Optics: varying focus at extremes of image circle ?

[Ken Lee]

Today I put a 150mm and 200mm lens on my 5x7 camera and focused on a flat subject around 10 feet away. These are not "wide" lenses but standard designs.

I noticed that with the lens wide open especially, the edges of the subject were out of focus when the center was in focus, and vice versa. Once I focused the lens for the corners, the corners were sharp: I hadn't reached the limits of coverage.

I tried the same experiment at infinity distance and still noticed a focus shift towards the extremes, although it was less and disappeared with a bit of stopping down.

I rarely shoot wide angle lenses. Is this normal ?

We expect a bit of darkening towards the corners due to the increased distance, but focus shift too ?

[Leigh]

The depth of focus of a lens is defined as the distance from the rear node to the film plane along the lens axis.

If you rotate that line about the rear node, the point of focus defines an arc, not a plane.
I think this is one reason the Image Circle diameter is usually less than the circle of illumination diameter.

I believe there are some aspects of lens design that can counteract this effect to a limited extent.

- Leigh

[polyglot]

The depth of focus of a lens is defined as the distance from the rear node to the film plane along the lens axis.

If you rotate that line about the rear node, the point of focus defines an arc, not a plane.
I think this is one reason the Image Circle diameter is usually less than the circle of illumination diameter.

I believe there are some aspects of lens design that can counteract this effect to a limited extent.

- Leigh

Sorry but this is just plain wrong. A "simple" lens, ignoring the various shortcomings they all have, has matching planes in focus, not spheres. If you have a planar sensor, you will have a plane in the scene in focus, the angle and placement of which obviously depends on your movements.

[Leigh]

Sorry but this is just plain wrong. A "simple" lens, ignoring the various shortcomings they all have, has matching planes in focus, not spheres.

Sorry, but my comment is just plain right. In fact it's even worse than the simple description I posted.

As a focused point on the subject recedes (moving away from the camera), the focal distance from rear node to film becomes shorter. So if you expect a planar subject to be in focus, you have a tough row to hoe, since the distance from rear node to image off-axis increases rather than decreasing as would be required to keep the image in focus.

Quoting from the reference that Dan provided, as regards image focus:
"With a real lens, the sagittal and tangential focal surfaces are in fact curved."
Full article here: http://toothwalker.org/optics/astigmatism.html

The focal distance from rear node to image is a fixed dimension.
As you move off-axis, that distance defines an arc, not a plane.

- Leigh

[polyglot]

Sorry, but my comment is just plain right. In fact it's even worse than the simple description I posted.

You're taking the effect of an aberration and stating it as the behaviour of an ideal lens. Of course real lenses without astigmatic correction have focus curvature, that's what we're talking about. If you buy a macro lens, the field is very flat. It doesn't rotate around any nodal points in the lens, it's just flat as it should be.

Just look at your first two sentence:

The depth of focus of a lens is defined as the distance from the rear node to the film plane along the lens axis.

If you rotate that line about the rear node, the point of focus defines an arc, not a plane.

Note that the important thing is the distance parallel to the lens' axis as per your first sentence. You don't then go rotating that around, because then it's not "along the lens axis" anymore.

Anyway, the Van Walree optics link is excellent. Everyone should read that instead of idiots like us arguing on forums.

[Mark Sawyer]

"Flat field" is a relative term. Most modern lenses are close to it, but the wider you go, the more you notice they aren't quite. Process lenses are optimized for flatness of the focal plane, but at the expense of speed and coverage. One reason manufacturers' coverage figures are conservative is that they won't count illumination after the field curves.

[polyglot]

You have field curvature. While the focus field should in theory be a flat plane, it is often not perfect and can sometimes be significant.

Petzvals have significant field curvature for example, and for those who also use 35mm cameras, some of the modern ultrazooms have huge field curvature, e.g. the Tamron 17-50/2.8.

[Tin Can]

I don't know the science, but it seems an empirical given that edges will not be the same focus. I see it all the time.

I do not worry about it as I keep my important content within 2/3 of the image area.

I also shoot a lot of Macro where this is evident, most likely because I am shooting beyond the suggested IC.

[Dan Fromm]

Ken, visit http://toothwalker.org/optics.html and scroll down to the table that shows the dependence of third order aberrations on relative aperture and distance off-axis.

Curvature of field is affected by both. Stopping down reduces it and it gets worse as the distance off-axis increases.

[Tin Can]

Nice link. Great concise info.

Ken, visit http://toothwalker.org/optics.html and scroll down to the table that shows the dependence of third order aberrations on relative aperture and distance off-axis.

Curvature of field is affected by both. Stopping down reduces it and it gets worse as the distance off-axis increases.