Originally posted by Marc Sabatella Whereas it is next to impossible to stuff up a test to make a lens look *better* than it actually is. If one given controlled test shows great sharpness, then that lens *is* capable of great sharpness, period, and no further tests will cause that to not be true.
I see several ways this might break down.
First, some lenses are good at a particular distance range, but less so in others.
A standard lens review might show a lens to be sharp
when focused on a test chart at a few meters,
but that lens could be much worse at infinity or in the macro range.
I see your point about the lens always being good at the intermediate range,
but that doesn't help a user who may be looking for sharpness at infinity.
Another case involves (typically wide-angle) lenses with strong field curvature.
A tester may refocus the lens between the central and the edge resolution measurements.
The lens will then look like it gives great sharpness,
but users will be disappointed if they're photographing flat objects.
A third case might arise with lenses that exhibit astigmatism,
where a test is picking up the "good" orientation, and disregarding the "bad."
There are yet other ways a lens could be made to look good,
for example post-processing the test image,
or Sigma's practice of showing 30 lp/mm MTF curves,
where other manufacturers show 40 lp/mm.
PS: Sorry if this is rambling off-topic too much.
They are called "threads," implying they can weave around a bit.