CO wrote:
Would you really use the 18-300mm on your D850? I just went on DXOMark and checked the perceived megapixels of the Nikon 18-300mm f/3.5-5.6 G ED VR on the high resolution D7100. It has 9 perceived megapixels. The Nikon 24-70mm f/2.8E ED VR has 20 perceived megapixels on a Nikon D800E.
The 18-300mm has an overall score of 15. The 24-70mm has an overall score of 29.
The information you show on the DXO comparator is correct, but I am pretty sure the interpretation is not, since the screenshot clearly shows the 18-300 has a score of 15 pmp, on the D7100, and we have Tony Northrup to that for this with his video where he "explains" perceptual megapixels (pmp).
What pmp refers to is the percentage of resolution that a viewer will perceive with a given lens, compared to a "perfect" lens on the camera. On a 12.3 mp camera, like a 12.3mp D90, the "perfect lens" would produce a 12.3 pmp score. The perfect lens on a 24.3 mp D7100 would similarly show pmp score of 24.3. This is the basis for the pmp scoring system.
in the real world, there are no "perfect lenses."
The Nikkor 18-300 shows 9 pmp on a D7100 or 9/24.3 pmp, or 37% of the resolution the camera is capable of.
The same lens on a 12.3 mp D90 has a score of 6 pmp, or 6/12.3, or 49% of the resolution the camera can produce.
Another way to look at the pmp score is comparing a real score to the theoretical one using a lower resolution sensor and a perfect lens. Used in this fashion, the 9 pmp score of the 18-300 on a D7100 shows that it would be the same as using a perfect lens on a 9 mp camera, and the same lens on the D90 would be like using a 6 mp camera with a perfect lens.
DXO has no data on any DX lenses being used on a full frame camera, because typically no one does that.
Do these results imply that the 18-300mm is somehow less sharp on a lower resolution sensor? Absolutely not - what it does infer is that it allows more than 49% of a 12 mp camera's resolution capacity to come through, and that putting the same lens on a camera with higher resolution will be less successful. The score for the lens on the 24.3 mp camera clearly illustrates that. If you were to compare the images from both cameras, you'd likely see very little actual difference at normal viewing distances.
But the other consideration is how does a "better" lens perform on the D7100?
Not surprisingly, the excellent Sigma 18-35 F1.8 returns scores of 9 pmp on the D90 (73%), but 17 pmp (70%) on the D7100. This does not imply that the lens is less sharp on the higher mp D7100.
So, would it make any sense to get the Sigma (or any other known to be very high resolution lens) for a D90? Probably, since the lens will be a better match for a subsequent higher mp camera upgrade down the road.
This is how the DXO Mark rating system should be used.
This is an article that explains pmp with some other examples:
https://petapixel.com/2012/12/17/perceptual-megapixel-mtf-charts-boiled-down-to-a-single-number/