Merge "Speed up iht8x8 by rearranging instructions. Speed improves from 282% to 302% faster based on assembly-perf."
Speed up iht8x8 by rearranging instructions.
fix a problem where an invalid mv used in search
Merge "Revert "Improved 8t filters""
Merge "Minor adjustment in unit tests"
Merge "Adaptive motion search control"
Adaptive motion search control
Merge "Clean up of the search best filter speed feature"
Merge "Fix VP9_mode_order[]"
Merge "New mode_info_context storage -- undo revert"
Merge "Deleted #include <inttypes.h>"
Merge "fix clang warning in rdopt"
Merge "Add neon optimize iht8x8 which is 282% faster than C."
Merge "First draft of vp9_short_idct32x32_add_neon."
Deleted #include <inttypes.h>
Add neon optimize iht8x8 which is 282% faster than C.
clang warnings : ref frame enum mismatch
Merge "clang warnings : remove split and i4x4_pred fake modes"
Merge "Changes in speed 2 settings"
clang warnings : remove split and i4x4_pred fake modes
First draft of vp9_short_idct32x32_add_neon.
Clean up of the search best filter speed feature
Merge "Improved 8t filters"
New mode_info_context storage -- undo revert
Merge "resolve clang issue : implicit convert tx_mode -> tx_size"
Changes in speed 2 settings
Merge "Remove redundant condition check in 32x32 quant"
Merge "vpx_mem: increase default alignment"
Merge "Enable and fix resize_test for VP9"
Merge "Remove the use of uninitialized_safe in encode_sb_"
Remove redundant condition check in 32x32 quant
Enable and fix resize_test for VP9
Merge "Small tweaks on the constant quality mode"
Small tweaks on the constant quality mode
Stop partition checking when distortion is small
Merge "Modify encode breakout for static frames"
Modify encode breakout for static frames
Merge "Enable accuracy/memory check for 16x16 transforms"
Merge "Rework 16x16 transform unit test"
Merge "Speed up idct16x16 by rearrange instructions."
Modified mode skip functionality.
Added extra check to rd_auto_partition_range()
test/idct_test: add missing vpx_integer.h include
Speed up idct16x16 by rearrange instructions.
Enable accuracy/memory check for 16x16 transforms
Merge "API extensions and sample app for spacial scalable encoder"
Rework 16x16 transform unit test
API extensions and sample app for spacial scalable encoder
Remove the use of uninitialized_safe in encode_sb_
Merge "Revert "New mode_info_context storage""
Revert "New mode_info_context storage"
Merge "Reduce the amount of extension in src frames"
Merge "Enable kf restrictions at speed 4"
Reduce the amount of extension in src frames
resolve clang issue : implicit convert tx_mode -> tx_size
Merge "New mode_info_context storage"
Fix overflow issue in 16x16 quantization SSSE3
vpx_mem: increase default alignment
New mode_info_context storage
Merge "Support a constant quality mode in VP9"
Enable kf restrictions at speed 4
Support a constant quality mode in VP9
Merge "fix loop filter setup_mask could reach out of bounds issue"
Merge "Speed up idct8x8 by rearrange instructions. Speed improve from 264% ~ 270% to 280% ~ 300% base on assembly-perf."
fix loop filter setup_mask could reach out of bounds issue
Merge "Enable 32x32 Transform unit test"
Enable 32x32 Transform unit test
Merge "Use saturated addition in SSSE3 of 32x32 quant"
Merge "resolve clang warnings : uninitialized vars in vp9_entropy.h"
Use saturated addition in SSSE3 of 32x32 quant
Merge "faster accounting of inc_mv"
Merge "make bsize requirement for SEG_LVL_SKIP explicit"
Merge "Added ClearSystemState in a unit test"
resolve clang warnings : uninitialized vars in vp9_entropy.h
Merge "wrap non420 loop filter code in macro"
Merge "Attempt to fix speed 4"
make bsize requirement for SEG_LVL_SKIP explicit
Speed up idct8x8 by rearrange instructions.
Added ClearSystemState in a unit test
Merge "Fixing problem with invalid delta_q reading."
Merge "Add neon optimize vp9_short_iht4x4_add."
Add neon optimize vp9_short_iht4x4_add.
Fixing problem with invalid delta_q reading.
Merge "Replacing init_dequantizer() with setup_plane_dequants()."
Merge "speed up inc_mv_component"
Merge "make vp9 postproc a config option"
Merge "Use correct bit cost while static-thresh is on"
wrap non420 loop filter code in macro
make vp9 postproc a config option
faster accounting of inc_mv
Replacing init_dequantizer() with setup_plane_dequants().
speed up inc_mv_component