- OBU_TD => OBU_TEMPORAL_DELIMITER
- METADATA_TYPE => OBU_METADATA_TYPE
- Prefix OBU_METADATA_TYPE enum vals with "OBU_".

BUG=aomedia:1046

Change-Id: I0c63d36b77905520e427e6b77fbf4cbedabc7e51

Modified the warped motion mode context based on neighbor's motion modes
and current block's mode.

Change-Id: I77ca35fab37ec640bb38661ff1799f643d5aafdc

subset-1:

new_dct4@2017-11-27T20:52:07.119Z -> new_dct8@2017-11-27T23:57:04.520Z

   PSNR | PSNR Cb | PSNR Cr | PSNR HVS |    SSIM | MS SSIM | CIEDE 2000
-0.0030 |  0.0566 | -0.1127 |  -0.0244 | -0.0078 | -0.0154 |     0.0026

Change-Id: I1fde36a5ed454a50acf81004a618fc0a0c8c9073

subset-1:

master@2017-11-27T19:24:03.517Z -> new_dct4@2017-11-27T20:52:07.119Z

   PSNR | PSNR Cb | PSNR Cr | PSNR HVS |   SSIM | MS SSIM | CIEDE 2000
-0.0113 |  0.0459 |  0.1285 |   0.0085 | 0.0005 |  0.0093 |     0.0352

Change-Id: I0a76037ea2a08071ca9c4013979cca3ee3efe55c

Change-Id: I7f108fce272b5bf416836d99430f07af801daada

Although four bits are enough to represent current distances since
Golden Frame Group is 16, for flexibility, we use 5 bits and allow
frame distance up to 32.

BUG=aomedia:1072

Change-Id: I9f413baffd656eb8bd54333ba31a4e33faefd57a

And do so unconditionally: It's harmless to allow the
types to be defined without CONFIG_OBU enabled.

BUG=aomedia:1046

Change-Id: I5b9a3a68e4e70b07137e381f05345d2ea609a09a

Multisymbol BR coding is simplified as follows.
1. Remove computation of level counts by using a template of size 8;
2. Context is derived by using a template of size 3.
3. lps and eob probabilities are trained.
4. Share contexts between TX_16X16 and above.

The number of probability values used in BR coding are reduced from 1152 to 378.

Change-Id: I0419127e871f9e566c2489aa4b1825c5364aec5a

Add a block level usage flag for skip mode. If no block has chosen the
skip mode, the frame level flag for skip mode will be set off.

This patch also includes a small code cleanup, including the check on
whether the best RD mode is aligned with skip mode, if yes, the best
RD mode will be replaced by skip mode.

This patch slightly improves the coding performance of ext-skip.

Change-Id: If06092d5e32f15e63dcb5f35d32e68bc0f827c2b

Change-Id: I584694374a2468e0dcfe6e4fdb2582e5cae051ef

Change-Id: I671009d26c7be031b0f742ae9cf841233edbc4c8

This produces a 0.03% drop in performance but reduces
the tx search space.

Change-Id: I27ae62403f059c4654e9237c39f15bed7f860c15

Add counters and newly trained initial probabilities for coefficient
coding.

Change-Id: I2647303c639b60018b8caacfefce77be26cae92c

Change-Id: Id6127524a02d308072408b3c3af0e4c604e5f527

Change-Id: I891a331ca7c8c59f493ecfe460a96de07e6537a1

Change-Id: I763e844875680961d9550bfd39a61a9caf2727b5

Record the best_txk_type in calc_rd_given_intra_angle()

Change-Id: I174709096d019d31567517b5aec6c2858ee68bc3

If the frame's monochrome flag is set, the decoded frame
will be given constant chroma planes, regardless of any
chroma data in the original frame.

Change-Id: I9ea078f02c279ae9ec1d23cc33f69c6abcb18b9f

Change-Id: I6eca9ab77cf48c5b027319a25c59b36b9c0c4313

Change-Id: Id695bbe1ac37750b98c144685c9a9fd3ab5ae091

Change-Id: I244a7285e7a21fbb4dbe8ad9f6ff7c966a0bd601

When a target has no C++ sources, but includes C++ object files
Xcode requires special treatment. A dummy source file is created
using the target's name to force the CMake Xcode generator to use
C++ as link language.

Append "_dummy" to the dummy source file's basename to avoid
collisions with the target's main source file when it
is target.cc.

Also adds a TODO to avoid adding the dummy source file when the
target being adjusted already has C++ sources.

Change-Id: I9795f6490e88eae6afc3a5901e6ee76a36cc801a

This prefix is not needed for static functions.

Change-Id: If15dfed10626d88afb9294985e70a6bcf3ea6848

Previously CDEF was implicitly disabled for a filter block if all its
sub-blocks were skip, and no bits for that block was signalled.  That
required the CDEF signal to be transmitted at the end of the block,
and it was not possible to begin the filtering of the block before
that.

This patch moves the signalling to just after the first zero skip bit
within a 64x64, 128x64, 64x128 or 128x128 block.  If sub-blocks are
skip, then no CDEF bits will be signalled, as before.  Also, the skip
condition flag has been removed so it's always known at the skip flag
whether a coding block is to be filtered or not.

Change-Id: I3adfa3d2d01e288b7db876198aa9985ee9f53917

Change-Id: I496c2a83b227ec88a6cd59a4c227b59eeeb0fc86

This patch brings back the coding performance of ext-skip to where it
should be. The bug was caused by a code refactoring when the previous
patch for ext-skip was merged.

Change-Id: I8fbefe8cee0b4a5340a24936a4e37f98609582e3

1. Turns off hash based rd for rectangular blocks temporarily.
2. A small fix for zero_rate aggregation.

Change-Id: I571e5b2b5441cc0d1b8b0a522d9ab23787441103

Also adds const correctness to palette pointer


Change-Id: I3ddde5a8ef0e2d8a5a5a5bc87378e33eed5d8945

Change-Id: I852e75e8f41f057c2cf3693cb3441e6640ac7e0d

Change-Id: Iff2780944e954973b9ee9ab3c04d1e6b44a7810d

Applies to speed >=1. Instead of searching all dual filter space
{R,Sm,Sh}x{R,Sm,Sh}, only check {R}x{R,Sm,Sh} followed by
{R,Sm,Sh}x{best of prev R,Sm,Sh}.

Saves ~6% of cycles by reducing av1_convolve_2d_sse2, with 0.023
overall psnr drop.

Change-Id: I82d7a6321b335293124a007ff4c87f0e260052e1

cfl.h is no longer included in the onyxc_int.h. cfl.h is now included in
each file that uses cfl functions.

Change-Id: I80a2e57159217b0ee20b2562fc9711d929472b8c

The CfL context is now stored inside MACROBLOCKD instead of
MACROBLOCKD only storing a pointer to the CfL context.
The intent is to avoid race conditions as MACROBLOCKD is stored
inside ThreadData. This change also simplifies CfL Initialization.

Change-Id: I991503716b21fc9aca60caddb2008b8bff397e6d

Change-Id: Ie6f6f5405cfe4c59dec1dc6dabc9ef5bcf2fb756

This is to get the high range within 32 bits.

The compression performance is neutral.

Change-Id: I475b2694f7cc7d786478c8821cf17982281dabce

This avoids copying a whole struct on return. This is important, because
this function is called frequently: once for every block that is transformed.

Overall, aomenc speed seems to be improved slightly: by 0.15% - 0.30%.

Change-Id: If44c351e1141c0fe0458daa2262aeb4afe44de46

dct_const_round_shift() takes tran_high_t as parameter and Sqrt2 is of
type tran_high_t. It is unnecessary to convert to int64_t then back
to tran_high_t.

Change-Id: I610d37b0772185863c406c6d77ad957d24feaa5e

Change-Id: Ie6f7dd3cfaa29f70f0052f87a8e82e9e1efa8507

This increases the deconstruction offset from the center of
each quantization bin. It gets -0.187% gains on the lowres
set.

Change-Id: Ie8cf34b600dd2caf4e874e35453d3176210914e3

Change-Id: I40daf7cbb7c7a8613ad1f9a3175d66f1f983ccce