Exactly - which is why I said "What matters is how much dynamic range is captured in the first place."
Applying an S-curve to linear digital data has the opposite effect; it compresses the DR at either end, so more input levels are squeezed (demagnified, if you will) into fewer output levels.
Do the reverse process, to linearize scanned film, and yeah, the stretching out of those levels gives you pretty nasty S/N. I obviously don't recommend this; I only remark that such manipulations can be done.
It is relevant because it illustrates that the effect of the curve shape can be changed at will - even with film. If I hadn't given that concrete example, someone would probably have (rightly) challenged me to back up my statement with evidence.
OK, at last you are finally telling me why you got so mad. I apologize for saying that you confused long exposure with high ISO performance. It was an honest mis-interpretation of your post, and there was no ill intent.
But I challenge you to show me any place in the thread where I called you "an ignorant". I certainly did not.
OK! Now we can be friends again! "Ignorant" is not a word you used it is an obvious result to characterize anyone that would confuse "long exposure" with high "iso performance" as you presented me.
1. The DR captured in digital is less in the highlights, not in magnitude, but the actual HLDR because it has less compression (ie film records more in less HL range), the rest which is retained (at the moment) with analog, is lost ...so its nothing left in digital to recover (yet), in the "lowlights", digital may be considered comparable or better (depending of how much noise is acceptable from the photographer) but because the total DR range is more (or less, or equivalent, depending on the noise criterion) it doesn't mean that you can move it by simply underexposing because the S-slope curves don't much and since S-slope is not linear, underexposing will create the problems earlier quoted and unnatural image. Even more, depending on the highlights needed to be recovered, some (much?) of it will still be left left out. I don't deny that are people that consider lowlights as important as highlights, but I don't (!) and most of the great photographers of the previous century didn't, if I was to judge importance, I would say "1 stop of HL for 2 stops of LL" but thats me. So DR measurement as total and up to an individuals standard for low light noise acceptance is for me irrelevant. The scientific experiment you quoted, I still think is irrelevant, they take pictures for a purpose, they don't do photography! This is very different, they want to research a certain fact that happens in a certain part of the light "phasma" so they set up equipment for that, it has nothing to do with correct exposure, its "their" correct exposure. Regards, Theodoros. www.fotometria.gr
P.S. Please excuse my poor English, I can do better, but I haven't practice it for a long time.