1. 11 Apr, 2017 5 commits
  2. 10 Apr, 2017 12 commits
    • Zoe Liu's avatar
      Add compound-singleref (compound mode using single ref) · ed69b09f
      Zoe Liu authored
      This new coding tool is targeted to complete following two tasks:
      (1) Remove the inter mode NEWFROMNEARMV;
      (2) Add new inter coding modes that provide compound prediction but
          using single reference frame - initially NEAREST_NEARMV from
          single reference.
      
      Change-Id: Ifadf91147d0b12c2f62cfd79a72ce0c37c4ee949
      ed69b09f
    • Fergus Simpson's avatar
      frame-superres: Move resize from encoder to common · d0565006
      Fergus Simpson authored
      The resizing functions in resize.h and resize.c are useful for the
      frame super-res experiment. These functions will be needed in both the
      encoder and decoder, so the files have been moved into av1/common.
      
      Change-Id: I66154b7ec0eade0df460c4f4cf8eaa5f663c8904
      d0565006
    • Debargha Mukherjee's avatar
      Reduce precision of shear parameters to 16 bits · 27f6e66e
      Debargha Mukherjee authored
      Change-Id: I9cd9362edbb7b642f4b632bf574abfe5b2159ff3
      27f6e66e
    • Sarah Parker's avatar
      Turn on SEPARATE_GLOBAL_MOTION by default · 27f3f988
      Sarah Parker authored
      Change-Id: I7e64307f4bd8821e0abfe804d00d022414759706
      27f3f988
    • Fergus Simpson's avatar
      scaling: Avoid SSSE3 convolution instructions when scaling · fc59389f
      Fergus Simpson authored
      The SSE3 convolve functions do not work when scaling is involved. If
      either x_step_q4 or y_step_q4 is not 16 scaling will occur and an assert
      fails.
      
      This patch creates a C version of av1_convolve, av1_convolve_c. This
      function in turn calls all C versions of what av1_convolve calls. New C
      versions of these called functions have been created where needed and
      the same for the functions they call. This means that when scaling is
      enabled no asserts fail. av1_convolve_c is called instead of
      av1_convolve when x_step_q4 or y_step_q4 is not 16.
      
      Change-Id: I604d2716e034e23a0553fb7004133d3075514a7a
      fc59389f
    • Angie Chiang's avatar
      Use block_idx rather than block_raster_idx · 752ccce7
      Angie Chiang authored
      1) block_raster_idx is actually raster order only when tx_size
      is TX_4x8.
      It's very specific, so we should put it near to the place it's
      actually used.
      
      2) Sync the meaning of block_idx on encoder/decoder sides
      
      Change-Id: I7d37a992cb773503e29f9c0d9d2586e580aa6173
      752ccce7
    • Jingning Han's avatar
      Enable cb4x4 by default · e335c671
      Jingning Han authored
      Change-Id: Icda2f6bae803fe73090261c81ac9a0f65d55d8f8
      e335c671
    • Jingning Han's avatar
      Fix high bit-depth distortion computation · 6a9dbefc
      Jingning Han authored
      Resolve the segmentation fault issue in nightly run unit test
      AV1/AltRefForcedKeyTestLarge.Frame1IsKey/0
      
      Change-Id: Ib27ac0d41767f39114905b7edb6c220bf62214a1
      6a9dbefc
    • Debargha Mukherjee's avatar
      Make global model acceptance depend on params cost · e832d578
      Debargha Mukherjee authored
      lowres:
      BDRATE (overall PSNR): -1.661% (up from -1542%)
      
      Change-Id: I612cb16f1d2362ab6375a5ef3258a4ed0777ac29
      e832d578
    • Jingning Han's avatar
      Temporarily disable multi-thread loop filter in cb4x4 · 52ece884
      Jingning Han authored
      The loop filter parameters change due to many experiments. The
      multi-thread loop filter is built on certain assumptions that would
      require changes in many table contents. Temporarily disable the
      multi-thread loop filter pass and only maintain the single frame
      reference during the development process.
      
      Change-Id: I33a6a7a6fc4299f8fc414ae7f28c5a463af1e12b
      52ece884
    • Debargha Mukherjee's avatar
      Reduce max mv allowed for local warp · d49c5c41
      Debargha Mukherjee authored
      Reduces value from 512 to 256.
      
      Change-Id: Ia1f924e8b0f38baee812740d7f90e6b95c6df15f
      d49c5c41
    • Jingning Han's avatar
      Make choose_partitioning support cb4x4 · 6fc515e8
      Jingning Han authored
      This resolves unit test AV1/AqSegmentTest failure issue in higher
      speed settings.
      
      Change-Id: I3dfe804bb47796f254bbd87e266d190fe75b85c8
      6fc515e8
  3. 09 Apr, 2017 1 commit
  4. 08 Apr, 2017 4 commits
  5. 07 Apr, 2017 17 commits
  6. 06 Apr, 2017 1 commit
    • Yue Chen's avatar
      Disable OBMC + compound · 8636da61
      Yue Chen authored
      Disable OBMC mode if the current block uses compound prediction.
      Results if we only disable this combination for small
      block sizes are also listed.
      
      Loss of gain (lowres/midres/AWCYHL)
      no compound :         0.230%/0.105%/0.15%
      no compound <= 8x8:   0.043%/-0.016%/0%
      no compound <= 16x16: 0.153%/0.029%/0.05%
      
      
      Change-Id: I94c9d704172dd02e53f3b2004a8032a4f1265057
      8636da61