GitLab

Support the transform block kernel coding for rectangular
transform block size in var-tx. This integrates txk-sel with
var-tx.

Change-Id: I9a8edd84812168f56c79b78cc9af34f6304b1d54

Define the syntax and entropy coding templates for
NCOBMC_ADAPT_WEIGHT. The actual values of the default
probabilities and the index tree structure need to
be fine tuned.

In this experiment all mv's in a superblock are sent
first as in the ncobmc case.

Change-Id: I68d50d3d27346c2847ea449a1168c6a99fbb4d3d

Change-Id: Idafef140d3425a9a9f66cb8864a804c4d2a89a70

The var-tx has its own suite of tx size/type RD search functions,
which recursively split the partition into square tx blocks.

The Daala-dist requires access to 8x8 pixels (both decoded and predicted)
since it measures the distortion for multiple of a 8x8 pixels.
Thus, if tx block is smaller than 8x8, it waits until all of sub8x8 blocks
are RD searched (with MSE) then replaces the MSE of 8x8 pixels with
daala-dist's calculated distortion for 8x8 pixels.

It is also applied to luma pixels only.

Change-Id: Ic4891e89b4ef05cf880aa26781d2d06ccf3142de

- If invisible pixels, av1_daala_dist() simply use source pixles for dst.
- Added av1_daala_dist_diff() which inputs diff signal instead of dst.

- Refactored daala_dist code so that av1_daala_dist() and _diff()
is called inside av1's distortion calculation functions, pixel_sse() and
sum_squares_visible().

Change-Id: Id857db52fe19856d92c46a9e84ac2962c01ae045

cb4x4 itself should not require these sizes.

This simplifies compatibility with other experiments, since we can
first make them work with cb4x4 (which is now on by default), and
then worry about chroma_sub8x8 and chroma_2x2 (which is not) in
separate steps.

Encoder and decoder output should remain unchanged.

Change-Id: Iff2a5494cab3b7d96f881e8bd9cd4bf18c817cfa

Without tempmv-signaling configured, using the previous frame's MVs
requires that the last frame was a show frame. With tempmv-signaling
configured, cm->show_last_frame is not checked when calculating
use_prev_frame_mvs. This patch adds that check and resolves mismatches
seen with random resizing and random superres.

Includes a couple fixes too - cm's last_width, last_height, and
last_show_frame were updated under different conditions. Now they're all
updated at the same time.

Change-Id: Ibdfb196cb6e9d002fd57cb4df10a899b60faac00

CfL performs an extra loop iteration during luma mode selection. Recent
changes have broken the extra iteration. Remove previous approach.

New approach adds the extra iteration right before uv parameter
selection. Interesting fact, If the best luma intra mode already has
worse RD performance than the best inter mode found so far (if any),
then the entire chroma intra search is skipped, including the extra 
iteration.

Results on Subset1 (compared to 3e18e4ae with CfL)

   PSNR | PSNR Cb | PSNR Cr | PSNR HVS |    SSIM | MS SSIM | CIEDE 2000
-0.3090 | -2.7271 | -2.3521 |  -0.3369 | -0.3463 | -0.3525 |    -1.1868

Change-Id: If67b0badd2c8ea25c61685483d39d622c1729b18

The tool of ext-comp-refs adds the uni-directional compound reference
prediction. In details, 3 pairs of uni-direcitonal compound references
are added for the comp ref prediction:
(LAST_FRAME, LAST2_FRAME),
(LAST_FRAME, GOLDEN_FRAME), and
(BWDREF_FRAME, ALTREF_FRAME).

This new tool of ext-comp-refs will eventually overwrite
one-side-compound and have the two coding tools to merge to one.

It achieves -0.35 ~ -0.55% coding gains in BDRate, compared against
AV1 baseline with the default experiments on, but without
one-sided-compound. It achieves -0.2% ~ -0.3% coding gains when
one-sided-compound is on. It achieves larger gains on higher
resolution.

Change-Id: Icbdb16e97b96aaebaf2213f5f72d5331e2e358eb

Reset xd->mi and x->mbmi_ext for the superblock after the first
scanning pass.

Change-Id: Iae9142ff2b1a2b576f54dc545b58fe37c97cecac

-0.242% gain on cam_lowres

Change-Id: Ib9ba2f7dfc7d10b8066bd3f70eeafa7faac8a607

This was initialized from a const and never modified. But was still
passed around and stored in multiple structs.

Removed these 'rddiv' member variables and now RDOPT() and RDOPT_DBL()
always use the const RDDIV_BITS directly.

Change-Id: I1a8dfd2c8fa857d466ad1207b4f0dd6ec07eafb8

This patch will not cause any performance change regardless of whether
speed_refs is on or off.

This coding tool is targeted to speed up the encoder side reference
frame selection process. The essential idea is to have two scanning
passes for each superblock of size 64x64 and this CL lays out the
initial framework but no reference frame selection is done yet:

First scanning pass - To simplify the partition and the mode
candidates (e.g. considering nearestmv / nearmv / zeromv only) and
identify the best reference frame prediction candidates;

Second scanning pass - Use the best reference frame candidate(s)
obtained from the first pass to encode the current superblock.

Change-Id: I11266d468de3077271a5e866eebd341a8014d136

Fix a bug where the encoder sets reference_mode =
COMPOUND_REFERENCE even if the frame contains some
4x4 blocks (but no larger blocks) that are single
reference. This occurs only in corner cases,
typically for very small frame sizes (e.g. 16x8).

BUG=aomedia:575

Change-Id: I5caedaa753a56e9f9b48995c1d2e0a0886aa4366

Now the single ref comp mode should work with WEDGE and
COMPOUND_SEGMENT. For motion_var, the OBMC_CAUSAL mode uses the 2nd
predictor if the neighboring block is single ref comp mode predicted.

This patch removes the mode of SR_NEAREST_NEWMV and leaves four
single ref comp modes in total:

SR_NEAREST_NEARMV
SR_NEAR_NEWMV
SR_ZERO_NEWMV
SR_NEW_NEWMV

Change-Id: If6140455771f0f1a3b947766eccf82f23cc6b67a

Daala-dist replaces the luma distortion of sub8x8 partitions with
its own distortion thus requires to split the luma distortion only.
Doing so, there has been a bug that INT_MAX64 value comes
when the sub8x8 parition is skipped. This happened because the existing
code does not initialize the rd_stats_y or tmp_rd_stats_y, i.e. rd_stat struct
for luma only in several places.

Change-Id: If229b53bb7a6cff0b8751138a32b1dcf02665624

This patch implements the post-encode and post-decode upscaling for the
frame superresolution experiment to work.

Upscaling happens after cdef and before loop restoration.

For now, this patch forces on random-superres.

The patch also cleans up some broken rate control hooks from VP9
days, to be brought back later when the resize and superres tools
are stable.

Change-Id: If0a8f69224dfaa0f4ae7703bd429ea2af953c7a6

Refactoring: split prediction+extension for each plane, so we can
handle luma/chroma supertx pred in different ways.
Compatibility fix: fix conflicts with cb4x4 and chroma_sub8x8, now
for chroma sub8x8 supertx, only the top-left(basic cb4x4) or the
the bottom-right(cb4x4 + chroma_sub8x8) predictor will be used
without any blending within a 8x8 unit.

Change-Id: I6cf7b12768a82d3c7e01811ada02de84af9bd8ac

Check the availability of the reference frames at the frame level at
both encoder and decoder, and if a reference frame is not available
for a specific video frame, remove the signaling of such reference
frame info at the block level.

This patch adds the consideration of the bit saving inside the RD
optimization loop.

Change-Id: I4c22f1b843b21c7d2b47e118c99c3ad615a3d4e4

This change does not impact the bitstream
  PSNR | PSNR Cb | PSNR Cr | PSNR HVS |   SSIM | MS SSIM | CIEDE 2000
0.0000 |  0.0000 |  0.0000 |   0.0000 | 0.0000 |  0.0000 |     0.0000

Change-Id: I6e131e91bad5efa345ed2542ae970eb6122eff51

This terminates the computation for the warp error once
the frame error exceeds the best frame error found
so far to avoid unneccessary computation.

Change-Id: I094a0b3e13f8b91610e051cb91d20a815879dd80

First keyframe BD-RATE objective-1-fast:
   PSNR | PSNR Cb | PSNR Cr | PSNR HVS |    SSIM | MS SSIM | CIEDE 2000
-0.3705 | -0.3232 | -0.3812 |  -0.3782 |     N/A | -0.3412 |        N/A

First keyframe BD-RATE twitch-1:
   PSNR | PSNR Cb | PSNR Cr | PSNR HVS |    SSIM | MS SSIM | CIEDE 2000
-0.2479 | -0.2477 | -0.2467 |  -0.2567 | -0.2486 | -0.2508 |    -0.2487



Change-Id: Iea6c895c6fe9e9764887a8968f6e5330903969d3

Remove the special rate-distortion optimization loop for sub8x8
block size from vp9.

Change-Id: I62c6cf537a54769f26f2d4938ebed5fed2c84741

The place where av1_daala_dist() is applied for sub8x8 partition is
moved from sub8x8 mode decision functions to rd_pick_partition().

BD-Rate change by daala-dist with '--disable-var-tx' is:
(AWCY, objective-1-fast, high delay mode)

   PSNR | PSNR Cb | PSNR Cr | PSNR HVS |    SSIM | MS SSIM | CIEDE 2000
15.1558 | 12.9585 | 14.4662 |  -3.8651 | -1.7102 | -9.2956 |    10.8686

In MSE probe mode:

  PSNR | PSNR Cb | PSNR Cr | PSNR HVS |   SSIM | MS SSIM | CIEDE 2000
0.0429 |  0.0435 |  0.1651 |  -0.0415 | 0.0850 |  0.0122 |     0.0546

Change-Id: I3b2ea916d41c48e433eb641adf44552e4725c198

Make the txfm_above and txfm_left be processed in the unit of
miniumum transform block size. Scale the transform block step
size with respect to the mode_info step size.

Change-Id: Iee4421e005db742cd4ff7899215560063e5f68e5

Resolve the enc/dec mismatch issue when filter intra is turned on,
in addition to chroma-sub8x8, ext-tx, and ext-inter experiments.

BUG=aomedia:570

Change-Id: Idf0875763ce24bd8d4022f7338fc86debf4dfd0c

BUG=aomedia:526

Change-Id: I5d9b86a36f412ded2d6f20e198d2f4de4f97aaeb

Change-Id: I05b2c8c116b5e3287be2703e840a28f6dda9c42a

Fixes AV1/AltRefForcedKeyTestLarge.Frame1IsKey test.

Change-Id: I457b5c1dacb9fd2294c650d11821d6a1930d04dc

Account for intra tx size case when coding block is in skip mode.

Change-Id: I030bc05d890c4a7c8d4628a7f9bdebaba6083a03

- Add av1_build_quantize() function so that quantization
  related unit test would be able to access quant/dequant
  table without starting an encoder instance and directly
  accessing cpi.

Change-Id: I8ba429e5deb7a4e7f967996aaec1d20bff0feb3c

Change-Id: I904283119d8f2c1099e6ec2953ea1c10c5e3b280

This commit resolves a rare enc/dec mismatch issue due to the
intracacy of using rectangular transform block size for intra
coded blocks.

BUG=aomedia:562

Change-Id: I10c52339a995a89bafa3a3330469ee3f2b784768

This patch forces DAALA_EC on by default and removes the dkbool coder.

Change-Id: Icd2ff08efd7bf467adf554344111473cb357adf8

This resolves a potential race issue in multi-threading test when
chroma-sub8x8 is enabled.

Change-Id: I443f4af10e55046b04d6d739f7594dd8bda5c830

This reverts commit 05c38874.

BUG=aomedia:553

Change-Id: Ibbfe3e357bbfb8cb7d07482755c47c64afd30092

Currently the "transform units" used to measure transform sizes
in some parts of the code are based on the smallest defined
transform size.

cb4x4 currently defines a 2x2 transform size, even when chroma_2x2
is not enabled, which means that the scale of the transform units
was always double that of MODEINFO units.

Several areas of the code were hard-coding this assumption instead
of converting from one to the other using appropriate constants.

Change-Id: Ibc55671aa5bc3ad272cb8a036f9c4f9621df85ab

reset_skip_context() was always clearing the entropy contexts for
all three color planes, using a block size that corresponded with
the luma plane.

However, when chroma_2x2 is disabled, then for sub-8x8 luma block
sizes, the corresponding chroma block size is always 4x4, and the
skip flag only affects the chroma blocks corresponding to the
upper-left luma block.

This patch makes reset_skip_context() reset the contexts that
actually correspond to the chroma blocks that are skipped (if any).
It also moves reset_skip_context() to av1_reset_skip_context() in
blockd.c, because blockd.h gets included before onyx_int.h, which
declares the required is_chroma_reference() function.
reset_skip_context() was too large and used in too many places to
be a reasonable candidate for inlining, anyway.

AWCY results on objective-1-fast:

cb4x4-fix-base@2017-05-11T06:26:50.159Z -> cb4x4-fix-reset_skip@2017-05-11T06:28:45.482Z
  PSNR | PSNR Cb | PSNR Cr | PSNR HVS |   SSIM | MS SSIM | CIEDE 2000
0.0301 |  0.1068 |  0.1463 |   0.0359 | 0.0260 |  0.0347 |     0.0479

A regression (near the noise range), but without this fix, the line
buffer size required by the entropy contexts will be doubled.

Change-Id: I12fa6e60d9c1c7c85927742775a346ea22b3193f

Create named constants for the minimum TX_SIZE used in the luma
plane, as well as the minimum allowed TX_SIZE for which we have to
explicitly code a selection (i.e., one larger than the minimum luma
TX_SIZE).

Then, use these constants consistently with tx_size_to_depth() to
index the tx_size counts in FRAME_COUNTS, and also consistently use
TX_SIZE named constants to index the tx_size_implied counts.

Failing to index these counts correctly can, e.g., cause tx_mode to
be chosen incorrectly.

Change-Id: I706a62a33e2282e67c97a68bade87fb8023ec13b

* set dependent tile to be default
* make dependent tile to work with supertx

Change-Id: I2bcc73c6e6df46ca0c2fbdae7c4f2944738fff79