Diffraction limiting

Started 2 months ago | Discussions
Ad12 Contributing Member • Posts: 539
Diffraction limiting

Hello!

On imaging resources, I noted my new XF 18-55 says diffraction limiting starts at about f16. My old lens for my Canon the sigma 17-50 they say it begins at f8. I must admit, the lens was usually super sharp, but did definitely soften up after about f10 in my experience quite a lot.

I always mistakenly thought diffraction limiting was kind of similar for lenses at a certain sensor size. this suggests not.

So is it the lens that really dictates this? Not interested in pixel by pixel worrying, just curious!

 Ad12's gear list:Ad12's gear list
Canon EOS Rebel T7i Fujifilm X-T30 Sigma 105mm F2.8 EX DG OS HSM Fujifilm XF 18-55mm F2.8-4 R LM OIS Canon EF-S 55-250mm f/4-5.6 IS STM +2 more
Lee Jay Forum Pro • Posts: 56,096
Depends on the definition
4

Diffraction is always a limit and it's always present.

"Diffraction limiting" could mean many things.  One way to define it would be when diffraction effects and aberrations are both reducing sharpness by the same amount.  In this case, a sharper lens would have a lower f-stop for "diffraction limiting".

-- hide signature --

Lee Jay

 Lee Jay's gear list:Lee Jay's gear list
Canon PowerShot SX50 HS Canon EOS 7D Mark II Canon EF 70-200mm F2.8L IS II USM Canon EF 8-15mm f/4L Fisheye USM Canon Extender EF 2x III +10 more
Great Bustard Forum Pro • Posts: 44,677
Diffraction...
6

Ad12 wrote:

Hello!

On imaging resources, I noted my new XF 18-55 says diffraction limiting starts at about f16. My old lens for my Canon the sigma 17-50 they say it begins at f8. I must admit, the lens was usually super sharp, but did definitely soften up after about f10 in my experience quite a lot.

I always mistakenly thought diffraction limiting was kind of similar for lenses at a certain sensor size. this suggests not.

So is it the lens that really dictates this? Not interested in pixel by pixel worrying, just curious!

...goes like this:

  • Diffraction softening affects *all* systems equally at the same DOF.
  • Diffraction exists right from wide open and increases as you stop down.
  • Lens aberrations exist right from wide open and lessen as you stop down.
  • The relative aperture at which diffraction softening dominates over lessening lens aberrations is often called "the diffraction limited aperture".
  • The "diffraction limited aperture" is only moderately a function of pixel size, where smaller pixels, all else equal, will achieve the "diffraction limited aperture" at a lower f-number. However, more smaller pixels allows you to see the effects of diffraction earlier, because the effect of diffraction is hidden in the *greater* blur of fewer larger pixels. Thus, more smaller pixels (all else equal) will *always* resolve more, stop-for-stop, than fewer larger pixels.
  • The "diffraction limited aperture" is, however, strongly a function of the particular lens, because the lens aberrations vary tremendously from lens to lens whereas diffraction is the same.
  • Typically, the "diffraction limited aperture" will be a stop later at the edges than the center, because the edges have more aberrations than the center.
  • Diffraction is often the least of your problems when stopping down "too far" -- stopping down will result in less light on the sensor (for a given scene and exposure time) resulting in a more noisy photo or may result in excessive motion blur (for a given scene and exposure), either of which may well be more damaging to the IQ of the photo than diffraction itself.

Thus, the "advice" that was given:

I noted my new XF 18-55 says diffraction limiting starts at about f16. My old lens for my Canon the sigma 17-50 they say it begins at f8.

is just plain wrong.

John Sheehy Forum Pro • Posts: 23,671
Re: Diffraction...

Great Bustard wrote:

  • Diffraction is often the least of your problems when stopping down "too far" -- stopping down will result in less light on the sensor (for a given scene and exposure time) resulting in a more noisy photo or may result in excessive motion blur (for a given scene and exposure), either of which may well be more damaging to the IQ of the photo than diffraction itself.

Also, you may get much more visible sensor dirt if you stop down more than you really need to.

Great Bustard Forum Pro • Posts: 44,677
Re: Diffraction...

John Sheehy wrote:

Great Bustard wrote:

  • Diffraction is often the least of your problems when stopping down "too far" -- stopping down will result in less light on the sensor (for a given scene and exposure time) resulting in a more noisy photo or may result in excessive motion blur (for a given scene and exposure), either of which may well be more damaging to the IQ of the photo than diffraction itself.

Also, you may get much more visible sensor dirt if you stop down more than you really need to.

By "sensor dirt", do you mean like which sensor is cheating on which sensor, and stuff like that?

MHshooter
MHshooter Contributing Member • Posts: 602
Re: Diffraction limiting

True diffraction limiting should begin with the lens near wide open, but very few camera lenses can claim it.

John Sheehy Forum Pro • Posts: 23,671
Re: Diffraction limiting
1

MHshooter wrote:

True diffraction limiting should begin with the lens near wide open, but very few camera lenses can claim it.

Unfortunately, many people hear the term "diffraction limited" and think of something negative. I often think after posting a phrase like "an f/2.8 lens that is diffraction-limited at f/2.8", that some people think that I'm saying that the lens is not sharp at f/2.8, when in fact, it is sharpest at f/2.8, and an extremely sharp lens.

Even as an attenuater of MTF, diffraction is not a hard wall, at all; it maintains high image frequencies at reduced contrast, rather than obliterating them at moderately low frequencies, as some people seem to assume. It also helps a lot with avoiding color moire. Even a sensor with what is considered a strong AA filter still gives color moire in the absence of diffraction or significant aberrations. An AA filter that had a large enough pattern of 4 points to anti-alias the red or blue channels would turn a point of light into 4 separate points in the luminance channel. Alternate AA patterns could prevent that, but then you are attenuating the finest luminance detail quite a bit, which would require more sharpening.

Many photographers who know just enough to engage in tech talk seem to be trapped in a paradigm where the target for quality imaging happens at the pixel level, and they feel like they are being hopelessly trapped by diffraction (as well as noise) as they upgrade to sensors with higher pixel densities, when the fact is that diffraction is immutable, and the higher densities just mean a more analog-like capture of what is always there, with more detail; not less, and more artifact-free detail that is more rugged in the face of any subsequent resampling.

Expecting pixel-level views to maintain sharpness and apparent noise-less-ness as pixel densities increase is not reasonable. People who want maximum sharpness should be looking towards sharper lenses with minimal aberration at low f-numbers, better focus, and better stability; not clutching onto lower pixel densities for their more pleasing, illusory 100% pixel views.

a_c_skinner Veteran Member • Posts: 9,701
Re: Diffraction limiting

"I always mistakenly thought diffraction limiting was kind of similar for lenses at a certain sensor size. this suggests not."

Actually it is pixel density plus sensor size. Within that it is determined solely by the aperture as in f-number. Lens makes no difference.

Within the normal aperture range you can ignore diffraction, it produces a very slight softness that yields to unsharp mask easily, though it does mean you shouldn't stop down beyond mid aperture unless you need the depth of field.

The real lesson here is that anyone can put poor advice on the internet with no editorial process, I know I do.

-- hide signature --

Andrew Skinner

 a_c_skinner's gear list:a_c_skinner's gear list
Fujifilm X-T2 Fujifilm X-E3 Fujifilm XF 35mm F1.4 R Fujifilm XF 14mm F2.8 R Fujifilm XF 55-200mm F3.5-4.8 R LM OIS +5 more
All Bushs Fault Regular Member • Posts: 156
I agree wholeheartedly with both of you

although I shudder to think of the cost (Zeiss Otus comes to mind).

 All Bushs Fault's gear list:All Bushs Fault's gear list
Canon EOS 5DS R Sony a7R IV Canon EF 50mm F1.8 II Canon EF 100mm F2.8L Macro IS USM Sony FE 200-600 F5.6-6.3
John Sheehy Forum Pro • Posts: 23,671
Re: Diffraction...

Great Bustard wrote:

John Sheehy wrote:

Great Bustard wrote:

  • Diffraction is often the least of your problems when stopping down "too far" -- stopping down will result in less light on the sensor (for a given scene and exposure time) resulting in a more noisy photo or may result in excessive motion blur (for a given scene and exposure), either of which may well be more damaging to the IQ of the photo than diffraction itself.

Also, you may get much more visible sensor dirt if you stop down more than you really need to.

By "sensor dirt", do you mean like which sensor is cheating on which sensor, and stuff like that?

Without mentioning names or hinting at particular individuals, I'd say yes.

It is said that you can tame the immoral behavior of dual-pixel sensors by dissolving the microlenses, which is a virtual lobotomy for each pixel.

MHshooter
MHshooter Contributing Member • Posts: 602
Re: I agree wholeheartedly with both of you

All Bushs Fault wrote:

although I shudder to think of the cost (Zeiss Otus comes to mind).

Leica 50mm f/2.0 Apo-Summicron.

Keyboard shortcuts:
FForum MMy threads