Posted by Dan Roberts on December 28, 2012 at 12:33 PM
I suppose like most people as my skills develop and my equipment inventory improves, I pay more and more attention to the minutia. So when I added a high rez body to my kit I started reading about pixel size and diffraction. This topic poses a dilemma because I got the high rez camera primarily for landscapes. Now I understand that at apertures of f/8 or smaller, diffraction degrades IQ. So my question to the forum community is how much does this affect photography in the real world? And I mean based on actual observed results with your own work.
Posted by Mark Stahlke on December 28, 2012 at 12:59 PM
Diffraction is always present at all apertures. It increases as the aperture decreases. I think it's more important to get the DOF I want so I never worry about diffraction.
Shooting landscapes on 4x5, I regularly uses apertures of f/22, f/32, and f/45. I don't notice any diffraction effects. I don't think there's any reason to obsess about diffraction.
However, The enlargement ratio of 4x5 for a print is much smaller than the enlargement ratio for a DSLR print of comparable size.
I agree with Mark, here. You would notice areas with poor focus, usually due to lack of DOF, long before you would see any diffraction effects. Also, if you print your work, the texture of the paper would obscure diffraction fringing, anyway.
In my experience, the sweet spot, aperture-wise, for small formats lies in the f8 to f16 range. Like Mark, I shoot 4x5 and regularly use the same apertures and have not noticed any ill effects from diffraction in 16x20 prints.
I think it's more important, especially in landscape work, if you use an aperture that gives you DOF you need for the sharpest image across the frame.
It's easy to get caught up in the technical minutiae; we've all done it, but in the end, it's best to stick to the basics.
Posted by RL Potts on December 28, 2012 at 7:37 PM
How does it affect photography in the real world? I shot for a long time with medium format film. I couldn't see much harm at f16 or 22, but f32 was bothersome. Large format folks do lots of work at f32-45. The D800 is totally different. You get to see everything in stunning clarity with excellent primes. Lots of folks use wides for landscapes, but it seems to me that these lenses start showing diffraction sooner. Longer lenses aren't too problematic at f8-11. Things certainly look better at f5.6 with a wide. My approach is DOF be hanged. I want sharpness at my chosen point of focus. That's how you get depth and pop. If you do a tight crop and print to A3, you will be able to see all of the factors you are discussing. I end up printing a fair number of images that way. It is nice to have that option.
Posted by John Mumaw on December 28, 2012 at 7:50 PM
Using focus stacking as it seems you might try (per a previous post) pretty much puts diffraction issues to rest. This is a pretty major leap in photography, IMO. 'Perceived sharpness' using small apertures is no longer an issue when real sharpness can be achieved.
It is so simple that I pretty much stick with f8-11 apertures, which is the best quality aperture for my lenses, and that's what is most important to me.
Posted by Dan Roberts on December 29, 2012 at 12:40 AM
Thanks for the replies, all. John, as you noted my other post on focus stacking was definately related. That's what I am trying to get at, is diffraction enough of an issue to warrant shooting multiple images at lower f-stop and stacking them? I recently got a D800E and have only had it on one shoot. Did some landscapes shot at small apertures and was really disappointed with the results. Used a decent lens and good technique including mirror lock to eliminate all camera shake so somewhat perplexed with the less than perfect sharpness. That's what started me questioning whether diffraction was the culprit. I guess ultimately I'll have to do some testing with various lenses to determine whether the difference justifies the additional work.
The digital world gets more amazing all the time
Posted by David Bostedo on December 29, 2012 at 2:08 PM
Dan - A few notes on diffraction :
-- It's a property of the lens aperture, and the camera attached does not in any way affect the amount of diffraction. All lenses produce diffraction blur that increases as the aperture size decreases. And I think all produce approximately the same amount of diffraction blur for a given physical aperture size. (I.e. there isn't any lens that's much better for diffraction than any other.)
-- If diffraction is visible in your photos, it's because you're resolving as much detail as possible - so much in fact that you've hit the diffraction limit, which is the maximum detail that a lens can possibly resolve. This isn't all that hard to do... you can see its effect yourself if you, say, tape a dollar bill to a wall, and shoot it at various apertures with a very fast shutter speed on a tripod (to eliminate other blurring effects).
-- With your D800, you'll see diffraction effects at larger apertures than with a lower resolution camera, because the pixels are smaller. But that image, diffraction and all, will still have more resolution than an image with a lower megapixel camera that isn't able to resolve the diffraction blur. For instance, if you see a little diffraction blur with your D800 at f/8, and you were to shoot the same scene with a D700 and it showed no blur, that's just because you've resolved so much more with the D800. At equivalent reproduction sizes, the D800 image will still be much better technically than the D700.
-- Diffraction blur even when you can see it isn't all that bad, compared to other sources of blur. Many folks are willing to trade some diffraction blur for greater depth of field. And depending on how you either pixel-peep or press your nose to a print, you may not even notice it.
Posted by Dan Roberts on December 30, 2012 at 12:20 AM
Thanks, David. I understand the theory around diffraction. That's why the original post was an inquiry regarding whether it is a practical issue. In spite of the science behind it, at the end of the day we're talking about the perception of sharpness which is subjective. Though there are software programs now that can help identify the sharpest image of a group. Which may be the way to ultimately resolve this subject.
Posted by David Bostedo on December 30, 2012 at 11:36 PM
...an inquiry regarding whether it is a practical issue.
I suppose to more directly address that : yes, it's a practical issue because you need to decide how to deal with diffraction, or just to ignore it. Either way, you should at least know which choice your making and why.
...we're talking about the perception of sharpness which is subjective.
I disagree with that if you're talking about what to do about diffraction. We're talking about actual, measurable (and visible), sharpness difference at a low level. Depending on your final output - be it screen or print - you might be talking about perceived sharpness. But at capture time, there's a real difference at the pixels. (Of course, that doesn't mean that the best course of action for you and your images isn't just to ignore diffraction.)
Posted by Mark Seaver on January 2, 2013 at 10:41 AM
Dan, In my macro work with a 100mm macro lens, I used f/22 - f/32 regularly to get maximum dof. With my 5D and viewing at 100% it was only at f/25 and f/32 that I could see diffraction effects and what I saw was a general reduction in the dynamic range, not an apparent loss of resolution. At f/25, the difference was subtle enough that in most cases it would not be noticed. Even at f/32, I could boost the contrast slider to get results that duplicate f/22 images. One thing to be aware of in this sort of analysis is that we have instruments that are capable of measuring differences that the human eye can't see, so the question becomes do you push for the best that can be measured or do you stop at the limits of human perception. The answer to that is philosophical.
Posted by Dan Roberts on January 2, 2013 at 2:31 PM
Thanks for the input regarding your own experience and perspective, Mark.
... One thing to be aware of in this sort of analysis is that we have instruments that are capable of measuring differences that the human eye can't see, so the question becomes do you push for the best that can be measured or do you stop at the limits of human perception...
This really gets at the crux of the practical question. There is an arguement to be made to simply make it as good as you can. But how much effort is it worth for a potentially imperceptible or barly perceptible difference? I do have some of the software to which you refer so I suppose I'll do some testing to wile away the winter hours anyhow. See how much better the software's eyes are than mine.
Discussion
11 repliesShooting landscapes on 4x5, I regularly uses apertures of f/22, f/32, and f/45. I don't notice any diffraction effects. I don't think there's any reason to obsess about diffraction.
However, The enlargement ratio of 4x5 for a print is much smaller than the enlargement ratio for a DSLR print of comparable size.
In my experience, the sweet spot, aperture-wise, for small formats lies in the f8 to f16 range. Like Mark, I shoot 4x5 and regularly use the same apertures and have not noticed any ill effects from diffraction in 16x20 prints.
I think it's more important, especially in landscape work, if you use an aperture that gives you DOF you need for the sharpest image across the frame.
It's easy to get caught up in the technical minutiae; we've all done it, but in the end, it's best to stick to the basics.
--P
A very solid point.....
It is so simple that I pretty much stick with f8-11 apertures, which is the best quality aperture for my lenses, and that's what is most important to me.
The digital world gets more amazing all the time
-- It's a property of the lens aperture, and the camera attached does not in any way affect the amount of diffraction. All lenses produce diffraction blur that increases as the aperture size decreases. And I think all produce approximately the same amount of diffraction blur for a given physical aperture size. (I.e. there isn't any lens that's much better for diffraction than any other.)
-- If diffraction is visible in your photos, it's because you're resolving as much detail as possible - so much in fact that you've hit the diffraction limit, which is the maximum detail that a lens can possibly resolve. This isn't all that hard to do... you can see its effect yourself if you, say, tape a dollar bill to a wall, and shoot it at various apertures with a very fast shutter speed on a tripod (to eliminate other blurring effects).
-- With your D800, you'll see diffraction effects at larger apertures than with a lower resolution camera, because the pixels are smaller. But that image, diffraction and all, will still have more resolution than an image with a lower megapixel camera that isn't able to resolve the diffraction blur. For instance, if you see a little diffraction blur with your D800 at f/8, and you were to shoot the same scene with a D700 and it showed no blur, that's just because you've resolved so much more with the D800. At equivalent reproduction sizes, the D800 image will still be much better technically than the D700.
-- Diffraction blur even when you can see it isn't all that bad, compared to other sources of blur. Many folks are willing to trade some diffraction blur for greater depth of field. And depending on how you either pixel-peep or press your nose to a print, you may not even notice it.
-- For a bit more detail, you might want to look at this Cambridge in Colour tutorial on diffraction
I suppose to more directly address that : yes, it's a practical issue because you need to decide how to deal with diffraction, or just to ignore it. Either way, you should at least know which choice your making and why.
I disagree with that if you're talking about what to do about diffraction. We're talking about actual, measurable (and visible), sharpness difference at a low level. Depending on your final output - be it screen or print - you might be talking about perceived sharpness. But at capture time, there's a real difference at the pixels. (Of course, that doesn't mean that the best course of action for you and your images isn't just to ignore diffraction.)
This really gets at the crux of the practical question. There is an arguement to be made to simply make it as good as you can. But how much effort is it worth for a potentially imperceptible or barly perceptible difference? I do have some of the software to which you refer so I suppose I'll do some testing to wile away the winter hours anyhow. See how much better the software's eyes are than mine.