<div id="linkdxomark">Notwithstanding the fact the professional photographers can work wonders in programs like Photoshop,
There are things you can do to increase the impression of sharpness, especially in a situation where you have the option of downsampling. But there is no way of actually increasing the amount of true detail from the original scene when it is not in the original picture.
how does DXOMark explain the fact that such a disparity in DXOMark sharpness scores can still result in photographs that show no appreciable difference as far as sharpness?
I don't know how they
explain it but there are possibilities that come to mind.
For one, sharpness usually drops off some toward the corners. I don't know how they arrive at the overall score but there are detailed measurements. A telephoto zoom might also be a lot sharper at about 2/3 of its focal range than at the very end of it, which a good photographer will take into account if possible in a given shooting situation.
You might want to compare the evaluations at photozone.de which I find somewhat more accessible and less complicated to understand at times.
Also, full frame digital cameras deliver a lot more pixels these days than are actually needed for a regular magazine print. As a rule of thumb, about six megapixels are actually sufficient for prints of any conceivable size. Viewing distance increases along with the print size, so fopr larger prints, you can generally just space the pixels further apart, aka decrease print resolution. Mostly, the extra pixels from the past 12 years or so of camera development just give more headroom for downsampling or cropping.
But even for a picture printed over a double page in a magazine at something like 12x16 inches and 300 dpi, technically, you'd only actually need 17.3 megapixels. So even with files from a seven-year old EOS 5D MkII (21.1 MP), that gives you the option of downsampling or centre-cropping from a larger file, even for the largest possible print situation in a magazine.
I have relied on DXOMark sharpness scores in buy several lens. Was I wasting my time?
Depends. There might be applications where sharpness is a lot more critical than print magazines. Generally, I find it overrated though. Most lenses are capable of delivering images perceived as 'sharp'. Once, I was hoodwinked into applauding the sharpness of a 4x7 print whose negative turned out to have been taken with a cheap old 110 film pocket camera ;-)
These days, I'm much more concerned with issues like consistency of performance across the field. Even distortion, because increasingly, manufacturers seem to go the MFT route of cheap lens designs undercorrected for distortion and then correcting that in-camera.
This makes for softer corners though and generally a slight reduction in field-of-view, which I find disappointing. I like to put in some effort in image composition and not necessarily place my main subject smack dab in the centre of the frame.