Diffraction

[eddymendoza]

<p>Hello all, I've been reading about image diffraction online for a few days and I think I may have toasted my brain with all the technical stuff. Could someone here please help me understand better what this is by posting an image with diffraction and one without and explain side by side. I'm afraid that's how I learned to learn and simply reading and looking at charts won't do it for me. I hope someone can help? I hate to be a pain.<br>

Side note: I am using an RZ 67 Pro with a Kodak DCS Pro Back Plus and two lens (Mamiya Sekor 110mm 2.8 and a 180mm 4.5)</p>

<p>Thanks</p>

<p>Ed</p>

[dan_fromm2]

<p>Make your own.</p>

<p>Shoot a subject with fine detail at every marked aperture with one of your lenses. The ideal subject would be near, not at, each lens' near focusing distance and won't be flat. But shoot a convenient brick wall too, same procedure.</p>

<p>Now look closely at the images, in order from the one taken with the lens wide open to the one taken with the lens stopped as far as it will go. You'll see the fine detail gain sharpness/clarity and the zone of apparent sharpness grow for a few shots and then sharpness/clarity will start going away. That's what diffraction does.</p>

<p>Repeat the exercise with your other lens.</p>

[Andrew Garrard]

<p>Eddy - to save you wasting some film... if you're just after the general effect of diffraction, I could suggest looking at a lens review on dpreview.com. Take, for example (reviews->more lens reviews->) the Canon 100mm f/2.8 L IS USM macro - a lens that's pretty sharp wide open, and which stops down to f/32. Have a look at the APS-C page (because the pixels for the APS-C camera are smaller). If you wave the mouse pointer over the chequerboards on the graph, you can see what the camera captured. Drag the slider at the bottom and you can change the aperture whose results are being shown. Drag it to f/32 and the chequers look blurry (and the chart goes orange to show the resolution has gone away).<br />

<br />

Sure, the actual effect of diffraction will depend on the lens, focal length, wavelength, etc. and the "blur" is the image convolved with a Fraunhoffer diffraction pattern (or something like it), but unless you're photographing stars on a dark night, what you'll see is a slightly blurry image. You could also google for pinhole camera images!<br />

<br />

If you want to know the specific effects on your own lens combination, I'm afraid your best bet is to try it, as Dan suggests. Myself, I'd think carefully before going much below f/16 (at least on 35mm), but then I do that anyway so that dust on my sensor isn't so visible. I hope that helps.</p>

[steve_levine]

<p>The softness from diffraction is present at all f stops. But it is only at the smaller f stops that the diffracted light becomes a greater part of the image forming light.</p>

[frank uhlig]

<p>Try to look up the diffraction calculator at </p>

<p>http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm</p>

<p>For your format (6 x 7) [but of course you use a digital back which does not cover all of 6 by 7 ... ?] and viewing an enlargement up to 12 or 18 inches a side at 25 cm (= 1o inches) in front of you, you will have no visible/viewable effect due to diffraction if you use f/32. With f/64 diffraction becomes a factor. But I doubt that your lenses have f/64 stops.</p>

<p>If you enlarge much more and want to print-peep at 25 cm arm length distance, you will notice diffraction, even at f/32, I am afraid.</p>

[dave_redmann]

<p><em>The ideal subject would be near, not at, each lens' near focusing distance and won't be flat.</em></p>

<p>Actually, a non-flat subject is going to demonstrate the combined effects of diffraction and depth of field, which are working in opposite direction. Sure, most real-world subjects (i.e., all subjects that are not totally flat and in the plane of focus) can be affected by both. But if you really want to separate out and see just the effects of diffraction, the subject needs to be flat and in the plane of focus, so that depth of field won't matter. Also, I'm not clear on what the subject-to-camera distance matters, except to exaggerate the effects of depth of field.</p>

<p>With a 6x7, if we're considering, say, a 16x20 inch print viewed at 3 ft, theoretically diffraction can start having a little effect around f/27, but as a practical matter, the lenses don't stop down enough to make it a big deal. But if you print 30 x 40 inches and really view more closely (say, 2 ft), theoretically you can get just an effect between f/8 and f/11--if you're using a microfilm that can record that fine detail and printing techniques that can reproduce it.</p>

<p>As you might guess, generally I think people worry too much about diffraction. If you're shooting a DSLR or 35mm film, avoiding f/32 or even f/22 is probably a good idea, unless you really need extreme depth of field. But generally, it's not a huge issue, IMO.</p>

 

[sumo_kun]

<p>I think all you need to know is that at smaller apertures like f22 and f32 etc, the image will get blurrier, although you get a bigger DOF. It won't be unusably blurry and certainly nothing as bad as a pinhole image but typically it will be blurrier than an image at f8 or so. The degree and exact aperture where this becomes apparent varies from lens to lens but as a general rule, f8 or f11 is usually the peak performance point which balances the aberrations with diffraction.</p>

[david_goldenberg1]

<p>At the risk of adding confusion to the circle, allow me to offer a slightly different perspective on this:<br>

Diffraction is the phenomenon in which light waves originating from different points add either constructively or destructively. It is intrinsic to the wave nature of light. Rather than adding blurriness to an image, constructive and destructive interference is what actually creates the image!<br>

The resolution of an image from an ideal lens (i.e. one with no significant aberrations) is limited by the wavelength of the light and the diameter of the aperture. This is not because of some artifact due to light scattering from the diaphragm (though that could be an additional factor). Rather it is because the size of the aperture limits the angle of over which light can be collected, and limits the ability of the diffraction effect to resolve closely spaced points. In other words, the size of the aperture limits the amount of information that the lens can collect and, therefore, its maximum resolution. In the world of microscopes and telescopes, if a lens is said to be diffraction-limited, this is a good thing: It means that the resolution is limited by the wavelength of light and the diameter of the aperture, rather than any optical aberrations. (Telescopes and microscopes don't usually have diaphragms to adjust the aperture, precisely because they would reduce the resolution.)<br>

Camera lenses (at least of the sort that most of us ever see) are rarely diffraction limited at their maximum apertures. This is because of the various aberrations, which are more difficult to control if the lens is to be used under a variety of conditions and built at a reasonable price and size. Stopping down the lens tends to reduce the aberrations, and this almost always dominates the reduction in optimal resolution, at relatively large apertures. At some point, though, the aperture becomes small enough that the aberrations are largely eliminated, and the lens becomes "diffraction limited". Further stopping down does not significantly reduce aberrations, but does limit the resolution further.<br>

As others have said, the competing effects of aberrations and the diffraction limit usually result in an optimum aperture for a given lens. Further stopping down will reduce resolution, but whether or not this is significant depends on a variety of factors, including the enlargement of the final image. <br>

In my mind, the diffraction limit comes in a third (usually distant third) in considering the aperture I use for an image, following: How much light do I need to expose the negative? and How much depth of field do I need? (It's also worth thinking about how much depth of focus one needs to allow for likely focus errors.)<br>

I hope this helps. I know that it isn't the usual (photography) textbook explanation, but it can be found in texts on microscopy, for instance. <br>

David</p>

[celasun]

<p>David,<br>

Thanks for one of the cleanest & clearest explanations I've read on diffraction limitedness of photographic lenses.</p>

[eddymendoza]

Thank you everyone! Lots of great information here. My next question is, how does one then use f stops like 32 and 45 to

get everything in focus and sharp as can be? Do I need to use large format instead? I suppose the lens in that category

are much better at dealing with f32 and above. Do I have to forget about using medium format then or are there other

tricks or techniques I can incorporate to get better resolution?

 

Thanks

 

Ed

[eddymendoza]

I have another idea that has been rolling around in my head. I own two large format cameras but have only used them

using Polaroid film just to get the instant pictures and of course the biggest problem for me is the changing of the sheet

film in a dark bag. I can't seem to do that part very well. I've seen these medium format roll film backs that fit large format

cameras but haven't yet bought one to try out. Has anyone here used these and if so how is the resolution? Other than

the smaller cropped image what else is there to hate? Does the large format lens perform as well on MF film as they do on

LF film or do you loose more than size.

 

Thanks

 

Ed

[frank uhlig]

<p>Eddy, the diffraction's visibility depends on the size of the print and the viewing distance (and primarily the f number).</p>

<p>So once you have a large print, you need to stand back sufficiently to see the whole and that settles the issue in museum exhibits.</p>

[steve_rasmussen]

<p>Eddy, I use f/32 and f/45 on my 6x7 on a regular basis because I shoot a lot of macro work. The slides turn out pretty sharp. If I were not satisfied I would move up to 4x5. Large format lenses are no better than MF lenses when it comes to diffraction. The bigger film makes the difference.</p>

[rodeo_joe1]

<p>There is an often stated "rule" that resolution is limited by diffraction to the tune of: max resolution (lppmm) = 1390/N; where N is the relative aperture of the lens. I'm sure this works for that non-existent object the perfect simple lens. However, it's obvious to anyone that's used a variety of <em>real</em> lenses that diffraction doesn't affect all lenses the same way.</p>

<p>Using an optical simulator program has shown me that diffraction actually improves the definition of many lenses by partially cancelling some of the effect of geometric aberrations. Other lens designs have a very restricted internal aperture, which has a greater diffraction effect than lenses which don't "neck" the ray path as much. Therefore we can't rely on the simplistic formula given above, and we're back to empirical "suck it and see" methods to determine how diffraction affects definition.</p>

<p>In any case, diffraction is not the terrible beast that some people claim. It shows itself as a gentle clouding of edge definition, a bit like acutance in reverse, rather than an out-and-out fuzziness. If you're printing digitally you can probably completely reverse the effect with a judicious bit of sharpening.</p>

[dan_fromm2]

<p>Joe, you must have been thrown many times and landed on your head most of them.</p>

<p>When I was chasing high performance macro lenses -- think Zeiss Luminar and strong competitors -- I used an Olympus stage micrometer as the target for acceptance testing. The one I used has 100 tick marks in 1 mm, every fifth tick is longer. All are the same width, all are much narrower than the spaces between them. </p>

<p>This is <em>not </em>equivalent to using a USAF 1951 target on glass. All the acceptance tests told me was whether the lens could resolve thin marks ten microns apart. That's roughly the size of features I needed to be able to separate, so the test was appropriate.</p>

<p>I didn't know the range of magnifications at which the lens under test was best so I shot at a range of magnifications. And I didn't know whether the lens under test was best wide open -- that's Luminars, generally not Tominons -- so I shot wide open, down one stop, ...</p>

<p>As I stopped down the little clouds of grain that were the tick marks' images on the negative got wider and wider, eventually merged. They were widest in the middle of the tick marks, thinnest at the very ends.</p>

<p>Same thing happens with the USAF 1951. Three bars become two, then one.</p>

<p>That's what diffraction does. It makes the blur disks wider. It doesn't make edges softer, it wipes out fine detail. </p>

[m_dawg]

<p>Large format only has more depth of field (in effect) when you use the Scheimpflug effect (lens tilt), which doesn't work except for some landscapes and some tabletop photography. If you want to take a photo of a lake or a plain, you can tilt the camera to get the entire plane in focus. If you want to take a picture of some trees or of a near/far composition with complex foreground elements across the frame (and any macro work at all) lens tilt is fairly useless.</p>

<p>And f5.6 on a digital back might have the same depth of field as f16 on large format so you need to stop down even MORE on 4x5, in general.</p>

<p>Most 4x5 lenses are designed to be used around f16-f64 and so diffraction is a huge issue when you use them with smaller formats. Most large format lenses are not very sharp. Unless you get a high-end scan, I would guess 6x7 shot with large format lenses is no better than a 10MP APS-C image--and digital backs are more like 645 and so would suffer more, but be sharper by virtue of being digital. Furthermore, some digital backs respond poorly to tilt/shift/rise/fall (they vignette severely). I would use a focus merging software.</p>

[rodeo_joe1]

<p>Actually Dan, diffraction can manifest in a number of ways. If you go down to the micron level then diffraction can actually create false bars or shadows in the middle of an even bar/space pattern, and so can give the illusion of<em> doubling</em> the resolution. Have you never seen the experiment where collimated light projected at a perfect sphere creates a small bright dot of light in the centre of the circular shadow on the other side? THAT's what diffraction does, creates additions and subtractions of light.</p>

<p>My description above was derived from observation of real prints from high-quality photographic lenses at various apertures, not by looking at an arial image via a microscope.</p>

<p>I'll thank you to stick to the subject in future and not make spurious personal attacks on contributers.</p>

[q.g._de_bakker]

<p>A bit late perhaps, but David Goldenberg:</p>

<p>Great explanation.<br />But <strong>not</strong> (!) of what diffraction is and what diffraction does.<br />Yes, the size of the aperture (and the wavelength of the light that passes through it) determines resolution. But it also sets the ratio of circumference vs area, and with it that of 'scattered', or rather: diffracted, rays vs unhindered rays.<br />Both phenomena have an effect on resolution. Aperture size alone by - as you describe well - limiting the difference in angle objects can be 'seen' from, from opposite sides of the aperture. Diffraction by throwing a veil of contrast (and with it resolution) limiting light over the image formed by the light not being bend at the limits of the bundle.<br />Two very different things.</p>

<p>Diffraction is the terrible beast that people claim it is, because it alone (i.e ignoring angular resolution limits set by the aperture size) will reduce resolution at a alarming rate.</p>

<p>Resolution effects' visibility, Frank, does indeed depend on viewing distance, print magnification, and such.<br />But (!) only in as far as at small enough magnifactions or large enough viewing distances, you no longer can tell that a print is unsharp. Or in other words: a print affected by diffraction will either be unsharp (when examined closely) or unsharp (when 'examined' under conditions that do not allow seeing whether it is sharp or not even if it is).</p>

<p> </p>