Diffraction limited aperture

Question, what do you guys think of the diffraction limit aperture on cameras like the 5D mkII (10.3) and now the new 7D (6.8)?

I'm a bit ignorant in this issue, I thought it is a thing to be concerned about.

But nobody talks about it, am I missing something? Going over the diffraction limited aperture isn't as big of a problem as I think it is?

Do a search. There has been a in depth discussion about that very thing in the last month or so. It's a big deal to pixel peepers and maybe 20% of the people here. For the other 80% here it is an issue to be aware of and work around but not a big deal.

I think aperture derived diffraction is more a function of the lens you are using than the camera. In general, I try to avoid f-stop values larger than f/16. I've tested by 40 with my 17-40, 24-70 and 70-200 lenses, and f/16 is about as large in value that I would go.

Sadly reading all the threads on diffraction will probably confuse you further or make you want to take up needlework instead of photography.

Simply put diffraction occurs when light passes an edge, in lenses that means when it passes through the diaphragm.
As ever smaller apertures are used the wavelengths of light are 'squeezed' closer together and start to interfere with each other. This is diffraction.

The camera body and size of the sensor, and the sensor resolution, have nothing to do with diffraction at all - it is simply a lens diaphragm phenomena.

What people worry about is seeing the effects of diffraction in the final image, which, like depth of field, is a product of another set of complicating factors.

Again put simply a higher resolution sensor will indeed show the effects of diffraction more than a lower resolution one, BUT the gains of the higher resolution will more than compensate for this in real-world applications.

Don't worry about diffraction - after all if it ruined images to that extent how would we do macro work where DoF is crucial and requires apertures of f22, f32 and even f64?

tgara wrote in post #8752124
I think aperture derived diffraction is more a function of the lens you are using than the camera. In general, I try to avoid f-stop values larger than f/16. I've tested by 40 with my 17-40, 24-70 and 70-200 lenses, and f/16 is about as large in value that I would go.

Sorry to be pedantic.............
but a 'large' aperture is surely a bigger diaphragm opening - like f2.8 is a larger aperture (opening) than f4.
What I assume you mean is that f16 is the smallest aperture you would use?

Larger f-numbers = smaller aperture
Smaller f-numbers = larger aperture

At least that's what I was taught ;and it hasn't changed since digital came in so far as I'm aware.

Cuervo79 wrote in post #8752074
Question, what do you guys think of the diffraction limit aperture on cameras like the 5D mkII (10.3) and now the new 7D (6.8)?

I'm a bit ignorant in this issue, I thought it is a thing to be concerned about.

But nobody talks about it, am I missing something? Going over the diffraction limited aperture isn't as big of a problem as I think it is?

It's a technical distinction that is used to frighten people about high pixel density cameras. The reality is that it means nothing like what people think it means.

Suppose you have a 30D that is 'diffration limited' at f/11 and a 7D that is diffraction limited at f/5.6. Now suppose you mount a perfect, aberration free lens on each camera.

The 7D will resolve perfectly (remember the lens is perfect) every possible detail down to the nyquist limit of the sensor. Once stopped down to f/5.6 the 7D will start to diverge from the nyquist limit because of diffraction. But note....at f/8 (for example) the 7D will still be resolving more detail than the 30D!

Only once we get to f/11 should the cameras be resolving the same detail, because by that point they are both limited by diffraction and the higher pixel density of the 7D cannot offer any theoretical advantage.

lol well I doubt I would take on knitting if I get an information overload, I would go to the "ignorance is bliss" route if it ever gets too bad...

so this means that as long as you get "plenty of megapixels" diffraction limited aperture is pointless?

PaulB wrote in post #8752154
Sorry to be pedantic.............
but a 'large' aperture is surely a bigger diaphragm opening - like f2.8 is a larger aperture (opening) than f4.
What I assume you mean is that f16 is the smallest aperture you would use?

Larger f-numbers = smaller aperture
Smaller f-numbers = larger aperture

At least that's what I was taught ;and it hasn't changed since digital came in so far as I'm aware.

Well, to be extra pedantic, he WAS referring to f-stop values, not the actual size of the aperture.

As you say:

Larger f-numbers = smaller aperture


... he WAS referring to f-numbers.

Cuervo79 wrote in post #8753331
so this means that as long as you get "plenty of megapixels" diffraction limited aperture is pointless?

No, what it means is that this discussion of diffraction is largely a non event for most users. Pick up your camera, and go shoot, and forget you ever heard about diffraction.

I'm interested in the subject, it may not be relevant for most shooters but it is an issue is it not?

Cuervo79 wrote in post #8753416
I'm interested in the subject, it may not be relevant for most shooters but it is an issue is it not?

No, it really isn't an issue. Re-read my 1st post in this thread.

I did read your post, but you didn't answer my response which was "so this means that as long as you get "plenty of megapixels" diffraction limited aperture is pointless?"

Cuervo79 wrote in post #8753437
I did read your post, but you didn't answer my response which was "so this means that as long as you get "plenty of megapixels" diffraction limited aperture is pointless?"

I guess I don't know what you meant by 'plenty of megapixels'. See if the following leads to what you meant.

In the end there are three things going on with absolute resolution in an image.

1. Lens resolving ability and aberrations. Succinctly, the better the lens the more things like diffraction and pixel count matter. Lenses that cannot resolve small details render all else moot.

Also note that what a lens does is resolve tiny details on a physical sensor. The smaller the sensor the more we have to enlarge those details to make a given print.

2. Pixel count. Sensors cannot resolve details smaller than the line pair sets (this is the Nyquist limit). So in theory more pixels = more resolution for a given sensor size so long as the lens can keep up (see 1).

3. Diffraction. For any one aperture size there is a diffraction limited circle of confusion (known as the Airy disk) on the sensor cause by the scatter of light passing through a small aperture. The smaller the aperture the more the light scatters and the bigger the disc. When the discs are bigger than the pixels, having more pixels on a sensor will not improve resolution.

So essentially, what diffraction limitation means is where the pixel count cannot provide any additional resolution because the diffraction is blurring details of that size or smaller.

If you had enough pixels then a camera could be 'diffraction limited' at f/1.0 and this would actually be a good thing. It would mean that sensor resolution could never be the limiting factor in any shot - every photo resolution would be strictly limited by a combination of lens aberrations and diffraction.

Regarding diffraction, so it doesn't improve resolution past a certain point, but doesn't it also start to "blur"? For example when taking a photo at F8 and then one at F32 not only will the DOF be different but the F32 won't be as sharp as the F8 is. Or am I missing something?

JeffreyG wrote in post #8753532
3. Diffraction. For any one aperture size there is a diffraction limited circle of confusion (known as the Airy disk) on the sensor cause by the scatter of light passing through a small aperture. The smaller the aperture the more the light scatters and the bigger the disc. When the discs are bigger than the pixels, having more pixels on a sensor will not improve resolution.

So essentially, what diffraction limitation means is where the pixel count cannot provide any additional resolution because the diffraction is blurring details of that size or smaller.

If you had enough pixels then a camera could be 'diffraction limited' at f/1.0 and this would actually be a good thing. It would mean that sensor resolution could never be the limiting factor in any shot - every photo resolution would be strictly limited by a combination of lens aberrations and diffraction.

This first of all relieved me, then within a second or two, confused the living s**t out of me.

I'll say from the outset that I'm not trying to argue anything you've said (or in 'net' terms, I'm not trying to flame your reply), more just seek some clarification, because it appears you've got more of a handle on the issue than I (at present).

I've read a lot about DLA, and I won't lie, I found it a pretty complex concept to get my noodle around. The most useful (I thought) article I read was this: http://www.cambridgein​colour.com …ffraction-photography.htm

Anyway, to summarise (wish me luck!) my understanding thus far...

1) Whenever you're focusing light to a point, diffraction occurs as it passes through the aperture - 'creating' (for want of a better word) an Airy disk.

2) The most resolvable light falls in the centre of the Airy disk (within the diameter of the first dark circle).

3) Problems occur when Airy disks overlap.

i) Perfect resolution occurs when the Airy disk falls completely inside the area of the pixel.
ii) Details are less resolved when disks overlap, up to half their diameter.
iii) Details are not resolved when disks overlap by more than half their diameter (and in extreme cases when the area inside the first dark circle from two Airy disks overlap).

So, if you have a nice big pixel, you can use a narrower aperture (creating a larger Airy disk) without diffraction causing a loss in resolution, as there is a greater pixel area to 'contain' the Airy disk. If you have a tiny-weeny pixel, you're constrained to larger apertures to keep the Airy disks small (and inside the area of the pixel) in order to resolve the same level of detail.

If we're comparing different numbers of the same sized pixels, the greater number of pixels delivers better resolution as we've captured more picture elements, and DLA is the same because the area of the pixels is the same - of course we'd have a different image because of the sensor area increasing (same size pixels, but more of them).

If then, we compare the same amount of different sized pixels, the larger pixels are preferable for overall image sharpness at a range of apertures (ie will resolve detail at narrower apertures), as the Airy disk can become larger (due to narrower aperture) before it exceeds the boundaries of a pixel, and begins to overlap with the Airy disk at adjacent pixels.

Of course the issue is complicated when comparing a full-frame to a crop sensor camera (as 5D2 and 7D), because of differences in field of view and depth of field. Assuming DOF is not an issue (creatively), and we're trying to shoot a subject which falls entirely within the focused area; we'd have to use different focal lengths to get a 'similar' image.

We'll say we're using a 24-70mm f/2.8 L - 30mm on the cropper, 48mm on the FF - both inside what is commonly perceived to be the range of peak sharpness.

DLA on the 7D sensor is f/6.8 and on the 5D2 is f/10.3.

So, if we shoot at f/8 on both bodies, we're inside the DLA for the 5D2 (so the Airy disks are inside the pixel area), but exceed that of the 7D (so the Airy disks are overlapping somewhat). Wouldn't fine detail therefore be better resolved on the 5D2? I appreciate (as above) pixel count has a role to play too, but not only does the 5D2 have larger pixels (which can 'contain' larger Airy disks), there's also 3.1 million more of them, so there's a win there too?

...And with that, I think it's time to head outside for a cigarette (yes, I know! )

Best regards,

(a head-scrambled) Rob.