Awhile back, someone posted a link to a site that shows info on the sharpest apertures for various camera sensors. If memory serves, on the 30d bodies I most often use, f9 is about the point of max sharpness as I stop down before diffraction begins to soften the image, and if memory also serves, the more pixels are packed on a sensor, the wider the aperture for max sharpness.
Does anyone remember that site, and could you post the link again? I had it bookmarked, but my OS melted down a month ago and I had to reinstall everything, so I lost it.
There was also a link posted here a goodly while ago to a DOF calculator, which I had bookmarked, used widely, and then lost.
Hey Steve, as you know I shoot most of my images at f/32 and don't seem to have a problem with diffraction. I looked at the link you posted for the diffraction and couldn't believe my eyes when I saw the image where you could view it at the different f/stops and see the difference in diffraction from f/8 to f/22. I didn't notice what lens they were using but I've never experienced diffraction that looked like that. I did my own test with a tamron 90mm macro and didn't have results that looked anywhere near as extreme as they did on this site. I posted the images on my blog if anyone would like to check it out. www.MikeMoatsBlog.com
Posted by David Bostedo on March 2, 2009 at 11:07 PM
A few points, guys....diffraction is not a property of any particular lens or sensor. Rather, it's caused by the size of the aperture opening, and distance from the sensor. This is the same for any lens (for a given focal length) and any sensor. (As you get smaller pixels, the amount of blur based on diffraction starts to show up sooner - but the actual diffraction of light is the same. You just don't see it because of larger pixels. Hence, if I remember correctly, f/11 or so on cropped cameras, and f/16 or higher on larger sensor cameras.)
I do think that it might be different for macro, since you're in such a unique case with the high repro ratios and close subject distances. I don't know of anything online that explains if that's the case, but your test does seem to show that there isn't much of a diffraction effect, even for f/32. If I find anything about that, I'll post it...
Hey David, I've read all the same tech stuff about diffraction and my mind goes numb. You may be right about the macro lens being different from other lenses, I couldn't say cause I only work in macro, but I have photographers telling me all the time you shouldn't shoot that way, but I've yet to have a magazine editor, art director, art consultant, art buyer, or contest judge ever tell me my images suffered from diffraction. As long as it works I'm gonna keep doing it..
Posted by Dan Borzynski on March 2, 2009 at 11:58 PM
Macro lenses are different from other lenses, from the point of view of diffraction. (David, I know you just recoiled in horror with that statement).
If you use a macro lens at normal focussing distances, it behaves like any other lens with respect ot diffraction.
However, macro lenses have the wonderful property that you can easily go beyond 1:1 with them and a few accessory items, or even get the Canon 65mm 1x-5x lens. Very, very roughly (and with current lens designs I have to emphasize the rough part), the effective aperature of a lens is
effective f-stop = marked f-stop x (magnification + 1).
At a normal focussing distance, say 0.05 of life size, or a head and shoulders portrait, the effective fstop at f8 is 8x(1 + 0.05) = f8.4, which really won't show much deterioration from diffraction going from f8 to f8.4.
On the other hand, once you get down to 3 times life size, and f16 to get something that vaguely resembles depth of field, you get an effective fstop of 16x(1+3) = 64, or f64 and the diffraction effects which go with it. That, when enlarged from a 35 mm based image source, will begin to show diffraction effects.
In theory, you'll get the same effects with any lens once you get to, say three times life size. It is just that macro lenses lend themselves so nicely to this type of abuse.
Personally, I would rather take the effects of diffraction over not having enough DOF, knowing that it will restrict how large I can print at before it becomes apparent to the eye.
Posted by David Bostedo on March 3, 2009 at 12:15 AM
...I know you just recoiled in horror with that statement...
Nah...I figured that not including subject distance in the "normal" diffraction equations/calculators was likely a simplification. I just didn't know in what way. Apparently, macro is one of those ways. Of course, you're right about getting good DOF - your results would be much worse than any diffraction effect.
Posted by Dan Borzynski on March 3, 2009 at 12:46 AM
I'd much rather think about things like this in terms of magnification. It gets rids of needing to deal with additionally variables like subject distance and focal length, both of which effectively give you the magnification anyway.
(From John Shaw's book on macro - I have to look this part up. Actually, I had to find the book first.)
Actually, the equation I gave above works for what is called a symmetical lens design. Most current macro lens designs are not symmetric. Take the your lens and measure the apparent diameter of the diaphram from both sides of the lens. Then take the ratio
(back measurement)/(front measurement)
which is usually less than 1. Plug this into the equation
effective f stop = actual stop x (magnification/(the above ratio) +1)
which will show that you have a higher effective stop and consequently more diffraction issues.
It gets ugly beyond 1:1, no matter how you look at it.
Mike, I'm with you 100 percent on what works or what doesn't work. If it works in the real world, it works. Period. That is the only test that matters.
One of the things I like least about the online revolution is that it's spawned a whole generation of pixel-peeing wannabe digital photography experts whose reviews and perspectives focus on a computer screen (the lowest quality threshold) instead of printing (the highest quality threshold). Now, in saying that I'm certainly not casting aspersions on folks in here, where professionalism and real-world ethos run strong. In fact, that's why I keep coming here, for just those valuable perspectives.
I do a lot of different types of photography, and many of them under conditions where aperture isn't too important. But some in which it is. If I'm shooting an event outside in bright conditions on my 1.6 crop bodies and I want more DOF but also require very sharp images, it helps me to know that my 70-200, for example, is at its sharpest on my bodies at f9 and then starts to fall off slightly.
Also, I notice in my own macro work (usually the Canon 100 f2.8 macro with or without extension tubes on a 1.6 crop body), that overall sharpness does fall off past f11. Whether that's purely a lens issue or a combination of lens/diffraction or purely diffraction I don't know. And on the most important level I don't care. I'm not so interested in the physics of why a thing happens, but in the results of the thing happening, and about how and when to control those things.
And, practically speaking in these cases, I care about it requiring more sharpening (and more sophisticated sharpening) for me to bring a macro image to its best for printing/publication at f16-f32 than it does at f8 or f11. It is a subtle falloff at first, but at f22 or greater, I notice it more. And it's not a technique difference. We're talking about rock solid tripod, mirror lockup and a remote release, as well as testing I've done in the home studio on completely static subjects with no movement of any type compromising the series.
I will say that in no case has f32 produced images that could not be sharpened effectively, simply that in my experience, with my gear, there has been a difference.
Boy, ask a lazy question, find your own answer, and the next thing you know an interesting discussion has ensued. Who knew?
Hey Steve, well said, and agree about what works in the real world, and do not care why or how or who says what, its about whats working with the equipment I'm using at the time. I will be the first one to admit that diffraction does come into play stopped down, but it isn't something that can't be adjusted with some sharpening. With all my workshops I have many different cameras and lens that people are working with and have never had a problem with with this style unless they are using a total crap lens. Correcting diffraction with sharpening isn't any different than dealing with other problems that we correct with PS, like HDR, poor exposure, cropping, cloning out distractions.
Well, from a practical side I always test a new lens to determine the sharpest aperture for that lens/camera combination using Bill Chambers excellent article "Getting the sharpness you paid for" in the NPN articles (see url below). I usually use that f-stop to limit diffraction effects. However, I now guess I need to go back and shoot a series of test shots at smaller apertures and see how well sharpening does to eliminate the problem as Steve has suggested. Eventually I hope to be able to stop testing my lenses and get back to taking real photos--Larry
Discussion
11 repliesAwhile back, someone posted a link to a site that shows info on the sharpest apertures for various camera sensors. If memory serves, on the 30d bodies I most often use, f9 is about the point of max sharpness as I stop down before diffraction begins to soften the image, and if memory also serves, the more pixels are packed on a sensor, the wider the aperture for max sharpness.
Does anyone remember that site, and could you post the link again? I had it bookmarked, but my OS melted down a month ago and I had to reinstall everything, so I lost it.
There was also a link posted here a goodly while ago to a DOF calculator, which I had bookmarked, used widely, and then lost.
Thanks for the help!
Edit: Ooops, here's a DOF calculator I turned up with a quick Google search, and a diffraction calculator, too. Lazy Steve! Bad Steve!
DOF: http://www.dofmaster.com/dofjs.html
Diffraction: http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm
http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm
I do think that it might be different for macro, since you're in such a unique case with the high repro ratios and close subject distances. I don't know of anything online that explains if that's the case, but your test does seem to show that there isn't much of a diffraction effect, even for f/32. If I find anything about that, I'll post it...
If you use a macro lens at normal focussing distances, it behaves like any other lens with respect ot diffraction.
However, macro lenses have the wonderful property that you can easily go beyond 1:1 with them and a few accessory items, or even get the Canon 65mm 1x-5x lens. Very, very roughly (and with current lens designs I have to emphasize the rough part), the effective aperature of a lens is
effective f-stop = marked f-stop x (magnification + 1).
At a normal focussing distance, say 0.05 of life size, or a head and shoulders portrait, the effective fstop at f8 is 8x(1 + 0.05) = f8.4, which really won't show much deterioration from diffraction going from f8 to f8.4.
On the other hand, once you get down to 3 times life size, and f16 to get something that vaguely resembles depth of field, you get an effective fstop of 16x(1+3) = 64, or f64 and the diffraction effects which go with it. That, when enlarged from a 35 mm based image source, will begin to show diffraction effects.
In theory, you'll get the same effects with any lens once you get to, say three times life size. It is just that macro lenses lend themselves so nicely to this type of abuse.
Personally, I would rather take the effects of diffraction over not having enough DOF, knowing that it will restrict how large I can print at before it becomes apparent to the eye.
Nah...I figured that not including subject distance in the "normal" diffraction equations/calculators was likely a simplification. I just didn't know in what way. Apparently, macro is one of those ways. Of course, you're right about getting good DOF - your results would be much worse than any diffraction effect.
I'd much rather think about things like this in terms of magnification. It gets rids of needing to deal with additionally variables like subject distance and focal length, both of which effectively give you the magnification anyway.
(From John Shaw's book on macro - I have to look this part up. Actually, I had to find the book first.)
Actually, the equation I gave above works for what is called a symmetical lens design. Most current macro lens designs are not symmetric. Take the your lens and measure the apparent diameter of the diaphram from both sides of the lens. Then take the ratio
(back measurement)/(front measurement)
which is usually less than 1. Plug this into the equation
effective f stop = actual stop x (magnification/(the above ratio) +1)
which will show that you have a higher effective stop and consequently more diffraction issues.
It gets ugly beyond 1:1, no matter how you look at it.
One of the things I like least about the online revolution is that it's spawned a whole generation of pixel-peeing wannabe digital photography experts whose reviews and perspectives focus on a computer screen (the lowest quality threshold) instead of printing (the highest quality threshold). Now, in saying that I'm certainly not casting aspersions on folks in here, where professionalism and real-world ethos run strong. In fact, that's why I keep coming here, for just those valuable perspectives.
I do a lot of different types of photography, and many of them under conditions where aperture isn't too important. But some in which it is. If I'm shooting an event outside in bright conditions on my 1.6 crop bodies and I want more DOF but also require very sharp images, it helps me to know that my 70-200, for example, is at its sharpest on my bodies at f9 and then starts to fall off slightly.
Also, I notice in my own macro work (usually the Canon 100 f2.8 macro with or without extension tubes on a 1.6 crop body), that overall sharpness does fall off past f11. Whether that's purely a lens issue or a combination of lens/diffraction or purely diffraction I don't know. And on the most important level I don't care. I'm not so interested in the physics of why a thing happens, but in the results of the thing happening, and about how and when to control those things.
And, practically speaking in these cases, I care about it requiring more sharpening (and more sophisticated sharpening) for me to bring a macro image to its best for printing/publication at f16-f32 than it does at f8 or f11. It is a subtle falloff at first, but at f22 or greater, I notice it more. And it's not a technique difference. We're talking about rock solid tripod, mirror lockup and a remote release, as well as testing I've done in the home studio on completely static subjects with no movement of any type compromising the series.
I will say that in no case has f32 produced images that could not be sharpened effectively, simply that in my experience, with my gear, there has been a difference.
Boy, ask a lazy question, find your own answer, and the next thing you know an interesting discussion has ensued. Who knew?
http://www.naturephotographers.net/articles1207/bc1207-1.html