1. 21 Dec, 2016 4 commits
  2. 20 Dec, 2016 2 commits
  3. 19 Dec, 2016 7 commits
    • Nathan E. Egge's avatar
      Delete duplicate cdf updating code. · b69cb528
      Nathan E. Egge authored
      The mv cdfs are updated below in calls to av1_set_mv_cdfs().
      av1-master@2016-12-19T17:27:14Z -> av1-set-mv-cdfs@2016-12-19T17:27:54Z
        PSNR | PSNR Cb | PSNR Cr | PSNR HVS |   SSIM | MS SSIM | CIEDE 2000
      0.0000 |  0.0000 |  0.0000 |   0.0000 | 0.0000 |  0.0000 |     0.0000
      Change-Id: I6061602a4f2cc91dadf79e39d513c9678b9d2075
    • Guillaume Martres's avatar
      Fix uninitialized entropy contexts · e50d9ddb
      Guillaume Martres authored
      When enable_optimize_b is false in av1_encode_intra_block_plane the
      entropy contexts were never initialized.
      No changes on metrics for objective-1-fast when no experiment is
      Change-Id: Ic68913f6400d2becbaec3cc14214a0257530ed0b
    • Jingning Han's avatar
      Scale reference motion vector search step size · fab16037
      Jingning Han authored
      This commit allows the dynamic motion vector referencing system to
      scale its search range according to the coding block size. This
      provides higher search resolution for smaller size coding unit.
      The cb4x4 mode improves the compression performance across all the
      test sets:
               avg     low    mid    high
      lowres   2.8%    2.4%   3.1%   3.0%
      midres   1.3%    0.3%   1.8%   2.7%
      hdres    0.9%    0.5%   1.4%   1.5%
      Change-Id: I1bc501506a9f2f06071c5274391f6bd053b235a7
    • Jingning Han's avatar
      Refactor loop filter frame border condition · ed8f3964
      Jingning Han authored
      Use the proper scaling factor to decide if a block is sitting on
      the frame border. This refactor does not change the coding
      statistics of the code base. It fixes an enc/dec mismatch issue
      due to out of boundary memory access in the cb4x4 mode.
      Change-Id: Ia1e999c0f4e4ef10aac6120e69c1fb10a738dd4d
    • Jingning Han's avatar
      Fill the token cost for 2x2 transform blocks · 6a9b2400
      Jingning Han authored
      Refactor the fill_token_cost() function to automatically compute
      the token cost arry for all transform block sizes.
      Change-Id: I2f44c9c08fb169bc14282ba48bce23577b1ab184
    • Jingning Han's avatar
      Allow 2x2 transform block forward model update · 8363063d
      Jingning Han authored
      This commit allows the forward model update for 2x2 transform
      block size.
      Change-Id: Ie08c401e488b3872be0d92640468857f0aa0d0b3
    • Thomas Davies's avatar
      Turn on delta_q by default. · 0090c8fb
      Thomas Davies authored
      Also make sure that qindex is clipped to the quantizer range.
      Change-Id: I3163da4b45e190f9ab34982d1bbbefa5cba7514e
  4. 17 Dec, 2016 2 commits
  5. 16 Dec, 2016 9 commits
  6. 15 Dec, 2016 16 commits