GitLab

Move computing the y_mode_cdf tables per coded intra mode symbol to
 computing them only when the probabilities are updated.

Change-Id: I8c43d09b8ef5febe2a3ec64bd51d28bd78ea73ed

Move computing the kf_y_mode_cdf tables per coded intra mode symbol to
 computing them only when the probabilities are updated.

Change-Id: I5999447050c2f7d5dbccde80bee05ecd1c5440ab

Change-Id: I90ef3b49b499c2ac9c24797467cb4eb194fdf23b

av1_clpf_frame() was always called with the same src and dst,
so we only need one argument and the code supporting different
src and dst was removed.

Change-Id: I70919f50e5cfb19c22eb4dff9ee7c0fa2697fad3

Change-Id: Ia44ea1798c3347bba2f59a0a1d6765852b783e2f

Instead of having CLPF write to an entire new frame and
copy the result back into the original frame, make the
filter able to work in-place by keeping a buffer of size
frame_width*filter_block_size and delay the write-back
by one filter_block_size row.

This reduces the cycles spent in the filter to ~75%.

Change-Id: I78ca74380c45492daa8935d08d766851edb5fbc1

This patch adds bit account infrastructure to the bit reader API.
When configured with --enable-accounting, every bit reader API
function records the number of bits necessary to decoding a symbol.
Accounting symbol entries are collected in global accounting data
structure, that can be used to understand exactly where bits are
spent (http://aomanalyzer.org). The data structure is cleared and
reused each frame to reduce memory usage. When configured without
--enable-accounting, bit accounting does not incur any runtime
overhead.

All aom_read_xxx functions now have an additional string parameter
that specifies the symbol name. By default, the ACCT_STR macro is
used (which expands to __func__). For more precise accounting,
these should be replaced with more descriptive names.

Change-Id: Ia2e1343cb842c9391b12b77272587dfbe307a56d

Purpose is to allow parsing of bitstream after a packet loss.

Change-Id: I7d1c010a465aa18914762cc1a3e61db377304c08

Simplify the code logic and make the codebase easier to maintain.

Change-Id: I867060614209f3d82426bb97e2e4d7d16db2097b

Make TX_4x4 as the minimum selectable transform size form bit-stream
syntax perspective. Use value 0 as the minimum supported transform
size unit for the block coding.

Change-Id: Ieb3167216a96aaddb0810a424ccaaf8ae896fec5

Change-Id: I4a6d84007bc17b89cfd8d8f2440bf2968505bd6a

Change-Id: Ia09edf92bf9f7ecacc65c232ac6e656cde236634

Move computing the segmentation_probs.tree_cdf table per symbol to
 computing it only when the probabilities are updated.

Change-Id: I3826418094bbaca4ded87de5ff04d4b27c85e35a

Use lookup table to replace the arithmetic computation for transform
block step.

Change-Id: Ie7cb33e75e3b90aa9e9b46f8d11fa1d03e1ebe85

Use the table access to retrieve pixel numbers from tx_size.

Change-Id: Ibf392dcfdf80606b6ab972d60550ed75cbe652db

Palette code originally by huisu@. Gives 11.68% BDRate improvement on
screen content set.

Last 7 small speedup patches from me, giving 10% speedup overall for
screen content set.

From aomedia/master:
Revert 57fa6267 : Remove color_index_map

From libvpx/nextgenv2:
c93e5cc3 : Bring palette back to nextgenv2
827e1b3f : Add test for screen content coding tools in end to end test
78b0bd0a : Complete (mostly) migration of palette mode
8a128c2a : Fixes for Palette mode
4ab00912 : Palette mode: record selected transform type
cbb8be76 : Set block size upper bound for Palette mode
bb0e6921 : Convert palette from double to float.
a0a23b7f : faster code: replace nested for loops by memcpy().
016a5daa : Palette code: simpler and faster duplicate removal
d0000208 : Palette code cleanup
6dde8018 : Use built-in qsort(); create remove_dup() method.
f746c103 : Handle centroid rounding inside palette.c itself.
d68c7b6d : Palette: count Y colors only for screen content.
f1906e96 : Palette code: remove the use of same if condition twice.

Locally made sure all combinations of "--enable-ext-intra" and/or
"--enable-palette" and/or "--enable-aom-highbitdepth" build OK.

P.S. Note for future reference:
EXT_INTRA experiment has only been moved *partially* to aomedia at this
point. If and when the rest is moved to aomedia, EXT_INTRA + PALETTE
combo needs to be moved to aomedia as well.

Change-Id: I8826780014ec8a88281e56c8258b7c966121b46f

Change-Id: Ic3a6855799be010e69aeab924b013679282ab191

Dead code.

Change-Id: I13ed4007865e45e41fc79dfb63cbd01500a1d3ae

Cherry-pick vp10 fd96aec9: Don't send segment probability updates when
 the map isn't updated.
This patch only effects --enable-misc-fixes.

Change-Id: Ia1acfc59e4f2e5383a26c786bdd90c02ac49df9f

Change-Id: Ifc065eb3eee0960be2e085e7e18478bf91bf56dd

Cherry picked from libvpx: 511da8cb

Change-Id: I4f4132235495adcd445690d8f060f607b755cacd

Move computing the partition_cdf tables per symbol to
 computing them only when the probabilities are updated.

Change-Id: I442f9230ba00be7f5d0558d7c38d7324ad009ee8

Move computing the inter_ext_tx_cdf tables per symbol to
 computing them only when the probabilities are updated.

Change-Id: I5e1e62f8eae8f6b2edbbd378beeb786649502c10

Move computing the intra_ext_tx_cdf tables per symbol to
 computing them only when the probabilities are updated.

Change-Id: I26d5e419e103093e98a7d896c196176305b50fc9

While we are at it:
- Rename some variables to more meaningful names
- Reuse some common consts from a header instead of redefining them.

Change-Id: I110491c071ae6f68378030493679d49fe670129a

- Change struct name to all caps SCAN_ORDER to be locally consistent.
- Rename struct pointers to 'scan_order' instead of hard to read short
  names 'so' and 'sc'.

Change-Id: Ia131f14320c1a18aa12654cdb87bf8008878950a

Move from computing the switchable_interp_cdf per symbol to
 computing once per frame when the probabilities are adapted.

Change-Id: I6571126239f0327e22bb09ee8bad94114291683e

For now this is in its own experiment for benchmarking purposes, but it
will eventually be folded into the main ANS experiment.

Change-Id: Icdb54b6ca3c0191ba717d6d10aaaffb11e2bebea

Change-Id: I371297e6ee000e6dc01ba1544763cbed429b0e5a

Remove the separate implementations of set_contexts() in encoder
and decoder.

Change-Id: I9f6e9b075532faae0f74f885d9443589254258a7

Change-Id: I61eb08862a101df74a6b65ece459833401e81117

Change-Id: I14e66bb18959a3d78d8befb9858c192901e0406b

Change-Id: I3ba5f7061f7a76a76e5c906132f33d4110a0dd87

Following supports are added in the decoder:
(1) Decoder support to show_existing_frame == 1;
(2) Decoder support to avoid the referring to non-reference frames;
(3) Decoder support to handle the case when cm->prev_frame gets retired
    in the reference buffer.

Change-Id: Ibe8590da8db26e18bcf149b9f155db71636ba890

For the experiment of EXT_REFS, add the handling of the context
probabilities for both forward reference frames (namely
LAST/LAST2/LAST3/GOLDEN) and backward reference frames
(namely BWDREF/ALTREF).

Change-Id: Ibd991673acf6ff2fc9a56b8805eca9173cbf5fdc

The decision framework in enc (without obmc mv refinement) and
reconstruction at both sides is implemented.

Performance gain: 2.229/2.000/1.915% lowres/midres/hdres
Encodinig time: +13%

Change-Id: I5ab634c17d7050f15ecf5d4145626d0eaa425058

Change-Id: I5d30e8f07373482d7d7a4aff5a2026e4314e32b1

PSNR gain
hbd lowres dataset: 0.811%
lbd lowres dataset: 0.895%

Change-Id: Id6cc426883de725adf1a630f4151d5018c4b819b

Change-Id: I5d36aba6870f0783577b80d46defda7604029382

This commit makes the encoder and decoder use the same motion
compensated prediction functions to simplify the codebase.

Change-Id: Ic492700bd003aec658e95cb39bcc2ba60bc6e2e4