End of Megapixel wars and EF-S buying decisions

Started Feb 23, 2006 | Discussions thread
Don Glenn Contributing Member • Posts: 854
Re: My read on the diffraction limit issue

I think it is more useful to focus on the resulting print. For a crop camera, you have more magnification, which means you have a lower F-stop diffraction limit right from the start (calculations for this are shown below).

The diffraction limit people or normally talking about here is based on a 100% view (pixel peeping), which is important I guess, but I don’t think it is the main issue we should be worried about. I think the real issue is determining when diffraction will cause enough resolution loss that it is equivalent to up-sizing a smaller source. At this point you are wasting processor time and write time, both affects shot to shot time limits, and storage space.

I also think the math people are using is not complete, because there aren’t enough details available to really calculate actual numbers. The Bayer filter and the AA filter will affect the math, and very little is known about the AA filter specs. If there is no AA filter, then the Bayer filter means that you have another 0.4 diameter width before is should effect the resulting image (diagonal green to green distance is the closest same color pixels - 1 for a total of 1.4 x pixel diameter). If there is a strong AA filter, then you start approaching a single pixel diameter. Knowing a little about the AA filter, I think it blurs one direction then the other, so the G-G diagonal may even be less of an issue since it isn’t a radial blur (resulting in something larger than 1.4 x pixel diameter). The calculations for this can go on and on.

So, I don’t think it is a big problem right now because we haven’t gotten close to failing the up size comparison test in my opinion (dslr’s).

Some simple Diffraction limit math, ignoring pixel dimensions:
The Circle of Confusion and the diffraction limit are directly related.

Diffraction Limit (a) / CoC (a) = Diffraction Limit (b) / CoC (b)

So, if the Circle of Confusion goes down, the Diffraction limit f stop will go down. The Circle of Confusion is based on target print size and some fussy number depicting human eye resolution and is proportional to 1/crop factor.

CoC(crop) = CoC(film) crop

So, let’s start with a diffraction limit of F22 for 35mm film (not to picky rule of thumb) and a target of a 1.6 crop.

Diffraction limit (1.6x) = Diffraction Limit (film) CoC (1.6) / CoC (Film)

Diffraction limit (1.6x) = Diffraction Limit (film) (CoC (Film) / 1.6) / CoC (Film)
Diffraction limit (1.6x) = Diffraction Limit (film) / 1.6 = F22 / 1.6 = F13.75

If you are targeting a smaller circle of confusion (cropping before printing, or a shorter viewing distance) then the diffraction limit will be lowered.

Post (hide subjects) Posted by
Keyboard shortcuts:
FForum PPrevious NNext WNext unread UUpvote SSubscribe RReply QQuote BBookmark MMy threads
Color scheme? Blue / Yellow