Is focus a plane to plane affair?

[Kevin Crisp]

The excellent little 15mm lens in Minox cameras matches up with a curved pressure plate. As you advance the film, the pressure plate backs off, then reapplies for the next exposure, putting a curve in the film for the next shot. It is an incredibly sharp lens. In huge enlargements it is the shortcomings of film that limit what you can do with it.

[Havoc]

It depends from lens to lens (design). Some lenses are designed to project a plane onto a plane. Others project a curved surface onto a flat plane (as good as possible). If you look at old panoramic group shots, you'll see the people are in a curved setup just to get the best focus (and with minimal distortion) on the film plane. Some cameras do have curved film planes like the Afga Clack 6x9. And I think it was Sony that patented a curved digital sensor for this reason (of cheaper lenses).

[Darren Kruger]

Have you looked into figuring out the hyperfocal distance for your intended lens and aperture combination? I have heard stories of old press photographers setting up their speed graphics that way (with a 135mm lens set to like f8) to be increase the chance of getting their shots quickly.

[lungovw]

Darren, I understand your point about using the hyperfocal, that would be a solution. Nevertheless I would like to go a bit further. I asked this question because I am planning to make, for some old lenses RR and the like, focusing scales and wireframe viewfinders and use them for pictures in which I can keep film and lens board parallel. That would be for portraits, for instance. Having a reasonable focus and depth of field my idea is dispensing with the ground glass in order to speed up the workflow. Using a laser measure I think that could be a nice set up.

[Alan Klein]

It depends from lens to lens (design). Some lenses are designed to project a plane onto a plane. Others project a curved surface onto a flat plane (as good as possible). If you look at old panoramic group shots, you'll see the people are in a curved setup just to get the best focus (and with minimal distortion) on the film plane. Some cameras do have curved film planes like the Afga Clack 6x9. And I think it was Sony that patented a curved digital sensor for this reason (of cheaper lenses).

Wouldn't setting the people up in a curve also makes them all look about the same size? Otherwise, the people at the ends will look smaller than they are. Also, it's a way of fitting a lot of people into a less wide-angle lens.

[pgk]

If it were not for field curvature, or conversely if the film or sensor could be arbitrarily curved to lie on the sphere of best focus, then optical designs would be much, much simpler.

And lenses using curved sensors have been with us for a very long time: http://www.earlyphotography.co.uk/site/entry_L129.html and have never caught on particularly well.

But to go back to the original diagram. If the drawing is extended to include the image surface all might become clearer.

[Nodda Duma]

Oh of course. Curvature in one axis has been around since the Brownie cameras, and observatory grade Schmidt telescopes had vacuum backs for lining up film with the curved field.

Every so often I would get a sensor supplier in asking about interest in curved focal plane arrays.

Academically interesting, but field curvature varies too much from one design to the next to be useful. The curve of the sensor would need to be tailorable to the specific project. Not really feasible at reasonable cost… even for the high tech stuff I work on.


Hence the use of the word “arbitrarily” in my statement above. Technology’s not there yet. It’s much more cost-effective to correct field curvature in the optics.

[Corran]

Interesting, I never thought about how field curvature and sensor curvature would need to be tailored on a per-lens basis. But it makes sense.

I'm reminded of a very expensive fixed-lens Leica that corrected all that in-camera, even baked into the 'raw' files. Someone hacked the uncorrected files and showed the lens actually didn't even cover the sensor fully, clipping a few pixels in the corner. Easier to crop a little bit and do software corrections than any heroics in the lens design I guess.

[Bernice Loui]

Lens optical ills can be corrected in the lens design or today by software in camera or during image processing.

Leica is far from the only camera company that is applying software correction for lens optical ills..

Canon's recent EOS rf 16mm f2.8 applies both in lens and in camera or post processing software corrections that in combo results in overall optical performance not possible by either alone at a given cost as a system. Expect far more examples of these combo lens/camera/post processing systems in the future..

One of many ways to apply apodization function and much more..
https://en.wikipedia.org/wiki/Window_function


Bernice

Interesting, I never thought about how field curvature and sensor curvature would need to be tailored on a per-lens basis. But it makes sense.

I'm reminded of a very expensive fixed-lens Leica that corrected all that in-camera, even baked into the 'raw' files. Someone hacked the uncorrected files and showed the lens actually didn't even cover the sensor fully, clipping a few pixels in the corner. Easier to crop a little bit and do software corrections than any heroics in the lens design I guess.

[Tin Can]

Little me conducting the World inside a camera