I’m not talking about artistism, but about basic treatment. Once you have more or less accurate basic chroma curve, you can add plenty of artistism later in the pipeline with fewer color arrifacts.
Yes, I replied quickly without really thinking about it, just from the top of my (empty) head!
Well, that’s still what I observe: if I move the right/left points but keep the symmetry, I feel like the overall color equilibrium is preserved, while departing from the center introduces color shifts.
1 Like
Sure they are. This is somehow continues the discussion I started about the filmic module’s colors. I may repeat here that my usual goal is to reproduce feelings from the original scene rather than get exact colors. I’d even say more: accurate reproducing colors will never get you the same feeling due to the nature of human perseption. So you’ve have to pick one of 2.
As for Lab tone curve, I use it constantly (and was a little bit frustrated when it defaulted to RGB as for some release, so I’ve to adjust the mode every time (probably it is changeable somewhere but I don’t personally find dt too friendly with setting your own defaults). So it’s one of my favorite tools, anyways. I normally use it to increase the saturation and – more important – color separation. So I make *a and *b steeper keeping the center point intact. It’s not always symmetric and I often make *b steeper than *a, especially for landscapes to separate green tones more. If I need something more I sometimes add some points to correct certain colors, but more often I use color zones for that. All depending on a scene.
Here’s an example using the b* curve in RawTherapee:
-
Original image (neutral + automatched tone curve):
-
Increasing the saturation of b* channel colors:
-
Decreasing the saturation of b* channel colors:
-
Asymmetric curve, cooling effect:
-
Asymmetric curve: warming effect:
-
Cooling effect using the middle point of the curve:
-
Warming effect using the middle point of the curve:
This is from a “dumb” user perspective, I don’t know if any of this is the correct way to do things, but it shows some interesting things IMO, regarding color balance.
One can also play with inverting the curve in one channel, to get a partial color inversion, for special effects. Fun stuff.
3 Likes
I also found another use case of the a* or b* curves: partial color desaturation.
In the following image of New York, if want to keep mostly blue and yellow colors, I can decrease the saturation in the a* channel:
Before:
After (look at the pinks and greens which are now less saturated):
If I try to keep the greens intact, the asymmetry introduces a slight color cast (in such a case, better use a S=f(H) or C=f(H) equalizer):
3 Likes
I was addressing this statement:
and agreeing on how it doesn’t affect the shadows and highlights in the way that L* does. Still affects the tone, as you said, as long as the curve is symmetrical about the mid-point.
1 Like
Yes you’re totally right, and I was wrong. I don’t know why, I thought that some time ago I was able to use the a* and b* curves to emulate the (in)famous “teal orange look” to render the shadows bluer and the highlights more orange. But my memory didn’t serve me well.
It still might have but not in the way you might have assumed. My guess is that the shadows already exhibited bluish shades, and the highlights, warmer; so an s-curve eased the image in that direction anyway.
Hey Claes,
Have you seen this “normal” edit of a landscape picture?
Ja, zeker 
1 Like