GitLab

This addesses a crash in a situation where the scaled convolve
function is called without a valid dst buffer in the conv_params
structure.

Change-Id: Ia4a0a1213f06447155d6c92aa9efc183d8c4a79c

This flag enables an extra bit in the frame header to indicate the
signaling choice of the reference frames in the frame header.

Change-Id: I67ef0cbb0673bf877189236b5060f2a854a05d11

Change-Id: I24e98062fec2cf73e294d34bf02419f7917a9bf0

Fix the rate cost count when the optimize_b route is set off in
the rate-distortion optimization process. Thanks to pohhsu@microsoft
for finding this issue and pointing the fix.

Change-Id: I54d9ec8102792a1575fc66bc6be2003e94eeb488

When a coding block runs VERT_A partition, the coding order does
not follow raster order. This requires a special handle on the
bottom left square block to disable its reference towards the
top-right corner. Prior to this change the codebase would disable
the bottom-right square, as well as the right rectangular blocks,
from referencing the top-right mv. This commit fixes such check
condition to allow the right rectangular block to access the top
right mv.

BUG=aomedia:1343

Change-Id: I87049f0cec8ed7557a87c3fdef83e01498bbcd75

Remove deprecated context models for drl index coding.

BUG=aomedia:1208

Change-Id: If255fa93d0c746738f0fc005464388e790c89b63

Change-Id: I40b7672ed98f2d1a49a7a9aabfd313ba64b3e8fb

The 4x2 transforms gives a compile warning with gcc 6.3.0 but appears
to be unused:

*((void *)&temp2+8)' is used uninitialized in this function

Change-Id: I8b08e05d0365dc117b5374ec00bddc6f7bd84bd3

Also some minor cosmetic changes and style fixes.

BUG=aomedia:1125

Change-Id: Ide24057736376accd2f9907ef87b5a725ab3f3c2

Change-Id: I103547d5bd4c9537cc0640a10d50f5c3cb6b7771

Make the convolve_2d_sr_avx2 to support various bit shift options.
Turn back on the SIMD functions for av1_convolve_2d_sr.

Change-Id: I38b29eee40831c8c88e574884205bc0fdf886916

Make the convolve_2d_sr_sse2 support all the bit shift tuning
settings in CONFIG_LOWPRECISION_BLEND and off.

Change-Id: I497105d9d7f0a0f9c5edcac55b366825ec13a2ed

Let the cdf update order match the actual coding order

Change-Id: Iaac0830b2530228ce1a4ea95d842c573dd3d0684

av1_read_coeffs_txb() is sped up by 10%

The overall decoder is sped up by 1%

The speed performance is tested by runing city_cif.y4m on speed1
with bitrate 1000

BUG=aomedia:1369

Change-Id: I0de5402a88fbb3ea46905a23b32627eba681a250

When the CMAKE_SYSTEM_PROCESSOR is aarch64, AOM_TARGET_CPU
is set to arm64.

BUG=aomedia:1385

Change-Id: I35676a3ad8a90dc814fe8acb2a9b56fbf60b3700

Some fixes and additions of asserts.

Change-Id: Ib5f706baf48acb2853f7fe55991efa95e6a848de

Change-Id: I474cd6d75074eb908e8fcf317c2524e3c7f72269

Avoid reading/writing out side of buffer. Triggered by ASAN.

Change-Id: I7de2a9f01cc13feb1c13556dfe77e9e6e7e55056

Although the clamping for the inverse transforms is still being
discussed, it is useful to have clear code.

The current code is not ideal - both av1_inv_txfm1d.c and av1_inv_txfm2d.c
have functions called clamp_buf(), but they do different things - one
clamps to the exact number of bits passed in, the other optionally
clamps to 16 bits.

This commit refactors things so that:

* clamp_* always clamps to the exact number of bits passed in
  To compensate, we push the "if bit <= 16: clamp to 16 bits"
  logic into the callers of the clamp_* functions

* We deduplicate some of the code

BUG=aomedia:1352

Change-Id: I1b9d85d6db5d291814c055a8113b7b054690b04c

Change from 2D to 1D to reduce number of tables.

Change-Id: I6569a757afae5a748e4d84df687468fe460eaa5a

Provide full width and height in functions' names.

Change-Id: I2439ddbce0ad37084547ece9f62c7d6ce12d87f9

Change-Id: Ie290cd3512a3cc20b56815048e3470bd9d78abb5

This macro is same as FILTER_BITS and the two can be safely merged.

Change-Id: Idbc7e6492c14070aa6eb054ce3c0a30bc89d41be

Change-Id: I2c77dd79c47901f8038ad7531679c3acc7b70dfe

Reduce the blk_skip array size to the maximum txb numbers in a
coding block. This makes the speed 1 about 5% faster for smaller
resolution clips.

Change-Id: I1812dc4d403e08f0566ead8a14c6046e505bb849

The maximum coding block size is 128 and the minimum tx size is
4. Using 32 per dimension to keep the txb coeff cost should be
safe.

Change-Id: Ie44fc581037f0d8270caec64543454701159eec5

Change-Id: Iadf001ae293ca50b805e8b8c569900f60a23943c

Commit ea190906 fixed a major bug in has_top_right(), but missed
out one special case:

Consider a 128x128 block using 64x64 intra predictions. These
intra predictions are applied in a 'Z' order, and so the bottom-left
64x64 unit has pixels available from the top-right one. But we were
mistakenly setting has_top_right() = 0 for this case.

More generally, whichever transform unit has its top-right corner
at the center of a 128x128 block, should have has_top_right() = 1.

Fix this by introducing an explicit check for the special case.

BUG=aomedia:1309

Change-Id: I690a292be6c1755c76bd428be94ab953dd71fbd2

The experiment is fully adopted.

Change-Id: I6cc80a2acf0c93c13b0e36e6f4a2378fe5ce33c3

Discontinue all VP9 style convolve rounding operations in the non-normative
parts of the encoder.

The function av1_convolve_2d_sr_c is forced instead of SIMD versions
of the same function, because of incompatibility when round_1 > 0.

In the -DCONFIG_LOWPRECISION_BLEND=2 -DCONFIG_HIGHPRECISION_INTBUF=1
setting, results on 15 frames of lowres (cpu-used=1) is -0.019% better.

Change-Id: I72154bd896357c352c944fb2cd3b25bafafba46a

When bit_depth is 8 the copy_rect function was setting the size to
2 bytes per sample. This causes a buffer overflow as each line copied
in the loop is twice the number of bytes it should be and the last
line writes off the end of the buffer.

BUG=aomedia:1389

Change-Id: Ib9fa11d1dd13806dedbce2cd47dd8d562007428d

Superres should not be used together with intrabc, for the same
reasons why the other loop filters are disabled when intrabc is
used. However, we actually read superres information first,
so the simplest change is to only read allow_intrabc if superres is off.

BUG=aomedia:1384

Change-Id: I55c4c8ace72ab3fac9417da5cf803883be7efb84

BUG=aomedia:1383

Change-Id: I8c3de068fd158b4706b578f7609c8ef939364525

Includes unit tests for conformance and speed.

SSSE2/CFLSubAvgTest
4x4: C time = 234 us, SIMD time = 152 us (~1.5x)
8x8: C time = 664 us, SIMD time = 208 us (~3.2x)
16x16: C time = 1687 us, SIMD time = 581 us (~2.9x)
32x32: C time = 6118 us, SIMD time = 2119 us (~2.9x)

AVX2/CFLSubAvgTest
4x4: C time = 250 us, SIMD time = 221 us (~1.1x)
8x8: C time = 683 us, SIMD time = 284 us (~2.4x)
16x16: C time = 1727 us, SIMD time = 1091 us (~1.6x)
32x32: C time = 6092 us, SIMD time = 2107 us (~2.9x)

Change-Id: I44ffedc683829d2c16089854ac43d4ddb4415bcd

BUG=aomedia:1366
BUG=aomedia:1368

Change-Id: I63f84dca86ca426b9c6927b056657741022d5f68

Move tcoeff reconstruction from the for loop of reading lower
levels to the loop of reading signs

This is to facilitate the potential simd optimization of
reading lower levels

Change-Id: I61e41ea95d30a948243fb13e3d708dd9860a058a

As pointed out by rsbultje, my previous patch to is_smooth()
(a883e6ea) was not quite correct. This is because, when we're
making a chroma prediction, the uv_mode for the above/left chroma
predictions is not necessarily in above_mbmi/left_mbmi. Instead,
it may be in any of several places, depending on subsampling and
the values of mi_row/mi_col.

The cleanest solution is to explicitly maintain pointers to the
above and left chroma blocks. Then we can simply look at those
pointers when we want to know the above or left uv_mode.

Also include a bit of refactoring of get_filt_type: It seems
to be recalculating what's already in xd->{above,left}_mi,
so just use those directly.

BUG=aomedia:1362

Change-Id: I0230474a50d43b78cb587a2b553da9ca78cec0c6

Fixed a bug in jnt_comp simd function.
Re-write the function based on existing av1_warp_affine_ssse3.
The function is almost the same as av1_warp_affine_ssse3, only
difference is the weighted computation of final value.

Change-Id: I65f2fadc9142f6c264a7d0e59250602636c9808b

Replace the last parameter cos_bit for all 1D inv_txfm funcions with
a macro define, as it is actually always equal to 12. By changing it
to const value, compiler can do further optimization.

Change-Id: If8a9fd99c7ac7eb6f485dafbce22b4803efda62e

Change-Id: I9ab260c5ca31fe7e06bfc0f806893463c5255c45