Nikon D850 - Ten Technical Questions

Started 5 months ago | Discussions thread
ChiArch New Member • Posts: 16
Re: Nikon D850 - Ten Technical Questions

Tony Beach wrote:

ChiArch wrote:

Tony Beach wrote:

ChiArch wrote:

Tony Beach wrote:

praxxim wrote:

8. What is the diffraction limited aperture?

There is no such thing as "diffraction limited." I use the term diffraction impacted because that more accurately describes what happens as diffraction increasingly impacts acuity. It's not as if there is a brick wall limit to how much resolution you have or not as you stop down the lens.

There is a physical limit.

Really? Show it. Here's what I've seen:

Look at here

That's why it is called diffraction-limited.

Then if we add twice as many pixels per area then you would not see any more resolution because it has been limited by diffraction, but I'm not seeing that. Instead what happens is that as diffraction increases resolution drops off for a given lens and sensor, if the sensor and/or lens resolves more then using a reasonable aperture (most 135 format lenses only go to f/22) you will see more detail in spite of it being impacted by diffraction.

You need to compare the same sensor below and above diffraction-limit aperture (DLA) to see the resolution difference, not two different sensors at two different apertures.

No, you are missing the principle I am describing. I understand the principle that with a given lens on a given sensor there is a peak aperture where stopping down any more causes a loss of resolution, but if you take a higher resolution lens and/or camera and stop it down more than a lower resolution lens and/or camera (even significantly more if there is a significant difference in resolution between the two) then you will still have more resolution on the higher lens and/or camera -- just as I showed in my example above.

Suppose there is a lens with ideal optical performance and a 1000px X 1000px sensor. When the F-number is smaller than DLA, the angular resolution recorded in the final image is limited by the pixel resolution of the sensor. That is the image on the focal plane is sharper than what the sensor can record.

Then the limitation is the sensor -- as it is in the example above with the D70.

When the F-number is larger than DLA, the angular resolution recorded in the final image is limited by diffraction. That is the image on the focal plane is soft (has a blurry edge), so the higher pixel resolution is unexploited. You still get a 1000x1000 file, but the image is soft.

Another way to think of it is that you may want to put the maximum amount of resolved data in the photo. You can choose between having a narrow DOF that is resolved up to the maximum the lens and camera sensor will allow, then you might have less in the photo that is acceptably resolved (for instance, nose and ears may be out-of-focus when you want them in-focus), so you stop down thereby increasing the total amount of resolved data even as that data becomes less resolved, and at some point you trade off too much optimally resolved data for what you gain in total acceptably resolved data.

The thing is diffraction is not a cliff, it's a slope. If you place that slope on a higher mountain then even as you go farther down the slope you are still at a higher elevation than you were on the lower mountain.

For example, at Photozone we can see that the Canon 35/1.4 USM L attains maximum resolution on the 5D Mark II at f/2.8 with an MTF 50 of 3869 LW/PH, and thereafter it is increasingly diffraction impacted. Also at Photozone we can see the Canon 35/1.4 USM L II attains maximum resolution on the 5DSR at f/2.8 with an MTF 50 of 5568 LW/PH, thereafter it is increasingly impacted so that by the time it is at f/11 it is "only" at 4135 LW/PH, but that is still resolving more than the other 35mm lens was capable of resolving at any aperture on the lower MP sensor -- again, same as what I showed in my example above.

For the links to those graphs, see:

Try a pinhole lens, then you will see.

Sure, at f/128 there is likely to be a diffraction limit, perhaps sooner, but as a practical matter that is not an issue for modern 135/FX format DSLRs. The implications for many photographers is that they are starting to think that if they aren't shooting at f/5.6 on the D850 then they might as well be using a D750 because at f/8 it isn't diffraction "limited," and that's just not true.

I didn't say DLA is a cliff. If you plot the resolution at different f-numbers of an ideal lens, you get a downward slope. If you factor the sensor resolution in, you get a roof line. This is the limit.

The resolution of the real world lenses get better when reducing the aperture is only because of they are far from ideal. That is they are below the roof line. Even the center of Canon EF 35mm f/1.4 USM L II (f2.8) on 5DsR is about 300 LW/PH lower than the roof line. You are not hitting the limit it does not make the limit false.

Whether you want/should go beyond DLA when shooting a picture is irrelevant to DLA itself. Of course, choosing a camera base on DLA is nonsense.

 ChiArch's gear list:ChiArch's gear list
Panasonic Lumix DMC-LX5 Nikon D700 Nikon AF-S Nikkor 16-35mm F4G ED VR Nikon AF-S Nikkor 24-70mm f/2.8G ED Nikon AF-S Nikkor 70-200mm f/2.8G ED VR II +2 more
Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow