OK, with cameras getting more and more pixels, diffraction limitation keeps lowering the optimal lens aperture to the point that on the Canon 7D its down to about f5.6. This constantly comes up in all the reviews but no one ever touches on the net effect of this phenomenon. So my question to the pixel peeping experts is to what degree do the sharpening routines in PS and other programs effectively compensate for diffraction when shooting at f8 or f16 or smaller apertures? In other words is this criticism really a red herring or something that cannot be corrected in post processing? I've never seen any mention either in reviews or other discussions as to whether or not sharpening can fix this. I usually try to just shoot at the optimal aperture for my 50D but occasionally do stop down a bit. To be honest with you I can't really see the difference when shooting at f8, 16, or 22 but I'm not sure if its just PS doing a good job or my 58 year old eyes playing tricks on me. Thanks--Larry
In a few minutes, I will be posting an image in the macro gallery shot this weekend at 4x and at f11 (shot an an effective aperature of f55). The sharpness of the large print in my cube is really limited more by having missed the placement of the focal plane by maybe 0.05 mm (yes, that number is about right) than by diffraction.
Yes, diffraction is real. Yes, it will limit the print size for the above image. But at the same time, softness due to not enough DOF is a far greater sin than softness due to diffraction.
By the way, a lot of Mike Moats work is shot at f32 and is printed to 24x32. His customers and clients don't care - they just know they are seeing high quality work. Granted, he usually isn't shooting at 4x but for the real world, f32 and large prints works.
Don't worry about (and stop taking everything you read on DPreview seriously )
In other words is this criticism really a red herring or something that cannot be corrected in post processing?
That's a good question Larry. Although I would say that it's a bit of a red herring whether you can correct it or not. As Dan pointed out, it doesn't always have enough of an effect to care about. In addition, you get the exact same amount of blur from diffraction regardless of the sensor or film used. Just because the sensor can actually capture enough detail now to see the diffraction effect doesn't mean that you should worry about it. Above f/16, though, I'd be very surprised if you didn't see an effect from diffraction, assuming you have the same shot at another aperture to compare it to.
That being said, where it does come into play more is in large prints, when you want the extra resolution in order to be able to enlarge something a lot. This is going to be highly subject dependent, and you'll see diffraction effects mostly on regular, fine detail.
Here're a couple of good pages with some examples :
Now as for the question of being able to correct it, I'd guess (just a guess really) that sharpening would help with diffraction the same as it would for any other regular blur - camera vibration (not linear movement), the AA filter on your sensor, or lens/air issues. That means it would definitely improve it, but can't actually recover everything that's been blurred. Smart Sharpen in PS, and other advanced algorithms probably do better than a USM or edge contrast type, but you can only get back what you lost to a certain extent.
Hey Larry, I can't speak for diffraction for landscape and wildlife shooters, but for macro I shoot all my images that require everything in focus at the highest f/stops on all three of my macro lenses. My best selling images are shot at the f/32. I add extra sharpening which takes care of the slight softness caused by the diffraction. I blow these up to 24X36 for my art shows and also to display for my workshop students and they look awesome. The diffraction on the larger mega pixel cameras is slightly worse, but I've had all cameras in my workshops and have been able to correct the diffraction with sharpening. I have a test I did for my blog, here's the link.
Mike--great test and it seems to answer my question. Your sharpening of the f32 image is really crisp and in many areas is sharper than the f8 image. So now I guess I'll get on eBay and find a 100mm macro ans see what I can do at f32. Thanks--Larry
Posted by Steve Sieren on May 22, 2010 at 11:32 AM
I would keep on shooting for DOF and balance out the need for lens quality. F18 and f22 I pick up most of it to where it's really noticeable but if you need the dof I find it worth it.
Discussion
6 repliesIn a few minutes, I will be posting an image in the macro gallery shot this weekend at 4x and at f11 (shot an an effective aperature of f55). The sharpness of the large print in my cube is really limited more by having missed the placement of the focal plane by maybe 0.05 mm (yes, that number is about right) than by diffraction.
Yes, diffraction is real. Yes, it will limit the print size for the above image. But at the same time, softness due to not enough DOF is a far greater sin than softness due to diffraction.
By the way, a lot of Mike Moats work is shot at f32 and is printed to 24x32. His customers and clients don't care - they just know they are seeing high quality work. Granted, he usually isn't shooting at 4x but for the real world, f32 and large prints works.
Don't worry about (and stop taking everything you read on DPreview seriously
That's a good question Larry. Although I would say that it's a bit of a red herring whether you can correct it or not. As Dan pointed out, it doesn't always have enough of an effect to care about. In addition, you get the exact same amount of blur from diffraction regardless of the sensor or film used. Just because the sensor can actually capture enough detail now to see the diffraction effect doesn't mean that you should worry about it. Above f/16, though, I'd be very surprised if you didn't see an effect from diffraction, assuming you have the same shot at another aperture to compare it to.
That being said, where it does come into play more is in large prints, when you want the extra resolution in order to be able to enlarge something a lot. This is going to be highly subject dependent, and you'll see diffraction effects mostly on regular, fine detail.
Here're a couple of good pages with some examples :
http://www.luminous-landscape.com/tutorials/understanding-series/u-diffraction.shtml
http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm
Now as for the question of being able to correct it, I'd guess (just a guess really) that sharpening would help with diffraction the same as it would for any other regular blur - camera vibration (not linear movement), the AA filter on your sensor, or lens/air issues. That means it would definitely improve it, but can't actually recover everything that's been blurred. Smart Sharpen in PS, and other advanced algorithms probably do better than a USM or edge contrast type, but you can only get back what you lost to a certain extent.
http://tinylanscapes.wordpress.com/2009/03/03/diffraction-in-macro/