Hello
After verification by a developer, and then by Copilot, I made some modifications and optimizations to the code, without any impact on the user experience.
Here’s a reminder of what was done in this pull request:
-
Introduced the excellent Michaelis-Menten algorithm, which is simple, efficient, and more user-friendly than GHS. However, it is not a replacement.
-
I added the following to the basic algorithm: a) internal saturation to avoid washed-out colors (following a user’s feedback); b) the ability for the user to optimize the black and white points. This is to better optimize the algorithm for contrast and out-of-gamut data; c) an LMS matrix for a more pleasing rendering; d) a display of these values (black and white points). e) I chose ‘RGB standard’ as the default mode (of course, I can add RGB Luminance mode).
-
I modified GHS in two ways (following a user’s comment): a) setting RGB standard as the default mode instead of RGB luminance (which, from what I’ve seen, without being able to test it, is the default mode for ‘Pixinsight - Astrophoto’); b) modified the way equivalent Luminance is calculated in out-of-gamut data (for RGB Luminance), to avoid certain color drifts in extreme cases.
-
I also modified part of the algorithm (entirely based on wavelets) that performs the “Contrast Enhancement” function in ‘Abstract profiles’ by avoiding areas of very low and very high highlights. In practical terms, this can prevent minor artifacts that could occur.
If no one objects, or suggests changes to the Labels or tooltips, or wishes to report other undesired behavior (saturation, etc.), I will push this PR in the coming days.
Thank you to everyone for contributing to evaluating and improving these algorithms.
Last executables
Michaelis-Menten
Jacques