Backfocus - Is the solution Physics?

[BLKRCAT]

Some of you have probably seen me posting around with problems that seem like backfocus. Especially wide open. (Rigid summicron 50). Working in the graphics industry and understanding quite well how planes of focus work in 3D I'm wondering how that translates into real world examples.

Below is what Im thinking could be the problem. On the left we see a camera focussed on the front of an object. When the camera is moved to compose the plane of focus moves as well thus inducing back focus. In extreme examples shooting wide open with a summicron or summilux at close focus I could see this being a problem.

My quesiton is do leica optics have a flat plane of focus or is it convex as to keep focus accurate as one composes an image?

 

[Robert Lai]

Besides macro lenses, most lenses have some curvature of field. If you center the subject to focus, then recompose, the subject may end up out of focus. The worst offender that I had was the SLR Nikkor AIS 35mm f/1.4. That seems to have the curvature of field of a goldfish bowl. For that SLR lens, it was necessary to focus with the subject where you wanted it to be. Of course, that is not possible with a rangefinder system.

If your lens is already back focusing a bit to start with, when you move the subject out of the center when composing, you may actually end up with the focus plane where you wanted it to be by the two errors compensating each other. That is, the back focusing of the lens is compensated by the field curvature.

 

[mfogiel]

Keep it also in mind, that when you focus with a rangefinder, you tend to look for contrast, which usually is to be found at the side edge of things, rather than at the front.

 

[Rob-F]

The error described by ZF1 has also been called Cosine Error, because the focus point is shifted by a distance equal to the cosine of the angle through which the camera was moved, times the camera-to-subject distance. When I try to focus on an impressionist painting, it can be too blurry to find anything to focus on. So, I focus on the frame. But if I were to swing the camera to one side to focus on the frame, I would get a cosine error. So, I sidestep far enough to get the frame in the rangefinder patch; then sidestep back to center on the painting; always keeping the camera axis perpendicular to the painting. Result: a sharp picture of a fuzzy picture.

 

[john_s]

The situation described by the OP is one way in which one can get an out of focus pic. Another is focus shift of (to varying degrees) wide aperture lenses when aperture is altered. Quite a different cause, although the result is similar.

 

[Mark C]

I've always wondered if this wasn't one of the reasons Mandler allowed as much curvature of field in many of his designs, most particularly the Summiluxes. Allowed the field to curve as much as he did let him produce very sharp, flare free, lenses working at 1.4 in the 50 and 75.

I have mosty ignored this issue when focusing off center, and found the field curvature pretty well compensates for the error with most of my lenses in most of my picture taking.

But, no, it is not the explanation for backfocus in general.

 

[BLKRCAT]

Thanks for the responses. Rob, interesting that it has a name. I kind of figured that in 3d land cameras are perfect. In real life it's not necessarily the same. Good to know that I can get this "variable" out of my mind.