Effect of Diffraction with Full-Frame Cameras?

<p>For those of you that have moved to a full-frame camera, have you noticed a difference in diffraction from the same lenses? It's very common for a lens to not work as well on a new camera, but has anyone seen lenses that work just as well, but at fewer or more settings?</p>

<p>The reason I ask is this: moving to a full-frame camera means I will often be closer to my subject than before, and the closeness means that I must stop down further to get the same depth of field. At a certain point, diffraction is going to be an issue.</p>

<p>I've thought about this a lot, but I'm not sure what will happen. Going to a higher pixel count on the same size sensor increases the effects of CA, diffraction, and all those little optical issues. That means that theoretically, a full-frame camera with the same or similar pixel count should have fewer optical issues. But on the other hand, full-frame cameras often have higher IQ, and are better at exposing flaws in lenses - leading to more optical issues.</p>

<p>Am I correct in assuming that a larger sensor with equal or similar pixel count will result in fewer optical issues from a really good lens (like a Zeiss or Leica), and more optical issues from a not-so-good lens? Is there anyone here with some experience on the matter?<br /><br />Also, please please please don't start arguing about sensor size and depth of field. I'm usually shooting at f/16 now, and I'll be shooting at f/22 if/when I get a full frame camera. That is all you need to know, and the end of the discussion. If you want to argue, I would kindly ask you you use one of the many other threads dedicated to the topic. Thanks.</p>

<blockquote>

<p>more optical issues from a not-so-good lens?</p>

</blockquote>

<p>Not if you print or display at the same size you were before. That's the key thing, here. If you try to take advantage of much higher resolution by going for much larger <em>output</em>, you're going to notice those things.<br>

<br />I'm not being glib to say that the solution to your problem is: a longer focal length lens, so that you can work at your familiar distance, and compose as you're used to.</p>

<p>I don't know any serious Leica owners who routinely shoot at f/22. Most Leica lenses are designed for optimal resolution between f/2 and f/4, and by f/11 begin to show signs of diffraction, film or digital.</p>

<p>This is a pretty useful page on this - with a calculator that you might find helpful...</p>

<p><a href="http://www.cambridgeincolour.com/tutorials/diffraction-photography.htm">DIFFRACTION</a></p>

The only reason why diffraction would not show itself as a resolution limiting factor (forgetting about not using the full resolution as recorded for now) is that other factors (lens aberrations) limit resolution more than difraction would.<br>In lenses that are designed to be optimal, i.e. have lens aberrations reduced below the diffraction limit, diffraction shows itself at all apertures.<br>You can't design diffraction away. It's always there, limiting the achievable resolution when the lens is wide open, reducing it every time a lens is stopped down (at a rate of about halving the resolution every two stops the lens is closed).<br>That's a rather hefty rate, and unless a lens has really bad lens faults that only go away slowly when stopping the lens down, you will see signs of diffraction, i.e. decreased resolution, every time you stop a lens down, no matter from f/2 to f/2.8 or f/16 to f/22.<br><br>So after having established that diffraction is always there doing its thing, but often hidden by even worse thingies, the thing that remains is that about not using the full resolution being offered (how good or bad that may be) to begin with. The more you are throwing away of what was on offer, the less you are going to notice diffraction's limiting effects (nor that of lens abberations). Conversely, the more you are going to make use of the resolution offered, the greater the chance you are going to hit the limit of what is on offer.

<p>After a few responses (Matt, Stephen), perhaps I ought to go back a bit and explain why I'm even asking this question.</p>

<p>Right now, I do most of my shooting with a Hasselblad, and some with a Linhof, almost always black and white. I find that more often than not I am working at f/16, typically in-studio. I use my NEX, usually with a 50mm lens, mostly to capture previews of new looks or lighting setups before I shoot film. There are some people who want colour images for their trouble, so I usually shoot some non-preview photos too.</p>

<p>Well, here's why the issues pop up. I need to be very close to my subject with the 4x5, and that means I need to stop down much further. If I were to nearly fill the frame with the subject's face, then f/22 or f/32 is not unreasonable. When I'm shooting at f/16, all my lenses/cameras look great. I don't currently own any Leica glass Stephen, but I have some Schneider for the NEX, and that works great too. But if I shoot those smaller format lenses at f/32 then diffraction becomes a pretty big issue. It doesn't make my previews any worse, but it does mean that those photos aren't very good for the model's use, and I need to make a separate setup for them.</p>

<p>If a full frame camera had fewer diffraction issues, that would mean that using one would make my photo shoots much simpler, as I could effectively reduce the amount of time I that had to spend setting up and changing things around. Instead of doing a setup just for the digital camera, I could snap off a few photos with it at any time.</p>

<p>I realize that this isn't the issue that I had originally mentioned; I just figured that since I had a closeness/aperture issue, it would just be easier to describe it the way I originally did. Sorry for any confusion.</p>

<p>John, thanks a lot for the link. According to that page, I am currently using my NEX lenses 1-2 stops smaller than I "should". Also according to that page, a larger sensor with the same pixel count will mean that I can stop down two more stops and get the same results as I have now, at least as far as diffraction is concerned. That seems to mean that if I had more pixels, but not double, in a full frame camera, I could probably expect diffraction to set in only one stop later, using the same lenses.</p>

<p>Thanks for all the help guys. And so fast!</p>

<p>I agree with Matt. Shoot at the focal length you need for the perspective you want. Don't try to force yourself into tighter perspectives with full frame.</p>

<p>Zack, larger formats will tolerate smaller apertures without running into a diffraction limit. By way of example, "Group f/64" was a group of notable west coast large format photographers (including Adams and Weston). I don't think f/64 is even available on any of my 35mm format lenses, and if it were, it would probably look awful. However, it worked quite well for the large format guys, because diffraction limitations were not as serious for them as they are for me.</p>

<p>Going the other direction, diffraction limits are quite tight for a tiny-format camera, such as a compact digital. My G-11 seems mildly diffraction limited around f/4, while my 5D is good down to f/16.</p>

<p>The result of all this, tidily, is that the maximum depth of field achievable in any format is about the same, just at a different aperture. Of course the minimum depth of field on the other end, at the same aperture, is shallower in the larger format.</p>

<p>Another point: Diffraction results in a Gaussian blur, which is the type of blur for which our digital sharpening tools are best suited (as long as we set the radius right). I've become a little more fearless on the diffraction-limited side in the digital age, because I know I can do a decent job of re-sharpening. So there are times (albeit rarely) that I will go ahead and shoot at f/22. Normally f/16 is where I call it quits, though (and f/11 in APS-C format).</p>

Just a side-note: though the diffraction limit is here discussed as something bad, being diffraction limited is in fact also the ultimate goal you can achieve: if your imaging system is diffraction limited, it is as good as it can ever be.<br>;-)<br><br>(Yes, i know that we're discussing increasing diffraction itself by changing the aperture size.)<br><br>Diffraction is the same, limits resolution to the same amount, for the same f-stop, regardless of format. F/64 on LF is as bad as f/64 on 35 mm format. The saving grace of LF is that you don't (have to) enlarge it as much as 35 mm format.

<p>I'll argue about anything I want. You do not own this forum. Arguably, you should take a class in Photography 101.</p>

<p>Sarah-<br>

Thanks for contributing. The issue isn't as much about focal length as about getting an accurate preview with my digital cameras for 4x5 capture though. If I could stop down my lenses further with a larger sensor before I suffered a loss in resolution, that would make the digital cameras a more accurate preview of what I can expect from the 4x5, while still giving me sharp files that I could give to the sitter that day. I should have said that in the original post; I stupidly thought that the much shorter explanation that I gave would have sufficed, but I just muddied the conversation. Sorry about that.</p>

<p>Q.G.-<br>

Way to see the bright side :) Although I would think that film and digital capture would be affected by diffraction differently. A 35mm piece of Tri-X has the exact same grain structure as a 4x5 piece, and thus the same resolution at the film plane. But since a full-frame sensor does not necessarily have twice the megapixels, those photosites may be larger, and theoretically less prone to diffraction. Or something.</p>

<p>Obviously no one can say anything with absolute certainty yet, as the D700 is (I think) the only full frame camera to exist alongside a crop-sensor camera from the same system and the same generation, and with a theoretically similar sensor design. Once the D600 and the A99 have been out a bit longer, I'm sure we'll be able to back up these conversations with actual data, instead of theory.</p>

<p>Wayne-<br>

Not taking the bait, troll. You do know that when you respond to someone's desire to not argue with an attempt to argue with him it makes you look like a jerk, right? You know that?</p>

<p>Zach look at William Neil's images. In his blog he gives you all the exposure data. I was surprised to see that many of his landscapes are taken at f22 when I would have thought that f8 would have been good enough. Apparently he pays careful attention to his initial input sharpening. He uses a full frame DSLR. I've been a lurker here since 2000. Or you might want to ask him yourself. </p>

<p>Thanks for de-lurking, Hugh! Neill does seem to use some very tiny apertures! It's gratifying to see that others have stumbled on the same solutions in dealing with diffraction. Like Neill, I've found input sharpening to be the critical first step in the process (and also turning off all sharpening in DPP). Again, it's the radius that's most critical, and to get the right radius, I do an extreme pixel peep and guesstimate the width of a "sharp" edge in pixels. I'm not primarily a landscape photographer and haven't often wandered very deeply into diffractionland, but perhaps I'll experiment a bit more now. I'm curious what can be achieved.</p>

<p>On full frame cameras like the D800 there are technical tests which show diffraction having a marked effect by f11, with f5.6 to f8 being the optimal point for best image quality on most lenses. I've got a D800 myself and am trying to keep to f8 for most landscape images only going smaller than that with macro images, where you will find much less depth of field than a DX camera would give. When I had a Pentax 67II I used f11 most of the time and that was the sweet spot for that format. Larger formats than that need smaller apertures still but full frame/ 35mm format usually won't need very small apertures, honestly in most cases a smaller aperture than f8 isn't going to be beneficial, be very sparing of f22 unless you really need every last bit of depth of field.<strong><br /></strong></p>

<blockquote>

<p>Right now, I do most of my <strong>shooting with a Hasselblad</strong>, and some with a Linhof, almost always black and white. I find that more often than not <strong>I am working at f/16</strong>, typically in-studio.</p>

</blockquote>

<p>Zack, I found your question about full frame digital curious. Diffraction is diffraction. <a href="http://www.hevanet.com/cperez/MF_testing.html">If you are shooting a Zeiss 80mm 2.8 Planar on a Hasselblad at f/16 you aren't getting anywhere near the best resolution from that lens</a>. And that is fine. Sometimes DOF concerns force you to shoot at f/16, but that is not optimal on a Hasselblad nor a Nikon. If I am not doing something like a portrait or taking a picture of some detail I generally stick to f/5.6-f/8 on all my cameras (I don't shoot large format).</p>

<p>My humble advice is put your DSLR on a tripod and using mirror lock up take a series of pictures of a landscape with a lot of complex details at various f/stops. It will be obvious to you where things start to fall apart. f/22 is basically useless to me on my film 35mm... Unless I really need to bring my shutter speed down.</p>