GitLab

Modified the copying code and the profiling showed better performance
than previous implementation.

Change-Id: I41f585e0b0eac7a0deb4dec197c178e412a48db9

Change-Id: I8604a1f644302f5e6249b89c96550e1d5715188f
warning: assignment from incompatible pointer type

Change-Id: I2405656d1151559420b101fc51ceae8a6ba5ca56

Change-Id: I32c5015bedef07f58413c679b4aaf30651a357b9

Change-Id: Id3124c0a4724e1c9e94960702445055da1e99277

This is to get an approximation of 64x64, 32x64 and 64x32 transforms.
Compression performance is neutral for lowres and midres sets.

Change-Id: I6faaca02170bfef802a6294410fd691f73914173

The performance drop introduced by inconsistency decision between
optimize_txb in rd loop and final encoding phase.

By making them use the same fast_mode,  will speed up encoder
without having the performance drop

Change-Id: If25e253cf99b6f17353b4031d03b57da50a1cd95

Change-Id: Ib5e79caa67ba6a061e8dc22491c86cde702be137

Change-Id: Ib0e0b5f7085d3c910dcd15a3628a001af864c8fc

Change-Id: Ic8670dcfed8efa37050816bbb8062f9092886ef7

After this fix
eob_extra provides lowres 0.1% gain

Change-Id: I5eec92ff89c38410acc6f1ab463528507783f359

Change-Id: I39bc0d7926f43bd683034bd7660d110e6552c388

lowres 0.14% gain

Change-Id: Id4d2bee555075c002b7996f0329e89c82657bfdf

Inter/intra matrices are the same. Reduce the dimensions
of the various matrices.

Change-Id: Ie5049af9195d9c48b6928143ac04ac0a5058d0d8

This will avoid overflow issue when coeff uses more than 8 bits

BUG=aomedia:954


Change-Id: Ib5f323b0d9db3141cc019a6aad019f09b8c4f9fa

Change-Id: Ia67beabb85eedd5da5eb69f434dde71f66f08006

This commit makes the following changes to level-map coding
	1. Encode and decode eob position first
	2. Context derivation for non-zero flags
	3. Single pass in optimize-txb

lowres 0.33% gain

Change-Id: Ia2b6cfd70425f49ac021454d42d5950bfcfab50a

restrict_compressed_hdr experiment is made redundant by
new_multisymbol.

BUG=aomedia:870

Change-Id: I86100d114ca641a453c32ffd09cabb736809abd1

This experiment has been adopted.

Change-Id: Ife4c18a59791268b7ac0de5a8a08e762a042cae2

The try_restoration_tile function now works by calling
av1_loop_restoration_filter_unit rather than
av1_loop_restoration_filter_frame. This is rather more efficient,
because it only works on the unit in question, rather than memcpy-ing
the rest of the frame.

Change-Id: I7cda078523202bec77e933e7b5bd2f34daae5865

This patch also does a certain amount of rejigging for loop
restoration coefficients, grouping the information for a given
restoration unit into a structure called RestorationUnitInfo. The end
result is to completely dispense with the RestorationInternal
structure.

The copy_tile functions in restoration.c, together with those
functions that operate on a single stripe, have been changed so that
they take pointers to the top-left corner of the area on which they
should work, together with a width and height.

The same isn't true of av1_loop_restoration_filter_unit, which still
takes pointers to the top-left of the tile. This is because you
actually need the absolute position in the tile in order to do striped
loop restoration properly.

Change-Id: I768c182cd15c9b2d6cfabb5ffca697cd2a3ff9e1

Use the valid reference frames' motion information to perform
motion field projection.

Change-Id: I942d62043c7a7bec820e4d13c1fdab8e04719735

This accounts for border extension necessary when image size is not
a multiple of SB size in either horizontal or vertical direction.

BUG=aomedia:647

Change-Id: I18ae9edf148a83f8b32a06ca31410e30ebaa5a47

Change-Id: Ib1145bf79222b40a11018fb64a6121aecfc3c152

Change-Id: Iad4afbe195d51eba880ad595d37a41676eb0c0b5

The earlier default was 25, which was reduced to 17 in
https://aomedia-review.googlesource.com/c/aom/+/28181

However, as temporal filtering may use 3 frames on either side by
default, we should have default lag in frames as 16 + 3 = 19.

The compression performance is exactly neutral between lag-in-frames of
25 and 17.

Change-Id: I5603f899183a0363eb4c9160a6cacc4e68585725

This is to reduce the size of the huge token buffer by at least more
than 10%.

BUG=aomedia:940

Change-Id: I2ee7f7f62e8e8fb819cf7fba40a7ca7f860acaa4

Change-Id: I75ce79db34789151cc4cc9183a550f59e1357bfe

Change-Id: Ie740120e44f123c5011df99c2915a4c9ddb9f974

After this patch, they use 1 byte instead of 4 bytes.

BUG=aomedia:940

Change-Id: Id00f2a0618e458ba7fb985ecc10d581f156a18cd

Let read_nz_map() etc. output to uint8_t*.

Temporarily introduced coeff_is_byte_flag as an argument of
get_nz_count() and get_nz_map_ctx() to handle different types of input
coefficients. It helps to not duplicate too many functions.
If possible, unify types to uint8_t* later.

Change-Id: Idbbe67b7ec563f8c9299daa0fa7d05e15c8295e6

We replace the division by shifts using already calculated 'x_scale_fp'
and 'y_scale_fp'.

That way, hardware can do it on-the-fly instead using a set of
programmable registers.

BUG=aomedia:343

Change-Id: I99d93dec5327e835ce288212ab75b8a75a8fe520

BUG=aomedia:896

Change-Id: I414a1d7c924609ceb1cbf3edabae74d7521b8e55

Change-Id: Ib1ec44f6276b78543296e7dae2f4f8d1b6dc303f

Change-Id: I2565dfaf0d077984e47e4bd9009a0c174c8124c9

Supports 4x16 and 16x4 partitions as well as adds a fix
for 422 and 440 chroma subsampling.

Change-Id: Ibeb2218c8cb1c648c3b2f6bce10281ccbcacf20f

Change-Id: Ie4c0bb153bbfd7dfc3845b8ab1e517e8f1beedd7

Calls to daala tx have been moved to idct.c

Change-Id: I1e84afd2bcbccaf58261f0bb1dd7d09c6c04014c

Since we are not using homography models this code is not
exercised.

Change-Id: Ieba6f5c6fb58067d0fa910a7b942f45b5a22df3d

Change-Id: Ib93f090f395e52bd69c59e1229e2e035a8772af5