> A spherical shaped film plane would be ideal to maximize the edge
> "sharpness", I guess I should say: to minimize edge softness (instead), as
> well as minimizing the fall-off. Practically, though, cylindrical is the
> shape most suitable to accomodate either film or paper. I believe that in
> one of the 2 editions of Renner Pinholo Photography books there is a picture
> of a guy that made a spherical film plate dug in the ground and B&W paper
> placed to conform the shape of the hole.
>
> Guillermo
A spherical film plane would eliminate one of the cosine squared factors in
illumination, but leave the other one. The pinhole still has a smaller
effective area as you move off axis. Even so, this produces more even
illumination than a flat camera, at the expense of introducing distortion.
I think that in a cylindrical camera like an oatmeal box, the illumination
is even more uniform because the edges of the negative get closer to the
pinhole, at least in one dimension.
Paper or film won't bend around a tight sphere, only a cylinder. One
photographic telescope design that astronomers use is called a Schmidt
camera, and it has a curved focal plane. In order to get sharp star images
from edge to edge, the film or glass plate is bent in a shallow convex
sphere and held by clamps. The 48" diameter camera at Mt. Palomar uses 14"
square plates 1mm thick. I'm told that plates are tested first before
being put in the telescope as some of them break when bent. To see a
picture of this telescope with famous astronomer Edwin Hubble, see:
http://opostaff.stsci.edu/~levay/presres/ehubble/jpeg/10_12-17.jpg
