The penultimate paragraph in the David Wells photo techniques article that I have been discussing is the following:
The point is to get a good exposure for a raw file, even if it looks too light as seen on the back of your camera. Do this to avoid a “combed” histogram which results when you take a digital file that looks good but is in fact underexposed. You are wasting the best tones (the ones to the right). In order to get a good print with a full range of tones you will have to go into a program like Photoshop, spread the tones out and push them to the right. As you do that, gaps are created as you spread the darker tones apart, moving tones that were to the left into the middle or the right side. The resulting gaps between the tones result in the dreaded combed histogram and a print with course tonal transition [sic], bad colors and noise.
I have problems with this explanation on two levels. The first, and most obvious, is that applying curves to shift tones in any direction (not just brightening them) can, with certain bit depths, cause the effect that Wells is describing. Therefore, if you are terribly worried about the effects of applying curves or curve like settings to your images during the raw conversion process, the best thing to do is not use the “expose to the right” methodology, but place the tones where you want them in the final image as you make the initial exposure, a la the Zone System. However, if you do this, you will leave yourself open to other potential problems during the editing process, which is why most people who teach digital photography recommend some sort of expose to the right strategy.
My larger problem is the idea that the “combed” histogram actually occurs in today’s workflows. I would like to take a few posts and explain the nature of the effect, why it used to be a problem, why it no longer is a problem in proper workflows, and why the way Photoshop implements its histogram tool makes it look like a problem even when it isn’t.