-
So, I tested it. Works as expected in 8 and 16bit integer, breaks everything in float.
BTW, is this expected to be calculated as:
(totalAlpha + (sumOfWeights / 2)) / sumOfWeights;
or((totalAlpha + sumOfWeights) / 2) / sumOfWeights;
EDIT2: Shouldn't we be leveraging KoColorSpaceMaths for this?
Edited by Wolthera van Hövell
Please register or sign in to comment