Diffraction question

[null]

You can place the aperture in various positions within the optics, before the front element, after the rear element, or in between any of the groups. Diffraction is an edge effect, placing the aperture at a point before the incoming light is narrowed down and collimated allows a larger diameter to be used.

[Bjørn Rørslett]

Of course an aperture can be placed anywhere, if you don't care for shading or vignetting due to the cutting off of peripheral rays. However, it is usual to put the aperture where it does the least harm, namely in the internodal space at the position where the ray bundle is (near) parallel.

[Dr Klaus Schmitt]

The dedicated photomacrographic lenses should always be used "wide" open as they are designed for that use.

ehemm, not in general. Quite a few of them work much better when stopped down a bit, ZEISS Luminars for instance are best 1 stop down.

[Bjørn Rørslett]

OK, my statement was too broad. I stand corrected. "More or less wide open" would be more apt.

[Roland Vink]

One thing I have wondered about, is diffraction affected by focal length? If for example we frame the same subject identically (eg a brick wall) using two lenses - a 1200mm lens at f/16 and a 28mm at f/16, will the diffraction be the same?
With the 1200mm lens, the physical opening is much larger (greater area : edge ratio) so diffraction is less. On the other hand the path to the image plane is much longer so any spreading of light due to diffraction is much greater. So is the overall effect the same between the two lenses?

[Bjørn Rørslett]

This is not easy to answer yes/no. Detail magnification will be vastly different and because things are rendered bigger with the long lens, more softening can often be tolerated. Traditionally long lenses offered much smaller apertures such as f/45 or f/64.

If one could arrange so magnification differences are eliminated, perspectives would be completely different, and again direct comparison would not be easy.

[Roland Vink]

Maybe I'm misunderstanding ... If the subject is framed identically with the two lenses (the 1200mm proportionally further from the subject than the 28mm lens) then the detail magnification is the same. Is the amount of diffraction also the same?

[smage85]

Maybe I'm misunderstanding ... If the subject is framed identically with the two lenses (the 1200mm proportionally further from the subject than the 28mm lens) then the detail magnification is the same. Is the amount of diffraction also the same?

Diffraction is the same assuming both lenses behave similarly to f stop vs diffraction.

Unless I'm wrong what will make the difference is that the 1200mm lens will have a more abrupt sharp to soft transition zone compared to the wide angle. This might increase perceived sharpness.

[Bjørn Rørslett]

Maybe I'm misunderstanding ... If the subject is framed identically with the two lenses (the 1200mm proportionally further from the subject than the 28mm lens) then the detail magnification is the same. Is the amount of diffraction also the same?

I have already discussed this, see above.

The massive difference in perspective will render a direct comparison very difficult. That is, unless the target depicted was perfectly flat with no depth to it at all. In the latter, hypothetical case, there should be no difference between the two lenses if the angle of incidence was perfectly normal.

[null]

Of course an aperture can be placed anywhere, if you don't care for shading or vignetting due to the cutting off of peripheral rays. However, it is usual to put the aperture where it does the least harm, namely in the internodal space at the position where the ray bundle is (near) parallel.

One would want to place the aperture in a position where the light has diverged in order to make the physical opening as wide as possible to implement the F-Stop. Effects of diffraction caused by the aperture mechanism of the lens the optical engineer can take that into account of the lens design. Looking at the retro-focus design of the "art" lenses, aperture placement could be done to reduce effects of diffraction.

[Bjørn Rørslett]

You cannot play two horses at once and expect both to win.

When the aperture stop is put in the non-optimal position, lens characteristics such as vignetting will suffer badly.

Besides, you cannot "enlarge" a given aperture number. This is a fixed relationship, combining focal length and entrance pupil size.

[null]

You can expand the wavefront and then place the Iris at that location.

Physical size for a simple, single-element lens of fixed diameter is given by focal-length/F-Stop. Complex lenses allow a degree of freedom in placement and size of the Iris. The Leica Elmar vs Zeiss Tessar is a good example of placing the iris in different positions: both lenses are the same optically, the Elmar places the Iris after the front element, the Tessar places it in front of the rear doublet. The Olympus RC and several other fixed-lens RF cameras place the Iris behind the lens, the MS-Optical 4/28 Perar places the Iris before the front element.

[David H. Hartman]

one more thing that puzzles me is that FX cameras get at least 1 stop more before diffraction starts to be obvious. i tested this on a D7000 and a D800 a long time ago so the pixel pitch is roughly the same.

I'm sorry I'm repeating my self...

The answer might be the "Enlargement Factor." To make an 8x10 inch print from a 24x36mm (FX) camera you'll need an 8x enlargement but for 16x24mm (DX) camera you'll need a 12x enlargement. The greater the enlargement the easier it is to reveal diffraction: make an 8x10 inch print from an FX image shot at f/45 and it will be surely be soft. Make a contact print from an 8x10 inch negative shot at f/45 and you won't likely notice. I believe the diffraction for each image shot at f/45 will be the same before enlargement.

Dave

[Bjørn Rørslett]

I think the discussion is going in circles. I tend to assume the optical designers know what they are doing, and put the aperture stop where it does the least harm to other characteristics of the lens. Thus, the typical position is within the internodal space.

[null]

I've had a number of Optical Engineers with me in the 80s, and work for me in the 90s. If you tell them what to optimize for, the design changes. If the size of the Iris is going to be the limiting factor in the design they will move it to a location to minimize that effect. Positioning the Iris where the cone is most constrained reduces the size of the mechanism, but maximizes diffraction. If light traveled as a cylinder through the lens, it would not matter. In most lenses, light travels as a cone that expands and contracts as it goes through the elements. This gives an opportunity to place the Iris in a position to minimize diffraction.