PDA

View Full Version : Diffraction large format vs smaller



George Kara
4-Sep-2009, 14:33
There appears to be a difference when diffraction lowers the performance of a LF lens at higher apertures vs. smaller formats. I assume it is because of the larger negative therefore the effect is spread out over a larger space but really dont have a clue.

Does anyone know this answer?

BradS
4-Sep-2009, 14:42
diffraction happens at different f/stops for different focal lengths because f stop is a function of focal length. Diffraction effects are directly related to the size of the slit (or aperture) through which the light is passing...

BetterSense
4-Sep-2009, 14:47
I believe that the amount of diffraction in terms of circle of confusion is constant for any given f/stop, regardless of the camera format, but that the larger circle-of-confusion of largeformat, allows this to be less of a problem due to the reduced enlargement that the image undergoes for final viewing.

I could be wrong though.

BradS
4-Sep-2009, 15:01
I believe that the amount of diffraction in terms of circle of confusion is constant for any given f/stop

This is incorrect (see above)



the larger circle-of-confusion of largeformat, allows this to be less of a problem due to the reduced enlargement that the image undergoes for final viewing.


Yes, this is partially correct but it is not the full explanation of why diffraction effects become apparent at different f/stops for different focal lengths (and therefore it appears, formats).

ic-racer
4-Sep-2009, 15:14
There appears to be a difference when diffraction lowers the performance of a LF lens at higher apertures vs. smaller formats. I assume it is because of the larger negative therefore the effect is spread out over a larger space but really dont have a clue.

Does anyone know this answer?

If you consider the absolute aperture size in millimeters, then the relative enlargement of each format is an equalizer. So, diffraction is essentially the same across format sizes for the same print size and the same aperture size in millimeters. This is mathematically derived from the sizes of 'bullseys of confusion' produced by diffraction through the aperture.

Jeff Keller
4-Sep-2009, 15:21
One of the best write ups I've seen is:
http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm
Play with the calculator near the bottom of the page.

For me the bottom line is:
aperture matters, focal length doesn't with fixed film size (ignoring DOF)
printed size matters
if DOF is inadequate or barely adequate, LF may be no better than 35mm, and diffraction can't be reduced by using a wider aperture

The advantage of LF (available when scene has limited DOF) comes from the fact that any given focal length covers a wider view and thus less enlargement is used to get the final print size.

Jeff Keller


I believe that the amount of diffraction in terms of circle of confusion is constant for any given f/stop, regardless of the camera format, but that the larger circle-of-confusion of largeformat, allows this to be less of a problem due to the reduced enlargement that the image undergoes for final viewing.

I could be wrong though.

Jeff Keller
4-Sep-2009, 15:31
I believe BetterSense is correct. The larger physical aperture of the longer focal length causes less angular divergence but the light travels further before forming the image, resulting in the same distance spread at the image plane.

See:
http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm

Jeff Keller

Quote:
Originally Posted by BetterSense
I believe that the amount of diffraction in terms of circle of confusion is constant for any given f/stop

This is incorrect (see above)

Glenn Thoreson
4-Sep-2009, 16:19
I wonder what the f/64 group would say about this. :D

Mike1234
4-Sep-2009, 17:47
IIRC, at least some of the f/64 group changed their methods.

Nathan Potter
4-Sep-2009, 19:39
Diffraction (properly Fraunhofer diffraction) is the result of light rays having to travel a slightly longer path length from the edge of the aperture opening than from the center of the aperture opening thus interfering at the image plane causing a radially diminishing set of dark lines. The diameter of the first dark interference pattern is called an Airy disc diameter and can be computed mathematically as D = 2.44(lambda)N. D = diameter in microns; lambda = wavelength in microns, N = f/no. So an f/5.6 lens of any focal length produces the same size airy diameter. It is the Airy diameter that is the diffraction limit.

To place the above relation in a comprehensible fashion (good luck) one needs a picture of how a subject image is resolved at the film plane. This need have nothing to do with a lens but only requires a pinhole, as in a pinhole camera where the pinhole is equivalent to the aperture. Note from the equation above as the pinhole gets larger (larger f/no) the diameter of the airy disc gets larger. The key here is conceptual. Every part of the pinhole opening sees each individual resolvable point in the subject and delivers it to the film plane with an airy disc diameter determined by the size of the pinhole only. Thus the number of resolvable points in the subject is determined by the size of the opening. And, if you will, every one of these resolvable points has its own diffraction limit. If we rewrite above so that D = 2.44(lambda)Fl/A in place of f/no we see that holding the focal length (Fl) constant and increasing the A (diameter of the aperture) we decrease the diameter of the airy disc and conversely if we hold A constant and increase Fl we increase the diameter of the airy disc. In the case of large format lenses the diameter of an airy disc is no different than with small format lenses but of course one has a much larger image to work with.

So the OP is on the right track - it's the larger negative that is golden with LF. However, LF maximum apertures are usually around f/5.6 rather than some at f/1.2 with miniature cameras. And of course microscope objectives with mm focal lengths can achieve very small Airy disc diameters especially using oil or DI water immersion (of course with a lens, not a pinhole).

Nate Potter, Austin TX.

George Kara
4-Sep-2009, 22:12
Thank you all for your contributions to this inquiry. IMO this forum is the best on the net for true in depth discussion about the craft of photography. While the subject is primarily from film shooters, the issues apply equally to all forms of capture.

Emmanuel BIGLER
5-Sep-2009, 04:39
Hello all !
In addition to all that has been said (and I agree with the group) we should add that using a large format camera allows us to get hight quality images (from film) with a minimum of magnification factor for the enlargement. Hence we can tolerate that residual aberrations and diffraction effects combined in our lenses correspond to a bigger circle of least confusion.
It happens that for a given lens design, for example take a classical 6/4 lens, 150mm for 4x5", 300 for 8x10" etc residual aberrations scale in direct proportion of the focal length, whereas diffraction effects depend only of the relative f-number. Instead of speaking in terms of diffraction disc diameters like Lord Rayleigh Himself, we can speak like the French scientists of the forties and fifties in terms of cut-off period. For a diffraction-limited lens the cut-off period is simply N.lambda where N is the f-number and lambda the wavelength (no Bessel functions and no 1.22 factor in this approach !).
You can take lambda =.7 micron which is the limit of what is actually visible by the human eye, and which is already beyond the sensitivity of many b&W films. Hence you get a cut-off period of .7 N microns, corresponding to a cut-off spatial frequency of 1400/N cycles per millimetre. This figure is really optimistic in real life photography except for the f/64 group where the lens is actually diffraction-limited ! ;-) However a repor lens like the beloved apo-ronar, at the centre of the field and stopped down to f/22-F/45 (according to the focal length of your apo ronar) is very close to a perfect diffraction-limited lens.

Since the residual aberration spot is bigger with long focal lengths whereas the diffraction spot depend only on the F-number, it happens that the best f-number (where both effects of diffraction and aberrations are equal) is not the same for different focal lengths.
I have done a compilation of the best f-number recommended by view camera lens manufacturers and we can summarize by the following rule of thumb for classical standard lenses :
N_best = f(in millimeter)/(8 millimeters)

for top-notch modern 6/4 lenses of the last generation and razor-sharp expensive "digital" view camera lenses the rule of thumb would be closer to f(in millimeters)/(11 millimeters)

The consequence is that a good standard 50mm lens designed for 35 mm photography should not be stopped down beyond f/5.6 or f/8, whereas a 150mm lens designed for 4x5 can be stopped down to f/16 to f/22, a 210 mm for 5x7 : f/22 to f/32, a 300 for 8x10 : f/32 to f/45. Assumed that the enlarginf factor is reduced in proportion, of course, for the same final print size.

Modern "digital" lenses of 70mm focal length should not be stopped down beyond f/8 in order to get the best performance ! And if we extrapolate to smaller silicon sensors, the situation is frioghtening, since a standard lens covering a sensor of 8mm in diagonal should not be stopped downs beyond f/1....

In other words, the diffraction effects are of the same magnitude for the same f-number, this is a seriuos limitation for small focal length required by small-size sensors ; but in large format since the final magnification factor is much less, we can stop down more and live happy with big diffraction effects !

Imagine a 8x10" contact print recorded at f/64. In those conditions a good lens is ususally diffraction-limited i.e. residual aberrations contribute much less than diffraction to the residual blur. The corresponding cut-off period is about 45 microns, this is close to 20 cycles/mm. By contact printing, the additional image degradation is minimal, but even if we get only 15 cy/mm in the final print, this is already twice as much as the human eye can see/resolve at 30cm of distance (10 feet), i.e. about 5 to 7 cycles/mm (except for jet-fighter pilots).

Hence, be happy, follow Saint Ansels' f/64 strict rules, and contact print your images, diffraction will still be invisible !!
Alternatively, you can also contact-print a 35 mm image frame (24x36mm), @f.64, seen from 30 cm / 10 feet, the image quality will be of the same quality as in 8x10" ... but on a stamp-size image ;-);-)

Kirk Fry
5-Sep-2009, 22:13
30 cm does not equal 10 ft. more like 12 inches.......

Nathan Potter
6-Sep-2009, 15:11
Well, add a zero to Emmanuels 30

Er; well 11.8 inches.

Emmanuel BIGLER
7-Sep-2009, 03:22
Outch ! Shame on me !
Confusing inches & feet ! ;-)
Next time I shoud really be extra cautious when "zooming with my feet" !!

rdenney
7-Sep-2009, 11:03
I have been chastised on various occasions for using too small an aperture.

But I'll defend myself this way: Insufficient depth of field usually causes more fuzziness than diffraction.

Thus, my practice is to adjust the camera movements to get things in focus as best as possible, stop down to achieve the necessary depth of field, and then limit my enlargement ratio based on the resulting image quality. If I had to stop down so far that diffraction became noticeable at high enlargements, I prefer to just limit those pictures to smaller prints.

Stated another way, use the largest aperture that provides sufficient depth of field. If, with a loupe on the ground glass, one can't see the effects of stopping down further, then diffraction is probably masking the effects of increased depth of field, so further stopping down becomes pointless.

Rick "nothing causes loss of sharpness like not being in focus" Denney