One should probably not overlook the subtle data processing performed in all modern cameras before the 'raw' files are written to the memory card. Nor should one overlook the blurring caused by conventional AA-filters. So, when DxO Marks "measures" sharpness of a given lens, they rightly consider the qualities of the system as a whole. That is, they always specify on what camera a given lens has been tested.
Further,
'Sharpness' isn't really a well-defined, measurable quantity in a strict, physical sense. But
'Contrast' and '
Resolution' are. If we take Resolution alone, one could argue that we already are at a point, where camera sensors outresolves lenses:
The
Raleigh Criterion tells us that for ideal, truly diffraction limited lenses the resolving power only depends upon the wavelength of light and the f-number of the lens as follows:
- Spatial resolution = 2.44 * wavelenght * f-number.
For example, with yellow light - wavelenght = 0,000580mm - one gets a spatial resolution on the sensor of = 0.004mm at f/2.8 and 0.008mm at f/5.6. Now that is pretty much the range of pixel sizes as we know them on today's APS-C and FF cameras.
And no photographic lenses are perfect, truly diffraction limited.