Detect invalid highbd iht input
Do nothing in vp9_highbd_iht#x#_##_add_c when input magnitude is beyond
20 bits. Note that, sign bit is not included here.

In the 20 bits, we use 12 bits for input signal, 7 bits for forward
transform amplification, and 1 bit for contingency in rounding and
quantizing

BUG=https://bugs.chromium.org/p/webm/issues/detail?id=1286

Change-Id: I332c6f68df4614fc2e7d2dc4c5bb0d0cff8a245c
2016-09-21 17:15:19 -07:00
James Zern
3f72509587 vp8: remove VP8_SET_DBG* control support
the --enable-postproc-visualizer configure option remains as a no-op as
do the control names and values for compatibility
+ remove the corresponding debug flags from vpxdec: --pp-*

Change-Id: I4a001cd9962b59560d7d6bda6272d4ff32b8d37c
2016-09-20 20:19:36 -07:00
James Zern
cec6433e41 vp9_idct: delete dead TODOs
Change-Id: Icdd5494f557d83026dc078bce37997a76aa288fb
2016-09-20 19:46:27 -07:00
James Zern
feb4313c5f Merge "vp9_rtcd: remove non-existent highbd convolve fns" 2016-09-20 22:07:09 +00:00
James Zern
5841929fde vp9_rtcd: remove non-existent highbd convolve fns
these were moved to vpx_dsp

Change-Id: I307b07ae05e2333277d4b7011cba36dcf8409959
2016-09-19 20:01:23 -07:00
Nathan E. Egge
de7f5ce9e5 Code class0 using vpx_read() / vpx_write().
The vp9_mv_class0_tree is a balanced tree with two leafs and can
simply be coded as a boolean with probability class0[0].

Change-Id: If294dac825a5f945371092c74aa8e3f84cd962b6
(cherry picked from commit be8a8ab62ebdd111c6f2e9a33b15630570671eba)
2016-09-19 10:50:39 -07:00
Alex Converse
01e2902521 Zero the whole rd_counts struct rather than the each member
Change-Id: I495aa9cec2b2b8f1ae69bdab8b3feeca76358472
2016-09-19 10:04:47 -07:00
James Zern
7a9e476072 Merge changes from topic 'clang-format'
* changes:
  apply clang-format
  .clang-format: update to 3.8.1
2016-09-16 07:11:33 +00:00
Marco
4c1a9fb8db vp9: Small code cleanup.
Remove the experiment LIMIT_QP_ONEPASS_VBR_LAG, as its
not currently used and no plan to use in near future.

Change-Id: Ib069f8d7225195be04b765d0ab477510dfba6a3b
2016-09-15 15:17:17 -07:00
clang-format
5f6d143b41 apply clang-format
Change-Id: I501597b7c1e0f0c7ae2aea3ee8073f0a641b3487
2016-09-15 15:07:53 -07:00
Sarah Parker
c892521b1d Fix missing write to opsnr in internal stats
Change-Id: I21c8ad0b5ed7f8d843cae45c18f5727bceb8f859
2016-09-03 12:15:32 -07:00
Yaowu Xu
594e53514b Merge "Fix formatting in internal stats for vp8 and vp9" 2016-09-01 23:55:23 +00:00
Yaowu Xu
454139ae13 Merge "Casts to remove some warnings." 2016-09-01 23:37:04 +00:00
Debargha Mukherjee
a6bc3dfb0f Merge "Refactor uv tx size with lookup arrays" 2016-09-01 16:46:32 +00:00
paulwilkins
3e9e77008c Casts to remove some warnings.
Added casts to remove warnings:
BUG=webm:1274

In regards to the safety of these casts they are of two types:-

- Normalized bits per (16x16) MB stored in a 32 bit int (This is safe as bits
per MB even with << 9 normalization cant overflow 32 bits. Even raw 12
bits hdr source even would only be  29 bits :- (4+4+12+9) and the encoder
imposes much stricter limits than this on max bit rate.

- Cast as part of variance calculations.  There is an internal cast up to 64 bit
for the Sum X Sum calculation, but after normalization dividing by the number
of points the result will always be <= the SSE value.

Change-Id: I4e700236ed83d6b2b1955e92e84c3b1978b9eaa0
2016-09-01 16:10:12 +01:00
Debargha Mukherjee
e6446b4b60 Refactor uv tx size with lookup arrays
Change-Id: Ife6a3d301c5faaba89d16d188d638631083511f7
2016-08-31 13:15:38 -07:00
paulwilkins
6fc07a217d Modified resize loop constraints.
Using a tighter resize constraint on undershoot seems to help
results (especially SSIM) as significant undershoot on a frame
seems to have more of a damaging impact than overshoot.

This patch has been tuned so that in local testing using the
derf set it is encode speed neutral for speed  setting 2.

Average quality result for speed 2 (psnr,ssim) were  as follows:-

 lowres  0.039,  0.453
 midres  0.249, 0.853
 hdres  0.159, 0.659
 NetFlix -0.241, 0.360

Change-Id: Ie8d3a0d7d6f7ea89d9965d1821be17f8bda85062
2016-08-31 12:45:49 +01:00
Paul Wilkins
129814fcb4 Merge "Adjust coefficient optimization and tx_domain rd speed features." 2016-08-30 16:54:40 +00:00
James Zern
2917737879 vp9_alt_ref_aq_set_nsegments: harmonize fn signature
Change-Id: I5f232664652a8dc3a71e43b8b1fa05ddb4a84ecc
2016-08-27 11:16:03 -07:00
Yury Gitman
507d272265 Move vp9_alt_ref_aq_private.h to vp9_alt_ref_aq.c
+ add a temporary dummy element to ALT_REF_AQ to avoid a warning about
an empty struct

Change-Id: Ib6e5c39ff62ad96eb4e3686d4882228a42b3843f
2016-08-27 10:53:41 -07:00
Jingning Han
dd2a475e43 Merge "Fix VS build warnings in vp9_alt_ref_aq files" 2016-08-26 17:19:12 +00:00
Paul Wilkins
badd32d914 Merge "Add ALLOW_RECODE_FIRST speed mode." 2016-08-26 15:46:45 +00:00
Jingning Han
84fccfe475 Fix VS build warnings in vp9_alt_ref_aq files
Change-Id: I5b19ec00a1eb8b148026f665d217c12eb50b614a
2016-08-26 08:43:36 -07:00
paulwilkins
dc42f343ae Add ALLOW_RECODE_FIRST speed mode.
This patch is to address concerns that changes to allow
recodes on the first frame in each ARF group do not give a
good enough speed quality trade off for speed 2. Though the
average impact  on encode speed is 1-2%, for some hard clips
it is > 5% rise.  For speed 1 this is less an issue and for Speed 0
the previous patch actually  improves speed.

Change-Id: Ie1bcefdbfdf846d3f4428590173f621465dffe3a
2016-08-26 11:43:47 +01:00
Sarah Parker
37e83789f1 Fix formatting in internal stats for vp8 and vp9
This corrects a formatting error introduced in:
I1e9d548ce445d29002f0c59ebfd3957a6f15e702
where spaces were used as delimiters instead of tabs.

The corresponding fix for vp10 is in
Ica3d625d6672b3c47e0e208b45eede29b9004030.

Change-Id: Ibc4eb8fd82e6b926ba259a679dc98557cadba9b1
2016-08-25 17:46:18 -07:00
Yury Gitman
292d221fed Create interface for the ALT_REF_AQ class
Current commit is just an API template  for the rest of the code, and
I will add inner logic later.

Altref  frames  generate a  lot  of  bitrate  and  at the  same  time
other  frames  refer to  them  a  lot, so  it  makes  sense to  apply
special  compensation-based adaptive  quantization scheme  for altref
frames. E.g.,  for blocks  that are  good predictors  for the  future
apply rate-control  chosen quantizer  while for bad  predictors apply
worse one.

Change-Id: Iba3f8ec349470673b7249f6a125f6859336a47c8
2016-08-25 10:55:14 -07:00
paulwilkins
635ae8bdc1 Adjust coefficient optimization and tx_domain rd speed features.
Previously Tx domain rd was used in all cases above speed 0.
Coefficient optimization was only enabled for best and speed 0.

This patch selectively sets these features at other speed settings
based on block complexity.

For the Netflix and HD sets in particular the quality gains are
large compared to the speed hit. At speed 1 the average psnr
gain in the NF set  is > 2.5% with one clip coming in at 18%
and some points almost 30%.  Average gains for the lower
resolution test sets are around 1%.

The gains are biggest at low Q so some further optimization
may be possible.

Change-Id: I340376c7b2a78e5389a34b7ebdc41072808d0576
2016-08-25 15:36:16 +01:00
Yury Gitman
d7c20079a6 Add --alt-ref-aq=<int> option
In the future this option will activate adaptive quantization special
for altref frames. Encoder will  create the adaptive quantization map
on the basis of lookahead buffers similarity which is the estimate of
the future motion compensation performance.

Change-Id: Ia0088b3babb0f9a4899c79d8d819947ba5a03df2
2016-08-24 15:49:25 -07:00
Jacky Chen
5260a6675e Merge "vp9: Refactor set_low_temp_var_flag." 2016-08-24 22:02:53 +00:00
Yury Gitman
c325fb748a Correct CHECK_MEM_ERROR macro (release builds)
The previous macro doesn't work with &cpi->common as a first argument

Change-Id: Iddf7a1f5d56d7abafd9b2b8707aa611d349e7a68
2016-08-23 22:46:04 +00:00
jackychen
8d4c0ec1f1 vp9: Refactor set_low_temp_var_flag.
No need to pass in force_split, since we should use sb_type in the
condition.

Change-Id: Ide27243ef46e017bbb98d676347fc566a6c828f7
2016-08-23 15:11:40 -07:00
Yunqing Wang
f6c5410cd4 Merge "Disable split mode in 4k video encoding" 2016-08-23 15:35:33 +00:00
Yunqing Wang
ef98f49cb0 Disable split mode in 4k video encoding
Disabled the split mode while encoding 4k video to speed
up the encoder.

Borg test result on 4k set:
Overall PSNR: +0.029%; SSIM: +0.009%.
Average encoder speedup at speed 2 is 2.5%.

Change-Id: I1519c658f07c3ac838affbe5aff0ed9b94f3f8f4
2016-08-22 19:46:44 -07:00
Yury Gitman
bf7a02a4cf Correct CHECK_MEM_ERROR macro
The previous macro doesn't work with &cpi->common as a first argument

Change-Id: Ic3f5c49a94cf8b17de6569811b957c963341bb58
2016-08-22 14:25:57 -07:00
Yunqing Wang
37169c0bd4 Merge "Adjust speed features for 4k video encoding" 2016-08-19 23:11:05 +00:00
Yunqing Wang
fe488cceff Adjust speed features for 4k video encoding
Adjusted speed 2 features to speed up 4k video encoding.
BDBR results from borg test:
PSNR: +0.313%; SSIM: +0.268%.
Average speedup: 8.5%

Change-Id: I1e2695a01fb3f3817c1df4480e184c2aed8f2eba
2016-08-19 09:30:32 -07:00
James Zern
149d082377 vp9_pickmode: quiet float conversion warnings
Change-Id: I591e4f958955b3f2edb2f95a83c54cd83c8ef075
2016-08-19 01:28:01 -07:00
James Zern
8b4c31584e vp9_alloc_context_buffers: clear cm->mi* on failure
this fixes a crash in vp9_dec_setup_mi() via
vp9_init_context_buffers() should decoding continue and the decoder
resyncs on a smaller frame

BUG=b/30593752

Change-Id: I9ce8d94abe89bcd058697e8bd8599690e61bd380
2016-08-19 00:18:11 -07:00
JackyChen
8be7e572a7 vp9 svc: SVC encoder speed up.
Bias towards base_mv and skip 1/4 pixel motion search when using base mv.
2~3% speed up for 2 spatial layers, 3~5% speed up for 3 spatial layers.
PSNR loss:
(2 layers) 0.07dB for gips_stationary, 0.04dB for gips_motion;
(3 layers) 0.07dB for gips_stationary, 0.06dB for gips_motion.

Change-Id: I773acbda080c301cabe8cd259f842bcc5b8bc999
2016-08-18 11:25:45 -07:00
Marco
7eb7d6b227 vp9 non-rd pickmode: Add limit on newmv-last and golden bias.
Add option, for newmv-last, to limit the rd-threshold update for early exit,
under a source varianace condition.
This can improve visual quality in low texture moving areas,
like forehead/faces.

Also add bias against golden to improve the speed/fps,
will little/negligible loss in quality.

Only affects CBR mode, non-svc, non-screen-content.

Change-Id: I3a5229eee860c71499a6fd464c450b167b07534d
2016-08-17 14:33:44 -07:00
paulwilkins
af3b0de732 Add casting to fix warning.
Frame bits can safely be stored int but group bits
(kf or arf) use 64bit.

Change-Id: I0800f2a28070f8749110a95721c116fc56987885
2016-08-17 11:18:07 +01:00
paulwilkins
ab7cd6d068 Add {} to try and keep Jenkins happy.
Change-Id: If1ca3cf83e058317c9751d7da6caa7cd75eb6845
2016-08-17 11:17:36 +01:00
paulwilkins
5d881770e5 Change default recode rule for good speed 0 and best.
Changes the default recode rule for Speed 0 and best quality
from ALLOW_RECODE to ALLOW_RECODE_KFARFGF.

Tested on the NF, hdres, midres and lowres test sets, this setting
when combined with patch I40cb559... now performs "as well" in
metrics terms (in fact it came out a tiny amount better overall)
but encode time is 9.6%  faster (measured as the average
from 27 mid rate local encodes on clips in the derf/lowres set.

Change-Id: I8c781c0cdfa3a9929cd9406d15582fce47d6ae3b
2016-08-15 10:52:54 +01:00
paulwilkins
de3b769524 Change to recode rules.
Allow recodes for the first inter frame in each arf group
even when the recode rule is set to ALLOW_RECODE_KFARFGF.

Small gains of 0.05%.

Change-Id: I40cb559d36a2bf0ebf5cf758c3f92e452b480577
2016-08-15 10:52:02 +01:00
Paul Wilkins
fe4dd4f43f Merge "Modified ARF group allocation." 2016-08-15 09:42:30 +00:00
Yunqing Wang
a413dbe594 Fix another motion vector out of range bug
This patch fixed a motion vector out of range bug:
vpxenc: ../libvpx/vp9/encoder/vp9_mcomp.c:69:
 mv_cost: Assertion `mv->col >= -((1 << (11 + 1 + 2)) - 1) &&
 mv->col < ((1 << (11 + 1 + 2)) - 1)' failed.

For blocks that returned without having full-pixel search, the original
MV limits were not restored, which caused the failure. Moved the set
MV limit function down to fix the bug.

Change-Id: Id7d798fc7214e95c6e4846c588f0233fcf1a4223
2016-08-12 09:27:58 -07:00
paulwilkins
656f4a88cf Modified ARF group allocation.
Small average gains in the range 0.05 - 0.1

Change-Id: I30e85c04be615cc84726427c5057388b20a6ff60
2016-08-10 14:22:01 -07:00
Alex Converse
941fe20336 Merge "Refactor mv limits." 2016-08-09 17:12:50 +00:00
Yury Gitman
c37d012ada Merge "Add cpi parameter for forcing segmentation update" 2016-08-08 21:29:42 +00:00
Yury Gitman
7a730d5901 Add cpi parameter for forcing segmentation update
Change-Id: I1b0bcb1ffe7604117bfaa0b9989d0e25ff04d28c
2016-08-08 13:20:42 -07:00
Alex Converse
6554333b59 Refactor mv limits.
Change-Id: Ifebdc9ef37850508eb4b8e572fd0f6026ab04987
2016-08-08 11:54:00 -07:00
Yunqing Wang
6a8d4631a8 Merge "Fix a motion vector out of range bug" 2016-08-08 17:59:50 +00:00
James Zern
19d2e73dea Merge changes Ice037acb,I806af11b,I344a7dd0,Ib7cb87fa
* changes:
  vp9: normalize vpx_enc_frame_flags_t usage
  args.c: add some explicit casts
  webmdec: quiet -Wshorten-64-to-32 warning
  test/decode_test_driver: rm unused deadline member
2016-08-06 01:20:52 +00:00
Yunqing Wang
2fb826c4d5 Fix a motion vector out of range bug
This patch fixed a motion vector(MV) out of range bug, which was caused
by not restoring the original values of the MV min/max thresholds after
the sub8x8 full pixel motion search. It occurred rarely and only was seen
while encoding a 4k clip for 200 frames.

BUG=webm:1271

Change-Id: Ibc4e0de80846f297431923cef8a0c80fe8dcc6a5
2016-08-05 15:23:05 -07:00
James Zern
7104833085 vp9: normalize vpx_enc_frame_flags_t usage
quiets -Wshorten-64-to-32 warnings

Change-Id: Ice037acb675d1d81bfedf2dfcfa91a8a29a19dfd
2016-08-04 23:37:49 -07:00
Johann Koenig
57f49db81f Merge changes I6ef79702,Id332c641,I354b5d22,I84438013
* changes:
  Use common transpose for vpx_idct32x32_1024_add_neon
  Use common transpose for vpx_idct8x8_[12|64]_add_neon
  Use common transpose for vp9_iht8x8_add_neon
  Use common transpose for vpx_idct16x16_[10|256]_add_neon
2016-08-04 22:30:47 +00:00
Johann
7103b5307d Use common transpose for vp9_iht8x8_add_neon
Change-Id: I354b5d22130d76b0eceda0748db1f871f58fa372
2016-08-04 20:13:03 +00:00
James Zern
70a7885a65 correct break placement
these should be placed within {}s when present

Change-Id: Ia775fac5373603e77360398f19b07958fb43f476
2016-08-04 13:00:14 -07:00
clang-format
3a4002b94d vp9_ratectrl.c: apply clang-format
after:
ff0a87c vp9 1pass vbr: Adjustment to gf interval.

Change-Id: I1296e53e601bf0c2b562e3a34082ac45c294a5f1
2016-08-04 11:57:00 -07:00
Marco Paniconi
9fdeeaf411 Merge "vp9 1pass vbr: Adjustment to gf interval." 2016-08-04 17:50:55 +00:00
Yaowu Xu
7a79fa1362 Fix msvc compiler warnings
MSVC 2013 complained about using 32 shift where 64 bit shift should be
used.

Change-Id: I7a2b165d1a92d3c0a91dd4511b27aba7709b5e55
2016-08-03 18:33:06 -07:00
Marco
ff0a87ce38 vp9 1pass vbr: Adjustment to gf interval.
Increase the minimum distance.
Reduces the overshoot somewhat on some clips,
small gain in avgPSNR (~0.1%) on ytlive set.

Change-Id: Id5ddde20c2907dbdb536e79542eff775019c142b
2016-08-03 15:36:27 -07:00
clang-format
08131055e4 vp9/decoder,vp9/*.[hc]: apply clang-format
Change-Id: Ic38ea06c7b2fb3e8e94a4c0910e82672a1acaea7
2016-08-03 14:29:31 -07:00
clang-format
8ff40f8bec vp9/common: apply clang-format
Change-Id: Ie0f150fdcfcbf7c4db52d3a08bc8238ed1c72e3b
2016-08-02 18:27:07 -07:00
clang-format
e0cc52db3f vp9/encoder: apply clang-format
Change-Id: I45d9fb4013f50766b24363a86365e8063e8954c2
2016-08-02 16:47:11 -07:00
Yaowu Xu
039f9e08f0 change HBD pixel value from uint8_t to uint16_t
This fixes a regression in 10/12 bit encoding results.

Change-Id: I438877352a41aae0a864a8d9979afe4aa2061d81
2016-08-02 11:01:39 -07:00
Yaowu Xu
dc5618f3bb Add pointer conversion for HBD buffers
This fixes a crash in HBD build.

Change-Id: I7f688f50227323e69bba65df0d56f4360f01771b
2016-08-01 15:56:43 -07:00
Alex Converse
004eebed31 Merge "Unfork 8-bit in HBD path in vp9_model_rd_from_var_lapndz callers." 2016-08-01 16:42:39 +00:00
Alex Converse
2c3807b89f Merge "Cache optimizations in optimize_b()." 2016-08-01 16:30:05 +00:00
Alex Converse
e446ffda45 Cache optimizations in optimize_b().
Move best index into the token state. Shrink it down to one byte. This
is more cache friendly (access are group together) and uses less total
memory.

Results in 4% fewer cycles in optimize_b().

Change-Id: I75db484fb3dc82f59928d54b659d79c80ee40452
2016-07-29 12:06:49 -07:00
Jacky Chen
462a7c9f0a Merge "vp9 svc: Enable different speed setting for each spatial layer." 2016-07-28 20:21:30 +00:00
Alex Converse
4508eb3123 Merge "Fix 64 to 32 narrowing warning." 2016-07-28 16:36:46 +00:00
Alex Converse
335cf67d8b Fix 64 to 32 narrowing warning.
- Solves potential integer overflow on 12-bit
- Fixes Visual Studio build

Change-Id: I26dd660451bbab23040e4123920d59e82585795c
2016-07-27 12:40:23 -07:00
JackyChen
47cc64cdf8 vp9 denoiser: Derefencing pointer should be after null check.
BUG=webm:1267

Change-Id: I899fc9e8d784c6eefcbe27945c619845adb7b6f0
2016-07-26 17:31:17 -07:00
Alex Converse
34201e50c1 Unfork 8-bit in HBD path in vp9_model_rd_from_var_lapndz callers.
BUG=b/29583530

Change-Id: Ia88a75f9572e08f228559ab84b8a77efb5aff0af
2016-07-26 21:57:58 +00:00
Alex Converse
1c85230344 Merge "Only consider visible 4x4s in pixel domain error." 2016-07-26 19:39:54 +00:00
jackychen
8ce67d714a vp9 svc: Enable different speed setting for each spatial layer.
This change only affects 1 pass cbr svc mode.

Change-Id: If0da87bb200f7e7762755340c40c8157cc7a16ca
2016-07-25 15:11:43 -07:00
Alex Converse
d6c5ef4557 Only consider visible 4x4s in pixel domain error.
BDRATE change
derf144: -0.327
lowres: -0.048
midres: -0.125
hdres: -0.238

Change-Id: I789aba9870b5c2952373a7dd4fc8ed45590c3c54
2016-07-25 21:44:06 +00:00
James Zern
7aa0c748b3 Merge "vp9: fix frame-level threaded decode shutdown" 2016-07-25 19:00:37 +00:00
Alex Converse
511bf49b7e Merge "Minor skip segment simplification." 2016-07-25 17:50:43 +00:00
Scott LaVarnway
ad5fea03e6 Merge "VP9: get_pred_context_switchable_interp() -- encoder side" 2016-07-25 11:58:24 +00:00
James Zern
f368f86df6 vp9: fix frame-level threaded decode shutdown
Shutdown all threads before reclaiming any memory. The frame-level
parallel decoder may access data from another worker.

BUG=webm:1259

Change-Id: I26856ebd1f77cc4a4545331baa19bbf3e01c4ea4
2016-07-23 10:59:15 -07:00
Alex Converse
9a62ecbd35 Minor skip segment simplification.
Change-Id: I34863fce1abe94f9539e9a5a6149ae1efb6501bd
2016-07-22 15:31:18 -07:00
Marco Paniconi
53db633349 Merge "vp9 1pass-vbr: Adjust gf setting for nonzero-lag case." 2016-07-22 21:27:05 +00:00
Marco
c06a4b9df2 vp9 1pass-vbr: Adjust gf setting for nonzero-lag case.
Change-Id: I230c586c6d5ae56ee9a6d37b7d9452351bb4bd80
2016-07-22 11:48:09 -07:00
Paul Wilkins
830fa866a5 Merge "Sample points to reduce encode overhead." 2016-07-22 09:27:34 +00:00
Paul Wilkins
063e4a2914 Merge "Noise energy Experiment in first pass." 2016-07-22 09:27:19 +00:00
Scott LaVarnway
c969b2b02b VP9: get_pred_context_switchable_interp() -- encoder side
Change-Id: I7217c90d5cf38c51b76759a2dc4f10070f3a40ac
2016-07-21 11:47:51 -07:00
jackychen
71f9cbcfc8 vp9: Fix the clang warning of unsigned int type.
Change-Id: I6308db16bd626fa5943925471e9171f567669350
2016-07-20 15:58:35 -07:00
Yaowu Xu
297b2a12d6 Fix encoder crashes for odd size input
(cherry picked from commit 98431cde07)

Change-Id: Id5c30c419282369cc8c3280d9a70b34a859a71d8
2016-07-20 15:02:13 -07:00
Yaowu Xu
690fcd793b Change to call vp9_post_proc_frame()
This commit changes the call in vp9 encoder from vp9_deblock() to
vp9_post_proc_frame() to ensure the data structures used in the call
are properly allocated. This fixes an encoder crash when configured
with --enable-internal-stats.

Change-Id: I2393b336c0f566665336df4f1ba91c405eb56764
2016-07-20 11:01:49 -07:00
James Zern
e3f7991f99 Merge changes Ia6004c08,I1954f9d6
* changes:
  cosmetics: Add a few explanatory comments
  cosmetics: Correct grammar/spelling in comments
2016-07-19 19:12:23 +00:00
Yury Gitman
e4ac882007 cosmetics: Add a few explanatory comments
Change-Id: Ia6004c08e6f5fd269a1bbd4df51ce9b76345150d
2016-07-19 10:39:00 -07:00
Marco
05fe0f20a6 vp9: Allow usage of lookahead for real-time, 1 pass vbr.
Allow usage of lookahead for VBR in real-time mode, for 1 pass vbr.

Current usage is for fast checking of future scene cuts/changes,
and adjusting rate control (gf interval and active_worst/target size).

Added unittests (datarate) for 1 pass vbr mode, with non-zero lag.

Added an experimental option to limit QP based on lookahead.

Overall positive gain in metrics on ytlive set:
avgPNSR/SSIM up on average ~1-3%; several clips up by 5, 7%.

Change-Id: I960d57dfc89de121c4824b9a9bf88d2814e74b56
2016-07-18 15:20:17 -07:00
Yury Gitman
bdfdd7d993 cosmetics: Correct grammar/spelling in comments
Change-Id: I1954f9d6e33abff9081fe7a5cf59d5497768e0df
2016-07-18 12:49:00 -07:00
James Bankoski
ce6678fdc9 Merge "addnoise : clear out static size for generated noise" 2016-07-16 01:48:07 +00:00
Jim Bankoski
cb957c302a addnoise : clear out static size for generated noise
Change-Id: I5d4343f2da9cd4b01dd37be7a048d159fec109d1
2016-07-15 15:52:45 -07:00
Jim Bankoski
da1bda0fb2 vp9_postproc.c : unused variable if not vp9_highbitdepth.
Change-Id: Ib89b128f23767934c40b5add3fcf9dbd875e82f9
2016-07-15 15:04:57 -07:00
Jim Bankoski
0dc69c70f7 postproc : fix function parameters for noise functions.
Change-Id: I582b6307f28bfc987dcf8910379a52c6f679173c
2016-07-15 08:27:34 -07:00
James Bankoski
7eec1f31b5 Merge "postproc: noise style fixes." 2016-07-13 22:04:47 +00:00
Hui Su
8dd3bef7ef Merge "Revert "Eliminate isolated and small tail coefficients:"" 2016-07-13 21:30:12 +00:00
Yaowu Xu
d6197b621d Merge "Fix encoder crashes for odd size input" 2016-07-13 20:05:09 +00:00
Jim Bankoski
e736691a6d postproc: noise style fixes.
Change-Id: Ifdcb36b8e77b65faeeb10644256e175acb32275d
2016-07-13 12:39:01 -07:00
hui su
248f6ad771 Revert "Eliminate isolated and small tail coefficients:"
This reverts commit ff19cdafdb.

Change-Id: I81f68870ca27a1ff683ee22090530b6997815fb2
2016-07-13 11:14:44 -07:00
Jingning Han
fed14a3e94 Merge "Disable trellis optimization when lossless is on" 2016-07-13 16:01:01 +00:00
James Bankoski
e93f2fdb83 Merge "postproc - move filling of noise buffer to vpx_dsp." 2016-07-13 15:31:17 +00:00
Jim Bankoski
2ca24b0075 postproc - move filling of noise buffer to vpx_dsp.
Change-Id: I63ba35dc0ae9286c9812367a531e01d79a4c1635
2016-07-13 07:35:25 -07:00
Jim Bankoski
b24373fec2 deblock: missing const on extern const.
Change-Id: I0df08f7c431daf939e266f008bf5158b0c97358b
2016-07-13 07:27:29 -07:00
Jim Bankoski
6f424a768e vp9_postproc.c missing extern.
BUG=webm:1256

Change-Id: I5271e71bc53cce033fb906040643dcdd5ccb2381
2016-07-12 17:47:49 -07:00
Yaowu Xu
98431cde07 Fix encoder crashes for odd size input
Change-Id: Id5c30c419282369cc8c3280d9a70b34a859a71d8
2016-07-12 11:11:26 -07:00
Jacky Chen
19c157afe2 Merge "vp9 svc: Reuse scaled_temp in two stage downscaling." 2016-07-12 17:59:09 +00:00
JackyChen
110a2ddc9b vp9 svc: Reuse scaled_temp in two stage downscaling.
This change eliminates redundant computation in the two stage
downscaling, which saves ~1% encoding time in 3-layer svc encoding.

Change-Id: Ib4b218811b68499a740af1f9b7b5a5445e28d671
2016-07-12 10:09:55 -07:00
Jingning Han
efccbc9fb5 Disable trellis optimization when lossless is on
Disable trellis coefficient optimization when the lossless mode
is turned on.

Change-Id: I9001bf626e86dc3c8c32331ede04fd39036e5f7c
2016-07-12 09:00:16 -07:00
Jim Bankoski
88e6951465 deblock filter : moved from vp8 code branch
The deblocking filters used in vp8 have been moved to vpx_dsp for
use by both vp8 and vp9.

Change-Id: I5209d76edafc894b550f751fc76d3aa6799b392d
2016-07-12 05:53:00 -07:00
James Zern
45ed7effed Merge "remove *debugmodes.c from the default build" 2016-07-11 23:49:29 +00:00
Scott LaVarnway
2e93fcf893 Merge "vp9_rd_pick_intra_mode_sb(): set interp_filter to" 2016-07-11 22:31:06 +00:00
paulwilkins
3a986eac57 Sample points to reduce encode overhead.
Only noise filter sampled points in first pass to reduce
any first pass speed overhead.

Change-Id: Ic80d4400e59146d1c3332336c4350faf28ff8b17
2016-07-11 11:45:52 +01:00
Scott LaVarnway
ed7786869a vp9_rd_pick_intra_mode_sb(): set interp_filter to
SWITCHABLE_FILTERS.  This is a partial fix for the build
issues with Change 357240.

Change-Id: I4e507c196175bae729a4f1397878ec8776b0146c
2016-07-09 09:47:34 -07:00
Yaowu Xu
5adb43b8be Fix non-highbitdepth coding path for HBD build
Change-Id: I38eb42b8d051924a7cd1ccc3421a4057cf6e170f
2016-07-08 11:26:34 -07:00
Marco Paniconi
20946cdd3b Merge "vp9: Adjustment of gfu_boost and af_ratio for 1 pass vbr." 2016-07-08 16:26:06 +00:00
Yaowu Xu
dc008cc17d Merge "Enable HBD support in real time encoding path" 2016-07-07 22:32:48 +00:00
Marco
cc431ad50a vp9: Adjustment of gfu_boost and af_ratio for 1 pass vbr.
Modify the gfu_boost and af_ratio setting based on the
average frame motion level.

Change only affects 1 pass vbr.

Metrics overall positive on ytlive set.
On average up by ~1%, several clips up by 2-4%.

Change-Id: Ic18c49eb2df74cb4986b63cdb11be36d86ab5e8d
2016-07-07 15:18:14 -07:00
Marco Paniconi
a75965fa94 Merge "vp9: Adjustment to mv bias for non-rd pickmode." 2016-07-07 21:07:37 +00:00
Jingning Han
2f28f9072e Enable coeff optimization for intra modes
This further improves the coding performance by
lowres 0.3%
midres 0.5%
hdres  0.6%

Change-Id: I6a03b6da210b9cbc261474bad4a103e0ba021c68
2016-07-07 12:25:41 -07:00
Jingning Han
44354ee7bf Use precise context to estimate coeff rate cost
Use the precise context to estimate the zero token cost in trellis
optimization process. This improves the speed 0 coding performance
by 0.15% for lowres and 0.1% for midres. It improves the speed 1
coding performance by 0.2% for midres and hdres.

Change-Id: I59c7c08702fc79dc4f8534b64ca594da909e2c91
2016-07-07 12:25:33 -07:00
Jingning Han
62aa642d71 Enable uniform quantization with trellis optimization in speed 0
This commit allows the inter prediction residual to use uniform
quantization followed by trellis coefficient optimization in
speed 0. It improves the coding performance by

lowres 0.79%
midres 1.07%
hdres  1.44%

Change-Id: I46ef8cfe042a4ccc7a0055515012cd6cbf5c9619
2016-07-07 12:25:33 -07:00
Jingning Han
541eb78994 Refactor coeff_cost() function
Move the operations that update the context buffers outside this
function. The coeff_cost() takes all input as const value and returns
the coefficient cost.

This makes preparation for the next coefficient optimization CLs.

Change-Id: I850eec6e5470b91ea84646ff26b9231b09f70a0c
2016-07-07 18:09:39 +00:00
Jingning Han
7c1fdf02cd Merge "Support measure distortion in the pixel domain" 2016-07-07 18:09:20 +00:00
Marco
f451b404ea vp9: Adjustment to mv bias for non-rd pickmode.
Replace the existing mv bias with a bias only for
NEWMV, and based on the motion vector difference of
its top/left neighbors.

For cbr non-screen-content mode.

Change-Id: I8a8cf56347cfa23e9ffd8ead69eec8746c8f9e09
2016-07-07 10:33:06 -07:00
paulwilkins
2580e7d63e Noise energy Experiment in first pass.
Use a measure of noise energy to adjust Q estimate and
arf filter strength.

Gains 0.3-0.5% on Lowres and |Netflix sets.
Hdres and Midres neutral.

Change-Id: Ic0de552e7b6763e70eeeaa3651619831b423e151
2016-07-07 14:50:21 +01:00
Paul Wilkins
f037cf80c9 Merge "Add experimental spatial de-noise filter on key frames." 2016-07-07 13:30:07 +00:00
Jingning Han
e357b9efe0 Support measure distortion in the pixel domain
Use pixel domain distortion metric in speed 0. This improves the
compression performance by 0.3% for both low and high resolution
test sets.

Change-Id: I5b5b7115960de73f0b5e5d0c69db305e490e6f1d
2016-07-06 18:25:17 -07:00
Yaowu Xu
884c2ddc48 Enable HBD support in real time encoding path
BUG=webm:1223

Change-Id: If83a613784e3b2a33c9c93f9ad0ba39dd4d23056
2016-07-06 14:18:37 -07:00
Debargha Mukherjee
adbad6092f Merge "Remove decode asserts from better-hw-compatibility" 2016-07-06 20:55:29 +00:00
Jacky Chen
aa6108382e Merge "vp9: Choose the scheme for modeling rd for 32x32 based on skin color." 2016-07-06 20:03:55 +00:00
Debargha Mukherjee
4b6e4e1813 Remove decode asserts from better-hw-compatibility
Safer to have the decoder operate normally and have
better-hw-compatibility only implement encoding changes.
Fixes some test failures.

Change-Id: I0dd70d002e4e893992f0cd59774b9363e6f7fe76
2016-07-06 12:26:38 -07:00
JackyChen
2678aefc48 vp9: Choose the scheme for modeling rd for 32x32 based on skin color.
For real time CBR mode, use model_rd_for_sb_y for 32x32 if the sb is
a skin sb to avoid visual regression on the slowly moving face.

Refer to the cl: https://chromium-review.googlesource.com/#/c/356020/

Change-Id: I42c36666b2b474ce5ee274239d52ae8ab400fd46
2016-07-06 11:12:03 -07:00
Min Ye
ff19cdafdb Eliminate isolated and small tail coefficients:
Improve hdres PSNR by 0.696%
Improve midres PSNR by 0.313%
Improve lowres PSNR by 0.142%

Change-Id: Icabde78aa9689f539f6a03ec09f712c20758796c
2016-07-06 11:08:23 -07:00
Jingning Han
51aad61c8c Merge "Remove txfrm_block_to_raster_xy() from vp9 encoder" 2016-07-06 16:00:18 +00:00