Upload & Sell: Off
The best I could find without spending to much time on it was a quickie portrait of my GF and her dog, but I think it illustrates the difference quite well. What we're looking for here is colour shifts when luminance varies.
I want the white backdrop and the black knitting fabric in her cap/cardigan to be neutral (this locks my WB point), and the rest to be quite saturated, but preferably without casts or shifts in colour. The saturation should be EVEN, so that the face doesn't look "blotchy" even if the colour is cranked way up.
My whitebalance is exactly the same in both pictures, and the clarity/vibrance/saturation too. The difference is all due to the curve being shifted before/after the saturation is modified... The difference is fairly subtle, and the flickr recompression sRGB doesn't help the case either, so I suggest you download the pictures and swap between them in your picture-viewer of choice if you can't spot it. First out: normal processing.
1. in the first example this section is left fairly "normal", almost standard settings.
2. Crank the saturation a bit to exaggerate differences later.
3. The grey histogram in this box is what the first three sections (WB, tone and prescence in the "basic" part of the pane) leaves after they've done their work.
4. And then the "curves" go to work on the grey histogram in (3.). Here, no change.
5. This, the top histogram shows what the final produce is, after passing "Tone Curve", "HSL", "Split Toning", "Detail" and "Vignettes. This is delivered to the output section of LR to be converted into your chosen output ColourSpace and maybe output sharpening and resizing before export.
Then all over again, but this time I first compress the picture a bit (lower contrast and exposure, higher brightness) before adding the same amount of saturation. Then i use an S shape in the curves section to shape the histogram back to fairy normal (the grey histogram in (3.) is noticeably narrower than in the first example). Note that the final histogram in (5.) look almost the same as in the first example. Note also that the partly yellow/orange cast in the face is gone even if I haven't touched the WB or colour controls. The shadowy part of the face, and the nose on the dog also shows another hue than the "normal" processing.
The important thing is that the differences between the two pictures are almost entirely in how the saturated areas behave, the dark knit, the backdrop and the grey parts of the dog's nose is almost identical.