GitLab

Get a list of palette base colors that are used in the above and
left blocks, referred to as "color cache". For each cache color,
signal if it is present in current block's palette, so that we
don't need to transmit their raw values.

When palette-delta-encoding is enabled, compression is improved
by 2% on keyframe and 1% overall for the screen_content testset.

Change-Id: I4cb027f1904aa9d0ab1c8f00ea9ee34bf5f16234

BUG=aomedia:529

Change-Id: I9fccd0d29d100c92152d33a74dc3df8b7d256bcb

This commit moves the av1_scale_if_required functions from
av1/encoder/encoder to av1/common/resize. They take the
scale_and_extend_frame_nonnormative function with them.

This will be useful for resizing received frames in the decoder for the
FRAME_SUPERRES experiment.

Change-Id: I83afc3d97ff8ebce5ecb2e1205ba6c271156cb31

BUG=aomedia:121

Change-Id: Iec2e98c9f80ef3a01476234e637c135ff0513efd

This avoids repeatedly computing the frame error in
refine_integerized_param.

Change-Id: If4af77b68a7e61a700bafa3f60940b9017e12b40

Change-Id: Ie5718de1eb70fd753c2bfd7df9581aa92e13984c

Change-Id: I9a9d7960b46dd80012fc8efdefb640939e1f1e48

This avoids use of preprocessor when possible, and removes a
duplicated legacy check on codec name.

Change-Id: I8f0d25cf69b51d3379f25c1a69f88c8d404c4e38

When the input bitstream is 8 bit, the output YUV should default to
8 bit, even if the internal storage format is 16-bit.

Change-Id: I77eea8a78a60411cfd707c7a4605365abc508200

Change-Id: I3e3d458aec7e26abc366992ac20f9fa838fe35bb

Move cfl_idx_to_alpha in the header to facilitate inlining.
Remove the forward MB_MODE_INFO forward declaration

Change-Id: Id33fb0228d88b6285252843e2345a0d3ae875cd2

Change-Id: I9d771edccd90c72ada72942e11cf3fbabb8146a9

Merge cfl_alpha_sign_u and cfl_alpha_sign_v into cfl_alpha_sign,
where:
    cfl_alpha_sign = (cfl_alpha_sign_v << 1) + cfl_alpha_sign_u;

Change-Id: I8c491e37d23f6ef85a33a8b2f8770b7b606d8e0a

To avoid writting out uninitialized signs for CfL alpha, checks are
added in the write_cfl_alphas function.

Change-Id: I832875e496a6318132c82bc1e5406dc95ae9c055

BUG=aomedia:533

Change-Id: Iffd43f859f958ce2ccae80e3616c368ffebfc8db

Use CONFIG_INTERINTRA to properly separate interintra from the basic
ext_inter experiment.
When macro REDUCE_INTERINTRA_MODES is 1, only 4 ii modes are enabled
so as to reduce the complexity overhead.
(Right now the flag is off)

Change-Id: Iec6f36a1000f181adbb822ad66c7d5b2625e9cc7

This unhandled case was causing an assertion failure.

BUG=aomedia:532

Change-Id: I46665c380ad171e2b9ddf4d41b6e9fff838f8b3e

Change-Id: Id14d83bdce8321bef81972a4f5530a50589d9578

Change-Id: I1c00925f83c6a858b0e799ddd90f241570a40575

Change-Id: I325f781759c1087815a38ff7c12ea31b8e444575

BUG=aomedia:530

Change-Id: I6889a9b7ba05785ea6d4c684e0e12dcc85fdca7c

For 'left' array indices -1 onwards are valid (left[-1] == above[-1] ==
the top-left reference pixel).
So, make sure that 'base2' is in that range too.

Change-Id: If5118d97d4ef85f6520363bd648bfe6948c1211f

Change-Id: I175d6ff03f31a2e0d2fe7cd1c3852210d6e0ddf5

Change-Id: I4269d5b808e8809199a24b7511a3dee029f31820

Previously, the projected positions of chroma pixels would effectively
undergo double rounding, since we round both when calculating x4 / y4
and when calculating the filter index. Further, the two roundings
were different: x4 / y4 used ROUND_POWER_OF_TWO_SIGNED, whereas
the filter index uses ROUND_POWER_OF_TWO.

It is slightly more accurate (and faster) to replace the first
rounding by a shift; this is motivated by the fact that
ROUND_POWER_OF_TWO(x >> a, b) == ROUND_POWER_OF_TWO(x, a + b)

Change-Id: Ia52b05745168d0aeb05f0af4c75ff33eee791d82

- Function level improvement (ms):
Functions       ssse3  avx2   Percentage
idct32x32_1024  794    374    52.9%
idct32x32_135   354    169    52.2%
idct32x32_34    197    142    27.9%
idct32x32_1     n/a     26    n/a

- Integrating in default scan order.

Change-Id: I84815112b26b8a8cb800281a1cfb1706342af57d

Add an SSE4 version of compute_cross_correlation() from
corner_match.c. This function is about 3.4x the speed of
the scalar code; determine_correspondence as a whole is about
2.5-3x the speed it was previously.

BUG=aomedia:487

Change-Id: I707b7cfd5c513c025d3ee7fb6a5f1fa335ecd495

As the block sizes are powers of two, we can index into the weights
array as sm_weights_array[bs] now.

This uses 2 * MAX_BLOCK_DIM memory, instead of NUM_BLOCK_DIMS *
MAX_BLOCK_DIM earlier.

Change-Id: I55bcedc188b8ed7def719c4d002c1fe2ec5e1b7f

Change-Id: I88606ad4e0841fb1ccd18a0e63338723157451e5

Change-Id: I405b4fca33d7ee6f6010874bde59e1e25877a401

This function will be re-used for lv_map's rd optimization

Change-Id: If015fe2319f6b3abdeeb3e5d3e23f705fc1454e2

EC_ADAPT stores the count at the last position. As such, the CDF_SIZE
macro adds 1 to the array size. This resolves valgrind warnings about
jumps on uninitialized variables.

Change-Id: I58d607a57756cb166a9ae5c4565537ce202183fd

Variable needed quoting. Without the quoting the error was somewhat
cryptic.

Change-Id: I76dd9715b9a91436a2b10f8cf7fcbd70ae672d23

BUG=https://bugs.chromium.org/p/aomedia/issues/detail?id=76

Change-Id: I6b51219ff28758b38379033dc3ab11bf288ac3ce

- CONFIG_EXT_TX 0 => 1
- Add CONFIG_SMOOTH_HV

BUG=https://bugs.chromium.org/p/aomedia/issues/detail?id=76

Change-Id: I44f42e6cda5433d6d2d0993778091308900debbb

BUG=https://bugs.chromium.org/p/aomedia/issues/detail?id=76

Change-Id: I3db6022d361ce5e8b3db3d49b5e71b4367c355a1

BUG=https://bugs.chromium.org/p/aomedia/issues/detail?id=76

Change-Id: Ia5622b531d17bb1c67fe4f4bc01e57dd3560021e

Introduced in Change-Id: If9d98e49cee63f40085b179652bfacc31ff462ea

BUG=aomedia:527

Change-Id: Idfa609dc05a5981d83fa5d77c8c51b02bdf9abef

Change-Id: I8f67a222c27d12fc43cdc2b0f568a3c429901801

Change-Id: I253a3dfdc4329c19fd8c9aca32971c62cea864b3