all photos ~ (embiggenable)
IT SHOULD COME AS NO SURPRISE TO ANYONE THAT seeking “expert” advice on the web is, at best, a needle-in-a-haystack enterprise, or, at worst, a fools errand.
Case in point, photography wise; re: using Curves in Photoshop. One can find a plethora of experts who will correctly point out that when you use the master RGB Curve tool in Photoshop to change density (brightness) or contrast, color saturation and hue balance shift unintentionally––to wit, Increasing contrast (S-curve): Pushing highlights up and shadows down makes bright areas brighter and dark areas darker, which heavily boosts color saturation. Colors can become oversaturated or clip. Decreasing density (darkening the overall image): Pulling the master curve down compresses values into darker tones, which visually increases color density and richness. Increasing density (brightening the overall image): Pushing the master curve up stretches tones toward white, which washes out and desaturates colors.
So far, so good, but …. that’s when after which the good advice goes south. What inevitably follows is a convoluted, multi-step procedure to solve the problem. However, apparently either outright ignorance is at play, or, perhaps the expert is just flexing his expertise muscles to impress the impressible.
In either case, while they are not flat out wrong, I can write that I never, not once, found a single “expert” who mentions the easiest, most assured, 1-step procedure to eliminate the unintentional saturation problem. Most likely that is the case because the answer requires venturing into a color space that most photographers know nothing about or, if they do know about it, they treat it as a place that confounds and confuses them to the point of treating it like poison … LAB–technically CIELAB–color space.
LAB color space contains 3 axis channels; L - lightness-aka: black values + A green / magenta values + B blue yellow / values. For purposes of this entry, we are only concerned with the L channel–the channel that illustrates only the black values, 0-255, in an image file. The L channel is not linked in any way to the color channels. So one can manipulate them til the cows come home with absolutely no effect on the color channels / color saturation. None. Zero. Zip. Nada.
So, 1-step; convert to LAB, work your magic, convert back to RGB. If, for instance, if you darkened the shadow end of the lightness curve, the the shadow areas in the RGB file will appear darker but the color, hue and saturation will be true to the original file. Easy, peasy. But wait …
Stop the presses––there is another “hidden: secret lurking in that LAB converted file; the L channel contains all of the grayscale data of the original image file, 0-255. So what have is a full-range BW image. Just delete the A / B channels, convert to Grayscale and there you have it… a full range grayscale file ready to be adjusted to any BW look your picture making heart desires.
It always amazed and confounded me that, when back in day when everyone was wailing and caterwauling about digital picture making being inept at creating BW photos, once again, nary an expert mentioned a word about the L channel in LAB color space - a pure, full range monochromatic image file just waiting for the taking.
Experts, my ass.