Ephotozine Review: Z 100-400/4.5-5.6 S VR

Leonard Shepherd wrote:

kenw wrote:

As a simple thought experiment you might test a lens on 12MP sensor and because the sensor simply can't resolve the very finest detail the MTF results would show the center, edges and corners perform nearly the same.

Sorry - you seem to repeat a fairly common internet myth.

No lens out resolves any sensor, and no sensor out resolves any lens.

If you use a higher resolving lens on a 12MP sensor - image resolution goes up.

While the highest image resolution is achieved with a combination of the highest resolving sensors and the highest resolving lenses, 12 MP is not a ceiling for image resolution.

Apologies for my brevity.  I was making no such claim, at least none of the ones you list in your post above.

But if you look at the MTF of a 12MP sized pixel vs the MTF of a 60MP sized pixel and note their shape you will see that in fact for a MTF50 test such as Imatest does where it actually measures the system MTF rather than just the lens MTF you will see that using a lower resolution sensor will in fact smooth or flatten the reported MTF50 field map as I described.

Indeed there is no “ceiling” for a sensor of any given resolution.  But ever sensor has an MTF itself (really we should say “pseudo-MTF” to appease the pedantic) which multiplies itself with the lens MTF.  Imatest effectively picks the resolution at which the contrast is 50% and reports that as its MTF50.  Since the MTF curves are not linear that means using a different resolution sensor does not simply linearly scale the reported MTF50 values.  They change relative to each other in the manner I described.

The myths you listed are pernicious so certainly worth taking the time to debunk them again!