Diffraction Test

[giannis]

Advertisement (gone after registration)

Just some clarifications that could be useful:

1. Diffraction is an optical property that depends solely on the aperture (f/number). The wider the aperture, the less the diffraction. You can think of diffraction as the maximum theoretical "ceiling" a lens could achieve at that aperture if it were a perfect lens. As the aperture gets narrower, the same perfect lens has a lower resolution limit. This is an absolute property and a wider aperture always implies higher resolution potential.
2. However lenses are not perfect. While the resolution potential decreases as you stop down, the "gains" in resolution due to reduced spherical (and other) aberrations outweigh the "losses" due to diffraction. There will be a point though, at some aperture, that stopping down further from it, the situation reverses and the gains in resolution due to aberration reduction will be less than the losses due to diffraction. This point is the diffraction limit of the lens. It's the point where, that particular lens, is at its best, resolution-wise. This is a purely optical property too that depends solely on lens design and nothing else (like sensor etc.).
3. You use that lens on a sensor. Say that particular lens has a diffraction limit at f/8. This means resolution at f/5.6 and f/11 will be a tiny bit less than at f/8. Then, depending on your sensor's pixel density, the sensor might not be able to pick up the difference between f/5.6, f/8 and f/11: for all intents and purposes, the sensor picks virtually the same resolution at those three apertures, despite the lens "providing" a bit more at f/8. This doesn't mean the diffraction limit of the lens changed, it's still at f/8. It means though, with that sensor, you get the same res at f/5.6, f/8 and f/11. Then you could use f/11 as the definitive point after which the system resolution gets worse (for practical purposes). You could call this the "system diffraction limit" if you will. This will depend both on the sensor's pixel density and on the len's diffraction limit.

Of course, all that is a fancy way of saying that if you care about the system resolution the only way to find out is to use that particular lens on that particular sensor and shoot a resolution target across the aperture range, then determining the best aperture for the system. Everything else is a rough approximation at best, misleading at worst. There are no absolutes, since each lens will have its own diffraction limit (based on its optical formula) and each sensor its own pixel density, so when  combined each system will have its own diffraction limit.

That said, all that is a bit silly to worry about in practice, especially in 35mm sensor sizes. Firstly, because no such small differences are gonna make or break an image, 135 format lenses don't stop down to ridiculously small apertures so as to actually make substantial difference for pictorial use. Secondarily, you don't even need to stop all the way down for DoF, cause there's plenty of DoF in 135 format even at moderate apertures. And finally, if you indeed need that extreme DoF, there's no other way to get it anyway, so you have to stop down and let it be what it may. There's no dilemma like in, say 4x5": "I need the DoF but f/90 will be impacted by diffraction, should I use f/90 anyway or some camera movements and f/45?". Well in 35mm no such option (usually) exists, so if you need the DoF just shoot at the aperture needed and stop worrying  

Edited March 17, 2022 by giannis
grammar

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[Dennis]

5 hours ago, giannis said:

Just some clarifications that could be useful:

1. Diffraction is an optical property that depends solely on the aperture (f/number). The wider the aperture, the less the diffraction. You can think of diffraction as the maximum theoretical "ceiling" a lens could achieve at that aperture if it were a perfect lens. As the aperture gets narrower, the same perfect lens has a lower resolution limit. This is an absolute property and a wider aperture always implies higher resolution potential.
2. However lenses are not perfect. While the resolution potential decreases as you stop down, the "gains" in resolution due to reduced spherical (and other) aberrations outweigh the "losses" due to diffraction. There will be a point though, at some aperture, that stopping down further from it, the situation reverses and the gains in resolution due to aberration reduction will be less than the losses due to diffraction. This point is the diffraction limit of the lens. It's the point where, that particular lens, is at its best, resolution-wise. This is a purely optical property too that depends solely on lens design and nothing else (like sensor etc.).
3. You use that lens on a sensor. Say that particular lens has a diffraction limit at f/8. This means resolution at f/5.6 and f/11 will be a tiny bit less than at f/8. Then, depending on your sensor's pixel density, the sensor might not be able to pick up the difference between f/5.6, f/8 and f/11: for all intents and purposes, the sensor picks virtually the same resolution at those three apertures, despite the lens "providing" a bit more at f/8. This doesn't mean the diffraction limit of the lens changed, it's still at f/8. It means though, with that sensor, you get the same res at f/5.6, f/8 and f/11. Then you could use f/11 as the definitive point after which the system resolution gets worse (for practical purposes). You could call this the "system diffraction limit" if you will. This will depend both on the sensor's pixel density and on the len's diffraction limit.

Of course, all that is a fancy way of saying that if you care about the system resolution the only way to find out is to use that particular lens on that particular sensor and shoot a resolution target across the aperture range, then determining the best aperture for the system. Everything else is a rough approximation at best, misleading at worst. There are no absolutes, since each lens will have its own diffraction limit (based on its optical formula) and each sensor its own pixel density, so when  combined each system will have its own diffraction limit.

That said, all that is a bit silly to worry about in practice, especially in 35mm sensor sizes. Firstly, because no such small differences are gonna make or break an image, 135 format lenses don't stop down to ridiculously small apertures so as to actually make substantial difference for pictorial use. Secondarily, you don't even need to stop all the way down for DoF, cause there's plenty of DoF in 135 format even at moderate apertures. And finally, if you indeed need that extreme DoF, there's no other way to get it anyway, so you have to stop down and let it be what it may. There's no dilemma like in, say 4x5": "I need the DoF but f/90 will be impacted by diffraction, should I use f/90 anyway or some camera movements and f/45?". Well in 35mm no such option (usually) exists, so if you need the DoF just shoot at the aperture needed and stop worrying  

Very interesting and informative explanation, thank you for sharing.

My short conclusion for my M10: shoot at f8 (my diffraction limit) if it's enough. Do I need more DoF? Well, shoot at f11 and forget about the limit  

[pippy]

10 hours ago, giannis said:

...all that is a bit silly to worry about in practice......because no such small differences are gonna make or break an image......and finally...if you indeed need that extreme DoF, there's no other way to get it anyway, so you have to stop down and let it be what it may....if you need the DoF just shoot at the aperture needed and stop worrying...

Absolutely!

Philip.