Right. That's because r = 0.61 × λ/NA gives you a feature size at which you still have contrast that's usable though small, while r = 0.5 × λ/NA (=λ/(2*NA)) gives you the feature size at which contrast disappears altogether.Charles Krebs wrote:r = 0.61 × λ/NA seems to be more widely used and accepted
As a small attempt at humor, I might say that Nikon's numbers come from a calculation that tells you the pixel size needed to minimally capture nothing at all (2 pixels per feature at 0% MTF).
On the other hand, there's another way to interpret the same calculation.
With perfect optics, the MTF is about 40% at twice the cutoff size (half the spatial frequency) and about 20% at 1.5 times the cutoff size. (See the graph HERE.)
So, Nikon's numbers also tell us the pixel size needed to give 4 pixels per feature at 40% MTF, and 3 pixels per feature at 20% MTF.
Somehow I find this second interpretation more comforting.
--Rik