GitLab

While building the decoder-only AV1, the unit tests need to be the ones
that only call decoder functions.

BUG=aomedia:395

Change-Id: Iac7b464aa222a177c06b2e037faa6717305cd59d

Handles a rare divisin by 0 case.
Also adds a check on global motion parameters to disable
if the parameters obtained are outside the range that the
shear supports. This fixes a rare assert failure.
Also changes the recode loop threshold somewhat.

Change-Id: I4c6e74b914ac653cd9caa0563d78b0a19a2a8627

SSE2 may be needed for nx4 and 4xn.

Change-Id: I3c10112447fdb5fe51a68bc2c6e2f2641b102723

high-latency, cpu=0:

  PSNR | PSNR Cb | PSNR Cr | PSNR HVS |    SSIM | MS SSIM | CIEDE 2000
0.0378 |  0.1946 |  0.1385 |  -0.1159 | -0.2058 | -0.2085 |     0.1353

low-latency, cpu=0:

  PSNR | PSNR Cb | PSNR Cr | PSNR HVS |    SSIM | MS SSIM | CIEDE 2000
0.2388 |  0.2234 |  0.3290 |   0.0623 | -0.1716 | -0.1704 |     0.2542

low-latency, cpu=4:

  PSNR | PSNR Cb | PSNR Cr | PSNR HVS |    SSIM | MS SSIM | CIEDE 2000
0.4089 |  0.3477 |  0.6132 |   0.1729 | -0.1905 | -0.1610 |     0.5522

Change-Id: I35b8596667d82a127847b209416ad83e3b839a9a

Change-Id: I56a35b0d3d76294cc7b3d601770f7dcef12a8bc9

Only blend with the first N neighbors at each side. If the size of
one dimenstion is 8/16/32/64, the max # of neighbors to overlap
with is 1/2/3/4.
Previously we disable obmc mode if there are too many neighbors.

Change of performance in AWCY, compared to disabling obmc if
at any side there are more than 2 overlappable neighbors.
HL improved by 0.02%
LL improved by 0.09%

Change-Id: I93d9a65c6c4aabf0b4a4946e2253d3e2ef21a662

Change-Id: I1d3a7dcf8be1fc8f8fc6a70a5660a8c68c60b5ea

Change-Id: Ibaaed850ddceba9c3db542eaf4a1c623ce6b412b

This reverts commit c538d681.

Reason for revert: should be av1_fwd_txfm(), instead of fwd_txfm()

Change-Id: I58657e54eb0a1c20c930b32cd53b6d05493eb8f4

Requires new cmake toolchain file arm64-linux-gcc.cmake:
$ cmake path/to/aom
-DCMAKE_TOOLCHAIN_FILE=path/to/aom/build/cmake/toolchains/arm64-linux-gcc.cmake

Above requires aarch64-linux-gnu-gcc toolchain is installed and its
bin dir added to PATH.

BUG=https://bugs.chromium.org/p/aomedia/issues/detail?id=76

Change-Id: I6c7e297a67415008583b7ac2cd8aea528787e6d0

Requires new cmake toolchain file  armv7-linux-gcc.cmake:
$ cmake path/to/aom -DCMAKE_TOOLCHAIN_FILE=path/to/aom/build/cmake/toolchains/armv7-linux-gcc.cmake

Above requires arm-linux-gnueabihf-gcc toolchain is installed and its
bin dir added to PATH.

BUG=https://bugs.chromium.org/p/aomedia/issues/detail?id=76

Change-Id: Id2d0022fd5b4615fef7ea0ef38980cbdfd957c0e

The forward transforms required for PVQ are only performed if the block
is not skipped.

Change-Id: I22de3c7855ed800578a60aa06992cebea8bda119

Change-Id: I99bb8a849a632334359441f43705e93f8455f4cb

Change-Id: I66cdb67f8a1c2072516a65822dcc838e516ba9d7

Change in this commit is suggesed by Tim Terriberry.

BUG=aomedia:413

Change-Id: I4f79fd30d65d984ea4a1fe069bb715937856721f

Change-Id: I219a0ca3af7a0f6b03391ce500ea097e6106d06e

Improves and simplifies costing for global motion to use the
actual bits used to communicate the global motion parameters.
Removes some of the old hacks.
This patch also includes necessary refactoring to incorporate
reference based coding of the motion parameters to follow.

lowres results: -1.489% (up from -1.333%)

Change-Id: I994dc97046011de4261633ccb0c9d71e374f2c5a

Properly map the mode_info offset to pixel offset in cdef to
support cb4x4 mode.

Change-Id: I1f876387cb327cbf6c5689a4bc5ac41a1b1b0f34

Explicitly use block size to decide between functions for 4x4 and
8x8 blocks.

Change-Id: I0248b42a5b97441261a6331ec00dbd60610c79c8

Replace the hard coded numbers with mi_size definitions. This
would support scaling the mode information size. No coding stats
are changed.

Change-Id: I4bd63ad0b254a1c198258e7bc68e610067c982c4

Add missing sources:
- binary_codes_{reader,writer}.{c,h}

Change-Id: I4eda322a4c07a61f9da7370fda3163f726ab531c

BUG=https://bugs.chromium.org/p/aomedia/issues/detail?id=76

Change-Id: Ia3f09540787b90688656ffe23ec455379102c7fc

Change-Id: I2290eea6dc0b57f8bd3416ceae52ed75c24dd6ab

Change-Id: Ia5b2c851d43e889b753933590f36dae61cf236d8

Replace hard coded numbers to support the 4x4 coding block
structure.

Change-Id: I3602e09c8cfae9765069848828b534d177a75774

Fix the calls to be compatible with the new interface introduced in a
previous commit#77368af8.

Change-Id: I6944c189d2d53978250df9dee460a2ba9b8ccc83

Change-Id: Ie18fd2b8a3caf3948748ee353fe41e37f5803ba3

This patch also makes it easier to tune the mapping later.

Change-Id: I69f49b9dd1fddb306931599a354526abe0232eb8

Fixes the nightly tests.

Change-Id: Ifbd83f452337c80914b359a4e3e78b00a159146c

Change-Id: I8cd35f4697bc5ee90cfcb1fd37042770d6b5a544

Code the extra_bits using up to 5 non-adaptive symbols,
starting from the LSB. The number of skipped bits is
truncated to fit into the multi-symbol groups.

Change-Id: I147b5b0c3815bd39c338ee9a4779638ae75d7ab5

Change-Id: Ief5df3d5b1b81f09190d34022a3cb7d500992da2

Replace multiple calls for partition_cdf with a unified variable
in ec-multisymbol. This siliences a compiler warning in the
ec-multisymbol.

Change-Id: I56b07579e89e8fba565c4960d081817c86bc0097

Replace the use of num_4x4_log2 with the actual block size as the
condition to check if a block is partitionable. This avoids conflict
issues when multiple experiments are turned on.

Change-Id: I666d52458590a6f88c3995a140d20b0d66d27bfb

A bit and a three bit literal have been added to the uncompressed frame
header for use by the frame superres experiment.

The bit is true if scaling is to be used, and is immediately followed by
a three bit literal that encodes the scale to use. If the first bit is
false, scaling is disabled and the scale factor numerator is set to the
denominator (ie. 1:1 scaling). No literal follows if scaling is
disabled.

The denominator has been defined as a constant 16. The literal is biased
by a defined constant of 8 - allowing fractions from 1/2 to 15/16
scaling in steps of 1/16 when scaling is used. Experimentation will be
needed to discover which of these are useful.

The bit and literal are immediately after the optional render_width and
render_height, so that the superres parameters can be written and read
just after the regular width and height, without interfering with the
render_size parameters.

This patch also adds an arbitrary write to make the scale 1:1, so as to
not trigger any scaling until it's ready.

Accompanying encode and decode helper functions are added.

Change-Id: I8caa6247c73f5c7f84ef1fde1e80eb9b20bde0e3

Change-Id: Id2bcdad731c27a71cca29205d37c781a3f99bb3d

Change-Id: Ieec83d0b067ec36c35878aee02eeba436ede14a5

Fix an assertion fail in rd_pick_intra_sub_8x8_y_subblock_mode().

BUG=aomedia:410

Change-Id: I22d4aaea0566336605230e09d9590bc235719685

Change-Id: I9780e4e4f736b92b676f8a03a445a88749d4022e

1) Add txb_entropy_ctx into MACROBLOCK_PLANE and PICK_MODE_CONTEXT

2) Add av1_get_txb_entropy_context() to compute the entropy context

3) Compute and sore the entropy context before av1_xform_quant()
return

Change-Id: Ia2170523af3163b9456f7c6a305c1e77ad2b23be