)]}'
{
  "commit": "d00a1befd9275278994f670bf2392aa189a649bb",
  "tree": "de708b5d0ee10ba7ebfffe70537640d5bf93c7d9",
  "parents": [
    "c5a71030d6adeb5f6cefc987134e335d0d79929a"
  ],
  "author": {
    "name": "Gerda Zsejke More",
    "email": "gerdazsejke.more@arm.com",
    "time": "Tue Apr 30 13:52:22 2024 +0200"
  },
  "committer": {
    "name": "James Zern",
    "email": "jzern@google.com",
    "time": "Tue May 28 20:24:03 2024 +0000"
  },
  "message": "Refactor ConvolveScaleTest unit tests\n\nThe filter values used in the scaling algorithm are specified in the\ndocumentation of the inter-prediction process (chapter 7.11.3.4. [1])\nwhich process contains the scaling algorithm. An important\ncharacteristic of these filter values is that all of them are even\nand we can leverage this information when optimising the\nav1_convolve_2d_scale function. However the unit tests do not follow\nthis pattern, the test currently uses calculations to generate filter\nvalues instead.\n\nTo enable us to take advantage of specific filter characteristics,\nadjust the unit tests to use the predefined filter values of\nav1_interp_filter_params_list defined in filter.h file.\n\nAdditionally, update convolutions parameters when testing with\ndifferent is_compound values.\n\n[1]https://aomediacodec.github.io/av1-spec/av1-spec.pdf\n\nChange-Id: I6ed1f7b269d139e221cbe35a573e06f0e2adaf59\n",
  "tree_diff": [
    {
      "type": "modify",
      "old_id": "76cf77ab073c36f98071366f01e17a389d34446e",
      "old_mode": 33188,
      "old_path": "test/av1_convolve_scale_test.cc",
      "new_id": "764ac2fac67a36c8f454a4860ce219594baa4759",
      "new_mode": 33188,
      "new_path": "test/av1_convolve_scale_test.cc"
    }
  ]
}