Fix for the following warning:
Implicit conversion loses integer precision: 'unsigned long' to 'int'

Change-Id: I4d4453401eca1136b0cbf21b630893f242d7bb55

Change-Id: I034558404d054b295908f7f45225facfb0cb15ee

Adds spatial prediction to standard segmentation instead of a separate
segmentation feature.
When using temporal prediction skipped blocks are flagged as mispredicted.

Change-Id: I0c32281286d3fbce66d339c9247bcc6516f37a63

Change-Id: I0f5592587798e9c19b3d145aa6cbe712bf3b0631

Linfeng Zhang authored Dec 05, 2017 and James Zern committed Dec 09, 2017

Change-Id: I11e57afd02b355e29be3ec4e216d2c6910120467

Linfeng Zhang authored Dec 05, 2017 and James Zern committed Dec 09, 2017

Extract the shared 2 neighbors { 0, 1 } and { 1, 0 }.

Change-Id: Ie1b52cb62519b53bfb5388af71ea637eab3fccc7

Linfeng Zhang authored Dec 05, 2017 and James Zern committed Dec 09, 2017

This is a preparation of following get_nz_map_ctx() refactoring.

Change-Id: I60d71863f39d8f271c44abee5855fd4d89b544b1

Change-Id: I687453ab28d72412e49b4cc2676b70425b80848b

Change-Id: Iae7c054def0c0d5b5af52263f0c4d2e1e346282d

For each superblock, select the best deblocking filter level and apply
filtering. The filter level is signaled to decoder using a delta based
scheme.

Change-Id: I53e32589cabac9e2a4e580808fdd39ac878fe8c6

Rather than having a set of parameters for each of the 6
COEF_BANDS, we have 1 for DC and 1 for AC coefficients.
No change in performance since all of the bands had the
same parameters.

Change-Id: I3665e7c1b21f117be776f371d87d64b097715735

Also, reduce scope of one local.

Change-Id: I41cb53528d4b7bc88eb343d8c943ed241230af82

With shift, convolve copy no longer needs 32-bit multiplication of
two 8-bit numbers. Thus we can implement it with sse2 instead of
sse4.

Change-Id: I63e8ba414383a24f820bad4a6c607f222ec40ec2

Change-Id: I72ae23a60f79256b207753c429c3fecf4db6bd38

Currently 4:1 transforms have max 2 split levels:
4:1 -> 1:1 -> 0.5:0.5.

This refactor enables split levels:
4:1 -> 2:1 -> 1:1,

by simply changing the tables in common_data.h.

The actual switch will be made in a subsequent patch.

Change-Id: I33f8d9ca5159ba3e7d02ced449ddf6f804a8f12a

Change-Id: Ic741f269d0bd5e5e295b55f95bfef05050bc31e5

Slightly change the way we save and reload frame contexts during
frame setup. For "normal" frames everything is the same, but for
error-resilient and/or intra-only frames, we now:

* Reset the frame context using setup_past_independence()
  (+ extra code if q-adapt-probs is enabled), as usual
* Store this frame context into a special slot in cm->frame_contexts
* Use that slot to fill in cm->pre_fc

The main difference from before is that (for error-resilient/intra-only
frames which are not key frames) we used to throw away the frame
context after setting it up, and would re-use whatever was set up
at the last keyframe.
This was fine when q_adapt_probs is disabled, but it caused an
inconsistency when combined with q_adapt_probs. See the attached
bug report for more details on that.

BUG=aomedia:1104

Change-Id: I9532b6b0e8ae29efbb4f059a0c67a73d7c7828ce

subset-1:

daala_tx@2017-12-07T22:33:52.954Z -> new_dct32@2017-12-07T22:34:37.310Z

   PSNR | PSNR Cb | PSNR Cr | PSNR HVS |    SSIM | MS SSIM | CIEDE 2000
-0.0073 | -0.0284 | -0.1499 |  -0.0561 | -0.0128 | -0.0279 |    -0.0386

Change-Id: Ib01f21aa9fc3c95f9d820331b43e70410de99958

Limit the prediction residual hash table access within the same
tile. This resolves a data race issue across tiles in multi-threading
encoding that triggers instable encoding results.

BUG=aomedia:1088

Change-Id: Ia4a415a0750bd60ee8ac4e56aa1cd39ec99e19c7

Change-Id: I09cf970a9be3a7dfe7e134bec5d778cda77f2682

Exit early on large prediction error. This reduces the overhead of the
tx skip prediction, especially for low quantizers.

BUG=aomedia:1106

Change-Id: Icad8e01cdeb2e8f4cf0befa7f5a89e088f3c17e5

This simplifies the profile lookup table and removes the requirement
of computing the ctx for new-quant. This change produces a
0.01% drop in performance on 40 frames of lowres.

Change-Id: Ifcbd7d6e427090e02bb43f6de15e551796aa74c4

Linfeng Zhang authored Dec 04, 2017 and James Zern committed Dec 07, 2017

Make a couple of functions inlined, and other small changes.

Change-Id: Icd11b20f0fa2ddb14621df401f95cf0ce2f6be28

In addition to k-means, consider using the dominant colors directly.

Improve keyframe by about 1% on the screen_content testset.

Change-Id: I08a932c322cfe36fb8def778d14f96d71c1017db

This experiment has been abandonned for AV1.

Change-Id: I3a6dfd2178aff2502d0ea7e9dd6d584b4db0eb1f

Check the bottom neighbor availability with tile boundary for the
intra prediction condition.

BUG=aomedia:1088

Change-Id: I9baa98f8f18da84f95fd83ceca5556cfe9d9d844

Change-Id: Ia790b7f12cf78f628e99bb7fe2e272771c1a68c7

This experiment allows two choices for the frame-level reference
mode flag: SINGLE_REFERENCE and REFERENCE_MODE_SELECT. It removes
the choice of COMPOUND_REFERENCE, as it has been barely used.

Change-Id: I8af18acd2fe3c0d4928f2b05f35aad0ebcb1556a

Changed convolve 1d function names to be more precise.

Change-Id: I9bc1e99f5032ed8a84c61616036735af23d9a587

BUG=aomedia:1065

Change-Id: I35eda84e249fc42e319604079c122df8ab101f90

BUG=aomedia:1065

Change-Id: I0951a276865a5d810eb04bbb5251ed5c1b417ca4

BUG=aomedia:1096

Change-Id: Ibbb448e217ae1dd9a096c23d01b08c9804583003

This is to keep the high range within 32 bits.

AV1FwdTxfm2d.CfgTest passes after this fix.

Change-Id: I2df463c4ec9260c544d68aad445b60cabe2b531b

The purpose of this change is to reduce the cycles needed for warped
motion parameter estimation.

Method 1:
If we remove the 2-bit bit-depth reduction(as in patch set 2), the
downshifting of A, Bx, By is also removed. The borg test result(over
the baseline) is:
             avg_psnr ovr_psnr  ssim
lowres:      0.023     0.020    0.071
cam_lowres: -0.009    -0.017   -0.031

Method 2:
In theory, the above change uses 2 more bits for elements of A, Bx,
By. In patchset 3, we modified LS_STEP to be 8(1 full pixel), and now,
the least 2 bits in A, Bx, By elements are always 0. Namely, 2-bit
bit-depth reduction are achieved without extra operations. The borg
test result(over the baseline) is:
lowres:     -0.004    -0.007   -0.023
cam_lowres: -0.031    -0.033   -0.045
This is a little better than patch set 2 result.

Method 2 is the final choice.

Change-Id: I945aaba412e2ea86b7d67e8a90741fdf395b94cd

Change-Id: Ia8321afd087f99371cdf07f3a03249580e09964d

This patch simplies the checking criteria for the two groups of
compound modes. It also makes the encoder side cdf update inside the
RD loop consistent with that in the bitstream.

Experimental results on Google test sets (30 frames of lowres and
midres) confirm this patch obtains identical coding performance.

Change-Id: I170eea91f7d2be2170df544cfc2c692b09aa82d6

Fix the comments on the precision of quantizers and tx coefficients
during a quantization process for different input depth and tx size.

I think the author really meant "de-quantized/de-coded coefficients" by
"quantized/coded coefficients". So, made it clear to avoid any possible
misunderstandings.

Change-Id: Ib92ac7dcfddcbe58cf3adfb9448497512381c1f5

Load four 8-bit input and process.

Change-Id: I9b3ba58ea3a03c6a8129379afa37c54a57e04501

Also, make them const when appropriate.

Change-Id: I96d544e2cc9a0bce4d52fd33e44a4eaa40edda3c

Can be up to >10% faster with bit exact results

Change-Id: I5f169673fd2d5af96f425f00d862f3c989228d2e