Added more asserts for dist-8x8 running in PSNR mode,
i.e. with the encoder option "--enable-dist-8x8=1"
instead of --tune=[cdef-dist | daala-dist].

The asserts checks whether a 8x8 distortion measured on reconstructed 8x8 pixels
is identical to the sum of distortions from sub8x8 partitions (or tx blocks for
var-tx case).

Change-Id: I14f2b24e674a9cbbe60e663449fc4e7f46f0e481

Introduces two new TX type pruning modes that provide better
speed-quality trade-off compared to the existing ones. A shallow
neural network with one hidden layer trained separately for each
block size is used as a prediction model. The new modes differ in
thresholds applied to the output of the neural net, so that they
prune different number of TX types on average.

Owing to relatively low quality loss PRUNE_2D_ACCURATE is used
by default, regardless of speed settings. Starting with speed
setting of 3 we switch to PRUNE_2D_FAST mode to get better
speed-up.

Evaluation results:
----------------------------------------------------------
Prune mode | Avg. speed-up | Quality loss | Quality loss
           |(high bitrates)|   (lowres)   |   (midres)
----------------------------------------------------------
PRUNE_ONE  |     18.7%     |    0.396%    |    0.308%
----------------------------------------------------------
PRUNE_TWO  |     27.2%     |    0.439%    |    0.389%
----------------------------------------------------------
PRUNE_2D_  |     18.8%     |    0.032%    |    0.063%
ACCURATE   |               |              |
----------------------------------------------------------
PRUNE_2D_  |     33.3%     |    0.504%    |     ---
FAST       |               |              |

Change-Id: Ibd59f52eef493a499e529d824edad267daa65f9d

Removes some stray CONFIG_CB4X4 config macros.

BUG=aomedia:921

Change-Id: Icc65e0b000f659d7fb18178c928a7bff7879f58c

A frame should be referenced with a unique delta,
which is not guaranteed when diff_len is greater than frame_id_length.

Otherwise, it's possible to get into a situation where a frame
can be referenced with many different delta_frame_id values,
although the decoder only accepts one of them.

Avoid ambiguities when delta-referencing a frame, and
avoid wasting bits by using a coding enforcing the restriction.

Change-Id: If7c611c9b6ec69662c5af540ade59f08bacda2ba

Let the writing/parsing logic do the +7/+2 offseting,
and let the rest of the code directly manipulate the actual values.

Change-Id: I460d5bf18446c028f326cb747b3c26c8d451637c

Change-Id: Ib38368d0aa7e13e53e8040c75a8564fbdb7646c7

Change-Id: I5bff0a68602a89ce480fec049c8b2c4bce44f6bb

Support recursive tx block partition.

On the screen content testset, 0.2% gain for keyframe encoding.

Change-Id: I623e6fbb910fef9c91617e02edf420019f67d189

This variable is unused.

Change-Id: I8ee9c588898417e956e7cd7993ff54cbb885285e

- Missing forced disable of fp_mb_stats.

BUG=aomedia:856

Change-Id: If789c95b5101c592747c73b8b39138d7e62f2132

- Again.
- highbitdepth and rect_tx are now compatible with pvq.

BUG=aomedia:856

Change-Id: I806e99938f1149fa088ab5abd857f03f502ee320

BUG=aomedia:474

Change-Id: I42569ab97b5246de98310031e1d2398876283527

This avoid compiler warnings

Change-Id: I3977df12442532a699ded8f66c92422c30cd34a8

Fixes crashes on x86-win32-vs14 build

Change-Id: I045dd0fe4e9af3bfb80223e291617b717cbcb231

D207E
Predictor SSE2 vs C   AVX2 vs C
4x4       ~2.7x
4x8       ~3.0x
8x4       ~7.2x
8x8       ~8.5x
8x16      ~9.4x
16x8      ~12.8x
16x16     ~13.0x
16x32     ~14.3x
32x16                 ~19.9x
32x32                 ~23.6x

D63E
Predictor SSE2 vs C   AVX2 vs C
4x4       ~3.8x
4x8       ~4.3x
8x4       ~6.4x
8x8       ~6.8x
8x16      ~8.6x
16x8                  ~9.0x
16x16                 ~9.6x
16x32                 ~10.3x
32x16                 ~9.1x
32x32                 ~11.0x

Change-Id: I87373804c9d53276bf4d7788c4ae0d13d01c00dc

Change-Id: I3dcf3dfb849aadc9a9e9d06bb6cf22dfe2acf073

Put it in the correct part of 'configure'

Change-Id: I7731b13bbdbb61d2824d664d777b780d872bb1c0

Fixes a build failure when using 'make dist'

Change-Id: I8f63724678ce68328ef9ad2328e9a5015d6faf48

As per spec, if delta quant is signalled at the superblock
and at the segment level both should be allowed and the
deltas should be summed to get the correct quant.

BUG=aomedia:897

Change-Id: Ia07265c944c9969befa16f194e02bf4e7a8dd142

BUG=aomedia:907

Change-Id: Ibe367eb34596e2d34d8c059e083b083e702c225e

1. Reduce the worst-case number of samples from 34 to 16.
2. Simplify the trimming strategy to not use multiplication and
division operations.

Overall, this modification reduces the complexity without affecting
the coding gain too much.

Turning on "ext_warped_motion", the Borg test showed:
             avg_psnr ovr_psnr ssim
cam_lowres:  -0.912   -0.920  -1.246
lowres:      -0.229   -0.219  -0.373

Change-Id: Ida21d1d3f22998c851b1270fe61a5e30f9e4ec38

Those are leftovers from the abandonned supertx experiment.

Change-Id: If3ea7db9ea30c6be7c4955ae0a6a17730983ce66

When --enable-coefficient-range-checking isn't specified, clamp the
coefficients at each stage.

This doesn't change the decoder behavior for existing AV1 streams.
However, some AV1 bitstreams that would have been rejected by the
decoder as illegal (range check failure) are now legal bitstreams.

There is no impact on video quality.

BUG=aomedia:30

Change-Id: Ibcf1683e5c2ae9f91a7f37b468c4bc72e98e22fa

Change-Id: I9d3f0380865790d8adcb85f03305b193bc1949d7

This experiment has been adopted, we can simplify the code
by dropping the associated preprocessor conditionals.

Change-Id: I4705652ac8231ee25f960e8096dc97c2aeed1d20

This experiment uses switchable interpolation filter for intra
prediction. Current gain is too small to be useful for AV1.
May explore more for AV2 in the future.

Change-Id: I24ae8c75dc71adf6271eabdb68a81b68a0869fa2

Implement upscaler block by block to guarantee a fixed number of
output samples for each input block.

Change-Id: I302972d3711549b9e217ca2c052a55bd04d91dae

BUG=aomedia:901
Change-Id: I28165b5fbf712da02a609d8070a476f3b319a2c9

When the reconstructed coefficient is below the original coeff
in magnitude, skip the level down search for the given coefficient.
With fast-mode fix, performance change is within 0.01% level. The
overall clip encoding process speed is up by 10%-15%.

Change-Id: I93b68e494aa995faa3093fdfa1bda0bcb0e2564a

Tom Finegan authored Oct 12, 2017 and James Zern committed Oct 15, 2017

BUG=aomedia:733

Change-Id: I7044c55611a3ea9555347f9c4fef03e3042560e2

Reduce the needed memory space for motion field estimation.

Change-Id: I9caa228aa804a06370a34715a1c3463564d8d86b

Since BLOCK_4X16 and BLOCK_16X4 are listed with higher enum value than BLOCK_8X8.

Change-Id: I36cdd4d33dfb2f3a01fa03c8eb8be26d8055dc43

Yi Luo authored Oct 12, 2017 and James Zern committed Oct 14, 2017

BUG=aomedia:900

Change-Id: Ifb6c437d8d2f309d0717fb50c402618475fe021f

Debargha Mukherjee authored Oct 13, 2017 and James Zern committed Oct 14, 2017

Change-Id: Ie02e188662119ee2719c79283641b9b74849cfab

BUG=aomedia:909

Change-Id: I775c02df623a9ded5f04fcbbbb398f8e5d17a31a

Align all the motion vector storage buffer in 8x8 block resolution.

Change-Id: Iea786db95b2274332a755dc60f66510c068e3dc5

Change-Id: I4ce263b2731676e8620f0dbd1da2cd79bfded18c

Change-Id: If42a8cf9de9b4be14b4f228c5bf39747eefc168d

Also, simplify flow of control and avoid mutability.

Change-Id: I4e971b784b87c7240a83752499a1fbd48cfcb220

Support higher hardware throughput. The coding performance loss
as compared to 15-bit precision is 0.05% for lowres. The loss is
smaller as frame size goes up.

Change-Id: I2e22b156f3178cf63689df306e9da13e0e4d205b