M9 diffraction limits

[ianspector]

Advertisement (gone after registration)

Monday was a great day in England, cold, very bright with a North Easterly wind. Wanting to cram as many pleasures into one day I decided to take my M9 with my Noctilux f/1 (although I still think of it as Steve Huff's Noctilux) in the Chipmunk for a flight. http://ianspector.smugmug.com/Airplanes/De-Havilland-Chipmunk/Flying/20100308ics-M91587/805716412_MkBZg-L.jpg

 

Shooting at 1000ft from Buckinghamshire to Dorset with the roof open I very soon realised that I was going to move the Noct off its normal f/1 setting. To my horror Even with the ISO at 160 some of my shots required f/8 at 1/4000. Whilst over Reading I could maintain f/4 and the shots could not have been sharper; at 100% I can see people walking in the car park. http://ianspector.smugmug.com/Airplanes/De-Havilland-Chipmunk/Flying/20100308ics-M91573/805704174_G6FAf-L.jpg

 

However as I headed further south and onto f/8 my images were very soft; even after sharpening as can be seen here http://ianspector.smugmug.com/Airplanes/De-Havilland-Chipmunk/Flying/20100308ics-M91574/805707613_m8nnx-L.jpg

 

I can only assume that given the smaller lenses diffraction is the culprit here. What are the accepted Leica lens limits before diffraction, and is it the same for all leica lenses?

 

I guess an ND filter is in my future.

 

Many thanks for your help.,

 

Ian

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[Advertisement]

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

[Guest trond]

I can only assume that given the smaller lenses diffraction is the culprit here. What are the accepted Leica lens limits before diffraction, and is it the same for all leica lenses?

 

Ian

 

Dear Ianspector,

 

Great shots, wish I cloud do that!

 

The theoretical diffraction limit only depends on the lens aperture and the resolution of the film/sensor.

 

On the M9 diffraction kicks in at around f10, if we define the Nyquist limit as the maximum resolution of the M9 sensor.

 

It may well be that you will start to see the effect of diffraction softening at f 8.

 

Best regards

 

Trond

[microview]

 

Great shots, wish I cloud do that!

 

Freudian slip perhaps.

[okram]

It is generally accepted that the aperture of 5mm is the diffraction limit, so f/10 for a Noctilux sound ok, but i would stop at f/8 if critical. At 16mm, for example, this limit would be between f/2.8 an f/4.

Diffraction exists at all apertures...

[Guest trond]

A very good explanation of diffraction limit of lenses, together with an online calculator, can be found via the link below:

 

Diffraction Limited Photography: Pixel Size, Aperture and Airy Disks

 

Best regards

 

Trond

[ianspector]

Dear Trond, Okram and Microview,

 

That is very interesting. Perhaps it was not diffraction that was causing my problem on the third shot. Having the roof open meant that I was not shooting through perspex but there was actually large amounts of glare from the sun in the haze. It took some serious S curves in Photoshop to lose the haze. I wonder if that has caused the image to lose its sharpness.

 

Perhaps I need to do some experiments under controlled conditions in the studio. Hanging an M9 out in the propwash is a bit hit or miss, especially as I have to bank at the same time to get the wing out the way. I did have a safety pilot keeping a lookout of course.

 

Thanks so much for your expertise.

 

Ian

[tdtaylor]

Advertisement (gone after registration)

Trond,

 

Thanks for the link. Very helpful. Always was aware of the effect, but not the science behind it.

[ianspector]

Yes, Trond, that link is really interesting. It also smashes a couple of my wrongly held beliefs. It seems that the focal length of the lens is irrelevant, and also the fact that my Leica lenses are physically smaller than my Nikon glass. The calculator does indicate that it probably was not diffraction at f/8, which is good to know.

 

Ian

[Leicaiste]

...To my horror Even with the ISO at 160 some of my shots required f/8 at 1/4000. Whilst over Reading I could maintain f/4 and the shots could not have been sharper; at 100%

I can only assume that given the smaller lenses diffraction is the culprit here. What are the accepted Leica lens limits before diffraction, and is it the same for all leica lenses?

...

I guess an ND filter is in my future.

...

 

Ian,

 

Did you tried the PULL 80 ISO ?

 

Lucien

[ianspector]

Dear Lucien,

 

No, it never crossed my mind to use 80 ISO as I did not anticipate any trouble. I need to go back to the RAW images and see if the detail was there before I started editing it. Certainly I would try in future if it turns out to be diffraction, but from the article in the link it seems that it would not show until closer to f/11.

 

Thanks for the suggestion,

 

Ian

[Jerry_R]

Isn't it that 'fast' lenses will meet diffraction earlier, than 'slower' counterparts?

[jaapv]

Not really. It is caused by the absolute size of the diaphragm. Smaller than (iirc) 3 mm and diffraction starts occurring.

[mjh]

Isn't it that 'fast' lenses will meet diffraction earlier, than 'slower' counterparts?

No, the lens doesn’t factor in the equation at all. This is all about the aperture. And the calculation, to a rough approximation at least, is quite simple: Multiply the aperture by 1.35 to get the size of the Airy disc in microns. For example, given the reasonable requirement that the Airy disc must be no larger than twice the pixel pitch of the M9 (which is 6.8 µm) you should not stop down below 2 x 6.8 / 1.35 = 10.

[lars_bergquist]

That technical explanation was not too realistic.

 

The usual lens aberrations (chroma, spherical etc) are always at their maximum at widest aperture. They decrease when the lens is stopped down, improving definition in the plane of best focus. (Stopping down also gives you more d.o.f., which is a completely different matter and will not be considered here.)

 

There is also diffraction, which is ALWAYS there. The idea that diffraction 'cuts in' at some definite diameter of the aperture -- 3mm or 5mm -- is nonsensical. As you stop down, an increasing fraction of the light is passing near the edges of the diaphragm blades, increasing diffraction and thereby deterioration of definition. If the 'classical' aberrations had not existed, lenses would be sharper (n.b. in the plane of best focus) wide open than stopped down. When I ran a wet darkroom, I sometimes softened rough grain by stopping the El-Nikkor down to f:11.

 

So when stopping down, we have a rising curve of definition -- due to reduction of aberrations -- and a sinking curve due to increasing diffraction. Please note that this occurs no matter what sensor or film you use, or even none. I could see it by eyeball on the enlarger baseboard! Where those curves intersect, the lens performs at its best (at best focus, natch).

 

Where is that point? MTF curves can tell us that, or would, if they were published for all f-stops. But in general, very fast lenses (Summilux, Noctilux) are best at about f:4, while Summicron lenses have their optimum at 5.6. After that, general contrast starts to drop. But the effect of one further stop is usually negligible.

 

These are modern, highly corrected lenses. Older lenses often had so much aberration that no matter how far you stopped down, the improvement in this department was large enough to make the matter of diffraction moot. Hence Ansel and Edward and their "f:64 group". And increased d.o.f. can of course often make for a visual impression of greater overall sharpness. But for best results at the plane of best focus, I would normally never exceed f:8.

 

The old man from the Age of Ansel

[Guest trond]

Dear Lars,

 

I agree with you, diffraction is always there!

 

I work with integrated circuit design and layout with pattern resolution of 20 000 lp/mm and beyound, we fight diffraction every day.

 

You should see those lenses, not cheap stuff!

 

My point is that diffraction is not visible on the M9 until f10 or so, because the softness due to diffraction is smaller than the pixel size.

 

Best regards

 

Trond

[okram]

Dear Jaap, it is around 5mm, not 3mm- if it would be 3mm, then You could stop 50mm to f/16 without diffraction.

 

 

Trond, f/10 at 50mm. At 35mm there is more diffraction at f/10, and so on.

 

1/f is a relative aperture.

 

This is, along the enlargement factor, why we use bigger formats.

[mjh]

Practical considerations about diffraction effects would depend on why one would stop down at all. If one stops down to keep lens aberrations at bay, one would want find the sweet spot between unsharpness because of lens aberrations on the one side and unsharpness because of diffraction on the other. The position of this sweet spot would indeed depend on the lens. But if you don’t really worry about your lens but stop down to increase the depth of field, or simply because you are already at the hightest shutter speed and just have to stop down to achieve correct exposure, then there are no relevant sweet spots. It becomes a question of how far you can dare to go before diffraction effects will compromise resolution, which in turn depends on the resolution of the sensor (or film, for that matter).

[mjh]

Trond, f/10 at 50mm. At 35mm there is more diffraction at f/10, and so on.

The size of the Airy disc doesn’t depend on the focal length, so the f/10 limit is true for all lenses. Or rather, it does depend on the focal length, but so does the relative aperture, so we already got that covered. Put differently, diffraction across all focal lengths is constant for a given relative aperture, but not for a given aperture diameter. So the 5 mm limit does indeed apply to 50 mm only; for 35 mm it would be 3.5 mm and so on.

[Advertisement]

Hallo,

Look here: Leica M

[Guest trond]

Dear Jaap, it is around 5mm, not 3mm- if it would be 3mm, then You could stop 50mm to f/16 without diffraction.

 

 

Trond, f/10 at 50mm. At 35mm there is more diffraction at f/10, and so on.

 

1/f is a relative aperture.

 

This is, along the enlargement factor, why we use bigger formats.

 

Dear Okram,

 

Diffraction only depends on the aperture and the wavelength of the light, not on the focal length, and not on the format.

 

It is less visible on larger formats due to the lower magnification necessary to produce a print of a given size.

 

Best regards

 

Trond

[Jerry_R]

No, the lens doesn’t factor in the equation at all. This is all about the aperture.

But in general, very fast lenses (Summilux, Noctilux) are best at about f:4, while Summicron lenses have their optimum at 5.6.

So how it really is?

I have read on ReidReviews that faster lenses become softer sooner due to diffraction, than slower lenses.

Lars is also mentioning that.

 

I have heard elsewhere too, that slow\fast doesn't matter when we talk about diffraction.

But no-one neglects, that in most cases faster lenses become softer sooner, than slower lenses...

 

So probably this is a fact, but root cause is not diffraction, but lens construction?