1. 06 Jan, 2018 6 commits
  2. 05 Jan, 2018 8 commits
    • Yushin Cho's avatar
      [dist-8x8] Remove asserts for sub8x8 distortions · 796d4e29
      Yushin Cho authored
      Remove the asserts in rd_pick_partition(), which check
      whether the sum of distortions of sub8x8 partitions matches to
      the 8x8 distortion measured in final reconstructed 8x8 pixels,
      because the sum of the sub8x8 distortions returned from rd_pick_sb_modes()
      is sometimes different from what is measured with final reconstructed 8x8 pixels.
      
      The asserts have been provided for the purpose of sanity check with PSNR
      and triggered with "--enable-dist-8x8=1" encoder option.
      
      Change-Id: I08e331900b19077cbb325ef5117688f05095dcc7
      796d4e29
    • Jingning Han's avatar
      Make txk-sel support cfl in key frame coding · 60db3c95
      Jingning Han authored
      Properly store and update the luma component result in the key
      frame rate distortion optimization process for chroma component
      prediction. The txk-sel provides 0.15% gains for key frame coding
      on top of lv-map/multi and all other default experiments.
      
      Change-Id: I694860607b318d0a84333ed928d3de85c7676623
      60db3c95
    • Jingning Han's avatar
      Replace full txfm rd search with luma writing · 159f72c3
      Jingning Han authored
      The cfl route needs the reconstructed luma pixels in the chroma
      RD search. Replace an extra route of txfm_rd_in_plane with the
      use of av1_encode_intra_block_plane to directly write out the
      needed luma pixels in the key frame rate-distortion optimization
      process.
      
      The compression stats remain identical. The key frame encoding
      speed is on average 3% faster for a few tested 1080p clips.
      
      Change-Id: I65760b657a6d4b45bed9c4f1eca0cff1a2c274a1
      159f72c3
    • Jingning Han's avatar
      Fix txk-sel rd search in the key frames · c5c37032
      Jingning Han authored
      If the transform block selects a transform kernel type that makes
      it all zero coefficients and hits the best rd cost, store such
      decision and force the transform block to be coded as all zero
      block in the final coding pass. This makes txk-sel improve the
      kf coding performance by 0.15% now.
      
      Change-Id: I4bc3c185b9c68daefb856276da45bbf9e317d02e
      c5c37032
    • Sebastien Alaiwan's avatar
      Revert "Turn lv_map lv_map_multi on by default" · 76658d2f
      Sebastien Alaiwan authored
      This reverts commit 2944f9d1.
      
      Reason for revert: it breaks encode+decode with error message:  
      "Warning: Failed to decode frame 1: Bitstream not supported by this
      decoder".
      
      BUG=aomedia:1193
      
      Change-Id: I43e0adae0f000982455ed4d838c0f8e841f77af0
      76658d2f
    • Zoe Liu's avatar
      Reduce the unnecessary array size · 316a6235
      Zoe Liu authored
      For partition types pruning in ext-partition-types, horz_rd and
      vert_rd are used to store the RD costs for the two blocks in
      horizontal partition and vertical partition. Hence array size of
      2 is sufficient for either array.
      
      Change-Id: I3471ac573665410af4c5b4af3c54bc60fe8c0195
      316a6235
    • Frederic Barbier's avatar
      Remove experimental flag of SIMPLE_BWD_ADAPT · f57a937c
      Frederic Barbier authored
      This experiment has been adopted, we can simplify the code
      by dropping the associated preprocessor conditionals.
      
      Change-Id: I09423f67995ea4bea6f1de71bfeef7f6f1c2ae19
      f57a937c
    • Debargha Mukherjee's avatar
      Resolve a static analysis unstable build · e2f6b166
      Debargha Mukherjee authored
      Change-Id: Idfb819b0500aa8a7f451fc57dc35ccdba63dd836
      e2f6b166
  3. 04 Jan, 2018 12 commits
  4. 03 Jan, 2018 6 commits
  5. 02 Jan, 2018 4 commits
    • Jingning Han's avatar
      Properly handle the tile load size writing · 2e7d66d9
      Jingning Han authored
      Properly handle the tile load size writing operations in ext-tile
      when large-scale-tile is on and off. This fixes an enc/dec mismatch
      issue when both ext-tile and obu are turned on.
      
      BUG=aomedia:1179
      
      Change-Id: I460e0cd55ba5d0bf5b75b5a5f6866c8774693207
      2e7d66d9
    • Jingning Han's avatar
      Fix ext-tile and obu · 5737b4c0
      Jingning Han authored
      Fix the broken unit test when ext-tile and obu are both turned on.
      The test failure was due to the merge of CONFIG_ADD_4BYTES_OBUSIZE.
      This solves the unit test failure cases when cm->large_scale_tile
      is on.
      
      BUG=aomedia:1179
      
      Change-Id: I7b4a5720e493db2a7d914fc9f8b886a67b1e85cd
      5737b4c0
    • Linfeng Zhang's avatar
      Make av1_cost_coeffs() inline · 2ba7d206
      Linfeng Zhang authored
      Change-Id: I5821e35d7cda96c20e22ab02bc2dfe8639025fde
      2ba7d206
    • Urvang Joshi's avatar
      Remove a TODO that's already taken care of. · 6e9c30b7
      Urvang Joshi authored
      Change-Id: I555fe53c90988d0e6fc392417b4758d558842e94
      6e9c30b7
  6. 01 Jan, 2018 1 commit
    • Zoe Liu's avatar
      Remove unused variables for ext-skip · 404d92e7
      Zoe Liu authored
      The removed variable was once defined for evaluating the use of
      distance-weighted compound prediction for the new skip mode, under
      the scenario when all reference frames are forward predicted.
      
      Change-Id: I114daa66d911ec8bb8ef220f7c519566037429b1
      404d92e7
  7. 31 Dec, 2017 2 commits
    • Dake He's avatar
      [lv_map_multi] adjust speed of prob adaptation · 5641635f
      Dake He authored
      Use a lookup table to deterimine speed of adaptation in update_cdf.
      Tests on 5 keyframes show improvement in BD-rate: consistently more
      than -0.2% for google test sets (lowres, midres, hdres).
      
      Change-Id: I6f8927fc6e876e22ebcf2f9340ed62e54cc79d28
      5641635f
    • Dake He's avatar
      [lv_map_multi] multisymbol coding of eob · 0db7d0eb
      Dake He authored
      1. code the range of eob in a transform block as a single symbol.
      2. initialize cdfs directly.
      3. initial cdfs are trained by using laplace prob estimates as in base
      level coding (CL 38638)
      
      Change-Id: I65c88d6dd0da3381feeec837928ac5bd81be4296
      0db7d0eb
  8. 30 Dec, 2017 1 commit