Change meaning of cpi->sf.first_step and rename.
Merge "vpxenc: send usage to stderr"
Merge ".gitignore: add gcov files"
Merge "Move vp9_counts_to_nmv_context to encoder"
Merge "Move vp9_full_to_model_counts to encoder"
Merge "make: add libvpx_test_srcs.txt target"
Merge "tests/*source: test file pointer before reading"
Merge "encode_test_driver: check for fatal failures"
Merge "Refactor intra predictor block"
encode_test_driver: check for fatal failures
tests/*source: test file pointer before reading
Refactor intra predictor block
Merge "Only do metrics on cropped (visible) area of picture."
Merge "Don't skip right/bottom border pixels in SSIM calculations."
Merge "Add averaging-SAD functions for 8-point comp-inter motion search."
make: add libvpx_test_srcs.txt target
Merge "Cosmetic changes in 4x4 fwd transform unit test"
Merge "Tune the rounding operations in 8x8 ADST/DCT sse2"
Merge "I420VideoSource: normalize framerate types"
Only do metrics on cropped (visible) area of picture.
Don't skip right/bottom border pixels in SSIM calculations.
Add averaging-SAD functions for 8-point comp-inter motion search.
Merge "intrapred_test: add virtual dtor to IntraPredBase"
Tune the rounding operations in 8x8 ADST/DCT sse2
Merge "Add SAD unit tests for all rectangular sizes."
Merge "Don't re-allocate comp_pred buffers for each call to comp motion search."
Merge "Removing unused code."
Cosmetic changes in 4x4 fwd transform unit test
Merge "Add 8x8 dct/adst unit tests"
Merge "Changed size of mb_mode_context to 8 bits"
Merge "Small mode_info_context cleanup in filter_block_plane"
Small mode_info_context cleanup in filter_block_plane
Merge "Removing find_seg_id and using vp9_get_pred_mi_segid instead."
Merge "Transforming scale_mv_component_q4 into scale_mv_q4 function."
Add 8x8 dct/adst unit tests
Merge "Use aligned buffer operations in 8x8/16x16 2D-DCT"
vpxenc: send usage to stderr
Merge "Enable sse2 implmentation of 8x8 ADST/DCT"
Merge "change to enable use_largest_txform feature"
Use aligned buffer operations in 8x8/16x16 2D-DCT
Enable sse2 implmentation of 8x8 ADST/DCT
Changed size of mb_mode_context to 8 bits
change to enable use_largest_txform feature
.gitignore: add gcov files
Move vp9_counts_to_nmv_context to encoder
Move vp9_full_to_model_counts to encoder
Merge "Remove unused vp9_build_intra_predictors_sb{y,uv}_s"
Merge "Remove unused vp9_model_to_full_probs_sb()"
Add SAD unit tests for all rectangular sizes.
Don't re-allocate comp_pred buffers for each call to comp motion search.
Merge "Fix loopfilter of leftmost 4x4 edges in SB"
Fix loopfilter of leftmost 4x4 edges in SB
Merge "Allocate memory using appropriate expected alignment in unit tests."
I420VideoSource: normalize framerate types
intrapred_test: add virtual dtor to IntraPredBase
Allocate memory using appropriate expected alignment in unit tests.
Merge "Add some unaligned test vectors"
Remove unused vp9_build_intra_predictors_sb{y,uv}_s
Merge "Remove emms - that shouldn't be there."
Remove unused vp9_model_to_full_probs_sb()
Transforming scale_mv_component_q4 into scale_mv_q4 function.
Remove emms - that shouldn't be there.
variance_test: use REGISTER_STATE_CHECK
Removing find_seg_id and using vp9_get_pred_mi_segid instead.
Add missing SECTION .text marker in assembly file.
Implement SSE2 block_error.
Merge "Add subtract_block SSE2 version and unit test."
Merge "SSE2/SSSE3 optimizations and unit test for sub_pixel_avg_variance()."
Add subtract_block SSE2 version and unit test.
Merge "Get some speed back for cpuused 1"
Get some speed back for cpuused 1
Merge "rename variables to avoid build error in MSVC"
rename variables to avoid build error in MSVC
Merge "Implement sse2 and ssse3 versions for all sub_pixel_variance sizes."
SSE2/SSSE3 optimizations and unit test for sub_pixel_avg_variance().
Merge "clean out libvpx-srcs.txt if built"
clean out libvpx-srcs.txt if built
Merge "Revert "test_libvpx: disable pthreads in gtest""
Revert "test_libvpx: disable pthreads in gtest"
Merge "Add unit tests for 4x4 ADST"
Merge "Cast value to avoid size_t/int warning on win64"
Merge "Renaming 'nmv' to 'mv' for several functions."
Merge "Function decomposition inside vp9_decodemv.c file."
Improving model rd with variance and quant step
Cast value to avoid size_t/int warning on win64
adds force partitioning greater than or less than block size
adds a set partitioning to speed features
partition by variance using var from last frame
convert all speed things to speed features
new partition via variance
fix to set up new speed feature
don't copy partitions for key frames or altrefs
Implement sse2 and ssse3 versions for all sub_pixel_variance sizes.
disable speed > 1 speed corrections in firstpass
new debug modes code
Merge "copy partitioning from last fame"
copy partitioning from last fame