[Lititz]

With my K-X I'm puzzled by an interaction between the aperture I select and the focus distance that autofocus picks (in Av mode). I see this especially with an ultrawide lens (Tamron 10-24 at the 10mm end, less so at 20mm+), and less noticeably with wide angle (Sigma 17-70 at the 17mm). Here’s what happens, using spot autofocus and the same targets at distances of 50 yards to 1000 yards (a large target (bridge) in the latter case), with no intervening objects for the AF to focus on:



At (for example) f3.5, the autofocus stops with the focus ring at a particular place—about at the 7 foot mark on the focus ring in the case of a far away subject. It stays at the exact same focus point if I change aperture to, say, 4.0 or 5.6. But if I change the aperture to 8.0, the autofocus shifts the dial to the infinity mark. I thought that AF always takes place wide open, so I don’t understand why, when moving to f8, it would suddenly recalculate and select a new, noticeably different focus ring position. If the autofocus uses the lens wide open, why would it consistently shift to a different result when I change aperture to f8? The same thing happens, but to a lesser degree, at longer focal lengths, and with the Sigma at the 17mm end, though at longer focal lengths the focus ring shifting takes place at about f11 (and is in the other direction on the Sigma—it goes from having selected a longer distance to a slightly shorter distance when I change aperture to f11).


Normally I’d attribute this to sloppiness in the Tamron and Sigma quality control. But the fact that it happens predictably (always at f8 on the Tamron) suggests not sloppiness, but something about the way the choice of aperture influences the way autofocus calculates its stopping point with wider angle lenses. (For example: does the smaller aperture setting actively tell autofocus that it’s got more depth of field to work with, leading to a different choice of focus point, with the transition to f8 being the trigger for this?) I’m puzzled as to whether this is supposed to happen or not, as it does affect exactly where in the overall depth of field the best focus occurs. Thanks much for any insight.

---

Last edited by Lititz; 07-16-2011 at 06:18 PM. Reason: clarify question to avoid confusion with the 11 autofocus points

[Lowell Goudge]

I have heard comments like this before. If you search the forum you may find another thread

In your case if I were to speculate, and this is purely speculation on my part, maybe the camera is programmed to adjust the point of focus if you are focused beyond the hyperfocal distance. Does the camera behave the same way in all modes or only automatic modes? Specifically what does it do in manual?

[Lititz]

Specifically what does it do in manual?

It does the same in manual as in aperture priority.

I suppose it could be adjusting the focus according to some formula, such that when I change the aperture the camera or lens in turn comes up with some supposedly optimal positioning of the subject within the overall depth of field range. For example, using a depth of field/hyperfocal calculator, I see that at 10mm and f3.5, a subject four feet away, focused for 4 feet, only has about 5% to 10% of usuable depth of field in front of it, but refocused to, say, 2.5 feet, there's more usuable "in focus" depth of field in front of the subject. But I'm not sure, and a bit surprised that the shift in focus point seems to always occur when moving to f8 from a wider opening; I'd think that there would be some variance in when the shift occurs, depending on the subject distance.

[Lowell Goudge]

ok, here is a question I should have asked.

how are your autofocus points and exposure points set up.

do you have the AF point on auto, or did you select central af point. and is exposure linked to AF point or independant?

the reason for asking is maybe (but this would really be a surprise) the camera is doing some form of scene evaluation.

[Lititz]

ok, here is a question I should have asked.

how are your autofocus points and exposure points set up.

Autofocus is fixed on the center spot; exposure is set to center-weighted, independent of autofocus.

Also, this change in selected focus distance (when I change aperture) happens whether I use autofocus or manual focus with confirmation (that is, the confirmation light blinks at different distances depending on f-stop). So I assume the phenomenon is not just due to slop in the lens gears. Maybe the camera AF sensors just consistently err in particular ways with ultrawide lenses, but I also wonder if it's doing some calculations and adjustments that would reflect the greater DOF at smaller apertures.