Use a separate MODE_INFO stream for each tile column
Get rid of "this_mi", use "mi_8x8[0]" everywhere instead
Make the static_segmentation feature work again
Merge "Prevent accidental changes to the previous frame mode_infos"
Merge "change to use vp9_idct_32x32_add"
Inlining and removing fwd_txm16x16 and fwd_txm8x8 pointers.
Merge "Cleanup of decode_coefs() function."
change to use vp9_idct_32x32_add
Merge "Removing print_prob_tree function and vp9_coeff_probs typedef."
Merge "Moving FILTER_BITS constant from vp9_convolve.h to vp9_filter.h."
Merge "Using constants instead of plain numbers."
Implement variance-based adaptive quantization
Merge "lint-hunks: remove runtime/int filter"
Merge "Updated encoder to handle intra-only frames"
Merge "Remove Windows-style newlines using dos2unix"
Merge "Make vp9_zero use cases of consistent format"
Updated encoder to handle intra-only frames
Merge "Remove unused variable vp9_64x64_zeros"
Merge "Fix a few indent format issues in buffer defs"
Merge "Remove unused comment"
Merge "Re-design all-zero-coeff block index buffer use"
Make vp9_zero use cases of consistent format
Remove Windows-style newlines using dos2unix
lint-hunks: remove runtime/int filter
Merge "Converted assertion to returning error"
Merge "Added trap for invalid key frame"
Merge "Prevent access to invalid pointer"
Merge "Add clamp to prevent out of bound access"
Converted assertion to returning error
Added trap for invalid key frame
Prevent access to invalid pointer
Add clamp to prevent out of bound access
Remove unused variable vp9_64x64_zeros
Fix a few indent format issues in buffer defs
Removing unused 8x4 transform from the encoder.
Re-design all-zero-coeff block index buffer use
Merge "add a test vector with frame parallel mode enabled"
Moving FILTER_BITS constant from vp9_convolve.h to vp9_filter.h.
Removing print_prob_tree function and vp9_coeff_probs typedef.
Using constants instead of plain numbers.
Cleanup of decode_coefs() function.
add a test vector with frame parallel mode enabled
Merge "vp9_receive_compressed_data: remove unnecessary indent"
Merge "Move token_cache from cost_coeffs to MACROBLOCK"
Move token_cache from cost_coeffs to MACROBLOCK
Merge "Moving libmkv library to third_party folder."
Merge "Making input pointer of any inverse transform constant."
Merge "Adding TREE_SIZE macro + cleanup."
Merge "Adjust icc compiler options"
Adjust icc compiler options
Making input pointer of any inverse transform constant.
Merge "Get libvpx to compile on VS2013."
Adding TREE_SIZE macro + cleanup.
Consistent names for inverse hybrid transforms (2 of 2).
Merge "Consistent names for inverse hybrid transforms (1 of 2)."
Get libvpx to compile on VS2013.
Merge "Replacing {VP9_COEF, MODE}_UPDATE_PROB with DIFF_UPDATE_PROB."
Merge "Change in rddiv parameter to make it a power of 2"
Consistent names for inverse hybrid transforms (1 of 2).
Merge "Adding const to the input argument of all 1D transforms."
Merge "Masking intra mode choice adaptively"
Adding const to the input argument of all 1D transforms.
Replacing {VP9_COEF, MODE}_UPDATE_PROB with DIFF_UPDATE_PROB.
Merge "Removing vp9_tree_p typedef."
Merge "Removing vp9_idct4_1d_sse2 function."
Change in rddiv parameter to make it a power of 2
Masking intra mode choice adaptively
Merge "Code cleanup"
Merge "Restore mode skip feature in sub8x8 rd loop"
Merge "Experimental rate control change."
Experimental rate control change.
Adjustment to allowed range in resize unit test
Merge "Disable recode loop."
Merge "SSE2 8-tap sub-pixel filter optimization"
Removing vp9_tree_p typedef.
Removing vp9_idct4_1d_sse2 function.
d207 intra prediction ssse3 using bytes
Merge "Giving consistent names to IDCT 32x32 functions."
Merge "Consistent names for FDCT functions."
Merge "d153 intra prediction (32x32) ssse3 using bytes"
SSE2 8-tap sub-pixel filter optimization
Merge "Fix typo in comment message"
Merge "change to avoid out-of-range computation"
Merge "Re-design rate-distortion cost tracking buffers"
Prevent accidental changes to the previous frame mode_infos
Fix typo in comment message
Consistent names for FDCT functions.
Giving consistent names to IDCT 32x32 functions.
Re-design rate-distortion cost tracking buffers
change to avoid out-of-range computation
Merge "Removing inv_txm4x4_1_add and inv_txm4x4_add function pointers."
Merge "Moving all scan/iscan code into separate vp9_scan.{h, c} files."
Merge "Giving consistent names to IDCT 16x16 functions."
Merge "Adding const to several pointers."
Merge "Adjustment to mv cost parameters"
Merge "Fix intra dist model of skip_encode feature"