Leica Q2

[BJL]

Dan, in practice, diffraction only significantly limits the resolution potential of a Bayer CFA sensor when the aperture ratio is about twice or more the pixel spacing in microns, so for example the iPhone XS main camera with 1.4 micron pixels and f/1.8 lens is fine. Maybe you are mislead by the oversimplified reckoning that the Airy disc diameter cannot exceed the pixel width, sometimes seen on Internet forums.

Anyway, in the main realm of our discussion, we are not close to diffraction limiting the resolution available with good lenses at their largest “good” apertures, which are no longer limited by the old rule of thumb about “two stops down from wide open”.

[Dan Wells]

By BJL's formula (which is more optimistic than many I've seen), a phone sensor will be diffraction-limited around f2-f2.8 (pixels ranging from 1 to 1.4 microns), with most compact cameras (other than large-sensor compacts) limited around f4.  Micro 4/3 and large-sensor compacts will be limited somewhere around f5.6-f8, with their 3-4 micron pixels. Larger sensor DSLR/mirrorless cameras will reach their limit between f8 and f11.

By that calculation, only a few lenses are diffraction-limited wide open with the sensors they are most commonly used with. Many compact superzooms are diffraction-limited, with perhaps the most egregious example being the Nikon Coolpix P1000 (1.4 micron pixels and an f8 lens!). A few inexpensive Micro 4/3 lenses may be diffraction-limited wide-open on the 20 MP sensor (there are several with f6.3 maximum apertures on the telephoto end). You might see diffraction wide-open on an APS-C or high-resolution full-frame DSLR with certain lens/teleconverter combinations, but not with any lens used alone.


I've seen visible evidence of diffraction (in tests, not so much in my own images) wider than that by a stop or so. Absent diffraction, a lens would tend to sharpen up as it gets stopped down, no? Most lenses on 40+ MP full-frame cameras are clearly sharpest at f5.6, beginning to soften by f8 - really high performance lenses are equally sharp from some wider aperture to f5.6, but they too soften a bit at f8. The softening at f8 doesn't show up on 24 MP FF cameras, but it does on 24 MP APS-C.

[BJL]

By BJL's formula (which is more optimistic than many I've seen), a phone sensor will be diffraction-limited around f2-f2.8 (pixels ranging from 1 to 1.4 microns), with most compact cameras (other than large-sensor compacts) limited around f4.  Micro 4/3 and large-sensor compacts will be limited somewhere around f5.6-f8, with their 3-4 micron pixels. Larger sensor DSLR/mirrorless cameras will reach their limit between f8 and f11.

That sounds about right, and good lenses for any of those ILC formats are available and comfortably usable at those f-stop limits.

Also, remember that diffraction effects scale in a way that imposes the same limit on image resolution (lp/ph at 50% MTF or whatever) at the same DOF (viewing same sized images from the same distance). For example, the f-stop giving "40MP" resolution in 4/3" is about half that giving the same 40MP in 36x24mm (maybe f/4 vs f/8) and both give equal DOF once one allows for the larger format using twice the focal length and half the degree of enlargement (along with four times the ISO speed if one needs equal shutter speed!)

So let's not make the mistake that the smaller formats are hampered in DOF by those lower limits on "f-stop giving full resolution".

BTW, for the 40MP high res. mode of the OM-D E-M5 Mk II and 50MP high res. mode of the OM-D E-M1 Mk II, Olympus limits aperture to f/8, presumably as its estimate of where diffraction limits start hurting.