Nerd Alert: The Lens Diaphragm and its Influence on Optics

[AntPDC]

Hoping one of you gurus can enlighten me on this issue.

Many of you will know that a debate has been going on for some time about the use of "full frame" lenses on crop sensor bodies. One of the arguments against is that the extra "magnification", ie cropping, might reveal previously unnoticed imperfections in full frame lenses' glass.

A counter-argument to this is that, if the "imperfections" idea were true, then stopping down the diaphragm restricts a lens' total output to an even smaller portion of its front glass elements, resulting in far greater degradation (if we were to accept the glass "imperfections" theory in the FF v crop sensor scenario).

With regard to the diaphragm itself, by blocking a substantial proportion of off-axis photons does it thereby limit a lens' output largely to its central area? Or does it allow photons from the entire front lens elements to reach the sensor in amounts proportional to the front elements' centre-edge distance? Obviously, the diaphragm affects exposure according to its f-stop setting, but I'm keen to understand more about how it works with respect to how it blocks light, and from which part of the lens it is blocking.

Hope that makes some sense. I realise this is an esoteric issue, and despite my research I've been unable to find any useful information on the internet to satisfy my nerdy curiosity. Even a link to an expert source would be appreciated.

[xpatUSA]

Hoping one of you gurus can enlighten me on this issue.

Many of you will know that a debate has been going on for some time about the use of "full frame" lenses on crop sensor bodies. One of the arguments against is that the extra "magnification", ie cropping, might reveal previously unnoticed imperfections in full frame lenses' glass.

A counter-argument to this is that, if the "imperfections" idea were true, then stopping down the diaphragm restricts a lens' total output to an even smaller portion of its front glass elements, resulting in far greater degradation (if we were to accept the glass "imperfections" theory in the FF v crop sensor scenario).

With regard to the diaphragm itself, by blocking a substantial proportion of off-axis photons does it thereby limit a lens' output largely to its central area? Or does it allow photons from the entire front lens elements to reach the sensor in amounts proportional to the front elements' centre-edge distance? Obviously, the diaphragm affects exposure according to its f-stop setting, but I'm keen to understand more about how it works with respect to how it blocks light, and from which part of the lens it is blocking.

Hope that makes some sense. I realise this is an esoteric issue, and despite my research I've been unable to find any useful information on the internet to satisfy my nerdy curiosity. Even a link to an expert source would be appreciated.

Welcome to the Forum, @AntPDC (Tony?). Whilst being a bit nerdy myself I would advise you to go here, join up and ask again. Look especially for replies from Jim Kasson or Jack Hogan, they know their stuff. Your question is a common one - but at far greater depth than is commonly addressed here. Our Mod knows a bit and one or two others but opinion here is likely to exceed fact.

Roger Cicala at LensRentals knows a bit too.

[AntPDC]

Thank you Ted. My experience on DpReview has been mixed. I'm a member, so I posted there too. I was under the impression that CiC was more evidence-based, which is why I posted here. And as you say, Roger Cicala is an expert and trustworthy: I've read him before and been impressed by his expertise.

Best wishes.

[Manfred M]

Let me try to give you the "simple" answer:

1. Faster lenses (small aperture numbers usually f/2.8 or less) will have a relatively large entrance pupil, which means light at the edges, when shot wide open (or stopped down a bit), have to travel through a lot more glass than light that is collected near the centre of the lens.

If you remember basic high school physics; different wavelengths of light change speed as they travel from one material to another (air to glass, for instance) and the light is refracted. This is the principle behind a prism and different wavelengths of light exit at different places. A lens designer wants all wavelengths of light to converge at the same point on the sensor, so we end up with fairly complex lens designs, especially with fast lenses. No lens is perfect, so we do get small errors as we get further away from the centre of the lens. Ultimately, this impacts the quality of the image.

Stopping down means light is collected more towards the centre of the lens and these issues are minimized.

2. Countering this is that light is bent (diffracted) when it hits the diaphram blades, so we get some scattering. The smaller the aperture, the more pronounced this effect is. Lens diffraction is just as noticeable as softening from a fast lens. I generally shoot expensive "pro" glass on a heavy duty tripod. I do know what a sharp image looks like.

Generally we write about a lenses "sweet spot"; i.e. where refraction artifacts are minimized and diffraction impacts are minimized. This tends to be around 2 or 3 stops from fully open and once we get smaller than around f/11, diffraction can become noticeable.

That being said, let's look at the "real world" as camera or subject motion, focus issues, etc. can quickly override these lens issues. A good, sturdy tripod and care in focusing works well, but that is not the way that most people shoot.


I have shot full-frame lenses on crop frame cameras and have gotten excellent results. That being said, i have not tested the quality of resolution using test targets to see which one is technically better. In real world photography, technical quality (i.e. lens sharpness) is one of many factors that I consider. In some types of photography, sharpness is not a significant factor, in others, it matters a lot.

Unless you are someone who makes large prints, then it is very much an academic exercise. Most images posted on the internet are down-sampled to around 2MP; there is so much detail and subtlety lost that lens resolution is unimportant. If you make large prints (like I do; typically 17" x 22" / A2 format), then you can tell, but not necessarily if one looks at a normal viewing distance (2x - 3x the print diagonal). In fact, I find that I will notice digital noise more than lens sharpness once we get to even moderate ISO values. I can tell when I am printing a crop factor or full frame lens, but I am looking at full print resolution on my computer screen when I do so.

I hope that this helps.

[DanK]

I'm hesitant to post, as I suspect that once I finally understand your question, I won't know enough to answer it, but I don't understand your question.

You started with this:

Many of you will know that a debate has been going on for some time about the use of "full frame" lenses on crop sensor bodies. One of the arguments against is that the extra "magnification", ie cropping, might reveal previously unnoticed imperfections in full frame lenses' glass.

This has nothing to do with aperture. Assuming you stand the same distance from the subject and use the same focal lenth, when you put a FF lens on a crop camera, all that happens is that a good bit of the image circle doesn't fall onto the smaller sensor. The performance of the part of the lens that does cast an image on the sensor is identical to its performance on a FF camera. There is no additional magnification. If the two sensors have the same pixel density, you would get essentially the identical image by shooting on the FF camera and cropping the resulting file to the size of the crop-sensor image.

The only place that magnification enters into this is if you frame the image differently. If you frame a subject to fill the frame on both cameras, you need less magnification to print the FF image. However, that will be a different image, with different perspective.

There is a good explanation of diffraction in the tutorials section of this website, in case you didn't look there: https://www.cambridgeincolour.com/tu...hotography.htm. However, there has been a good bit of disagreement on this forum over the years about how much practical difference it makes. As Manfred pointed out, other factors can swamp the effect of diffraction.

As a practical matter, for the work I do, I don't find that diffraction usually matters.

[AntPDC]

Let me try to give you the "simple" answer:

1. Faster lenses (small aperture numbers usually f/2.8 or less) will have a relatively large entrance pupil, which means light at the edges, when shot wide open (or stopped down a bit), have to travel through a lot more glass than light that is collected near the centre of the lens.

If you remember basic high school physics; different wavelengths of light change speed as they travel from one material to another (air to glass, for instance) and the light is refracted. This is the principle behind a prism and different wavelengths of light exit at different places. A lens designer wants all wavelengths of light to converge at the same point on the sensor, so we end up with fairly complex lens designs, especially with fast lenses. No lens is perfect, so we do get small errors as we get further away from the centre of the lens. Ultimately, this impacts the quality of the image.

Stopping down means light is collected more towards the centre of the lens and these issues are minimized.

2. Countering this is that light is bent (diffracted) when it hits the diaphram blades, so we get some scattering. The smaller the aperture, the more pronounced this effect is. Lens diffraction is just as noticeable as softening from a fast lens. I generally shoot expensive "pro" glass on a heavy duty tripod. I do know what a sharp image looks like.

Generally we write about a lenses "sweet spot"; i.e. where refraction artifacts are minimized and diffraction impacts are minimized. This tends to be around 2 or 3 stops from fully open and once we get smaller than around f/11, diffraction can become noticeable.

That being said, let's look at the "real world" as camera or subject motion, focus issues, etc. can quickly override these lens issues. A good, sturdy tripod and care in focusing works well, but that is not the way that most people shoot.


I have shot full-frame lenses on crop frame cameras and have gotten excellent results. That being said, i have not tested the quality of resolution using test targets to see which one is technically better. In real world photography, technical quality (i.e. lens sharpness) is one of many factors that I consider. In some types of photography, sharpness is not a significant factor, in others, it matters a lot.

Unless you are someone who makes large prints, then it is very much an academic exercise. Most images posted on the internet are down-sampled to around 2MP; there is so much detail and subtlety lost that lens resolution is unimportant. If you make large prints (like I do; typically 17" x 22" / A2 format), then you can tell, but not necessarily if one looks at a normal viewing distance (2x - 3x the print diagonal). In fact, I find that I will notice digital noise more than lens sharpness once we get to even moderate ISO values. I can tell when I am printing a crop factor or full frame lens, but I am looking at full print resolution on my computer screen when I do so.

I hope that this helps.

Thank you Manfred for your comprehensive response. As to what I do with my images, like you, I sometimes print big - on canvas usually. I also have a YouTube channel to which I post my photography as an adjunct to its mains purpose, namely classical music. There, I use "Ken Burns effects": zooming and panning the image - sometimes at the extremes of acceptable sharpness.

[AntPDC]

Thank you so much DanK, and for the link too! And I take your point about "disagreement" with regard to these issues, but shall nonetheless follow the link you kindly provided.

[AntPDC]

Great! Optics expert Bill Claff over on DPReview gave me a link to his website which explains, visually, everything I needed to know:

http://www.photonstophotos.net/Gener...,AxisO,OffAxis