Rayleigh Limit and CCD Resolution

[PeterCris]

http://luminous-landscape.com/tutorials/resolution.shtml

The above mentions the following:

"The Rayleigh criterion – based on human visual acuity – isn’t adequate
for estimating the resolving power of a lens that projects images on
a sensor. The sensor needs more contrast and separation between
Airy disks than the human eye. Foveal cones aren’t like pixels."


This is why they suggested that we must match pixel size with airy disc
size and recommended 2 pixels per airy disc to obey Nyquist sampling
theory where sampling must always be twice or better than the signal.
But guys, in astrophotography, they used even more than Rayleigh
Limit by using the maximum wavelength/Diameter which is the
radius of the central disc in radian (so they used 4.7 pixels across
an airy disc for maximum resolution) even in planetary imaging
which is no difference from terrestrial photography since you are
after the most details in the object. So I think even details in
the Rayleigh limit (which is wider than the astrophotography
resolution) can still be resolved in the DSLR ccds as long as you
can make out any differences (no matter how tiny) that would entail
separations between two detail regions. Hence, the above website
must suggest 4 pixels per airy disc and not just 2 pixels.

What do you think?    

Pete

[ErikKaffehr]

Hi,

Just a couple of issues:

1) Photography is not about resolving patterns but seeing them. What we perceive as sharpness has little to do with resolution but much with edge contrast. Edge contrast can be improved by sharpening but not without artifacts.
2) A color pixel is synthesized from four bayer filtered sensels. That makes the resolution issue more murky.
3) We may have an ideal situation where two patterns are projected on two proximate sensels. If both patterns were bright we would need a dark valley to separate them. If the items would be equally bright we would not be able to resolve them without the additional pixel needed for the dark valley.
4) Fill factor also matters. How much of the pixel area is actually sensitive to light?

My suggestion is that smaller pixels are better from standpoint of resolution, but it seems obvious that diffraction effects start to affect todays DSLR at f/11 or so. With 4/3 sensor optimal aperture seems to be below f/8, depending on small pixel pitch and excellent lenses.

With smaller pixels the need of low pass filtering is reduced so it would be possible to get rid of the AA-filter.

Best regards
Erik

http://luminous-landscape.com/tutorials/resolution.shtml

The above mentions the following:

"The Rayleigh criterion – based on human visual acuity – isn’t adequate
for estimating the resolving power of a lens that projects images on
a sensor. The sensor needs more contrast and separation between
Airy disks than the human eye. Foveal cones aren’t like pixels."


This is why they suggested that we must match pixel size with airy disc
size and recommended 2 pixels per airy disc to obey Nyquist sampling
theory where sampling must always be twice or better than the signal.
But guys, in astrophotography, they used even more than Rayleigh
Limit by using the maximum wavelength/Diameter which is the
radius of the central disc in radian (so they used 4.7 pixels across
an airy disc for maximum resolution) even in planetary imaging
which is no difference from terrestrial photography since you are
after the most details in the object. So I think even details in
the Rayleigh limit (which is wider than the astrophotography
resolution) can still be resolved in the DSLR ccds as long as you
can make out any differences (no matter how tiny) that would entail
separations between two detail regions. Hence, the above website
must suggest 4 pixels per airy disc and not just 2 pixels.

What do you think?    

Pete