1. 04 Aug, 2017 13 commits
  2. 03 Aug, 2017 10 commits
  3. 02 Aug, 2017 12 commits
    • Angie Chiang's avatar
      Add txmg experiment · ad653a39
      Angie Chiang authored
      This experiment aims at merging lbd/hbd txfms
      
      So far this exp uses hbd transform on lbd path.
      The performances I observed are
      lowres -0.089%
      midres  0.065%
      (negative means performance drop)
      
      Started from here, two main things are needed to be done.
      1) Fix overflow due to quantizer noise
      2) Generate a 16-bit version from the hbd txfm
      
      Change-Id: I35bb1fc0cbb78decad2570ff5826ed665f739752
      ad653a39
    • David Barker's avatar
      Fix inconsistency in compound-segment masks · fc256542
      David Barker authored
      The value of 'mask_base' passed to diffwtd_mask is currently
      38 for the lowbd path and 42 for the highbd path. Going off of
      the mode name (DIFFWTD_38), presumably these are both supposed
      to be 38, so change the highbd path accordingly.
      
      Change-Id: I5fb0099c4b8b3ca3c4f211562401b12012f5c002
      fc256542
    • Urvang Joshi's avatar
      FRAME_SUPERRES bugfix · 0b90540e
      Urvang Joshi authored
      encode_width was used where encode_height should have been used.
      
      Change-Id: I722c8e80f95366e9d07aacddc729fe30d3f5af48
      0b90540e
    • Tom Finegan's avatar
      Remove dead experiment flag: onthefly_bitpacking · 3bc237eb
      Tom Finegan authored
      CONFIG_ONTHEFLY_BITPACKING no longer guards any code. Remove
      the flag from the configure and CMake builds.
      
      Change-Id: Id5605155bdedbf540fe5b9cea3899e8de5ee1062
      3bc237eb
    • Tom Finegan's avatar
      Organize aom_config_defaults.cmake and update docstrings. · 745e5d02
      Tom Finegan authored
      Change-Id: Ibf73c7fa69adfe83f6d57546b7a8f9da67638048
      745e5d02
    • Yi Luo's avatar
      Setup frame/tile boundary when frame/tile geometry changes · 10e23004
      Yi Luo authored
      Change-Id: I44bc9d8887526a5ee92bf79730fa3ce6c73b160b
      10e23004
    • Angie Chiang's avatar
      Integrate convolve_round with chroma_sub8x8 · b9a822b4
      Angie Chiang authored
      Change-Id: I9a1b5b6016cd1afbc52cdac4469acb79c412e475
      b9a822b4
    • Zoe Liu's avatar
      Enable flex-refs by default when altref2 is on · 438b3ae7
      Zoe Liu authored
      Compared against baseline with default enabled tools (except for
      ext-tx and global-motion for speed concern):
      
                       altref2 -> altref2 + flex-refs
      lowres: avg_psnr -0.395% -> -0.460%
      midres: avg_psnr -0.418% -> -0.478%
      
      In particular, flex-refs improves the coding performance for the
      following 3 clips while no impact on all other clips:
      
      bowing_cif.y4m:    avg_psnr  0.023% -> -1.022%
      pamphlet_cif.y4m:  avg_psnr  0.454% -> -1.111%
      snow_mnt_480p.y4m: avg_psnr -0.162% -> -1.948%
      
      Change-Id: I612c1ae5feb1f07d8bd5aaf67e21a076445e10b9
      438b3ae7
    • Todd Nguyen's avatar
      Add gaussian model to bgsprite. · 7b581827
      Todd Nguyen authored
      Changes:
      - Add two new functions build_gaussian and build_mask to bgsprite.
      (Unused for now).
      
      Change-Id: I298be87d9d6833c39e944e447ad8fa121d7d83de
      7b581827
    • Todd Nguyen's avatar
      Refactor bgsprite. · 1fd99c2f
      Todd Nguyen authored
      Changes:
      - Refactored alignment and blending out of stitch_image function
      - Variable name changes in stitch_image to increase readability
      - Changes to temporal_filter (under CONFIG_BGSPRITE flag) in preparation
      for blending of bgsprite and temporal filtered arf.
      
      Change-Id: I3c4e76f26d6586684079ae7e666881b6ce14b2db
      1fd99c2f
    • Wei-Ting Lin's avatar
      Fill ncobmc mode and motion mode costs using cdf · 060f7f2c
      Wei-Ting Lin authored
      Also remove cost update in encodeframe.c as the cost
      is now updated through the function av1_fill_mode_rates
      in rd.c
      
      Change-Id: Ib670fd078320c706747429bb86e0265bc08356fd
      060f7f2c
    • Angie Chiang's avatar
      Use 10 bits to represent adapt_scan probabilities · a506eb61
      Angie Chiang authored
      The performance drop slightly by using 10 bit probabilities.
      lowres: -0.048%
      midres: 0.007
      hdres: -0.06
      
      Change-Id: I5ba7b5607802d084a599b779e5745f88b31e2cbe
      a506eb61
  4. 01 Aug, 2017 5 commits