Moving {sb, mb, b, ab}_index from MACROBLOCKD to MACROBLOCK.
Merge "Replacing (raster_block >> tx_size) with (block >> (tx_size << 1))."
Merge "vpx[dec|enc]: Clean up target OS based IO focused preproc abuse."
vpx[dec|enc]: Clean up target OS based IO focused preproc abuse.
Merge "Localizing NEARESTMV special cases in the code."
Merge "Cleaning up vp9_quantize_b_c() function."
Merge "Cleaning up read_mv_probs() function."
Merge "Adding read_reference_mode() function."
Merge "vpxenc: Remove support for IVF input."
Merge "[BITSTREAM]Fix row tile mode_info pointer setup"
[BITSTREAM]Fix row tile mode_info pointer setup
Merge "Correct a couple of typos"
vpxenc: Remove support for IVF input.
Merge "Improve vp9_idct4x4_1_add_sse2"
Merge "Disable zeroblock forcing for lossless coding mode"
Merge "Unifying tile decoding for both direct and inverse tile order."
Disable zeroblock forcing for lossless coding mode
Merge "Remove TEXTREL from 32bit encoder"
Merge "Removed unused rate parameter."
Replacing (raster_block >> tx_size) with (block >> (tx_size << 1)).
Merge "Add back vp9_short_idct32x32_1_add_neon which is deleted in cleanup I63df79a13cf62aa2c9360a7a26933c100f9ebda3."
Remove TEXTREL from 32bit encoder
Merge "Fix the variable naming in encode_block"
Removed unused rate parameter.
Merge "Miscelleneous changes in detokenize for speed"
Merge "disable avx/avx2 for old versions of MSVC"
Fix the variable naming in encode_block
Unifying tile decoding for both direct and inverse tile order.
Merge "Move SVC per-frame loop from sample app into libvpx proper"
Merge "Using pd->dqcoeff instead of pd->qcoeff in the decoder."
Using pd->dqcoeff instead of pd->qcoeff in the decoder.
Move SVC per-frame loop from sample app into libvpx proper
Merge "Replacing mi_{width,height}_log2 with num_8x8_blocks_{wide,high}_lookup."
Merge "Allow test data path to be set by preprocessor symbol"
Replacing mi_{width,height}_log2 with num_8x8_blocks_{wide,high}_lookup.
Merge "Removes stack allocation of token_cache"
Allow test data path to be set by preprocessor symbol
Merge "Missing _ means no sse3 for vp9_h_predictor_32x32."
Missing _ means no sse3 for vp9_h_predictor_32x32.
disable avx/avx2 for old versions of MSVC
Removes stack allocation of token_cache
webmenc: Clean up the truly egregious style issues.
Merge "Move WebM writing support out of vpxenc.c."
Merge "optimize decode_bool operation"
Move WebM writing support out of vpxenc.c.
Cleaning up vp9_quantize_b_c() function.
Add back vp9_short_idct32x32_1_add_neon which is deleted in
Adding read_reference_mode() function.
Cleaning up read_mv_probs() function.
Merge "Removing old code."
Localizing NEARESTMV special cases in the code.
Merge "Fixes to buffer update for temporal layers."
Merge "Unified approach for backward probability update."
Miscelleneous changes in detokenize for speed
token_cache changes in decoder
Merge "Remove one shot q experiment"
Merge "Fix the use case of plane_block_idx in sub8x8 RD"
Merge "Avoid mv cost check for invalid mv values"
Merge "Remove unused members from VP9_COMP"
Avoid mv cost check for invalid mv values
Unified approach for backward probability update.
Splitting partition_probs array into two arrays.
Remove unused members from VP9_COMP
Merge "Remove unused member variables from VP9_COMP"
Merge "Removing 'new' probability calculation from convert_distribution()."
Remove unused member variables from VP9_COMP
Merge "Add second ref frame check back in rdcost hist"
Fix the use case of plane_block_idx in sub8x8 RD
Add second ref frame check back in rdcost hist
Merge "vp9 ssse3 d207_predictor_32x32: add missing GLOBAL()"
vp9 ssse3 d207_predictor_32x32: add missing GLOBAL()
Removing 'new' probability calculation from convert_distribution().
Merge "Convert filter kernel choice to lookup"
Convert filter kernel choice to lookup
Fixes to buffer update for temporal layers.
Merge "Cleanup. Adding const to function pointer arguments."
Merge "Cleaning up read_skip_coeff() function."
Merge "simplify read_coef_prob()"
Merge "Removing is_intra_mode() function."
Cleaning up read_skip_coeff() function.
Merge "Reducing the number of foreach_transformed_block() calls."
Cleanup. Adding const to function pointer arguments.
Reducing the number of foreach_transformed_block() calls.
Merge "mb_lpf_horizontal_edge AVX2 optimization"
mb_lpf_horizontal_edge AVX2 optimization
Merge "Fix x_offset_q4/y_offset_q4 calculation"
Remove one shot q experiment
Merge "Enable all-zero coeff block index for sub8x8 blk"
Merge "Reducing the number of recursive calls."
Replacing (SWITCHABLE_FILTERS + 1) with SWITCHABLE_FILTER_CONTEXTS.
Enable all-zero coeff block index for sub8x8 blk
Reducing the number of recursive calls.