[AnHonestComment]

My understanding is like this:

Imagine that you get one pixel for each hole in a chain link fence — the average color of light coming through that hole.

If you take a second fence, that’s out of alignment with the first — you get better resolution because each hole in the original fence overlaps with several in the second fence, and you can examine their overlaps. The “size” of pixel of the two fences together is smaller than the size of the fence grating.

A bunch of different overlaps, and some fancy statistics, lets you actually do that with space pictures.

In essence, you reduce from the normal pixel size to the area of overlap between pixels. But if you only have the snapshot averages, you have to guess at what the original was — so statistical result.

[Cancan82]

This was a really great explanation. I love smart people who can dumb it down for me. Thank you. Very cool.