I understand that the magenta cast in blown highlights is the result of saturated sensor cells for all R- G- B- channels in combination with white balance. Let’s turn the white balance iop off to get a neutral color in this area and also decrease exposure to avoid clipping:
As far as I know, the sensor cells for all channels should be saturated for the white area (the sun), which should result in maximum R- G- B- values in the whole photo. But if we measure the RGB values, we can see that the R and G values for the yellow area (with unsaturated B values) are greater than the R and G values of the sun. How is this possible?
Thank you for your response but frankly I don’t understand what you mean. If you mean this is caused by turning white balancing off, turning it on doesn’t change this difference (at least for G channel in this photo).
whitebalance is on by default. it usually has coefficients like (2.0, 1.0, 1.8) for rgb. this makes your nicely clipped at (1,1,1) values magenta. if you don’t do wb it’ll be 1 1 1 but everything else is green (as you noted).
Actually my question is not about the wb or the green cast. My question is specifically about the max RGB values in the photo. When wb is off and values are clipped at (211,211,211) (the sun), how is the yellow area (as measured in the second screenshot) something like (213,224,175)?
I think it has to do with the type of demosaicing and colour space conversion you are using, which would cause certain RGB values to spike and exceed the gamut or luminance. In dcraw, it depends on which highlights mode you choose: clip, unclip, blend 2 or blend 3+.
