dcpomatic/test/video_level_test.cc, branch foo DCP-o-matic DCP tools https://git.carlh.net/cgit/dcpomatic/atom?h=foo 2021-05-25T19:35:12Z Move video level conversion for RGB from FFmpegImageProxy to Image. 2021-05-25T19:35:12Z Carl Hetherington cth@carlh.net 2021-05-24T22:57:16Z urn:sha1:62f9b78a2eb5f0fc6b9028264bac6ad501d83309 Since FFmpeg does not do video level conversion for RGB sources when we (sort of) ask it to in Image::crop_scale_window() it seems to make more sense to compensate for that by calling full_to_video_range() in the same place (rather than in FFmpegImageProxy). Clamp results correctly when shifting video levels, and account for that in tests. 2021-05-25T19:35:12Z Carl Hetherington cth@carlh.net 2021-05-24T22:32:39Z urn:sha1:71f3520ea315fb65ff151c99aa0b64fc8dccdb1d Use decoder_factory() and a lambda to clean things up a bit. 2021-05-25T19:35:12Z Carl Hetherington cth@carlh.net 2021-05-24T21:16:07Z urn:sha1:cc0f518888dcb750018458d784cc844aef478854 C++11 tidying. 2021-04-29T18:53:49Z Carl Hetherington cth@carlh.net 2021-04-29T07:14:20Z urn:sha1:39fb8198febde1937019db1c300ec363aab5aa56 More test fixes. 2021-04-23T15:43:56Z Carl Hetherington cth@carlh.net 2021-04-23T15:43:56Z urn:sha1:151f5c81fade29e9bebea9904fd85975351b7b78 Fix test for random noise addition. 2021-04-23T13:22:26Z Carl Hetherington cth@carlh.net 2021-04-23T13:22:26Z urn:sha1:4d0df359036f8a15b85ae64060959480a7ff695f Verify a whole bunch of DCPs made by unit tests. 2021-02-23T19:09:39Z Carl Hetherington cth@carlh.net 2021-02-21T19:28:59Z urn:sha1:41262015eb2bb1fc3da8585883420975de381a65 Test reference tweaks. 2021-02-15T02:16:16Z Carl Hetherington cth@carlh.net 2021-02-15T00:15:19Z urn:sha1:4f09ec8b35c5e4c32d17b340c6840699c85f280a Use a predictable RNG when adding noise to low-bitrate images. 2021-02-01T22:44:36Z Carl Hetherington cth@carlh.net 2021-02-01T16:46:22Z urn:sha1:49b655d3e1937018c46ed7f6a62e1157b247e426 The recovery code assumes that encoding a given frame from the same data will always give the same output; it always encodes frame 0 again to set things up properly and if that frame is a different size to the previous run everything breaks. More enum class additions. 2021-01-31T02:14:24Z Carl Hetherington cth@carlh.net 2021-01-31T02:14:24Z urn:sha1:8fedaaa75c4586a4cc7ffb393bd71d1fdb091dc8