Fix a spatial svc bug related to scaling
Condition use of minmax in variance partition on speed setting.
Merge "VPX: x86 asm version of vpx_idct32x32_1024_add()"
Merge "vp9 denoiser: Fix to re-evaluate mode selection."
vp9: Update to noise estimation for denoising.
vp9 denoiser: Fix to re-evaluate mode selection.
VPX: x86 asm version of vpx_idct32x32_1024_add()
Merge "vp9 denoiser: Re-evaluate ZEROMV after denoiser filtering."
Merge "Speed up tm_predictor_8x8"
Merge "configure: simplify x86 asm dependencies"
Merge changes Iaf8cbe95,I6748183d,I2a49811d
Merge "vp9 non-rd pickmode: Fix logic in reference masking."
Merge "bitreader/writer: Change shift to signed"
Merge "Deduplicate some high bit depth tables"
vp9 denoiser: Re-evaluate ZEROMV after denoiser filtering.
vp9-denoiser: Fix to reset frame_stats.
vp9 non-rd pickmode: Fix logic in reference masking.
Merge "VPX: Removed unnecessary pmulhrsw in IDCT32X32_34"
Merge "VP9: Only zero counts when !frame_parallel_decoding_mode (2)"
VP9: Only zero counts when !frame_parallel_decoding_mode (2)
vp9-svc: Fix to allow setting qp-max/min per spatial and temporal layer.
VPX: Removed unnecessary pmulhrsw in IDCT32X32_34
Revert "Speed up h_predictor_4x4"
Merge "Use Interlocked calls in win32 once() implementation."
Merge "Speed up h_predictor_4x4"
vp9 denoiser: Bias to last reference for temporal filter.
Merge "vp9-svc: Fix the setting of is_key_frame."
Merge "Fix a signed shift overflow in vpx_rb_read_inv_signed_literal."
Merge "Fix unsigned overflow in rd_variance_adjustment."
Merge "VPX: x86 asm version of vpx_idct32x32_34_add()"
vp9-svc: Fix the setting of is_key_frame.
Merge "Changes to best quality settings."
bitreader/writer: Change shift to signed
Fix a signed shift overflow in vpx_rb_read_inv_signed_literal.
Fix unsigned overflow in rd_variance_adjustment.
Merge "Speed up tm_predictor_4x4"
vp9-svc: Fix to key frame counter for spatial layers.
Merge "Changes to exhaustive motion search."
Fixed a few sanity checks.
VPX: x86 asm version of vpx_idct32x32_34_add()
Changes to best quality settings.
configure: simplify x86 asm dependencies
Enable resize test(down&up) by changing the bitrate.
Use Interlocked calls in win32 once() implementation.
Changes to exhaustive motion search.
Reduce sampling time for noise estimate.
Merge "libs.mk, testdata: rm redundant test of LIBVPX_TEST_DATA"
Merge "Add AVX vectorized vp9_diamond_search_sad"
Merge "VP9 noise estimation: add frame level motion metrics and adjust thresholds."
VP9 noise estimation: add frame level motion metrics and adjust thresholds.
Revert "Update to noise estimation."
Adjust variance threshold for 16x16 split at low resolutions.
Merge "Non-rd partition: reduce variance threshold low resolutions."
Merge "Adjust varianace threshold for high noise condition."
Adjust motion threshold to limit cyclic refresh.
Merge "vp9 denoiser: Add another noise level to denoising."
Merge "Revert "VPX: x86 asm version of vpx_idct32x32_34_add()""
Adjust varianace threshold for high noise condition.
Add AVX vectorized vp9_diamond_search_sad
vp9 denoiser: Add another noise level to denoising.
Revert "VPX: x86 asm version of vpx_idct32x32_34_add()"
Revert "VP9: Only zero counts when !frame_parallel_decoding_mode"
Merge "Add bias to zero/small motion for noisy source."
Merge "convolve_copy_sse2: replace SSE w/SSE2 code"
Merge "VPX: x86 asm version of vpx_idct32x32_34_add()"
libs.mk, testdata: rm redundant test of LIBVPX_TEST_DATA
Non-rd partition: reduce variance threshold low resolutions.
VPX: x86 asm version of vpx_idct32x32_34_add()
Add bias to zero/small motion for noisy source.
convolve_copy_sse2: replace SSE w/SSE2 code
VP9 noise estimate: no noise estimate if frame size change.
Merge "VP9: add unit test for realtime external resize."
Merge branch 'javanwhistlingduck'
Merge "VP9 dynamic resize: enable resize unit test(DownUp)."
Merge "VP9: Only zero counts when !frame_parallel_decoding_mode"
Merge "VP9 dynamic resize: increase waiting time after key frame."
Merge "Use same bias (against non-zero mv for big blocks) for speed 5."
Merge "Expand unconstrained nodes in pack_mb_tokens and loop on zeros."
VP9: add unit test for realtime external resize.
VP9 dynamic resize: enable resize unit test(DownUp).
VP9 dynamic resize: increase waiting time after key frame.
VP9: Only zero counts when !frame_parallel_decoding_mode
Use same bias (against non-zero mv for big blocks) for speed 5.
Merge changes I8c83b86d,Ic53b2ed5,I4acc8a84
Merge "Use accurate bit cost for uv_mode in UV intra mode RD selection"
Merge "Revert "Add AVX vectorized vp9_diamond_search_sad""
Expand unconstrained nodes in pack_mb_tokens and loop on zeros.
Use accurate bit cost for uv_mode in UV intra mode RD selection
Revert "Add AVX vectorized vp9_diamond_search_sad"
variance_test: create fn pointers w/'&' ref
sixtap_predict_test: create fn pointers w/'&' ref