>Perhaps it has something to do with capture settings.
Yes; things like color balance will have a big effect for example. You have to realize that to AS! - without any additional meta-data - a raw bayer pattern is just a bunch of grayscale pixels that sometimes looks to have a bit of a grid like appearance, meaning that on average every pixel on every odd row and ever odd column has a slightly different brightness and or noise characteristics then other odd/even combinations. With all traditional bayer patterns you have two green pixels, and one red and one blue for every 2x2 block in the image, guessing the two green pixels is most of the time not too difficult - they should have VERY similar noise AND intensity characteristics, but the red and blue are more likely to result in mistakes.
So either use the SER format that stores this bayer pattern in the metadata of the file, or force the correct bayer pattern.
Few more bugs fixed, including one important one for those wanting to stack tifs or pngs or jpgs in some funny formats (there was a memory leak when converting some formats):
>Emil, I guess I meant clients or users or...... Anyway thanks for getting back to me.
Ooh, clients sound interesting too
(I'll be decreasing the update time from now on, but please keep sending bug reports)
Edited by MvZ, 30 March 2017 - 05:24 PM.