GitLab

Change-Id: I13eed9cb6950ea4fbdd586d43b73ac0cc2d78d33

Change-Id: Ib1088c3fc80952074e098385fe5eb81742e7dc59

Change-Id: I2f4e3fc877c03a5bee7f7fd1dc50e6a693697647

av1/common/allcommon.h
doc.mk

Change-Id: I7e08c9131ab1c0d7e7854f7e70b90397d041143a

Change-Id: I32b6e2e6c8454ffb64e4a4ceb87070d175f05fe9

Change-Id: I086b79462b0933cf9dc1101ff71cbc71c7da2738

enums for BLOCK_SIZE, TX_SIZE and PREDICTION_MODE.

Note: These were converted to #defines earlier to save on memory:
https://chromium-review.googlesource.com/#/c/269854/

But we, instead, use attribute 'packed' (see here:
https://gcc.gnu.org/onlinedocs/gcc/Common-Type-Attributes.html#Common-Type-Attributes)
to ensure that these enums use the smallest possible integer type,
and so use smallest memory when used in structs/arrays etc.

Change-Id: If1fc136686b28847109c9f3a06f8728165e7e475

Move experimental config from debug_util.c/h to aom_util.mk to avoid
empty object.

Change-Id: Id7978ed6a342262bddaa4df8b53115e750fa1c2c

Change-Id: I73d46229f0feea43cbe933e51da997833cce032b

Streamilines the functions and data structures to make it
easy to add new restore options.

Change-Id: Ib00638a5749e6c38c2455f3e3142b1025e6e0624

This was introduced in a cleanup in
I1e07ccab18558dfdd996547a72a396abe02ed23d

Change-Id: If6ac798d838a1ad392981f4e5970778207c3cb0b

Change-Id: I186e90d99cd54e66d38159b7cb55a881226b1568

That's where it lives in aom/master.

Change-Id: I38f405827d9c2d0b06ef5f3bfd7cadc35d5991ef

This makes room for typedefing some other struct to aom_writer.

Change-Id: I1e82de1320da00b3e41c90b14f2df45e7628aa89
(cherry picked from commit d69161f8)

Change-Id: I9e74bdb94c5640aca025b11b6676e8a8c008f47e

Includes a major refactoring/enhancement to support
tile-adaptive switchable restoration. The framework can be
readily extended to add more restoration schemes in the
future. Also includes various cleanups and fixes.

Specifically the framework allows restoration to be conducted
on tiles such that each tile can be either left unrestored, or
use bilateral or wiener filtering.

There is a modest improvemnt in coding efficiency (0.1 - 0.2%).

Further enhancements will be added subsequently to improve coding
efficiency and complexity.

Change-Id: I5ebedb04785ce1ef6f324abe209e925c2d6cbe8a

These are in response to a post-commit review in
Ib6664df44090e8cfa4db9f2f9e0556931ccfe5c8

Change-Id: I1e07ccab18558dfdd996547a72a396abe02ed23d

Change-Id: If22018f8911d9d7ee99c2127bdfcc56e42b0e2d7

cherry-picked #ecd07473 from aom/master

Change-Id: Id8f45d9c11406fc301b39801c5228ccd6aa2d5d6

This is avoid that -1 overflows uint32_t.

cherry-picked #c48106da from aom/master

Change-Id: Ic3d99b1985cdb0a28cc83f8291422f5aba5a5a6d

This is to avoid overflow at uint32_t.

cherry-picked #000098a0 from aom/master

Change-Id: I549d2d13d0577fd05d57303a438fbc8034755e45

This commit changes to use uint32_t for cost (always non-negative),
and promote to int64_t before calculation of the savings.

This fixes an integer overflow.

cherry-picked #a3028ddf from aom/master

Change-Id: I71c2580d188cc79d2d8069241d0353cf331b5c83

Backwards compatible with old API.

cherry-picked #02ae3dd3 from aom/master.

Change-Id: I65aa43f84bb9491e8cca73fe444094c2622b0187

cherry-picked #14ed7a61 from aom/master

Change-Id: I077b0e97186bdd292f925e08966a2ca3cf8c250d

This commit changes to use int instead of unsigned for a variable used in
inverse quantization.

Change-Id: I8f0ff5f80c9e68d52425265ef177357c65ead1e2

Encoder speedup ~2% with ext-tx + rect-tx

Change-Id: Id56ddf102a887de31d181bde6d8ef8c4f03da945

this matches style guidelines and stabilizes successive runs of
clang-format across the tree. remaining types should be address in
successive commits.

Change-Id: I6ad3f69cf0a22cb9a9b895b272195f891f71170f

A minor cleanup and an enhancement to return y_skip correctly
from sub8x8 intra mode search.

Change-Id: I87256d3cc5f57a2fd7b837d461cc1a7f06e01a1b

This makes the code in select_tx_size_fix_type match the
corresponding code in pack_inter_mode_mvs.

Change-Id: I69bcc0dc6fdd733091fafe9188a3f7397e1e613f

common/av1_fwd_txfm.[hc] are encode-only; add a TODO to relocate them

Change-Id: I28cf8d0b22632b04066bcb72f3d2252ee7eb153e

Some minor adjustments to tile size and bilateral filters.

About 0.1% improvement for midres and hdres, very small change for
lowres.

Change-Id: Ia94f68a926867dfd67da1a8795fd8de0ddd8e2d6

This allows for a clean subtraction of 1 along the transform
matrix diagonal and also makes the order of the parameter list
a little more intuitive.

Change-Id: I6a5d754af41b8d1292f241f9b21473160517d24f

This changes a remaining VP9_HIGHBITDEPTH to AOM_HIGHBITDEPTH

Change-Id: I35efaf9528de660fb69104792a563dba5c41f329

fixes test linkage

Change-Id: I15a7b32551fddc5e78e3035e9d2e94a57ff9f1d2

Change-Id: I8e08ebc8cbb2d1a1f97c8ef0c9237d8dfe0df208

Change-Id: I66ea5a819ab54ecb5327eee20f798d7d7f0833d3

Allow using it in aom_dsp.

Change-Id: Ide7d58b6d11f8a45d473fc13bf730ba5bccb5516

Change-Id: I75740e221deb3872647bd480ae506ba68800e8c7

Bitstream syntax:
For a rectangular inter block, 'rect_tx' flag is sent to indicate if
the biggest rect tx is used. If no, continue to decode regular
recursive tx partition.

Change-Id: I127e35cc619b65acb5e9a0717f399cdcdb73fbf0

Uses an array to map block sizes, y tx sizes, and subsampling
factors to various transform sizes for UV.

Results improve by 0.1-0.2%

Change-Id: Icb58fd96bc7c01a72cbf1332fe2be4d55a0feedc