GitLab

Uses a better non-normative scaler to scale reference frames.

Change-Id: I44b689c8328d6e6fd467c70c9b6b789874818153

Use the uniform way to compue the cost of symbols in AV1.

Results on Subset1 (compared to 8a516a8f with CfL enabled)

   PSNR | PSNR Cb | PSNR Cr | PSNR HVS |    SSIM | MS SSIM | CIEDE 2000
-0.0357 | -0.0854 |  0.0305 |  -0.0422 | -0.0097 | -0.0171 |    -0.1042

Change-Id: Ie908fc7d20c480634002c78027b070223b3ea96d

Alpha's biggest fraction is 1/8, so Q3 does not change the bitstream.

Results on Subset1 (compared to 503aca74 with CfL enabled)

  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: I1fe5b2ace97179d5f950d7406a4f3d391924f89d

Change-Id: Id8789a2c352a7c7940e5cabff39acd0d7a584cac

Set LOG_SWITCHABLE_FILTERS accordingly to USE_EXTRA_FILTER

Change-Id: I453fe201a66ec725061899337b17d517c57b9e7a

Change-Id: Idfcc4e19143a45e0729829301acb1e3b711faff0

The allow_lowbitdepth shouldn't be exported to the API at all.
This is a temporary hack to allow API users such as the analyzer,
inspector, and simple_decoder.c to function correctly.

Change-Id: Id171cc0c24d78216647e949a1dc30a81f4cc826d

Change-Id: Ief1fc77294397c32121dc666e09a9a802ff5b7d4

Should always use AOM_ICDF() macro for compatibility with multiple
symbol coding related experiments.
This commit has no effect on metrics.

Change-Id: Iab2c8f2a1234a36ebd1ec4307079d5344214608b

A python script is provided to accumulate stats from counts.stt
files generated from multiple encoding examples. Command line:
python ../../tools/aggregate_entropy_stats.py [dir of stats files]
 [keyword of filenames] [filename of final stats]

Plus cosmetics on output format of optimized prob tables. The comma
after the last element of each dimension is removed, i.e.,
before: {
          { 128, 128, 128, },
          { 128, 128, 128, },
        }
after:  {
          { 128, 128, 128 },
          { 128, 128, 128 }
        }.
The reason is to avoid moving '}' to a second line after applying
clang-format.

Change-Id: I3b764f4c89b7aefbaf48ac00b8c83713caf75a5d

Change-Id: I9bc925ab9b93f5318eb3f1aa3401ff3de2851fef

Change-Id: If7e9ef6509a1be821484f081152bec47ea55cb00

Change-Id: Ibb3a2c8c855fffe6fd75a94da09425d784709431

Change-Id: I9fe592d6fc4bd2cffea8cbc5ac80138f28182ef8

Removing these flags make the EC_ADAPT experiment an integral part of
 the draft AV1 bitstream definition
This commit has no effect on metrics.

Change-Id: Ice78520935e8bfa9d25cf4b8384a1b872069d09c

BUG=aomedia:629

Change-Id: I929e6e1db8491092f3e165d0990f9e10920fb2d3

AWCY results, 1 frame, tune-content=1

objective-1-fast:
PSNR (Y, Cb, Cr) : -0.26%, -0.24%, -0.26%
PSNR-HVS         : -0.26%
MS-SSIM          : -0.25%

screenshots:
PSNR (Y, Cb, Cr) :  -2.28%, -2.29%, -2.29%
PSNR-HVS         :  -2.30%
MS-SSIM          :  -2.26%

twitch-1 :
PSNR (Y, Cb, Cr) : -0.73%, -0.73%, -0.73%
PSNR-HVS         : -0.75%
MS-SSIM          : -0.74%

Change-Id: I7aac475710ab3895778d7ff05d9113f5713133b1

Remove deprecated tree coding when EC_ADAPT is on.

AWCY, objective-1-fast, High Latency:

PSNR (Y, Cb, Cr) : -0.10%, -0.40%, -0.37%
SSIM: -0.11%
PSNR-HVS: -0.07%
MS-SSIM: -0.06%

Change-Id: Ic2d436e037312e1af238a8586983f1288a9a95a8

Use eax instead of rax.

BUG=aomedia:546

Change-Id: Ie3f704557d428d84531311a6bc690293465a622c

Change-Id: I696effc72117b24739947be80ae298de8c6c0e2a

Change-Id: Iea00482779cd3fdc7f354d5c39684dd0752d0849

AWCY Results
High latency -0.18%
Low latency -0.28%

Change-Id: Iaa69eb6a8f039448e5772cf85bec82f51038c8b2

Change-Id: I9c8201c7053dba11e9ac0f02c1bffbb4a322e654

Here we have an LGT to replace ADST for intra residual blocks, and
another LGT to replace ADST for inter residual blocks. The changes
are only applied to transform length 4 and 8, and only for the
lowbitdepth path.

lowres: -0.18%

Change-Id: Iadc1e02b53e3756b44f74ca648cfa8b0e8ca7af4

The block level DC_PRED computed by CfL goes down from Q6 to Q0. This
will allow to reuse existing assembly for DC_PRED and also reduce the
requirements on the multilpy required to scale the reconstructed luma
values

Results on Subset1 (compared to f9684d222 with CfL enabled)

   PSNR | PSNR Cb | PSNR Cr | PSNR HVS |    SSIM | MS SSIM | CIEDE 2000
-0.0347 |  0.0229 | -0.1326 |  -0.0420 | -0.0057 | -0.0072 |    -0.0644

Change-Id: I6ba82cc9e04fa4ab7c8ec40a7856deb273881748

Match cos_bit precision actually used in the decoder (max: 13).

Change-Id: I609f59a9653d74b569420bb0d402ba4f5352aa76

Change-Id: I2833fa31f9e76c412d62538796e6931ff86f9842

Change-Id: I114a963838295fcbf1ea7687cf6e5431976ee942

Change-Id: I6534f21631b0f5cf66034c4ced9ef2db408b996c

BUG=aomedia:630

Change-Id: Idf39c702dfee4fe35f2b0e768cf60f27ed7db08b

- Add utility function for appending link flags to targets.
- Use it to set linker only opts for emscripten inspect target.
- Update README.md with emscripten instructions.
- Build everything with -O3 when no build type is specified.

Change-Id: I69a2083369d2880335a8162132f190377a4a85fb

Use a variable instead of an array.
Make the horizontal function the same structure as the vertical function.

Change-Id: Iabfebe823ed9641cda0b5a0973c4dad4ecaef1a7

Regression from commit 299ff04f

Change-Id: Ib2ab5d5f831b4805d5054cb19f8ae176232067ea

Use higher precision offsets for more accurate predictor
generation when references are at a different scale from
the coded frame.

Change-Id: I4c2c0ec67fa4824273cb3bd072211f41ac7802e8

This is intended to be a no-op when scaling is not
enabled, but is expected to result in more accurate
prediction when references need to be scaled.

However note all xs, yx, subpel_x and subpel_y values
are now at higher than 1/16th precision.

Change-Id: I4b22573ea290a31fc58ead980bb0d5e5a9e89243

Change-Id: I640173fecea7b97a13257d07e09862714cb88c14

Change-Id: I7d720a755f44f0c0094c9262d3e22ac0fd0edba5

This sets the cmake build config to match default paramters in the
 normal make build system.

Change-Id: I66c6dca1ad7433d90d6f92ab1bc0e568f38c1f23

In the case where:
* global-motion and ext-inter are both enabled
* cb4x4 is disabled
* We have an 8x8 macroblock which is partitioned into smaller blocks

the function is_nontrans_global_motion has a bug where it will
always act as if that macroblock does not use global motion. This
patch fixes that logic bug.

Change-Id: I712d7bab6d646726d97aa37630399b4bcbee7757

This fixes the test failure since af7846ec.

Change-Id: I0905a473841e546b13c30702219efbb0b8b092c6