[Marc Sabatella]

So what happens when you stop down past that certain point?

That's when images get softer. They get sharper before that point, softer after. Specifics depend on the pixel size, focal length, and specific lens. But a typical lens might be kind of soft at f/2.8, better at f/4, better at f/5.6, best at f/8, not as good again by f/11, worse still at f/16, and by f/22 be back to where it was at f/2.8, for instance (numbers just made up, but seem close to plausible based on the few lenses I've actually tried looking for this effect with - mostly telephoto lenses).

[philbaum]

All interested parties can look at pp. 17-23 of "Photo Techniques", Jan/Feb 2009. The article is "Diffraction: Resolution taxed to its limits" by Lloyd L. Chambers.

The author shot resolution charts with Canon EOS 1Ds Mark III & Nikon 85mm f/2.8 D PC-Micro-Nikkor,
as well as Nikon D3 & Nikon 85mm f/2.8 D PC-Micro-Nikkor.

This is his comment about his "real world" shots of San Francisco: "By f/16, however, the trend is clear: contrast drops and the image begins to look flat, though actual resolved detail is still present. Apertures f/22, f/32 and f/45 worsen progressively."

Note that K20D's pixel density is greater than both the Canon and the Nikon tested.

Gosh, I quoted the same article above in my reply. Nice to see that i'm not the only one that read the article and think its an important topic :-)

Its interesting that in some magazine articles recently, there's talk about how 10 to 12 megapixels is some sort of sweet spot for DSLRs. Hopefully aps sensors won't go the way of PS cameras.

[asdf]

Gosh, I quoted the same article above in my reply.

D'oh!

Nice to see that i'm not the only one that read the article and think its an important topic :-)

Its interesting that in some magazine articles recently, there's talk about how 10 to 12 megapixels is some sort of sweet spot for DSLRs. Hopefully aps sensors won't go the way of PS cameras.

At least these limits create a demand among the members of Group f/64 for Pentax's larger format DSLRs of the future.

[Cosmo]

That's when images get softer. They get sharper before that point, softer after. Specifics depend on the pixel size, focal length, and specific lens. But a typical lens might be kind of soft at f/2.8, better at f/4, better at f/5.6, best at f/8, not as good again by f/11, worse still at f/16, and by f/22 be back to where it was at f/2.8, for instance (numbers just made up, but seem close to plausible based on the few lenses I've actually tried looking for this effect with - mostly telephoto lenses).

Cool, that's what I thought. The other person was saying that was wrong.

[Marc Sabatella]

No, I think you misinterpreted their comment. Yu might wan to re-read the exchange. You had had written that images get softer when you stop down to a specific point. That's not accurate - softness doesn't happen as you stop down *to* a specific point; softness happens when you stop down *past* that point. And that is what the other person was pointing out.

[falconeye]

Cool, that's what I thought. The other person was saying that was wrong.

Since this seems to be of interest for you. You may want to click the following link:
Pentax SMC-FA 31mm f/1.8 AL Limited Review / Test Report
and watch the blue bars in graph entitled "MTF50 in LW/PH".

The lens is the Pentax FA 31 Ltd. and the chart already stops at f/8. But it is a good illustration of what Marc said. Any other lens would produce the same shape of curve.

[dylansalt]

Yet this pic was taken @ F11 and shows zero hint of softness and compared to F8 on my 18-55 kit lens, sharper - so I reckon let your eyes do the judging



Regards

Dylan

[asdf]

Yet this pic was taken @ F11 and shows zero hint of softness and compared to F8 on my 18-55 kit lens, sharper - so I reckon let your eyes do the judging

Regards

Dylan

I don't think anyone has claimed that you can see any difference on a computer screen on small scaled-down images. You'd see the difference by f/22, while pixel peeping.

[jeffkrol]

Yet this pic was taken @ F11 and shows zero hint of softness and compared to F8 on my 18-55 kit lens, sharper - so I reckon let your eyes do the judging



Regards

Dylan

Yes and the whole "diffraction limiting" thing is much more complicated anyways.. not to mention MTF 50 maybe unnecessarily stringent in the digital age..
But it not an unambiguous indicator because aliasing is related to sensor response, and MTF at Nyquist is the product of sensor response, the de-mosaicing algorithm, and sharpening, which can boost response at Nyquist for radii less than 1. Aliasing effects may become serious over 0.3. There is a tradeoff: the more effective the sensor anti-aliasing, the worse the sharpness.
Imatest - SFR results: MTF (Sharpness) plot
and:
As others have pointed out the system response is the product of the responses of each component. There is still a point called the diffraction cutoff frequency where the lens does set an absolute limit to the system resolving power but it is at a much finer spacing than typically cited "diffraction limits". When we consider that we can sharpen the lens output to recover information where the MTF is less than 50% we find that we can get useful information with about half the pixel spacing (which is four times the pixels) as the usually cited limit. Here is a post where I calculate this limit:

Over 500 Megapixels: Open Talk Forum: Digital Photography Review

For the mathematically inclined:
Hi Joseph,

As Jay pointed out while the premise of your argument that there is a spacing at which the diffraction prevents all aliasing is correct, the pixel pitch where that occurs is significantly finer than you are using above.

First consider the ability to resolve two lines at the Rayleigh criterion separation of 1.22*wavelength*f-number: At that spacing to discern that there are two lines it is necessary to have a sample at each line plus one sample in between to detect the dip in level that shows there are two lines present and not one. This means the pixel pitch needed to detect all the information is half the Rayleigh criteria spacing.

Second consider the MTF due to diffraction and the Nyquest limit. The MTF due to diffraction of a circular aperture is 2/pi*(acos(f)-f.*sqrt(1-f.^2)) where f is the frequency as a fraction of the diffraction cutoff frequency which is 1/(wavelength*f-number). To avoid all aliasing we would therefore need to sample at twice the cutoff frequency which is a spacing of (wavelength*fnumber)/2. For a practical spacing limit where there is measurable information content we could choose a level where the MTF is not yet zero but still high enough so that after sharpening to counter the low pass effect of the diffraction MTF we would be left with a useful signal to noise for the recovered detail. For a sensor with a low ISO input signal in the thousands of electrons per pixel the noise could easily be less than 100 electrons so a high contrast detail that gets reduced by an MTF level of 10% will still have a useful positive SNR. Using the 10% criteria for the MTF level yields a frequency of approximately the Rayleigh criteria of 0.82 of the cutoff frequency so the Nyquest spacing becomes (1.22*wavelength*f-number)/2 in agreement with the first consideration above.
Reference for diffraction MTF:
http://www.mellesgriot.com/products/optics/os_2_2.htm
http://forums.dpreview.com/forums/read.asp?forum=1018&message=30447131

[Cosmo]

No, I think you misinterpreted their comment. Yu might wan to re-read the exchange. You had had written that images get softer when you stop down to a specific point. That's not accurate - softness doesn't happen as you stop down *to* a specific point; softness happens when you stop down *past* that point. And that is what the other person was pointing out.

I said that in my first post, but corrected myself in my second.

[Digitalis]

Yet this pic was taken @ F11 and shows zero hint of softness and compared to F8 on my 18-55 kit lens, sharper - so I reckon let your eyes do the judging



Regards

Dylan

The Kit lens isn't what I would call a high quality optic. At 55mm f/11 is the peak of resolution for that particular lens, but for faster lenses they hit their peak much earlier, which is why I use them.

[ryan s]

Too many numbers and calculations going on in this thread...skimmed page 1...got bored...went to page 3.

I just got this email (and I hope the images show up...let me know if not and I'll rehost them) :

DIFFRACTION VERSUS SATISFACTION!

These past few months, the Q/A in many of our classrooms has been abuzz with concerns from our students about "shooting pictures at the aperture of f/16 or f/22."

Seems a couple of those "big" Photography Forum websites have unleashed some really OLD NEWS that when a lens is set to the smaller apertures, such as f/16 or f/22, lens diffraction is more noticeable; in laymen's terms, lens diffraction means a loss of contrast and sharpness.

I figured now is as good a time as any to set the record straight about 'lens diffraction' and want you to know that I have been shooting many of my landscape/cityscape images in this manner for years. So without further ado, here is what I have to say-"Shooting at F/22 can be a GREAT IDEA!" Any worries about loss of sharpness and contrast are just as over-blown as the Y2K fears!"

In 35+ years of shooting commercially, I can't ever remember a time when a client said "Bryan, whatever you do, don't shoot at F/22" nor can I ever remember a single instance where either Getty or Corbis called me to say, "Bryan, DON'T send us any of your pictures for our stock files if they were shot at F/22." And the reason I can't remember is because it NEVER HAPPENED and it NEVER WILL!

The aperture of F/22 produces a massive depth of field, in particular when combined with a wide-angle lens. And IF you have any ounce of creativity in your DNA, you will want to strive to use some foreground interest in your overall composition, since it will be the foreground interest that will create the illusion of depth and subsequent respective in your composition. And the only way to record sharpness from front to back when using an immediate foreground is to use F/22, the smallest lens opening which in turn produces the greatest depth of field a.k.a. the greatest are of acceptable sharpness.



Both of these images were shot with the same lens, Nikkor 12-24mm at the 16mm focal length and they are both shot at the same exact exposure in terms of its quantitative value, but oh my, is there ever a noticeable difference in their overall sharpness. The first image (left) was shot at the "dreaded" aperture of F/22 and the second image (right) was shot at the "highly recommended aperture" of F/8.

I don't know about you, but we here at the school prefer the image taken at F/22 because it shows the overall area of acceptable sharpness that we really need to convey here-FRONT TO BACK! In the image taken at F/8 clearly, we don't have sharpness front to back! So, with the proof staring at you right on your screen, what do you think? Are you going to begin to embrace the use of F/22? You will if you have any intention of being a creative photographer, because quite simply, you will NEVER record the great foreground to background landscape shots unless you choose f/22!

And just for the record; these next two examples are the same shot of this very scene but at a 200% magnification. The difference between the sharpness in either shot is nil but I will be the first to admit that there is a wee bit more contrast in the bark of the tree with the shot taken at f/8, BUT again this is at a 200% magnification. This wee bit loss in contrast is also something that I and so many other discerning photographers can live with!



The long and the short of it; the question of using F/22 was NEVER an issue during the days when we all shot film and it should NOT be an issue today. Diffraction is a real event, but it should never get in your way of shooting those compositions that demand extreme depth of field. Satisfaction is your reward, so get out there and get creative at f/22!

All My Best!

Bryan F. Peterson/Founder
The Perfect Picture School of Photography
Online Photography Classes: The Perfect Picture School of Photography

So with that...I leave this thread only to check on it if others can't see the pics. Have fun with all the mind-numbing number crunching. Good that some people care, cause I sure don't

[simico]

I know the kit lens isn't the best for this test (and probably the test isn't perfect either), but I don't have anything better at the moment.
Took 5 shots of the same scene with aperture set to F5.6, F8, F11, F22, F29 - everything else (except shutter speed, of course) was the same.

Picture showing the full image and the 100% crop areas


100% crops, aperture from left to right: F5.6, F8, F11, F22, F29
(for some reason it cannot be linked as picture)

The softness due to diffraction at F22 and F29 is clearly visible when viewing at screen size or on downsized images, too.

[ryan s]

The f/5.6 and f/8 shots look to be oversharpened.

Still...f/22 shots would be fine with careful USM added last in post processing.

[Marc Sabatella]

They might be "fine" in some sense, but it's also very obvious they aren't as good as f/8. That's a very good demonstration of soemthing that is also very easy to demonstrate for oneself.