GitLab

Change-Id: I62a8fd48593992d1451b96006efb1b43c048d4ba

Initialize the counter to be zero.

Change-Id: I6752a881eda032d9bcc803f724e53fdf79e95e75

Change-Id: I01c97d6200e3f4d17c6b38095ca7c8c31967a2ce

Change-Id: I751a6719e08d18ea45ee902b783bf3b3c830cc58

Change-Id: Ibdf1de139e8ce63b0be2d0cd1fd098f06739d8cc

Change-Id: I03adc36f5e5f839d23c4f50a864d317d008e30a6

This assert was wrong for tx sizes of 4x4 (and 2x2), and was causing
some failures on AWCY like this one:
https://arewecompressedyet.com/?job=debargha-adopted-0927%402017-09-28T00%3A03%3A48.310Z

Change-Id: Ib75ba7ec6282c700cbe1742b9871a5489eb7e18c

Directly adding libm to the libaom target when building
non-MSVC and non-APPLE targets.

BUG=aomedia:831

Change-Id: Ia1f9cf6b79c4b4d15297969ff9ef4b9831113f4f

Change-Id: Ib5e5346fb4231551739edefdf5ff09628ca71221

This patch removes the restriction that both rect_tx and var_tx must be
 turned off when any daala transform is enabled.
For now, all square transforms are done using daala with the rectangular
 transforms done using AV1.

subset1:

master@2017-09-19T20:41:26.136Z -> daala_tx-dst4@2017-09-19T21:06:47.190Z

   PSNR | PSNR Cb | PSNR Cr | PSNR HVS |    SSIM | MS SSIM | CIEDE 2000
-0.0135 | -0.0201 | -0.0049 |  -0.0619 | -0.0144 | -0.0521 |    -0.0264

Change-Id: Icceb3b0c5bcc40d01fd773b20a2a6588179c0ee6

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

Change-Id: Ic077963f72e8cc2ae9872b58c8a0241988384110

tx_size is transmitted using CDFs now, so these tables are obsolete.

Also, remove a related tx_size_totals table.

Change-Id: I949b46c1c9499bc556e68c77adbdb5aa3288c9a2

BUG=aomedia:39

Change-Id: I89858a8a26547fea4a9d37855a28c0098273316f

BUG=aomedia:851

Change-Id: Ib188f4b640a58c47147cda977a589e527f765a54

This isn't needed anymore, now that the super frame index is
transmitted before the frames.

Change-Id: I8dfb514e59f2df4c2d6697f0ef27fe591fc3b4a2

This experiment was provisionally adopted on 2017-9-12

The ResizeCspTest unit test is disabled when intra_edge is enabled,
as the PSNR of the first frame increases from 77dB to 100dB.
This results in a failure when later frames are compared with the
first one.

Change-Id: I4e9629d9657b01b5d3e64f2dfcfd82e47292c9c6

on windows build

Change-Id: I841eb96b26155dd8fcba008884437035abc7e806

Change-Id: I7aa27af5e8839a9acd756d8c3c12d977012f09ca

Change-Id: Ifa983d83a509cdfad78f6400df7d60c8f5b4f68c

Change-Id: I17c2193202b7b5299c31f8387ad22409ac6f9c79

Change-Id: I179037007c3f0e35ef66b0dcad84540f7d8cf098

Change-Id: I9ad1986d3f2332e81e80d687663ffd89e4efa1db

This bug (introduced in 8af64ae5) effectively turned off switchable search causing a
drop in coding efficiency.

Change-Id: I6db0cf48371638673fcf951158eef32a6db42545

There were two places where invisible pixels are included
when calculating distortion with dist-8x8 is turned on.
Those are bugs and fixed by this commit.

1) When av1_dist_8x8() is called in dist_8x8_sub8x8_txfm_rd()
- New assert() are also added to detect any failures in the future.

2) Inside av1_dist_8x8() and av1_dist_8x8_diff(), with tune-metric == PSNR.

Change-Id: I078d1feb3bb010c0e34041640bb69a400125817a

Change-Id: Ie4e2f04a6fa27480e500469a0d822e539b7cb066

Do the same for tools.

BUG=aomedia:828

Change-Id: I284dd44ae57c461012ae2b034a23a3cad9361af0

This is the first tool in the CMake build, so some extra
noise is involved:

- Setup tools list vars and handling.
- Add tools support to the dist rule.
- Move usage_exit.c generation to CMakelists.txt to allow
  use by the aom_entropy_optimizer target.

BUG=aomedia:834

Change-Id: I55239e89353033349ac1038b8d3d1aa8a8f23e27

This hides behind --enable-ext_partition_types_ab. When that flag is
set, the mixed partitions change from

   +-+-+---+          +-+-+---+
   |   |   |          | | |   |
   |---+   |    to    | | |   |
   |   |   |          | | |   |
   +-+-+---+          +-+-+---+

(that's PARTITION_VERT_A; other partitions change similarly).

Since 128x32 and 32x128 block support hasn't yet been merged, this
patch also temporarily disables these partitions for 128x128 blocks.

Change-Id: Ic479c0fc129f590b8ad56d72dc98ba79ae1bd1cf

The updates in this CL are for the scenario when ext-comp-refs is off.

For --disable-global-motion --disable-ext-tx --disable-ext-comp-refs
--disable-convolve-round, the patch obtains a small BDRate gain for
PSNR: lowres -0.08%; midres -0.05%

Change-Id: I47d1ba93318e2904b5714f62ffd6af056f4eb588

Move functions related to frame index in natural order setup into
a root experiment.

Change-Id: Iebc71514b951dbb658cf0e1c3d42bb2b180d830f

The frame index in natural display order is needed by several
experiments, hence move it as a root experiment. The additional
overhead cost associated with frame_marker is below 0.05% level.

Change-Id: Ie643a6bc1f5a400ea6ff088751a676bb56ffee76

Change-Id: Id967afa4537461c8cff7e114bea68d2557f9358a

Function speedup (i7-6700)
Predictor  ssse3 v. C
4x4        ~2.1x
4x8        ~2.4x
8x4        ~4.1x
8x8        ~5.4x
8x16       ~6.1x
16x8       ~5.9x
16x16      ~6.4x
16x32      ~6.7x
32x16      ~7.4x
32x32      ~8.0x

Change-Id: I52b8ebf8193e76f4ea1137cbad5ad7fa109d86d8

This should relief the concern of latency incurred by generating
scan order

The performance on lowres and midres remains neutral

Change-Id: If155f055540126ee834f5be1ab4b23013090ee89

CfL now consistently uses clip_pixel. This change does not alter the
bitsteam.

Change-Id: Ifb6631e786eb1a9cf704c039ed9e6bfa45894143

This change does not have any impact on the bitstream.

Change-Id: Ib7cf344eb1e682cdc2c5abdc0621d65e70689ab4

This change does not impact the bitstream. It is intended to simplify
High bit depth support in subsequent commits.

Change-Id: If22822dbb3329fc0b4a5e343772d63761d46d3f1

replace clamp64() with clamp32u() where applicable

Change-Id: I3fc97d576b3235eeda5d26a6a9692b5e51e016f3

Change-Id: I3aa0348c4f5337998c394f6489f0fffe8eddaf47

- Fix a typo.
- Add the mingw toolchain files.

Change-Id: I3a3b8157457500757059a27dcc10c63a452760f7