GitLab

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: I09cf970a9be3a7dfe7e134bec5d778cda77f2682

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

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

BUG=aomedia:1065

Change-Id: I0951a276865a5d810eb04bbb5251ed5c1b417ca4

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

Remove COMPOUND_AVERAGE from compound_type_cdfs since it is now grouped
to compound_idx. However, COMPOUND_AVERAGE is still used elsewhere.

Change-Id: Ie0d460aabf9252e80eb4130cfef9aaf0efc3969d

Divide compound inter prediction modes into two groups:
Group A: jnt_comp, compound_average
Group B: interintra, compound_segment, wedge

Change-Id: I1142da2e3dfadf382d6b8183a87bde95119cf1b7

This patch fixes and enables rectangular intra transform
sizes for 4:1 partitions (that were turned off before).
4:1 partitions can now use rectangular intra predictions with
2:1 rectangular transform sizes.
BDRATE lowres (single keyframe): -0.612%

Change-Id: I6f062f7c08aae8eeb0a55d31e792c8f7e3f302a2

Modified the warped motion mode context based on neighbor's motion modes
and current block's mode.

Change-Id: I77ca35fab37ec640bb38661ff1799f643d5aafdc

Add a block level usage flag for skip mode. If no block has chosen the
skip mode, the frame level flag for skip mode will be set off.

This patch also includes a small code cleanup, including the check on
whether the best RD mode is aligned with skip mode, if yes, the best
RD mode will be replaced by skip mode.

This patch slightly improves the coding performance of ext-skip.

Change-Id: If06092d5e32f15e63dcb5f35d32e68bc0f827c2b

cfl.h is no longer included in the onyxc_int.h. cfl.h is now included in
each file that uses cfl functions.

Change-Id: I80a2e57159217b0ee20b2562fc9711d929472b8c

The CfL context is now stored inside MACROBLOCKD instead of
MACROBLOCKD only storing a pointer to the CfL context.
The intent is to avoid race conditions as MACROBLOCKD is stored
inside ThreadData. This change also simplifies CfL Initialization.

Change-Id: I991503716b21fc9aca60caddb2008b8bff397e6d

Current encoding of single ref assigns bit 0 to LAST_FRAME and bit 1
to LAST2_frame, whereas the encoding of compound ref assigns bit 1 to
LAST_FRAME and bit 0 to LAST_FRAME. This patch unifies the design and
makes the binary symbol assignment for LAST/LAST2 consistent
regardless of the single/compound ref scenarios.

This patch incurs a small syntax change but should not have noticeable
coding performance impact. Once BUG 973 is fully addressed, default
cdfs for reference frame coding will be updated.

BUG=aomedia:973

Change-Id: I1f3027965c0bb2d343bc8ad873f7c0015123f151

BUG=aomedia:78

Change-Id: Id85cf22302317d33b593345fd31090a6620f4101

This experiment has been abandonned for AV1.

Change-Id: Ib9a8eea6b0ad5580dc81836be8fc5e898bb9bb45

Larger than 8 bit input seems to introduce different rounding methods in
different distortion calculation functions, which causes assertion failed
for dist-8x8's sanity check.

BUG=aomedia:1027

Change-Id: I1f607f32d4017d8f9abbbe0b09d9dbb01a63fe9c

When the flag DISABLE_VARTX_FOR_CHROMA is on chroma is
constrained to always use the largest transform size
for the prediction unit size.
This is meant to simnplify the logic for transform size
selection for chroma with hopefully no loss.

Results:
lowres 30 frames, speed 1: -0.038% (a slight improvement).
lowres 30 frames, speed 0: 0.000% (noise level difference).

Change-Id: I14dd5b1983d908bd98e59b7d252e11f5755c97e6

The change can be temporarily reverted by setting ALLOW_128X32_BLOCKS
to 1 (for easy "is this better?" testing: we'll get rid of the option
once it's clear what we're doing).

Since this means we have to modify the "num_partition_types"
calculation in functions like read_partition, we move it into a helper
function called partition_cdf_length(). Doing this makes the bodies of
read_partition and write_partition somewhat simpler than they were.

Change-Id: I3573f62f71dc3344aed2dcb3e423e034eb56a7c5

Change-Id: I9854d5ec193dadaa455c209b1482ead895982cb7

temporal segment map prediction from LAST frame instead of previous

Change-Id: I2f803581c2a6ff72ae428783e596aa3281492073

Use 4x2 processing unit.
Reduce # of modes from 6 to 5.

Change-Id: I3c12e18084636de0e279c9102a8b212342faf4c7

This was missing before. Hopefully it will improve costing.

Change-Id: Id7dbe963fe5e26798b853419804bdd128537a5d9

Change-Id: Icca39f1d037a3aca4540e35b70fdfafeae2b094e

In TX size search the rate and distortion values are often computed for
residuals that have already been seen before, with exactly the same
entropy context. Such repeated computations occur when the residue
signal is the same for 2 or more partitions for a given prediction mode.
By saving previous RD search results we can achieve some encoder
speed-up at the cost of higher RAM consumption (~20MB).

Test results with all default experiments:
Speedup      15% (measured on 4 lowres 30-frame sequences)
compression   0%

Change-Id: I8d9b8dd483cf862f0864b50ae80466beac978290

A new block mode, referred to as skip_mode is added. If a block is
coded as skip_mode, it will be inter-coded, with its references and
motion vectors derived from its neighboring blocks with zero-residue.
Otherwise, the block can be coded in the current intra or inter mode.

The computational load on skip_mode evaluation at the encoder should
be kept minimum. No transform size / type evaluations are needed.

Change-Id: I5aef0159c7d5ecd64258510835903375d6c536d6

Change-Id: Ifa2cdc2d2230dfa11396ee3e547653180f96b795

In large scale tile coding(namely, large_scale_tile = 1), forced
all frames to generate a bit-exact uncompressed frame header.
This patch modified parameters that could change from one frame to
another.

Change-Id: Ibe72519da0b8a4f5a4ef30a4303ad7d7e4992a65

Update cdf only when it is allowed.

Change-Id: I95ab08ff4220abdd277a7cd110b69c82364dddcb

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

Change-Id: I6ac62c2825eabcba8f854cfa25c84638d9a73872

This matches the code in read_partition in decodeframe.c and almost
gets rid of the last usage of the "partition_prob" tables which can
now be deleted.

Change-Id: Ibe0e947007d593d8e5f4ab10a38cd920856e566d

This experiment has been abandonned for AV1.

Change-Id: I3501181d096873dec35ea58e4b55530fa76e683f

Remove the previous experiment and now use the same name for a
simpler experiment that only enables 4:1 transforms for 4:1
partitions when ext_partition_types is on, and that which was
previously enabled with the USE_RECT_TX_EXT macro.

Change-Id: Iccc35744bd292abf3c187da6f23b787692d50296

This is still a rather inefficient black+white encoder, since it carefully
computes some chroma data, but just doesn't write it. However, at least the
bitstream is now monochrome.

Change-Id: Ie8a89bf329e7b41441032fb0d9e9011385bc12ff

I think the result is a little easier to reason about (you now talk
about a property of the block, rather than the behaviour that should
be enforced). It also matches the code in read_partition in
decodeframe.c

Change-Id: I13ba06b1504fa153b8b6b60fa14b373483639718

Change-Id: I96e5ff72caee8935efb7535afa3a534175bc425c

This reverts commit 060e192b.

Change-Id: I5700d351a3cbb682ec49a0efb9cca4d0e83f9a3a

This experiment has been abandonned for AV1.

Change-Id: I07be0876b21ee20a1eb2490bb98e4f84978ff001