Diffraction limits

greenspun.com : LUSENET : Large format photography : One Thread

With a 150 what will be the point at which diffraction is a problem with 16x20 enlargements?

Same question for SA 90/8 F-stop range on shutter is 8 - 64

Thanks in advance Steve.

-- Steve Nicholls (gl1500@chariot.net.au), November 25, 1999


To a degree, this is somewhat subjective. It has a great deal to do with what you can see at any particular magnification. A rule of thumb I use (this applies to all focal lengths), but which I sometimes find myself having to break in order to achieve the depth of field I need, is as follows: Use no smaller aperture than f32 when working in 4x5, f45 when working in 5x7 and f64 when working in 8x10. Again, I've made 4x5's at f45 with no objectional result and I regularly make 16x20 enlargements of these negs.

-- Robert A. Zeichner (razeichner@ameritech.net), November 26, 1999.


I would agree with Robert's suggestions. But if you wonder about the math, diffraction limit (defined by the Rayleigh criterion of resolution)in the center of the field is independent of focal length. This is the case bcause although angular resolution depends only on lens diameter, linear resolution is focal length times angular resolution. So here is what you get assuming 0.55 micron light (center of visual range):

Rayleigh Criterion for max resolution:

f/16 93 lp/mm f/22 68 lp/mm f/32 47 lp/mm f/45 33 lp/mm f/64 23 lp/mm

Wider angle lenses will show more rapid resolution fall off across the field due to the cosine effect on linear resolution.

If you assume you need 5 lp/mm in the final print, and a 4x enlargement, you conclude that even at f/64, you sneek in under the limit.

On the other hand, MTF contrast decreases rapidly with f/stop. Coupled with the decreasing contrast response of film at higher frequencies (say 20-40 lp/mm), the visual "sharpness" will be worse than the numbers suggest.

Since most modern lenses achieve diffraction limits by f/16 or f/22, you should shoot with the maximum opening you can given depth of field constraints, but you don't have to worry about f/45 if you need it. By f/64 the effect will be observable.

-- Glenn C. Kroeger (gkroeger@trinity.edu), December 21, 1999.

Moderation questions? read the FAQ