That's only a part of the story.
Yes, because I did not want to write a several pages long article here

.
Nevertheless its a very important part because it demonstrates a different approach Kodak has chosen with several of the latest films.
The high frequency information transmission capacity of a given emulsion is effectively limited by the granularity (quite severely) and/ or where the MTF response falls off (as that enhances the visibility of the granularity).
That is theoretically correct. But it is overlooking the fact that in real world tests with lens+film even at a lower object contrast ratio of 1:4 (my standard test) the granularity of fine grained CN films, BW films and especially colour positive films is so fine that you can surpass 100 lp/mm easily. So you don't have significant resolution limitations because of the grain in this film class.
Therefore if you design an emulsion that has extremely high low frequency MTF response, and a slightly faster roll off at high frequencies it'll look both much sharper and finer grained than one that has a less strong MTF response at lower frequencies and a longer roll-off at higher frequencies.
Correct in theory. But not relevant here in this case (topic was Ektar 100) because Ektar is not designed in that way, no extremely high MTF at low spatial frequencies.
In the real world, despite what high contrast resolution tests might suggest, there are very real limitations as to how much useful resolution a given opto-mechanical system can record on film, but getting the highest possible MTF below 40 cyc/mm will look dramatically better perceptually than any of the claims over whether an extra 10 lp/mm at high frequencies makes a difference.
1. I don't do high contrast resolution tests. My tests are all done at a relatively low object contrast of only 2 stops (1:4). Because that is extremely relevant to practical photography, as you have many very important details of that object contrast in about all photos.
2. The possible resolution values with film have very significantly increased in the last years with the new, much improved lens designs. Former knowledge about system resolution values of lens+film from 40 years ago is partly outdated. Old books cannot tell the whole current story.
3. In the case of the comparison of Provia 100F to E100, Provia delivers both a significantly higher MTF at low spatial frequencies (therefore looks sharper in direct comparison), and about 10% higher resolution values at higher spatial frequencies (and that is seen in projection with the higher magnification).
4. Concerning Ektar 100: The difference to the best CN films is 20-30 lp/mm, and that difference is visible in bigger prints.
And that is what really matters in making convincing big enlargements - not whether a film can resolve a high contrast target at 110 or 120 lp/mm (the micro-fine detail difference is of no matter to viewers who aren't aerial recon analysts),
Again, my test results are
not based on high contrast targets, just the opposite.
And for big enlargements you also have to consider the following: At viewing distances of about 25cm the human eye can resolve about 6-10 lp/mm in a quantative way, so the test persons are able to count the lines precisely. But if you then show them a chart with 15-20 lp/mm, they cannot count the single lines anymore, but lots of test persons are still able to see a quality difference: They see that the test object has a higher resolution and more details than the first object, they just cannot quantify the difference. And that is also one of the effects you see in bigger enlargements (prints, projection) with higher resolving materials.
You can see the effect I summarised in your results too - while the midtone gradient is being brought down to a reasonable level, the toe and shoulder remain very sharp - very much a fundamental function of the emulsion - and not something that process alteration is probably going to alter significantly. In the right circumstances it can be aesthetically successful. With neg/pos, it's possible to get round that problem more effectively (squeeze more on to the straight line), but for reversal it's a problem.
I would not draw any conclusions in that direction from only two snapshots I've posted here, and in a different context. The overall contrast on that day was extremely high, so high that any photographic medium would have had problems. And that are scans, not all data on the original transparencies could have been recorded.
For reasonable conclusions you have to make direct side-by-side comparison tests under exactly the same test conditions
For the record, Ilfosol 3 is much sharper than Rodinal, and finer grained - but because of the heightened sharpness the grain is more visible than something like D-76.
If someone wants both excellent sharpness and very fine grain, then the way to go is the route of the latest innovations in developers, like SPUR HRX, JOBO Alpha and SPUR Omega X (will be introduced soon).
Much better performance in that regard than the old(er) formulations, for which production costs and high-volume production capabilities played a significant role as well (a bit simpler formulars).
Best regards,
Henning