I'm doing my best to understand the sharpness scores and falling way short. I've got a Ph.D. in a science discipline and know a thing or two about psychophysics and measurement, so just maybe the problem is with DxO rather than me. In any case I would like to understand these scores so if someone could help me out I would be appreciative.
Rather than mixing-up lenses and bodies as DxO does for these super-telephotos, let's take two Nikon lenses that DxO has tested on the D800, the 500mm f4G and the 85mm f1.4G. DxO gives the 500mm a sharpness score of 25 whereas the 85mm gets a score of 40. This is what appears to be a whopping difference favoring the 85mm. But when I got to NikonUSA and examine the MTF charts for these two lenses, the 500mm f4g is pretty much stuck at a perfect value of 1 (http://imaging.nikon.com/lineup/lens/singlefocal/Telephoto/af-s_500mmf_4g_vr/
) whereas the 85mm f1.4g doesn't get higher than .9 (http://imaging.nikon.com/lineup/lens/singlefocal/normal/af-s_nikkor85mmf_18g/
). I'm not sure how Nikon is measuring MTF, but presumably they do it the same way for both lenses, so according to Nikon the 500mm is sharper than the 85mm.
Okay, DxO is reporting sharpness in a unit they're calling Perceptual Megapixels, which they believe is intuitive (NOT) and bypasses the complexities of swooping curves on a graph that is inherent in a MTF chart. But DxO says that the P-Mpix starts with the MTF measurements.
So how can a lens for which the manufacturer reports higher MTF performance get a substantially lower P-Mpix sharpness score than a lens for which the same manufacturer reports lower MTF performance? A Nikon 85mm is nearly twice as sharp as a Nikon 500mm? They are both incredibly sharp lenses, but I use them and find the 500mm a bit sharper, consistent with Nikon's MTF charts.
So what's up here? I've read the DxO technical documents and still don't get it.