Merge "Step towards making the 2-pass cq mode perceptual"
pickmode: only cost the skip flag once per prediction block
Merge "Remove obsolete segment skip checks from tokenization."
Merge "Adjust motion component of prediction decay."
Merge "Trap very short arf group just before a kf."
Merge "VP9: Combine TileData with TileWorkerData"
Step towards making the 2-pass cq mode perceptual
third_party: Roll libwebm snapshot.
configure: Disable webm_io when target is VS 7/8/9.
Merge "Correct comments for scan order neighbors"
Merge "vp9-denoiser: Avoid copy-block when denoising is at LowLow level."
Merge "vp8-denoiser: Avoid doing the mcomp if we don't denoise."
Remove obsolete segment skip checks from tokenization.
Adjust motion component of prediction decay.
Merge "vp9: Some adjustments to scene-cut detection."
Trap very short arf group just before a kf.
vp9: Some adjustments to scene-cut detection.
Merge "vp9-1 pass vbr: Use average QP for active_worst_quality, for inter-frames."
Revert "third_party: Roll libwebm snapshot."
VP9: Combine TileData with TileWorkerData
Correct comments for scan order neighbors
third_party: Roll libwebm snapshot.
Revert "third_party: Roll libwebm snapshot."
Merge "VP9: refactor read_switchable_interp_filter()"
Merge "Fix crash when mkvparser returns a NULL BlockEntry."
VP9: refactor read_switchable_interp_filter()
Merge changes I7a1c0cba,Ie02b5caf,I2cbd85d7,I644f35b0
Merge "Fix high bit depth mask and variance reference function"
Fix high bit depth mask and variance reference function
third_party: Roll libwebm snapshot.
vp9-denoiser: Avoid copy-block when denoising is at LowLow level.
vpx_fdct16x16_1_sse2: improve load pattern
vpx_fdct16x16_1_c/msa: fix accumulator overflow
vpx_fdctNxN_1_sse2: reduce store size
vp8-denoiser: Avoid doing the mcomp if we don't denoise.
Merge "vp9-noise estimation: Increase threshold for Low-level."
vp9-noise estimate: Adhust threshold for Low level detection.
Fix crash when mkvparser returns a NULL BlockEntry.
vp9-noise estimation: Increase threshold for Low-level.
vp9-denoiser. Code cleanup and bugfix.
Merge "dct32x32_test: create fn pointers w/'&' ref"
dct32x32_test: create fn pointers w/'&' ref
vpx_fdct32x32_1_msa: fix accumulator overflow
Merge changes I8e9692fc,Ifc06110c
Merge "vp9-denoiser: Adjust usage of skin and consec_zeromv."
dct32x32_test: add PartialTrans32x32Test, Random
vpx_fdctNxN_1_c: remove unnecessary store
vp9-denoiser: Adjust usage of skin and consec_zeromv.
vp8 denoiser: Some adjustments to usage of skin and motion.
VP9: Refactor vp9_decode_block_tokens()
Merge "disable vp9_diamond_search_sad_avx"
Merge "vpx_ports/mem_ops.h: clear some -Wconversion warnings"
Fix some static analysis issues with resize
Merge "VP9: Eliminate up_available and left_available"
vp9-1 pass vbr: Use average QP for active_worst_quality, for inter-frames.
disable vp9_diamond_search_sad_avx
vpx_ports/mem_ops.h: clear some -Wconversion warnings
vpx_fdct32x32_1_c: fix accumulator overflow
Merge "vp8_peek_si_internal: quiet static analysis warning"
VP9: Eliminate up_available and left_available
Merge "Properly propagate out of memory errors."
Merge "Increase min-max q range for normal inter frames."
Properly propagate out of memory errors.
vp9-denoiser: Incorporate consec_zeromv in skin detection for denoising.
Merge "vp8-denoiser: Use the same skin detection for model=1 as in vp9."
Revert "third_party: Roll libwebm snapshot."
Merge "third_party: Roll libwebm snapshot."
vp9: 1 pass vbr mode at speed 5: switch to use mv.search to NSTEP.
third_party: Roll libwebm snapshot.
vp8_peek_si_internal: quiet static analysis warning
Merge changes Id0a490bc,I1b4a0504
Merge "Force recode for bad rate misses."
Merge "Fix some HBD scaling bugs."
vp8-denoiser: Use the same skin detection for model=1 as in vp9.
VP9: Eliminate extra mv clamp in decoder
Merge "Add a TODO for using tile rows while encoding with multiple threads"
vp9/inverse_transform_block_inter: move eob check
vp9/inverse_transform_block_intra: move eob check
Merge "Allows sub-second frame rates in vp9"
Merge "Prevent encoder crash caused by row tile dependencies"
Add a TODO for using tile rows while encoding with multiple threads
Merge "vp9: Non-rd partition selection: use consec_zeromv to set sb_is_skin."
Merge "Simplify the loopfilter synchronization logic in VP8 encoder"
Prevent encoder crash caused by row tile dependencies
Simplify the loopfilter synchronization logic in VP8 encoder
Increase min-max q range for normal inter frames.
Force recode for bad rate misses.
vp9: Non-rd partition selection: use consec_zeromv to set sb_is_skin.
Merge "vp9: Improvement to skin detection."
vp8-denoiser: Update on the sum_diff threshold.
Merge "altref_test: mark AltRefForcedKeyTest as large"
altref_test: mark AltRefForcedKeyTest as large
vp9: Improvement to skin detection.
vp9: Fix the threshold for gf update in cyclic refresh.
Merge "Use whole pixel only at speed 8 screen content."
Merge "Reuse computation in build_nmv_component_cost_table."
Merge "Fix using screen content tuning in all the cpu speed tests."
Merge "vp9-svc: Bugfix on usage of svc_force_zeromv in non-rd pickmode."
Fix using screen content tuning in all the cpu speed tests.