Lens Diffraction
- Thread starter Rick M
- Start date
Interesting topic.
I have to wonder though if most of the time soft that folks like to contribute to diffraction is more a result of a slower shutter speed being used with the smaller aperture.
Reading that diffraction affects red, green and blue differently, loosing sharpness in one color would seem to be the definite indication that diffraction is culprit.
I have to wonder though if most of the time soft that folks like to contribute to diffraction is more a result of a slower shutter speed being used with the smaller aperture.
Reading that diffraction affects red, green and blue differently, loosing sharpness in one color would seem to be the definite indication that diffraction is culprit.
It was a very good article indeed. Goes way deeper than most, showing better conclusions. It is indeed a dumb notion to NEVER exceed f/8. There are many times greater depth of field helps MORE than diffraction hurts.... even f/40 sometimes. And sometimes not. Try things, look at your results.
It had only one weak point, at least I still have much trouble with this:
How does he manage to get all the Airy disks aligned on the pixels so well?
I think they are more likely to be on a corner than the center (four corners, only one center).
The point is, the lens aperture creates the diffraction, and it is what it is, regardless of the sensor. The best job the sensor can do is to reproduce it well.
A higher resolution sensor resolves detail better, diffraction too, but the pixel size does NOT affect diffraction. The area of the diffraction disk is a factor, but it is not about pixels.
It was a good article, and his examples did say as much, yet he shows these dumb graphics. Would have been better without them.
It had only one weak point, at least I still have much trouble with this:
How does he manage to get all the Airy disks aligned on the pixels so well?
I think they are more likely to be on a corner than the center (four corners, only one center).
The point is, the lens aperture creates the diffraction, and it is what it is, regardless of the sensor. The best job the sensor can do is to reproduce it well.
A higher resolution sensor resolves detail better, diffraction too, but the pixel size does NOT affect diffraction. The area of the diffraction disk is a factor, but it is not about pixels.
It was a good article, and his examples did say as much, yet he shows these dumb graphics. Would have been better without them.
I did some test shots on text months ago each time closing the lens further and on my D3300 it was pretty fast the text started to get fuzzy. I can't remember if it was f/8 or f/11. The D750 I could close to f/16 before the same started to occur there.
It plays little role except in macro. I no longer keep closing down even when diffraction might be a minimal issue. It's that once you check the actual DoF gain in real world sizing instead of thinking about it as "more", you realise it is hardly worth any sacrifice.
It plays little role except in macro. I no longer keep closing down even when diffraction might be a minimal issue. It's that once you check the actual DoF gain in real world sizing instead of thinking about it as "more", you realise it is hardly worth any sacrifice.
I'm saying that sacrificing anything for 1.5 instead of 1mm when that difference hardly covers the bug's eyebrows is an exercise in futility.
Not a popular notion.
See the BOTTOM two pictures (of blue ruler) at Diffraction limited images? Really?
It's an Inch ruler.
Try to identify the benefit of DOF.
Try to identify the harm of f/40 diffraction.
Last edited:
Wayne, let's quit talking about DoF as an abstract thing and put it into real world numbers.
Here's the reality:
In macro at 1:1 when I use f/8 with my cam I have 0.5mm DOF -if I remember well. (Feel free to correct me if I'm wrong)
At f/16 I double that and again double that at f/32. Thus at f/32 I have a whopping 2mm.
Now let's convert that to sensor pixels (even when my pixels don't really go into that direction). 0.5mm would be 80 pixels. 2mm would be 320 pixels. ON 6000.
That's the sad reality.
The truth is that the moment our subject in macro is large enough for us to distinguish its features with the naked eye, it is most likely too large for any DoF we can squeeze out of our cam.
DoF in normal photography is a whole different matter than DoF in macro where we talk about ranges that minuscule it is almost laughable.
DOF certainly is small at macro, but your NUMBERS are the abstract thing in this case. Founded on vague assumptions, like CoC approximations (just someone's notion), and enlargement and viewing size (variables), and even that a 105mm lens is 105mm at 1:1 (it ain't). The real world result is the important thing.
See the two pictures referenced just above.
If my numbers are wrong, you have to tell these guys: Macro Camera Lenses
I rarely have a measuring tape with me when shooting macro so I base mine on theirs. I'd say they're about right but if I'm wrong, so are they.
Last edited:
All of photography is trading one thing for another. In this case the whole purple thing acceptably sharp or one end of it a little sharper maybe. Could back up, but distance loses detail also, so another trade.
Diffraction is good to know about, but most folks don't run into it before they run into soft for some other reason.
Diffraction is good to know about, but most folks don't run into it before they run into soft for some other reason.
Of course, there is no free lunch is photography but I prefer to step back and use soft to get close again in post which also costs but it's a much cheaper lunch.
I shoot my lens at its sharpest and it provides me more options than shooting her at her weakest.
It's what many forget; if you close down, the lens drops in performance rapidly (which is partly because of diffraction).
If you take the Nikon 105mm f/2.8 as an example. On a D610 it reaches some 75% sharpness at f:5.6. A little less at f/8. Just below 70% at f/11 and it doesn't even reach 50% at f/32. That's a serious trade. Evidently for online use it matters less since scaling to those sizes adds an illusion of sharpness but the real shot did suffer plenty.
To add:
If I shoot at 1:1 f/32 or at half of that magnification at f/8, crop the shot and double in PS, do you think I'll lose 1/3th of my sharpness? If your answer is yes, you should close down.
But to each their own.
Last edited:
The results speak for themselves. The greater depth of field simply improves the picture, makes f/36 be a better picture overall (in this macro case). There is no question about it, because of course, it's been common knowledge for decades. Digital pixels don't change anything. If it were a 20mm lens at 15 feet, f/36 might not be true, but it certainly is true here.
We should try some things when we need depth of field, and compare the results. If it helps, it helps. Don't listen to the naysayers.
The OP video did not mention focal length, but an old rule of thumb I heard many years ago (about diffraction) is to stay at a fstop not exceeding focal length / 4 (for sharpest results). This basically defines a 4 mm aperture.
Ansel Adams founded his f/64 group in the 1930s, for the purpose to promote sharper pictures.
f64 x 4 = focal length 256 mm. His 8x10 view cameras typically used 325 mm to 600 mm lenses, so he was safe.
The 105mm lens at 1:1 is necessarily 210 mm... /4 = f/52.
I have more samples posted at Diffraction limited images? Really?
Last edited:
Thanks, but that was already my plan.
Lenses are not digital, and they create the diffraction. Not everything is new.