Make it complicated? I can do that
That is not true.
The reality is that pixel density has virtually no effect on diffraction. The theory, involving creative formulae, is not reflected in real world images. I have never seen an image with a higher pixel density show lower actual sharpness than than one with lower density, which would be the case if this was true.
Eg, Canon 40D vs 7D, 10mp vs 18mp. No question at all that the 7D is sharper.
You simply can't use that as a comparison because you are not comparing like for like. To make a true comparison all other factors have to be held equal and only sensor size varied. With all other things held equal the larger sensor will produce a sharper image. The simplest way to view this is to forget digital and think instead of film and do a reductio ad absurdium.
With lenses now optimised for crop sensors and technology advancing in many areas related to image processing there are a myriad of other factors to take into account, but sensor size is an important factor.
HoppyUK said:
Squawk said:
As a result, the wider the lens, the better the focus, with a caveat. The caveat is that lenses are made to tolerances, and it's a lot easier to manufacture a small lens than a big lens. The reason some lenses don't perform optimally wide open is simply manufacturing imperfections, not optics, the wide open aperture exposing the biggest bit of lens to the light.
None of that is true either. I would try and explain why, except I cannot make out the reasoning.
My fault, I explained it poorly by being so brief and ignoring aberrations, trying to focus entirely on diffraction, so lets go back to the start and be a bit more thorough.
There are essentially two factors that limit a lens. Aberration and diffraction, and for the purpose of this discussion I'm going to ignore aberrations at the start in order to concentrate on diffraction.
Cameras, or more specifically lenses, are limited by diffraction. Even if you could produce a "perfect" lens, its resolving power would be limited due to diffraction. The limit is proportional to the size of the objective lens. We see reference to this all the time when discussing astronomy, the constant search for manufacturing processes to produce bigger and bigger mirrors to get better resolution.
On the link I provided earlier, given again
here, you can see a table that actually lists the diffraction limited performance of various f-stops.
How are diffraction limits and aperture related? When using a smaller aperture you decrease the apparent size of the lens. Although the lens itself stays the same size the area of lens in use is decreased, and as we noted earlier, the size of the objective lens is directly proportional to the resolving power of the lens. The inference, wider aperture = higher resolution.
HoppyUK said:
Here is a graphic of diffraction in action, clearly showing what happens at high f/numbers
http://www.dpreview.com/lensreviews/canon_50_1p8_ii_c16/page3.asp
I'm missing something here because I don't see a contradiction with anything I said.
HoppyUK said:
Switch between full frame and crop format to see how that changes the aperture level at which diffraction becomes a limitation. Pixel density on the two cameras used, Canon 450D at 12mp and 1DsIII with 21mp, is very roughly similar.
I don't see how that's a useful comparison for the same reason I listed above, all other factors must be kept equal to give a valid test and that simply isn't possible with these cameras.
HoppyUK said:
On this test you can also see how diffraction effects the centre of the image sooner than the edges.
Can you explain to me how you infer that from the diagram, I'm missing it and would like to understand. I can see that the resolution at the image edges is inferior, but can't see how you infer this to be due to diffraction.
HoppyUK said:
Of course it is actually affecting the whole image equally, but raising the f/number still improves edge sharpness, by reducing other aberrations at a higher rate than diffraction reduces it.
This brings us to the subject which I glossed over in my initial post, aberrations. Aberrations cause the edges of images to be out of focus or distorted. They affect the center of the image too, but not to the same degree. Reducing a lens size (higher f-number) reduces aberrations, but at the cost of introducing more diffraction.
Aberrations and diffraction are the reason we pay for good glass (or more accurately, why I will do when I can afford some). There are fundamental laws of physics/optics that cannot be overcome, but manufacturers strive to get as close to those limits as they can. Where diffraction is concerned a device is said to be "diffraction limited" if it meets the theoretical limits of resolution.
I don't profess to know much about aberrations save for a few basic types. AFAIK spherical aberrations are the prime cause of soft image edges, with barrel distortion a secondary issue.
On the link you posted Hoppy it seems that there is a marginal drop off in performance (sharpness) with the lenses wide open, with optimal performance at around 1 stop from fully open. Sharpness then drops as the lens is stopped down. The implication would be that diffraction is the limiting factor other than with the lens wide open.
HoppyUK said:
Did somenone say that you could make this subject complicated?
That would be me
