Insert multiple arfs in a gf group to emulate multi-layer backward
reference frames structure. At maximum, two extra ARF's are inserted
in a gf group.

It improves the RD performance by 0.317% in Avg in lowres dataset.

Change-Id: I62c32e1b0f25b978484dd113b319bebcd959bf60

Change-Id: Ibdf1c2d422b9e644eba76fc200c8c10217394036

vp10_fht{16x16,8x8,4x4}_msa and the iht were disabled with this config
in:
4ab19eac Fix assertion failures in mips+msa setting

Change-Id: Ic675258b89ca490e8021c887b705c68428925129

- gcc x86_64 build binary is about 47 KB smaller.

Change-Id: I9e5f41fc9c5c75aec453f8b8567e228a6a6cd71d

This was originally subtracting 1 from the wrong element in the
parameter set.

Change-Id: I790aafc505f7a8fe7bb00d7d6c62549487a0980f

Originally we can have a BRF right before an overlay frame (in
display order), which might be unnecessary since we already has a
quality backward reference frame (ARF).
This patch avoids such a coding structure and improves the RD
performance by 0.086% in Avg in the lowres dataset, and 0.153 in
Avg in the midres dataset.

In the lowres dataset, significant gains are obtained for the
following sequences:

mobisode2_240p: 0.563%
keiba_240p: 0.440%
bus_cif: 0.336%
soccer_cif: 0.333%

And the performance drops only in the following four video sequences:

motherdaughter_cif: 0.028%
bqsquare_240p: 0.017%
basketballpass_240p: 0.015%
bowing_cif: 0.006%

Change-Id: Ic94f648ba8e52eb0014933d484fb247610a9ae05

Mannually cherry-picked:
15791332 Use OD_DIVU for small divisions in temporal_filter.
03122298 Replace divides by small values with multiplies.
9c48eec7 Removing divisions from od_dir_find8()
0950ed82 Merge "Port active map / cyclic refresh fixes to vp10."
efefdad7 Port active map / cyclic refresh fixes to vp10.
1eaf748c Port switch to 9-bit rate cost to aom.
0b1606e7 Only build deringing code when --enable-dering.
e2511e15 Deringing cleanup: don't hardcode the number of levels
8fe5c5d6 Rename dering_in to od_dering_in to sync with Daala
4eb1380d Makes second filters for 45-degree directions horizontal
7f4c3f58 Removes the superblock variance contribution to the threshold
3dc56f93 Simplifying arithmetic by using multiply+shift
cf2aaba9 Return 0 explicitly for OD_ILOG(0).
49ca22aa Use the Daala implementation of OD_ILOG().
85187243 Fix compiler warning in od_dering.c.
485d6a69 Prevent multiple inclusion of odintrin.h.
51b7a998 Adds the Daala deringing filter as experimental

Note that a few of the changes were already in libvpx codebse.

Change-Id: I1c32ee7694e5ad22c98b06ff97737cd792cd88ae

Manually cherry-picked following commits from AOMedia git repository:
bb2727c2 Sort includess for "clpf.h"
c297fd00 Add quantisation matrix range parameters.
0527894f Add encoder option and signaling for quant matrix control.
4106232b Turn off trellis coding for quantization matrices.
4017fca3 Modify tests to allow quantization matrices.
1c122c24 Add quant and dequant functions for new quant matrices.
95a89994 Enable CLPF
f72782bb Fix a build issue
73bae509 Add quantisation matrices and selection functions
33208d20 Added support for constrained low pass filter (CLPF)

Change-Id: I60fc1ee1ac40e6b9d1d00affd97547ee5d5dd6be

This was causing a compiler warning from -Wcast-qual.

Change-Id: Ie525ffe20be4f38ced68fb0c4141e36400eb0717

* changes:
  remove tools/vpx-style.sh
  README: add a note about clang-format

Adds forward, inverse transforms and scan orders.

Change-Id: Iab6994f4b0ef65e660b714d111b79b1c8172d6a8

The projected coordiantes in projectPointsTranslation were
being shifted by the incorrect precision.

Change-Id: If6040bea9e5187020d85c6095d85c7ff5786b7f9

update ftfy.sh to use clang-format

Change-Id: I8ac740c5b3842beed2b8878fbe506f381f4c57e4
(cherry picked from commit 958ae5af9c892e5328ec0363d1a69afbfe0e0907)

Change-Id: I835401e3befffcbc68e7d2bdd2fd556a19948e91
(cherry picked from commit 15f29ef0922c6584dea69ac75367edb2f6934bb3)

clang-format authored 8 years ago and James Zern committed 8 years ago

top-level *.cc were missed in the original change
debug_util.c was checked in with some warnings

Change-Id: I72999bf94d734ffc127bf6f96a8d17f9c313d5a0

clang-format authored 8 years ago and James Zern committed 8 years ago

Change-Id: I58a42ced5b8a4338524434ff3356850b89aa705a

clang-format authored 8 years ago and James Zern committed 8 years ago

Change-Id: I01d8241eba3ccaf4d06c00a51df2d17c126f6f9d

rather than sizeof(type)

Change-Id: I63755e4ca3810bec2d31013bebcc363c5c9f56ed

clang-format authored 8 years ago and James Zern committed 8 years ago

Change-Id: I1138fbeff5f63beb5c0de2c357793da12502d453

- On E5-2680, park_joy_1080p, 5 frames, baseline encoding time
  reduces about 0.8~1.0%.
- Credit goes to Erik Niemeyer (erik.a.niemeyer@intel.com).

Change-Id: I69f191d5a4e4b96a5f9ffd8286e484b69d565c01

Yi Luo authored 8 years ago and James Zern committed 8 years ago

fixes mix of malloc & delete[]

Change-Id: I89a1de0614234bf8b3dbe4aacfe71f75f39d08ff

clang-format authored 8 years ago and James Zern committed 8 years ago

Change-Id: Ie4d7ecb2f692c1c43eff1242e1f00e7fbae00e57

- Add unit tests to verify the bit-exact.
- Speed unit test, function improvement: about 8%-23%.
- On E5-2680, park_joy_1080p_12, 25 frames, --kf-max-dist=1
  encoding time improves from <1% to 3.5%

Change-Id: Ic16368885bb253db0200c3a6db143ab1a0b7fc26