Geoff D F wrote:
Here is my gratuitous contribution about diffraction - yes it is real and applies to all light traveling past a physical edge that blocks light, aperture blades of lenses included. It is not something that can be avoided with better lens design, within a given format. However, larger formats will suffer from less diffraction at any given aperture, but they will also have less depth of field.
At the same time it is wrong to have a blanket rule that says one should not shoot past the diffraction limit. As G Dan has pointed out, sometimes one will want to trade the extra depth of field for a slight softening of the image. A good example is macro work where before focus stacking was available a lot of subjects were shot on FF at f16-f22 or even f32.
To put this softening into perspective, if you look at a website like Optical Limits which tests for lens sharpness at different apertures, you will notice most Fuji APSC lenses tend to show softening beyond f8, ie f11 tends to show less resolution than f8. However, even at f11 many lenses show better performance than they do wide open, and most lenses these days show quite useable performance wide open. F16 might show further softening but again this might still be comparable to a wide-open shot in terms of sharpness.
When I shoot on Fuji, most of the time I don't go past f8 because I mostly don't need it, but on the occasions where I might have accidently dialed in f11 to f16 I haven't noticed diffraction badly softening the image.
There are also a lot of other things that affect perceived sharpness. Sharpening in post-processing, and adding contrast/clarity add sharpness, while adding saturation can reduce it. If shooting film, the choice of film stock makes huge difference. ...Show more →
An excellent summary of the realities.
Here's a link to the adaptall-2.com site. It has the MP numbers for the 17 Tamron, which has its best center performance at f/4 & 16. Even at 16, it got 67lp/mn, sharper than any aperture on the Canon 50/1.8, which topped out at 63lp/mm IIRC, presumably the metal mount one that used to be popular on here.
So no, you dont necessarily have to shoot at 5.6 or some wider aperture to get best performance. It is not a law of nature. Another one that does well stopped down is the 80-200/2.8 adaptall-2 if you care to look, and I'm pretty sure the FD 20-35 L had its best performance f/11 & 16
AmbientMike wrote:
Here's a link to the adaptall-2.com site. It has the MP numbers for the 17 Tamron, which has its best center performance at f/4 & 16. Even at 16, it got 67lp/mn, sharper than any aperture on the Canon 50/1.8, which topped out at 63lp/mm IIRC, presumably the metal mount one that used to be popular on here.
So no, you dont necessarily have to shoot at 5.6 or some wider aperture to get best performance. It is not a law of nature. Another one that does well stopped down is the 80-200/2.8 adaptall-2 if you care to look, and I'm pretty sure the FD 20-35 L had its best performance f/11 & 16...Show more →
Yes, this is great because we all shoot 25 year old Tamron Adaptall lenses on our current mirrorless cameras. Wow. 🙄
What you fail to understand is that:
1) diffraction is a real, incontrovertible physical phenomenon.
2) it becomes visible on the Fuji X 40mp sensors at f8 — but you need a good lens, one capable of resolving the sensor to see it.
3) some lesser optical lenses may be enough sharper at a smaller aperture than the diffraction limit for that system, that their resolution gain outweighs the resolution loss from diffraction, but this is an artistic choice, and moreover it doesn’t improve the base performance of that lens.
4) one might choose a smaller aperture that shows diffraction because the gain in DoF outweighs the loss from diffraction, especially true if using longer lenses, but again this is an artistic choice.
5) neither number 3 or 4 are any kind of proof or support that diffraction doesn’t exist or doesn’t matter — whether it matters remains up to the individual artist.
I have still a couple of Adaptall lenses, 28/2.5 & 70-210, and good they are too even after 35 years. When I wanted really wide I tried the Tamron 17 against the Tokina, bought the Tokina, and still have that too.
As far as I can see their mtf jeasurementsaredone on full frame film, s9difracraction would-be noticeable later than on a aps-c sensor, and other factors come in too such as tgeslight curvatureoffilm in tge gate vs sensor stack thickness, etc.
The 17 Tokina is not as good as the 14mm Fuji on my Fuji cameras. Showing some degradation in the corners vs its film performance on Nikon, likely due to the above mentioned issues with film era short focal length lenses on digital sensors tacks. It was even worse on tge Sony A7 when I had it.
AmbientMike wrote:
Here's a link to the adaptall-2.com site. It has the MP numbers for the 17 Tamron, which has its best center performance at f/4 & 16. Even at 16, it got 67lp/mn, sharper than any aperture on the Canon 50/1.8, which topped out at 63lp/mm IIRC, presumably the metal mount one that used to be popular on here.
So no, you dont necessarily have to shoot at 5.6 or some wider aperture to get best performance. It is not a law of nature. Another one that does well stopped down is the 80-200/2.8 adaptall-2 if you care to look, and I'm pretty sure the FD 20-35 L had its best performance f/11 & 16...Show more →
Every lens is a little different; some will show little diffraction even at very small apertures because diffraction is a factor of a lot of things (how the light hits the lens, the lens itself, and the sensor/film format).
So diffraction itself is a real thing but there isn't a specific numerical limit where it will set in on every lens (nor did anyone say there was, to my knowledge?).
The general rule of thumb still holds that the more you stop down, the likelier you are to see diffraction, all other things being equal.
They made lenses that were sharp at f/16, decades ago. So theres clearly no law saying it can't be done. Or that it can't be sharper at f/16 than f/8
Yes, like I've been saying, a lot of lenses are sharpest at 5.6 lately, and they obviously give better performance at 5.6 than f/8 on 40mp (or 24 or 16mp.) Or you'd be saying they're better at f/8. Which proves nothing.
If a lens happens to be sharper at f/4 than 5.6, does that mean you are to shoot at f/4 on other lenses, because there is some natural law saying you lose sharpness past f/4? No, that's ridiculous
AmbientMike wrote:
They made lenses that were sharp at f/16, decades ago. So theres clearly no law saying it can't be done. Or that it can't be sharper at f/16 than f/8
Yes, like I've been saying, a lot of lenses are sharpest at 5.6 lately, and they obviously give better performance at 5.6 than f/8 on 40mp (or 24 or 16mp.) Or you'd be saying they're better at f/8. Which proves nothing.
If a lens happens to be sharper at f/4 than 5.6, does that mean you are to shoot at f/4 on other lenses, because there is some natural law saying you lose sharpness past f/4? No, that's ridiculous ...Show more →
There may be large format lenses that are sharper at f16 than f8, but I am not aware of any FF or smaller format lenses that show an improvement at f16 compared to f8. I've probably looked at hundreds of lens tests at Optical Limits and other sites and never seen that. I suppose it may be possible if the lens suffers from a lot of aberations so that performance still improves past the diffraction limit, but then that would be due to a poorly designed lens not some super design that has managed to avoid diffraction.
That is not to say F16 is not useable in some situations.
Geoff D F wrote:
There may be large format lenses that are sharper at f16 than f8, but I am not aware of any FF or smaller format lenses that show an improvement at f16 compared to f8. I've probably looked at hundreds of lens tests at Optical Limits and other sites and never seen that. I suppose it may be possible if the lens suffers from a lot of aberations so that performance still improves past the diffraction limit, but then that would be due to a poorly designed lens not some super design that has managed to avoid diffraction.
That is not to say F16 is not useable in some situations....Show more →
If you can find a Modern Photography test of the FD 20-35 L, IIRC that had best performance at f/11 & 16. Also the 17 Tamron i posted a link to is excellent in the center at f/16
Optical Limits is mostly newer lenses, even my 11-16 and 180 Tamron are new enough to be best at around f/5.6 & f/4, respectively. I really like both, doesn't make much sense to make a landscape lens best at 5.6 imo, and macro best f/4. Although I use the 180 hh mostly, so its worked out well
AmbientMike wrote:
To post the usual garbage that get posted about diffraction? To try to correct me since I don't agree on said garbage?
You did in fact say that "at some point any lens starts to get less sharp if stopped down" (here)
This is because of diffraction.
I'm sure there are plenty of lenses that are still very sharp at f11, f16 or even f22. But a decent piece of advice is that one shouldn't stop down that far unless needed.
mdude85 wrote:
You did in fact say that "at some point any lens starts to get less sharp if stopped down" (here)
This is because of diffraction.
I'm sure there are plenty of lenses that are still very sharp at f11, f16 or even f22. But a decent piece of advice is that one shouldn't stop down that far unless needed.
Ive already posted one lens test showing best performance at f/16. Why not use it, if you want sharp images.
Found the site that used to have FD 20-35 L MP test, link is still there, lens test is gone, though. Another site said f/11 is best for landscape, basically like I said, best performance at f/11 & 16. Whether they tried f/16 given the diffraction paranoia i can't tell you
AmbientMike wrote:
So you're claiming this law of physics only applies to digital? Youre so full of it
No, I didn't say that. But the diffraction will be different for aps-c over full frame (about one stop) and tests of shortfall length lenses on film are generally not good guidebook perf9rmance on digital, and that will also depend on individual sensor construction and cover glass thickness. You should know that, plenty of discussion on FM over the years about this.
AmbientMike wrote:
Ive already posted one lens test showing best performance at f/16. Why not use it, if you want sharp images.
Found the site that used to have FD 20-35 L MP test, link is still there, lens test is gone, though. Another site said f/11 is best for landscape, basically like I said, best performance at f/11 & 16. Whether they tried f/16 given the diffraction paranoia i can't tell you
I agree -- if the lens is sharpest at f/16, and you are able to use it at f/16, then go ahead and use it.
gyoung143 wrote:
No, I didn't say that. But the diffraction will be different for aps-c over full frame (about one stop) and tests of shortfall length lenses on film are generally not good guidebook perf9rmance on digital, and that will also depend on individual sensor construction and cover glass thickness. You should know that, plenty of discussion on FM over the years about this.
Gerry
I realize there are differences between film and digital, but if there's a law about sharpness I can't imagine it only applying to one or the other
I suspect any issues due to glass or other thickness etc on the sensor are more likely to cause problems at 1.2 than 16, they might get covered by stopping down.
Ive gotten a lot of excellent images using older film lenses, and nikon in particular didn't quit producing film lenses once digital took over. I don't think Canon or Leica did, either
gdanmitchell wrote:
Seriously folks, what possible value is coming from continuing this back and forth? There's nothing new to add and no minds are going to be changed.
Dan, someone on the internet is WRONG and it absolutely has to be fixed!!!
