(07-17-2020, 05:55 PM)Ofnuts Wrote:(07-17-2020, 09:18 AM)Ben321 Wrote: Most graphics software performs processes on the perceptual pixel values, and this should be the default, unless the user expressly selects a different mode of operation. If it doesn't do this, I consider the software buggy.You are making a virtue of necessity. Processing perceptual values is done because they can't do better or don't know better.
I use Gimp for processing scientific images. The source files have are BMP files, and the values stored in BMP (and most other formats) are intended to be treated as perceptual values. As such, when Gimp loads them in, it uses these values as perceptual values. Yet Gimp will perform any convolution filtering in linear number space, which ends up with wrong results. You see, when I perform filtering on a scientific image in such a manner that should cut a certain value in half, I expect to be able to store that value back into a BMP file and load that file in another program (even of only something as simple as a hex editor) and see that the value was actually cut in half. If the value is NOT actually cut in half, (as is the case when Gimp performs perceptual to linear transformation before filtering, and linear to perceptual transformation after filtering) then the values stored back into the BMP file are INCORRECT, as they are DIFFERENT than the values that SHOULD be present after that filtering of the scientific data in question.