D60's ( APS-C )sensor is too small to stop down below f/11?

If we consider the AF was spot on, I find the "added DOF" irrelevant for the test. Just consider the test target is a piece of paper.

Sound paradoxical, considering what we knew about pixel pitch <---> resolution behaviour. The fact is that the airy disks would have the same size on both the 5D and 20D at the same focal length. I other words, the image would be equally crap on both sensors. And if there still would be something to extract from the soft images produced by the small aperture, clearly the 20D would win, due to higher sampling frequency (smaller pixel pitch).

My conclusion is that there shouldn't be any difference between the FF prints and the 1.6x crop prints, as long you follow the rules and don't try to imitate the 5D with your 20D (i.e using as wide lenses as on 5D or printing as large as from 5D).

Got that right!

The effect was first recorded in the 17th century by church astronomers (tracking the beam of light from a small hole high in the cathedral wall, cataloguing the sun's movements; too small a hole in the wall and the spot on the floor became very fuzzy; bigger hole had a well defined edge and was easier to track) A little diversion from the main topic but fascinating...

Thanks Chris.

I know it's dry, and I can see that it's a poor fit to the forum's interests as a whole. Still, I feel an obligation to warn people that cameras like the Sony DSC-V3 can produce diffraction-degraded images (like those seen in the clock tower photos at http://www.dcresource.​com/reviews/sony/dsc_v​3-review/) and, to equip anyone who is interested with the means to determine for themselves when diffraction will become visible in a camera they are considering for purchase. Not even the high end cameras are completely immune from the effects of diffraction.

Mike Davis

I think it's funny that posting to a thread you want to disappear has the direct opposite effect.

I have a brief question about this. I haven't done the math so forgive the simplistic question.

Let's say Canon comes up with a 1.6x sensor that has the ability to capture, say, 5000 megapixels (something stupidly large). As such, all this diffraction stuff kicks in very soon, at all apertures you might care to use (f/2, f/1.8 whatever you choose).

If your only goal is to print 6x4's and view from 10" then there is no issue is there? The issue would only occur if you wanted to generate much large prints (and you might attempt to, with a 5000 megapixel sensor!).

As such, this 5000 megapixel beast will be NO worse than a 3 megapixel CMOS if printing 6x4's at 300dpi and viewing at 10"?

I am thinking that physics will not allow a APS-C sized sensor to hold much more then 200 million pixels (could be 250MP - close enough). somewhere around that point the pixels drop below the wavelength of light and it would then be physically impossible for them to record the light itself. It would of course theoretically still be possible to record the passage of a light wave, just not record the photon itself.

Now back to the thread itself. I am really really glad I did not read or even see this thread before I bought my D60. While I understand the principles and agree they are important to understadn in getting the highest quality from the images taken. At the same time I have learned that getting a bad picture is better then no picture.

So I have relearned something I think I was taught many years ago, but there will only be some rare occasions where I might use this knowledge. Of course I tend to shoot at the large aperture end of the spectrum where vignetting, light fall off and getting focus right on the money are the problems. But every once in awhile it is nice to have the knowledge needed to go the other way.

Meanwhile, I think this thread should be kept forever, it is has been on POTN longer then I have been and I am trully impressed by that.

Your last phrase is the most important, and your question is not simplistic at all. There are two phenomenon that limit the size of a print: The resolution of the sensor (whether it be film or digital) and the resolution of the optical system. Neither tells the whole story. Those who determine print size solely on the basis of the number of pixels on their sensor will eventually run into optical limitations as pixel densities increase, and those who base it solely on the optics may well run into pixel limitations. Only when one system is vastly better than another will this not be the case.

But that has not proved to be true in practice. Nearly all cameras on the market are limited by the sensor in some cases and by the optics in other cases. The more recent portions of this thread talked about one of those limitations--diffraction--and its effects on very small sensors with high pixel densities in particular.

If one wants absolutely the best possible image quality, then one will seek to find the best balance between the effects of diffraction and the effects of lens faults and too limited a depth of field in any given situation. The solution space may be inadequate for prints beyond a certain size, in which case one will need camera movements, a larger format, or some other method of managing that balance point. Knowing that relationship will help keep disappointment at bay, for those who don't mind modeling things in math rather than learning them by experience.

I will never forget the disappointment I felt when my first 4x5 negatives, shot at f/45 using short lens, would not support a 20x30" print when viewed up close. I learned that f/32 gave me the needed results for such prints with that lens, and if that didn't provide enough depth of field, then I needed to make use of camera movements to adjust the depth of field to meet my priorities. I wish I'd spent a bit of time thinking about apertures before that first disappointment--every field session with that camera was difficult and consumed opportunities that were hard to repeat.

The standard you propose--4x6" prints viewed at 10 inches--is not particularly demanding and widens the solution space between the effects of diffraction and the effects of lens faults and too limited a depth of field. In practice, you probably will not experience any problem with this issue using most any digital camera.

But can one enlarge to the full potential of a 3x5mm 6-megapixel sensor? Not without seeing the limitations of the optical system, particularly diffraction. Thus, the maximum enlargement on such cameras will be limited by optical issues moreso than by pixel density.

When we get to the point where such densities pose no problem of convenience, then it won't matter. Even if a pixel only sees something marginally useful by itself, we will be integrating large numbers of pixels together when we make any reasonable print--"reasonable" meaning any print that is supportable by the optics. The errors of too small a pixel will average out.

But too many pixels still poses a challenging problem of memory card size, processing power, storage, and so on. So it makes sense to choose a pixel density that is appropriate for the format and optics--and for the objectives of the photographer. Pixel density and format (which is a surrogate for pixel count given a particular pixel density) go hand in hand when choosing a camera.

Rick "summarizing some salient truths for those unwilling to read the whole thread and delve into the arithmetic" Denney

I find this thread rather interesting, I had never heard of anything like this before. In the end it doesn't make much difference to me, I am still just going to shoot things, but none the less its still interesting to learn about.

When I read this thread (entirely) for the first time some years ago, I found it very interesting and informative (although I didn't get all the math involved and skipped some ) ... but I considered it - as many others here - as an issue of mainly academic importance.

However, this recent addition by the original poster brings all very much to the point in a short, hands-on example with pictures...

zilch0md wrote:
Here's another example - from a page reviewing the Sony Cybershot DSC-V3:

http://www.dcresource.​com/reviews/sony/dsc_v​3-review/

Scroll down to the pair of images showing a brick clock tower. One was taken at f/4. The other was taken at f/8.

Do you find the difference to be significant?

... and shows that it is a very practical problem ... at least when you are using a small-sensor, high-megapixel camera (e.g. like the Pro1 or G6). And I am sure that many users of those cameras will use f/8 - based on the common belief 'stopped down = always sharper' unless they are told that this is not always the case.

So I would also like to second others' expressions that - while the title may be a bit exxagerated - this thread does have its value for the forum and deserves its existence very much ... especially as it continues to draw out knowledgeable people (I use RDenney here as one example that comes to my mind) to comment and add other insights.

And - let's face it - the average fool would be bored by page 2 and open the next 'My new L is coming' thread ...

Best regards,
Andy

I have no problem with this thread - if you all just would keep on topic!

I'm sure people have used d60's, 20d's 300d's 350d's etc and stopped down to f/11 or lower without the world coming to an end.

It's like arguing that using a 1.6 sensor and a 100mm lens doesn't give you the equivalent of a 160mm lens on a ff sensor or it does or it doesn't or it's the field of view that's different and so on and so on and so on.

I have taken pictures at f/11 or lower on my 350d and printed the pictures at 8x10 and some even A3 and they have looked fine, other people think they look fine, nobody has ever said "NOOOOOOOOOO it's just not possible!!!" and fallen to the floor covering their eyes.

Likewise when I use a 300mm zoom on my 350d I'm just glad that my pictures "appear" more zoomed in than if I used the same lens on a 35mm camera.

I'm not worried about the technicalities, I just know that compared to my friends film rebel using the same lens I have more reach to all effective purposes.

I just hate posts like this because people starting with an slr read them and get SCARED and will probably never use f/11 which is just crazy.

Any newbie stumbling onto this topic would go and lock their camera in a vault and cower in the corner if they made it to the end.

I'm more concerend with just going out and taking photos.
I use f/11 or lower when I need a very large DOF
I can look out of my window with my 350d and 300mm lens and things appear closer than when i use my firends film rebel and the exact same lens.

And I really don't need to look into it any deeper than that.

The subject has come up on other forums (forii?) and it seems that most often, it's presented as though there is a diffraction "wall" or "limit". Diffraction occurs, and occurs to a greater extent with a smaller aperture. This isn't anything new. The fact is that diffraction's visible effects don't suddenly appear in an undaunting fashion when the magic f-stop is selected. Rather, it's a gradual change that becomes noticeable as the aperture grows smaller and/or as the enlargement grows greater. Film shooters dealt with it. Digital shooters deal with it. Most shooters ignore the issue as it isn't that significant most of the time.

My beef with this whole thread has been the original overstated premise that a D60 (and presumably a 10D, not to mention a 20D/350D/30D) cannot be used with an aperture smaller than f/11. It can, and successfully. Does diffraction become visible? Possibly, but it certainly isn't a hard limitation. There are other issues that are usually of greater importance. If you shoot close-up macro, you're going to shoot at f/16 or f/22 or even f/32 because depth-of-field in close photography is a much more significant issue. If your lens'es sweet spot is f/16 at the wide end, you're most likely going to shoot that landscape at that aperture to avoid those soft corners.

Diffraction happens, of course. And it is measureable. And visible. Sometimes. But it's presence should not cause one to falsely fear shooting at smaller apertures than the title of this thread would indicate as being the limit.

Just as a reminder of the *first* line of the *original* post :
" If your goal is to achieve a resolution of 5 lp/mm in an 8x10 print, you will not be able to stop down further than f/11...."

If that's not your goal, why read the thread?

Given the gazillions of posts about sharpness (such as : My new lens or camera's pictures of the angels on my pinhead are soft, what do I do now?) it's clear many people *are* concerned about anything that affects sharpness. Diffraction can limit this sharpness. Hence the thread. Simple as that.

What does lp/mm mean? I am guessing its something per millimeter, but what?

Line per millimetre I believe!

Dave