GitLab

Scale up the mb_mode_info resolution from 8x8 to 4x4 unit.

Change-Id: I60ee20f5fc353a3c16ced14b55422fc5e4ea1e06

Change-Id: I0e1954ca281df60e56c3b1a5dcd54c47c4256649

The pixel number remains the same.

Change-Id: I40e54f4c0893db8f9452624f8167ead994b71f1d

Change-Id: I99d3abd47480ebd73fc0034fac82663c003f7bb8

Relative to previous ext-inter:
lowres: -0.177%
     or -0.029% (with USE_RECT_INTERINTRA = 0)

* When predicting interintra modes, the previous code did not provide
  the intra predictor with the correct context during rdopt. Add an
  explicit 'ctx' parameter to the relevant functions, to provide this
  context.
  This fixes a nondeterminism bug, which was causing test failures in
  *EncoderThreadTest*

* For rectangular blocks, build_intra_predictors_for_interintra needs
  to overwrite part of the context buffer in order to set up the
  correct context for intra prediction. We now restore the original
  contents afterwards.

* Add a flag to enable/disable rectangular interintra prediction;
  disabling improves encoding speed but reduces BDRATE improvement.

Change-Id: I7458c036c7f94df9ab1ba0c7efa79aeaa7e17118

Map the block offset to pixel positions in the unit of minimum
transform block size. This change retains the same coding
statistics. The encoding speed is slightly faster (<1%).

Change-Id: I2641d2b1e317ef4cdf477718c446ef025b8eef41

Scale the counting scheme to operate in the unit of 2x2 transform
block size and 4x4 coding block size.

Change-Id: I7c4155c3b29df109278187d5ce0aa78d85a6245e

Make this scalable to minimum transform block size.

Change-Id: I13b0ac841df4e87e6c6bcd8be358d817b4be4328

Temporarily increase the buffer size from 2 to 4 in the sub8x8
rate-distortion search for cb4x4. These functions will be deprecated
after cb4x4 is fully enabled.

Change-Id: I6aecaf55bd21a9416713cf31e75e7b4d9b68f83c

This filter is meant to replace the bilateral filter,
but is currently added in addition to the bilateral filter
for testing.

Change-Id: Ia529701e69833d47c11b3367d5bf804eb8498079

Replace hard coded numbers with table fetch. Support scalability in
transform block operating unit.

Change-Id: I4a5e67576e72de4936ac52d1401412ca6217ed6d

Refactor to remove the assumption that transform block operates in
the unit of 4x4. This generalizes the codec to support 2x2 transform
block later.

Change-Id: If1448a915f5765d409394c914af68f44b0749bc2

BUG=aomedia:84

Change-Id: Ieb4bd4893970a910e7277d49956a389ed19be777

It is called relatively rarely and doesn't need to be inlined.

Change-Id: I4ee7f95548f008f2ee29da807aaca54b9a25aecd

The final ANS state gets further compacted because aliasing the super
frame marker is not an issue.

Change-Id: I26208accb117a6748abb6f1c32c28fadbc48de09

For some historical reason, coefficients of interpolation filters
in filter.c is defined in int16_t but use 8 bits only.
On the other hand, SIMD like to use int8_t to represent those
coefficients for economical reason.
The conversion from int16_t to int8_t will introduce
"possible loss of data" warning.

We use cast to work around this problem.

Change-Id: I42ef0a2a815a8f93c579f7984c9b957c1aed6d6c

Use the common function between encoder and decoder.

Change-Id: I0a85ddc864db21971cd97ab811792f9bbc7e96bb

Use the atom transform block as the unit to count transform block
orders. This avoids the assumption that the codec goes in unit of
4x4 block.

Change-Id: I2c73c19c1349843656c940a79547a5baebbda997

Change-Id: I6ff6916a76a1ebbd964ececb1d81e9df59234d92

Improves the refinement search for global motion.

Change-Id: If6160ebf0f5309a0d7fb648ec5efb0778766cb87

Control the loop over transform blocks in the unit of smallest
transform block size. This would provide scalability to support
smaller transform block sizes.

Change-Id: I6a8167cf91f71c5d4af4e244c628f378866a607c

Change-Id: Icbb4e686b4a5e7fdd1de76d412f43fdf1bf8830a

Use direct table access to fetch the transform block width and
height in the unit of pixel.

Change-Id: I80a90729bc0d078f918ffafff612fb29afa0e761

Directly use c code to compute the distortion in pixel domain for
2x2 transform block.

Change-Id: If70049f517d2086c462fcf9a02a12d6fa73adf54

BUG=aomedia:99

Change-Id: I36c3c3728500e0eefb7c56720a467e8754b85009

add an operator<<() for instances of
function_equivalence_test.h:FuncParam

Change-Id: Ifced8401443085bdbdccb97718564b0372b264b7

Change-Id: I3da3d48ade0286d430fa22aebe1c3fc8d3a68995

Change-Id: I509bbba0d84c1d378044e2c612dd48cd8f99848d

Bring 2x2 intra prediction online for chroma components.

Change-Id: Ia56af9101b2a977691bca4156a6dcf89e644b4a7

Change-Id: I094d7787b5ababa4855a0c57064c09c6fa02aede

If the coding block size is 4x4, map the uv transform block size
to 2x2 in non-444 format.

Change-Id: I5767df7c6c1f73938f97745d604207caad6f7d68

Support static scan order for 2x2 transform block.

Change-Id: I5518072a572dac39992dc96a5b2fb146f018a267

Fix compile error.
Make it work when ext_interp/dual_filter experiments on.

Change-Id: Icba532b46856245f962c8e0c67a3159ab7d997d9

Correctly use FILTER_INTRA_MODE enums instead of PREDICTION_MODE enums.

Change-Id: I3d89c0c5a926b9f8603b8492e7c8c9b32b41d4cc

Updates to enums and mappings as needed.

Change-Id: Ie3a71031c74a5d4caa5308f0e210ed23c2361e95

This should have no effect on the bitstream format (see also no related
encoder change). This is like moving code from the top of the loop to
the bottom of the loop.

This change allows us to:
* Make sure we consume the final renormalization byte after the last
symbol in an ANS partition.
* Move back toward a single renormalization operation for some ANS modes
since we know the bounds of the state mutation algorithm that got us out
of the valid state range.

Change-Id: Ia80246fd0ed805aa61b913a362546b3f08e4d79c

clang format gets confused with this table due to #if statements.
Turn off format check on this table.

Change-Id: I21e9069581749d6e01501d01ba779f56207ca84a

This was caused by a refactor in Ie105a7baf6a2c2258d3ef117e727957e4393f51b

Change-Id: I3a1d958d1e8759b9c1c215ea48763ee8deb62cf3

Let aom_convolve8_### SIMD implementation support any block width.
Turn on SIMD optimization when interpolation filter types on two
directions are different.

This will reduce 30% of encoding time when dual_filter and ext_interp
both on.

Change-Id: I539dbb2737f01835034b7269656a15b2058fa3cc

Change-Id: Ibb695cc12381ecddced85891810a3c97e2fca3cf