Does focus shift vary with distance?

[bjanes]

Just goes to show that one should be very careful about asking for advice on the Internet. But how do you actually know that Rossyd's information is correct? Just because he addressed your question most directly? In any case, I believe he has the correct answer to your inquiry.

When evaluating the validity of information on the net, it helps to know the reputation of the author and if his posts have been accurate in the past. For example, if I see a post by Bart van der Wolf, I take his information seriously. Since my reputation is lesser than Bart's, I supply a link to a reliable source when possible. With regard to focus shift, I referenced Diglloyd.

Bill

[elliot_n]

Thanks, Pieter.  I am aware of the variable curvature of the focus plane with some designs - the lens I am specifically interested in, the Nikon 28mm f1.8G has such a wavy MTF curve.  But I am specifically interested in focus shift at the center of the lens in this case.

I've used the 28/1.8G extensively (it's my favourite lens) and I have seen no sign of focus shift. When I read online reports describing the phenomenon, I did some tests at 6ft (on a D800) and could see no sign of the focus plane moving as I stepped down from f1.8 to 2 to 2.8 to 4. Perhaps at shorter testing distances the phenomenon is visible?

[AFairley]

I've used the 28/1.8G extensively (it's my favourite lens) and I have seen no sign of focus shift. When I read online reports describing the phenomenon, I did some tests at 6ft (on a D800) and could see no sign of the focus plane moving as I stepped down from f1.8 to 2 to 2.8 to 4. Perhaps at shorter testing distances the phenomenon is visible?

I've read both ways, which I find somewhat confusing since I don't see how "sample variation" should result in that outcome.

 Since positing, I ran some tests at 18' using a blown up version of the Lens Align target and got a +1 adjustment at f8 and f5.6, but +6 at f2, according to its target analysis.  That would be consistent with a lens that back focuses when stopped down.  Sort of academic for me, since my final practice is to tweak AF by looking at the DOF in pictures I take rather than running focus tests at (for me) extremely close distances.  I usually shoot at f5.6 or f8, so that is what I end up optimizing for.

[David Eichler]

When evaluating the validity of information on the net, it helps to know the reputation of the author and if his posts have been accurate in the past. For example, if I see a post by Bart van der Wolf, I take his information seriously. Since my reputation is lesser than Bart's, I supply a link to a reliable source when possible. With regard to focus shift, I referenced Diglloyd.

Bill

Oops, I got confused. Actually Rossyd did not address the subject of the OP's question, and I am not sure why the OP thinks he did. I am not really sure anyone really addressed the question: does the degree of focus shift with different apertures change with distance? You and one other correctly identified the cause of the focus shift: spherical aberration. However, I don't think anyone actually answered the question, including the Digilloyd article that you cited (unless I just missed it).

[Rhossydd]

Oops, I got confused. Actually Rossyd did not address the subject of the OP's question..... (unless I just missed it).

You did miss the answer. It was a simple question that allowed a simple answer.

Does aperture change induced focus shift vary with distance?  I.e., can a lens exhibit focus shift at far distances but not close distances or vice versa?

Yes, it can change.
If you think you have an issue with auto focus accuracy and have a camera that allows you to fine tune the AF accuracy, do have a look at Reikan FoCal. It will perform tests on issues like this and the results are well presented for you to make an informed choice of what might be the best settings to choose.

As there's only a single value for AF adjustment on most cameras, you have to make a choice about what value to apply. If you have a set of test results, from FoCal or elsewhere, you can see where there might be an error and can choose an appropriate value for the type of photographs you routinely take.
For distant landscapes, you'd probably opt for a correction value at a long distance and small aperture, if you routinely shoot much closer and with wider apertures you would be better to choose a value for those settings.
If absolutely obsessive about it you could create a table of correction values for all settings to use.

[David Eichler]

You did miss the answer. It was a simple question that allowed a simple answer.

As there's only a single value for AF adjustment on most cameras, you have to make a choice about what value to apply. If you have a set of test results, from FoCal or elsewhere, you can see where there might be an error and can choose an appropriate value for the type of photographs you routinely take.
For distant landscapes, you'd probably opt for a correction value at a long distance and small aperture, if you routinely shoot much closer and with wider apertures you would be better to choose a value for those settings.
If absolutely obsessive about it you could create a table of correction values for all settings to use.

I believe your answer does not fully address the question and does not answer the question about autofocus. I believe there is currently no autofocus system (at least none available in any production model camera) that will compensate for focus shift due to spherical aberration if it is focusing with the lens wide open and then the lens is being stopped down for the exposure.   As best I can infer from reading the online descriptions of spherical aberration, the answer to the first part of the OP's question appears to be that the degree of focus shift does not vary with distance.

[kers]

...I usually shoot at f5.6 or f8, so that is what I end up optimizing for....

If you uses f5.6-f8 and your subject is street landscapes than with wideangles i would concentrate on the extreme corners since getting those sharp is usually a problem.
It will be always a compromise: Centre sharpest or everything about sharp.
At least that is what i see with the three wideangle lenses i have.

[AFairley]

If you uses f5.6-f8 and your subject is street landscapes than with wideangles i would concentrate on the extreme corners since getting those sharp is usually a problem.
It will be always a compromise: Centre sharpest or everything about sharp.
At least that is what i see with the three wideangle lenses i have.

Good point

[Rhossydd]

I believe your answer does not fully address the question

There's no perfect answer. Don't get side lined by theoretical assumptions about what causes the issues seen, it's never just down to a single factor like spherical aberration.

The issue is there and is effected by a huge number of factors. However the only practical answer is to do tests to see what problems actually show up on an individual system and with that knowledge use any strategies available to alleviate it. Very simply put; at best you may have some AF micro adjustments you can set, but just being aware of any problems gives you the opportunity to correct them with more manual interventions.
It's also possible that any issues of this nature might not actually effect your photography anyway.

[BrianVS]

Does aperture change induced focus shift vary with distance?  I.e., can a lens exhibit focus shift at far distances but not close distances or vice versa?  I'm wondering whether an auto-focus correction value for a focus-shifting lens determined for a specific aperture will be good over a range of distances or only at the tested distance.

The OP specifically asked about aperture induced focus shift.

That is caused by spherical aberration.

It is that simple.

For lenses that exhibit focus shift as aperture is stopped down: The focal length of the lens is not constant across the full aperture of the lens. As you close down, you lose the contributions from the outer edges of the objective.

The formula for focal length has distance to the subject as part of the equation.

That means that focus shift increases as you get closer to the subject.