libdcp/src/rgb_xyz.cc, branch cpp17 DCP handling library https://git.carlh.net/cgit/libdcp/atom?h=cpp17 2023-10-21T22:42:22Z Remove logging of clamped values. 2023-10-21T22:42:22Z Carl Hetherington cth@carlh.net 2023-10-21T22:20:53Z urn:sha1:687e46f8770ee946ce149e2b893f3626052b6881 I can't remember seeing this logged for ages and it appears to cost nearly a second on the rgb_xyz benchmark on AMD Ryzen 7 5700U. Use an integer LUT for PiecewiseLUT2, hence removing a lrint and a multiply from the rgb -> xyz loop. 2023-10-21T22:05:28Z Carl Hetherington cth@carlh.net 2023-10-21T21:05:56Z urn:sha1:e171a0bbaf0dedbab261279561888e7259a40df7 Rename lut -> double_lut. 2023-10-21T22:05:28Z Carl Hetherington cth@carlh.net 2023-10-21T20:40:17Z urn:sha1:5c711929b4f1670ef756a7e54ea59d207c7253a6 Add rgb_to_xyz for packed int16_t. 2023-10-18T10:53:58Z Carl Hetherington cth@carlh.net 2023-09-03T22:48:10Z urn:sha1:353ed18e817980eaa3afac6cf5d26b1f187f8803 Adapted from a patch from Aaron Boxer for grok support. Split rgb_to_xyz() into two parts. 2023-10-18T10:53:58Z Carl Hetherington cth@carlh.net 2023-10-18T08:39:26Z urn:sha1:621926b6bfc15b6bb0d87c03b00bf831b46e7b14 Extract, test and slightly adjust the piecewise LUT for inverse gamma. 2022-05-08T22:21:21Z Carl Hetherington cth@carlh.net 2022-05-08T21:47:12Z urn:sha1:6de2bb3751b449abf06d44e4dda5e41b66df3f62 Fix scaling ranges to be the same as they used to be. 2022-05-06T21:47:55Z Carl Hetherington cth@carlh.net 2022-05-06T08:32:02Z urn:sha1:001b62c79d8f87badc5597d0335caa0f92dc42fd Use 2 LUTs for output gamma to improve accuracy (DoM #2242). 2022-05-05T19:09:40Z Carl Hetherington cth@carlh.net 2022-05-05T19:02:40Z urn:sha1:2a6b1bb060439453913e496840a3102bbff534df Allow LUTs to be created with a particular range. 2022-05-05T19:09:40Z Carl Hetherington cth@carlh.net 2022-05-05T18:56:14Z urn:sha1:cbbb9e63b7ec7a5c34f186ceafe5da6be8faa739 Use std::vector for LUTs. 2022-05-05T19:09:40Z Carl Hetherington cth@carlh.net 2022-05-05T18:31:26Z urn:sha1:ad7244de981a7dd0b9b4f8f3d62d4704f1968012