Thanks!
And here's the point I was making:
Today's color film is R-G-B and it has a sort of built-in compression with resulting "compression artifacts" -- that is, omitting colors from the greater range our eyes can see:
After that, anything that's been captured into the digital media is also compressed with it's own compression artifacts -- sub-sampling (4:4:4, 4:2:2, 4:1:1, etc.) and redundancy elimination (mpeg, et al). Media distribution compresses that even more due to space, cost, and technology constraints. So the media mush we end up playing with is pretty pitiful (fortunately, things are getting better and cheaper). The closer we stay to the original film, the better off we are.
In dealing with fixing up the inherent faults, color-space conventions that are useful within limits. The previously mentioned HSL is common and is intuitive. But it is only a representation that works best in small adjustments. When pushed beyond that usage, it produces effects that are actually misrepresentations.
Using HSL to analyze a picture, when saturation is max'ed, ...
reveals "damage". (Interesting that this looks allot like over-compression flattening.) But, those RGB numbers compared to the non-saturated picture ...
shows that increasing the saturation changed the proportions between the components of RGB to create a false impression of the type of picture damage (missing chroma rather than flattened color).
snicker demonstrated that "impossible" fixes could be made in RGB because "damage" in one or more of the parallel RGB channels could be patched with better condition parts from the others channels. (He was working on "crushed" or "blown-out" channels to recover detail, but any problem in between is likewise addressable.)
Anyway, from the RGB perspective, a fix might not be worth the effort. My quick tweak of the original picture's RGB gammas produced a nice color correction/adjustment without a hint of the "max'ed saturation damage":
I thought it might help with fixing the "hopelessly damaged" where it mattered.