Diffraction f-stop limits of IQ4150, IQ3100 and H6D-100c backs

[Smoothjazz]

Robert,

Very well stated- a good summary!

I tend to agree at f11 being my limiting f-stop at which I try not to go beyond.
The question in some cases is should I keep my f-stop at f8, which is ideal, and focus stack a few images, or stop down to f11 and beyond and take one image.

Cheers,

John

[Aram Hăvărneanu]

Besides, when you spend $10,000 to go to that awesome location to shoot are you going to rely on math or known data based on YOUR gear??

Ideally I would rely on math. Others might do differently, but not me.

[Shrev94412]

Robert,

Very well stated- a good summary!

I tend to agree at f11 being my limiting f-stop at which I try not to go beyond.
The question in some cases is should I keep my f-stop at f8, which is ideal, and focus stack a few images, or stop down to f11 and beyond and take one image.

Cheers,

John

For me that would depend on the situation. I always try and keep my ISO as low as possible to maximize dynamic range and limit noise. The other factor is if anything in the scene is moving, hence higher shutter speed, or if I am dragging the shutter to blur some water. In Landscapes, I just do what I have to do to make the Exposure triangle come together, staying within f/5.6 to f/11 and then If I have to focus stack then so be it. The Phase XF makes focus stacking VERY easy. If using another system I typically focus stack manually.

Just my craft....does not mean it’s the best way. I will say critical focus throughout the image is extremely important to me as when you print it up 90 inches (228cm) everything bad tends to be obvious.

[BJL]

Ideally I would rely on math. Others might do differently, but not me.

As a mathematician by profession, I trust the math but often not the interpretations that people put on it. (That is almost my professional motto.) For example, some basic math and physics can easily tell us the aperture ratio at which the diameter of the circle where the first null of the diffraction pattern occurs (twice the "Airy radius") matches the spacing of the photosites — but that is not a sudden wall between "diffraction" and "no diffraction". There is always at least some slight diffraction smearing (some light goes beyond the Airy disk; I tiny bit goes even far beyond) and there is still rather little at that threshold, with the amount increasing continuously as aperture ratio increases, so that the diffraction factor in MTF at any spatial frequency decreases continuously with aperture ratio. The practical question is when is there too much diffraction, with the criteria for being "too much" including possible trade-offs against DOF, longer exposure time needs, worse aberrations at lower aperture ratios, etc.

I doubt there is a mathematically proven formula which accounts for all of that, so we also need to look at some photographs.

[ErikKaffehr]

Hi,

Jim Kasson has made a very good series of images showing the effects of diffraction: https://blog.kasson.com/gfx-100/a-visual-look-at-gfx-100-diffraction-blur/

Here is a combination of two images f/11 on the left and f/5.6 on the right:

Full size image.

It was shot on the GFX 100, which has the same pixel characteristics at the IQ4150 with Fujifilms best lens the GF 110/2. That lens performs best at about f/2.8 to f/4 on axis.

The f/11 image is softer than the f/5.6 image, but it still contains all fine detail. So full detail can probably reconstructed with sharpening. With the same sharpening f/5.6 will be visually sharper.

We don't need to go to extreme lens designs or to very small pixels to see the effects of diffraction.


The plots here show MTF data for the Sonnar 180/4 for f/11, f8 and f/5.6. To my big surprise the Sonnar 180/4CFi performs best at f/5.6. So what this illustrates is really that diffraction reduces sharpness even on 6.8 micron cameras.

A side note is that I would normally shoot f/11 on the Hasselblad/P45+ combo I have and f/8 on my Sony A7rII. Comparisons I have done at those apertures used to show that the two systems are quite similar.

I have included the A7rII with my 90/2.8 G macro in those plots, probably at f/5.6. Using f/11 on the Hasselblad/P45+ combo may be a pretty decent explanation why  I do not see better results on MFD than on the Sony A7rII.

Real life is not that simple. The 180/4 Sonnar is the best lens I have. Also, the images here were chosen from 40 exposures shot with a focusing rail. In real world images, focusing and needed DoF may play a major role than diffraction.

An interesting point with Jim's images that the f/5.6 image has a lot of aliasing artifacts. The Siemens star is very prone to those artifacts, on normal subjects the artifacts are still there but they are far less obvious.

So, what is my take from this:

• Buying the best lens money can buy, it is wise to use it at optimal aperture to get what you have paid for. That aperture will be pretty wide.
• It is quite possible to stop down significantly, without actually loosing detail.
• Stopping down may eliminate some sampling artifacts.

The best way to see this is essentially considering MTF.

Now, all lenses made on this planet have in common that they are affected by diffraction and that diffraction is dependent of f-stop only.

But, it can be a useful idea to split diffraction out of MTF for the lens.

We can write the MTF of a system as:

MTFsystem = MTFdiffraction * MTFlens * MTFsensor

So, MTFdiffraction drops essentially with f/stop

• MTFdiffraction drops essentially with f/stop
• MTFlens is limited by residual aberrations and most aberrations decrease stopping down. So, MTFlens is increasing stopping down.
• MTFsensor is just a function of the active area of the pixel. Smaller pixel opening yields higher MTFsensor.

So, if any factor goes zero MTF will be zero. But as long as all factors are > 0 each factor will contribute to sharpness.

If we take a sensor with say 3.8 micron pixels, MTFdiffraction == 0 will probably occur about f/16. At larger apertures we can probably compensate with sharpening. But sharpening always comes at a cost...

Just to say, these things have been well know back in the 1950-es. The reason we discuss it that much today is that we can blow up images to incredible sizes by just clicking a button in Lightroom and the absence of film grain.
Best regards
Erik

[Smoothjazz]

I agree with you Shrev; every time I test my lenses carefully in the field, the optimum aperture is usually right around f8 or f8.5- which is my preferred aperture. At f11 a little bit of diffraction will start to be evident. Focus stacking at f8.5 is my usual approach to landscape photography.