GitLab

AV1 encoder's logic for handling current_delta_lf_from_base value has
a hole, which could cause enc/dec mismatch due to deblocking filter
level mismatch.

On the encoder side, when delta_lf is used. current_delta_lf_from_base
value is only written into the bit stream when the following condition
is true:

((bsize != BLOCK_LARGEST || skip == 0) && super_block_upper_left)

i.e., it is only written into the bit stream when current mi is the top
left corner of a super block, and it is not skipped or block size is not
64x64. this is applied to both INTRA and INTER blocks in write_mbmi_b().

And the following code is used to handle the case when the above condition
is not true.

if (!dry_run) {
    mbmi = &xd->mi[0]->mbmi;
    if (bsize == BLOCK_64X64 && mbmi->skip == 1 && is_inter_block(mbmi) &&
        cpi->common.delta_lf_present_flag) {
      mbmi->current_delta_lf_from_base = xd->prev_delta_lf_from_base;
    }

The idea is that if the above condition is not true, then the
current_delta_lf_from_base should be reset to the previous delta lf,
which is prev_delta_lf_frame_base.

However, there is a hole in this logic. For INTRA 64x64 blocks, even it
is skipped, the current_delta_lf_frame_base is not reset back to the
previous delta_lf. So, in this case, decoder side won't receive any new
delta_lf from bit stream, so it will use previous delta_lf. encoder side
will use the new delta_lf value. this mismatch will cause final loop
filter level calculation inconsistent in the deblocking stage and thus
cause encoder and decoder mismatch.

BUG=aomedia:826

Change-Id: I90c1dcca5da106e4f3e0bcba1ed539c1cd99c370

This experiment offset the filter tile grid 8 pixels upwards.
Deblocked pixels (rather than CDEFed pixels) are used for the
2 lines above and below the filter processing unit. The 8 pixel
offset is the offset produced by deblock/cdef. This way the
loop_restoration does not need additional line buffers in a
single pass hardware implementation.

Change-Id: I89e0831dc28413a5d3e02d7a426ce2885ab629d7

They're unused, disabled in the prototype setup, but still built and
complicating the already convoluted ifdef mess in TX experiment
configuration.

Don't leave dead code in the sourcebase.  That's what SCM is for.

Change-Id: Idb2adf597ac064c7b5027df8af1cf65054984aa4

Function speedup sse2 v. C
Predictor  V_PRED  H_PRED
4x8        ~1.7x   ~1.8x
8x4        ~1.8x   ~2.2x
8x16       ~1.5x   ~1.4x
16x8       ~1.9x   ~1.3x
16x32      ~1.6x   ~1.4x
32x16      ~2.0x   ~1.9x

This patch disables speed tests to save Jenkins build
time. Developer can manually enable them by using,
--gtest_also_run_disabled_test flag in test command line.

Change-Id: I81eaee5e8afc55275c7507c99774f78cc9e49f9a

Change-Id: I6d75d3955bd8dd2218fb07b36c10361e032eb66e

$GIT_EXECUTABLE was being passed as the
value of CMAKE_INSTALL_PREFIX to pkg_config.cmake. Pass
$CMAKE_INSTALL_PREFIX as intended.

BUG=aomedia:814

Change-Id: I86126f472447eaa6a2196dca3a8c68fb4de9415a

To facilitate HBD support, we extract the loop that builds the
prediction from cfl_predict_block.

This change does not alter the bitstream.

Change-Id: Ie88d1c5872a90303b7fa0692fed8c2855e1d315f

Pixels are subsampled when they are stored in the CfL prediction
buffer. This allows to avoid having two buffers. No impact on the
bitstream.

Results on subset1 (Compared to parent)

  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

https://arewecompressedyet.com/?job=cfl-avg-on-fly%402017-09-23T21%3A38%3A04.440Z&job=cfl-Sub-On-Store%402017-09-24T15%3A01%3A41.161Z

Change-Id: If051e34d6d7078d77609542305a2afebef5cb177

This is just a code cleanup, moving the calculation of cm->tile_rows
and cm->tile_cols into get_tile_size, which is already calculating
cm->tile_width and cm->tile_height. The patch also gets rid of a
spurious AOMMIN and replaces it with an assertion.

Change-Id: I46666c4a197ac26d4b3746d9ea6575dc4af0d570

This patch slightly reorders the fields in yv12_buffer_config and then
uses anonymous unions in order to make it possible to write code that
iterates uniformly over planes.

The patch also ports some code (mostly in yv12extend.c and
aom_scale.c) to show how this can make things more concise.

This should make no difference to the coded results. I think it's
unlikely to have any significant performance impact (the reordered
fields in a yv12_buffer_config only come to 17*4 = 68 bytes in total,
so almost fit in a normal sized cache line).

Change-Id: Iebb46344500b9df82915f34cfd193e189d712062

BUG:aomedia:824
Change-Id: I67bfd16a9ba994055ccbaf624c58117c7939d5c8

CDFs are used for coding and costs, so there is no need to
base the CDF on the obsolete tree and remap filter indices.

No change to common test conditions BDR.

Change-Id: I441dd21b127bc07ba33eef8ddcfd147eef21d0d4

This change enables using 13 taps for luma plane deblocking and 5 taps for
chroma plane deblocking when pixels are in flat area.

The aim for the experiment is to make sure that luma line 57 and chroma
line 29 of the current superblock is not changed by the deblocking process
of the superblock below. Previously this was already the case for luma
line 56 and chroma line 28 (but not for 57 and 29).

This experiment is part of an effort to reduce the overall line buffer
size for DEBLOCK+CDEF+LR. With this change it is possible to CDEF line
-8 to +55 direcly on the output of deblock (which require line +56 and
+57 to be final).

Change-Id: I7779a08d6ad5683bf35c3372b1526786eaac8472

This has been cleared by Tapas.

Change-Id: Ie62ffe84d2b285ae83939e7bcc354c2964fb5938

Also, print the actual out-of-bounds values,
instead of just printing the input buffer.

Change-Id: Ia6b73d98e41c3285d8fd9f39963f9d556d2d95f7

Change-Id: Ic1aa7dfd4da4e9e29318665d5ecff7ea75fd41e9

Re-enables most of the previously disabled tests.
The ones that are still disabled expect resize to be triggered
through rate control, which is no longer supported in Av1.

Change-Id: Ie5e9ba3eb0843cd44ff1ac988500081470ba0fe2

Instead of storing the transform block average, it is immediately
subtracted from the subsampled pixels. This change does not alter the
bitstream and it reduces CfL complexity.

Change-Id: Ia5038b336abf1ec01e295b235734318906d3bae6

Change-Id: Ia33e9962fce962e4f45281537d3e90b4d88543a5

With convolve round, tmp_buf32 should be set to tmp_buf rather than
tmp_dst because we want the actual pointer value rather than the funky
rescaled one when the frame is high bit depth.

Change-Id: I614ca1ebb4b6db8c4957fdced9dc2009a7ac107e

This fixes global motion read/write mismatch.
Minimal test case would be 3 encoded frames with:
    1) frame0: KEY frame
    2) frame1: INTER frame with refresh_frame_flags not 0x00 and
not 0xFF, say 0x01 (refresh first slot)
    3) frame2: INTER frame with all reference indices equal to 3
(i.e. no reference to frame1)

When decoding frame2 decoder will use cm->prev_frame=frame1
for gm params decoding, then switch to cm->prev_frame=frame0
for the rest of the frame.
This makes frame2 dependent on frame1 which is not used as reference.
And also it is different from what encoder does.

BUG=aomedia:750

Change-Id: I42ba19c03d909286af12ba3149910693a12249e8

When using uv_mode the corresponding enum is the UV_PREDICTION_MODE not
the PREDICTION_MODE enum.

Change-Id: I44b19e4db5947a2d44281430b34ed61b82f999eb

The buffer width and height are updated after adding to padding data to
the CfL buffer. This way, padding is not repeated. This change does not
alter the bitstream

Change-Id: I7e709fc4e18b03a516b1345099b334c9d33d71aa

Change-Id: I612517c6218c561ee94888c8c14298964851484a

Change-Id: I24d129aceb66e00388bedf98bc8e46524c0812eb

Enable the RESTRICT_COMPRESSED_HDR experiment when
NEW_MULTISYMBOL is on and apply when there is no
backward update also.

Change-Id: I5f6812df2cb3f9bc32fe533b2689fd51235aedbc

and to CONFIG_LIST + CMDLINE_SELECT. the experiment list should now
match cmake

Change-Id: I724d35b3a16ed5c641207525949093bc0caf4a49

Change-Id: I8080a860c32d465810a960bf25f4299c40122937

Add function get_binary_prob_new in aom_entropy_optimizer.c.
This function add "+1" to both 0 and 1 counters such that it won't
generate extreme probability when both counters are small

Change-Id: I759d6e1cfedaf1c36281b523c26e7a7d1cad147c

When deciding whether the top-right or bottom-left blocks are
available, we currently always act as if we're using 128x128
superblocks. This means that, when using 64x64 superblocks,
we sometimes conclude that blocks are available when they haven't
been decoded yet!

This typically happens at, for example, mi_row=15, mi_col=16
(for bottom left), where we're at a 64x64 boundary but not
a 128x128 boundary.

This patch fixes the issue by checking based on the signalled
superblock size.

Note: Most of this patch is just threading 'cm' through the
intra prediction process, so that we have access to cm->sb_size
in has_top_right() and has_bottom_left()

Change-Id: I126964c510aafffc870e7cd8b3e64a46abb14b3a

This patch is designed to complete the CONFIG_MAX_TILE experiment.
It adds the following features:
- Explicit tiles of same width can be signalled with command line --tile_width
- Explicit tiles of same hieght can be signalled with command line --tile_height
- max_tile now works with the depenedent horizontal tiles experiment
- additional checking added to ensure maximum number of rows/columns <= 64
- one fix to the tile group code for when the number of tiles is not a power of two

Change-Id: I749a68fdb3dcdd8d1ced2b3558f3817e4b832c06

inside the encode_super_block function, the tx_size decision logic is
different in the dry_run==0 and dry_run==1 case when segment is
lossless. this change fix the inconsistence.

BUG=aomedia:819

Change-Id: I2f4c5434d8c4ef722ac885a1ffd2d253863ae6d9

This fixes decoder's capability to start decoding
bitstream from non-KEY intra_only frame.

BUG=aomedia:816

Change-Id: Iecd5add418590119f59a51acd7208eb08cb17042

Add lowbd unit test functionality to intrapred_test.cc
Function speedup against C (i7-6700):
Predictor   DC     LEFT   TOP    128
4x8        ~1.4x  ~1.4x  ~1.7x  ~1.9x
8x4        ~1.2x  ~1.6x  ~1.6x  ~2.6x
8x16       ~1.4x  ~1.3x  ~1.4x  ~2.1x
16x8       ~2.0x  ~1.8x  ~2.3x  ~2.1x
16x32      ~2.0x  ~1.9x  ~1.8x  ~2.2x
32x16      ~2.0x  ~2.0x  ~1.9x  ~2.2x

Change-Id: I33db512020ca3c6853a9205a8079f3d00134f584

Since Subsampled values are in Q3, one does not need to take the
average, as the sum of 4 values is the average in Q2. As such, shifting
by 1 to the left, results in the average in Q3. By removing the
intermediate step in Q0, rounding error is removed.

Results on Subset1
   PSNR | PSNR Cb | PSNR Cr | PSNR HVS |    SSIM | MS SSIM | CIEDE 2000
-0.0204 | -0.1945 | -0.1474 |  -0.0170 | -0.0237 | -0.0178 |    -0.0973

https://arewecompressedyet.com/?job=cfl-baseline%402017-09-06T17%3A41%3A38.041Z&job=cfl-NoAverageInSub%402017-09-06T17%3A59%3A00.035Z

Change-Id: I8c6d4f71ec0e6e3923e254d79b83127b3734699e

Result from luma subsampling is left-shifted by 3. This avoids having to
do it during averaging, in alpha search and when building the
prediction. This change does not alter the bitstream.

Results on Subset1
  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

https://arewecompressedyet.com/?job=cfl-baseline%402017-09-06T17%3A41%3A38.041Z&job=cfl-SubsampleQ3%402017-09-06T17%3A42%3A01.252Z

Change-Id: I6e89eac6496f7c36e46364c9223fbcbca6759032

Exactly one of tmp_dst/tmp_dst_/tmp_dst2 is used in a call to
av1_make_masked_inter_predictor. Work out how many bytes are needed by
the largest one possible and just allocate that.

Change-Id: Ib8fabaf1f461ce1192c5657a3d2b5a4f7fa8d7b1

When calculating the above overlapped prediction, we adjust
xd->mb_to_bottom_edge to have the correct value for the prediction
we're about to make. This prediction is half the height of the
current block - *unless* the block is 128 pixels high, in which
case the prediction is only 32 pixels high.

This patch implements the special case for 128-pixel high blocks.

(all of this applies analogously for left predictions too)

Change-Id: I2780e8d4a6f8a9f82f540f2c2dfde691d07298e2

Change-Id: Id6f588876b1adf68abd8c17d7eef2223d552e457

BUG=aomedia:812

Change-Id: I635d620fc6d350391ecb7885cc119eaef0af39bc