Nikon D800E vs. D800: Which is more Diffraction Limited?

[chicagonature]

The Technical Guide to the Nikon D800/D800E was just published and I believe it is saying that the D800E (the one with effectively no anti-aliasing filter) is more Diffraction Limited than the D800. Here's the quote, "The D800E offers better resolution at apertures where diffraction is not an issue. The effects of aperture may therefore be more noticeable than with the D800, and care may be required to avoid loss of definition due to diffraction." Does this makes sense from a physics point of view? I mean, shouldn't it be just the opposite of what Nikon is saying because the Nikon D800 is causing the Airy Disk to be a little bit larger due to the increased blurring caused by the anti-aliasing filter? Plus, Nikon uses the word "may" and they show images from the D800, not the D800E. It's confusing.

I have a D800E on pre-order (along with the 14-24mm lens) and I have an opportunity to switch to the D800 without losing my place in line. My application is this: Landscape photographs with the nearest foreground subjects located VERY close to the lens, requiring very small apertures. With my Pentax 645NII medium format film camera, I frequently shoot at f/32 using a 33mm lens (the same as using a 20mm lens at f/20 on 35mm full-frame camera). However, with the D800/D800E, they say it's diffraction limited at f/11. Still, I want to choose the best possible camera for the job.

(Ultimately, I'm simply trying to do as well or better with digital than I am with medium format film. I suppose with the 645 camera at f/32, for all of these years my pictures may be well beyond the diffraction limit anyway and I never really notice it.)

I appreciate help from those who better understand this topic.

Thanks!

[herbert]

I think Nikon is right. The D800E will have a larger diffraction limited aperture when it is calculated based on the pixel level resolution. If based on the desired print resolution then there will probably not be as noticeable a difference.

Diffraction causes light passing through a hole to spread out into a cone. This falls onto your sensor pixels. When the cone is larger than a single pixel then the diffraction will cause light to fall on more than one pixel. So for a sensor with no anti-aliasing filter the diffraction limit is specified by the amount of diffraction that causes a point source of light to fall on more than a single pixel.

When you put an anti-aliasing filter on a sensor this spreads out light falling on one pixel to the surrounding pixels anyway. So you already have light falling on multiple pixels. So for a sensor with an anti-aliasing filter the diffraction limit is specified by the amount of diffraction that causes a point source of light to noticeably fall on more than the original number of pixels. How to define 'noticeably' is varied.

This equates to having a bigger pixel size to define the diffraction limit. So with the filter on you can state that the diffraction limited aperture is smaller (higher f-number).

However the exact specifications of the anti-aliasing filter are needed to calculate the effect. Also you need to specify how you will define a noticeable difference. Due to the vagueness of this I believe that diffraction limit calculators do not factor in the anti-alias filter and simply use the raw pixel size.

In reality the smoothing of the filter and the gradual expansion of the airy disk due to diffraction will cause a light smudging at the pixel level. This will increase as you go beyond the diffraction limit. However I believe that your photo will benefit more from the increased depth of field than it will suffer from the pixel level lack of sharpness.

When you shoot on film there is no pixel size. In this case the diffraction limit calculation is based on the acceptable circle of confusion, i.e. how much can a point of light be spread out before the human eye can notice it.

You can try a diffraction limit calculator here:
https://www.cambridgeincolour.com/tu...hotography.htm

This calculator will show you when your image will be diffraction limited for a real print at a real viewing distance. This is a far more useful limit than the theoretical limit based on pixel size. It will also allow you to work out the limit for your medium format film.

If you have a view camera you have the advantage of moving the plane of focus using tilt. This can overcome softness problems with high f-numbers by allowing you to use the lens in the sweet spot. I don't want to encourage you to spend too much money but you could look at the Nikon tilt-shift lenses in order to satisfy your landscape shooting needs.

I am sure you will enjoy either D800. Digital is a lot more convenient than shooting film. But if you are not sure you could wait and rent one (but unfortunately lose your place in the buying queue).

Alex

[Black Pearl]

Essentially what Nikon are saying is the diffraction limit will be more visible on the E body as you will be more acutely aware of the drop in performance than on the standard body where some of the observed sharpness has already been blurred by the filter.

Regardless of the version you choose and assuming you are going to be very critical then the image viewed at 100% isn't going to be pretty at f22. You might want to look at something like Helicon Focus to extend your DoF. Not only will this give you unlimited DoF but it will allow you to shot at the very best aperture on each lens without compromise.

[chicagonature]

I think Nikon is right. The D800E will have a larger diffraction limited aperture when it is calculated based on the pixel level resolution. If based on the desired print resolution then there will probably not be as noticeable a difference.

Diffraction causes light passing through a hole to spread out into a cone. This falls onto your sensor pixels. When the cone is larger than a single pixel then the diffraction will cause light to fall on more than one pixel. So for a sensor with no anti-aliasing filter the diffraction limit is specified by the amount of diffraction that causes a point source of light to fall on more than a single pixel.

When you put an anti-aliasing filter on a sensor this spreads out light falling on one pixel to the surrounding pixels anyway. So you already have light falling on multiple pixels. So for a sensor with an anti-aliasing filter the diffraction limit is specified by the amount of diffraction that causes a point source of light to noticeably fall on more than the original number of pixels. How to define 'noticeably' is varied.

This equates to having a bigger pixel size to define the diffraction limit. So with the filter on you can state that the diffraction limited aperture is smaller (higher f-number).

However the exact specifications of the anti-aliasing filter are needed to calculate the effect. Also you need to specify how you will define a noticeable difference. Due to the vagueness of this I believe that diffraction limit calculators do not factor in the anti-alias filter and simply use the raw pixel size.

In reality the smoothing of the filter and the gradual expansion of the airy disk due to diffraction will cause a light smudging at the pixel level. This will increase as you go beyond the diffraction limit. However I believe that your photo will benefit more from the increased depth of field than it will suffer from the pixel level lack of sharpness.

When you shoot on film there is no pixel size. In this case the diffraction limit calculation is based on the acceptable circle of confusion, i.e. how much can a point of light be spread out before the human eye can notice it.

You can try a diffraction limit calculator here:
https://www.cambridgeincolour.com/tu...hotography.htm

This calculator will show you when your image will be diffraction limited for a real print at a real viewing distance. This is a far more useful limit than the theoretical limit based on pixel size. It will also allow you to work out the limit for your medium format film.

If you have a view camera you have the advantage of moving the plane of focus using tilt. This can overcome softness problems with high f-numbers by allowing you to use the lens in the sweet spot. I don't want to encourage you to spend too much money but you could look at the Nikon tilt-shift lenses in order to satisfy your landscape shooting needs.

I am sure you will enjoy either D800. Digital is a lot more convenient than shooting film. But if you are not sure you could wait and rent one (but unfortunately lose your place in the buying queue).

Alex

Hi Alex,

Thanks so much for the very thorough response!

If you were to bottom line it, are you saying I'd be slightly better off with the D800 than the D800E if I'm going to use it at smaller apertures?

I am also wondering about sharpness and how softness from diffraction differs from depth of field softness. Since digital is 3-5 times sharper than film (at a given format size, according to this article: http://www.northlight-images.co.uk/a...mits_long.html) and that quite possibly the 36MP sensor out performs the 645 film, I wonder if I can shoot at a larger aperture and still achieve the same perceived sharpness. For example, say I shot a landscape at f/10 and focus at the hyperfocal distance instead of setting the lens to f/20 and using that hyperfocal distance, the subject in the middle will definitely be sharper because of the vastly less diffraction. BUT, how will the out-of-focus area of the f/10 shot compare to the the same area in the f/20 shot impaired by diffraction? Could it be possible for the f/10 shot to appear sharper than the f/20 picture even at the nearest and farthest ends.

Digital is SO sharp that softer areas in the corners, say, will sometimes appear terrible, but only in because of the side-by-side comparison. However, if you'd compare this soft corner (side-by-side) to the center sharpness of film, the difference wouldn't be that bad in comparison. So, maybe the softer out-of-focus area shot at f/10 would be sharp enough.

What do you think?

Mike

[chicagonature]

Essentially what Nikon are saying is the diffraction limit will be more visible on the E body as you will be more acutely aware of the drop in performance than on the standard body where some of the observed sharpness has already been blurred by the filter.

Regardless of the version you choose and assuming you are going to be very critical then the image viewed at 100% isn't going to be pretty at f22. You might want to look at something like Helicon Focus to extend your DoF. Not only will this give you unlimited DoF but it will allow you to shot at the very best aperture on each lens without compromise.

Hi,

Thanks for your answer.

I actually have an even better focus stacking program: Zyrene Stacker. And, I've thought of this. However, there are two problems. Shooting one picture is hard enough. But, it seems that it'd require 3 shots. Yet, the biggest challenge is dealing with the blobby blur from the closest out-of-focus subjects that block or obscure the distant subjects. I shoot very detailed prairie landscapes and this magnifies the problem. I shot a dewy dragonfly (from the side, wingtip to wingtip) by stacking 27 shots and that was a problem. Hey, the shot was worth the work, but it was tedious. However, a detailed landscape might present the same problem. But, I do have the software. It's just a matter of trying it out.

Mike

[Colin Southern]

What do you think?

Hi Mike,

I think photographers worry far far far far too much about diffraction;

- Many worry about diffraction, and yet do a poor job sharpening an image - and yet correct multi-pass sharpening will have a far greater effect on the image.

- Many worry about diffraction, and ruin the image with sub-optimal depth of field due to not stopping down sufficiently (ie they "ruin the image trying to save the pixels")

- In the real world, nobody examines a large print with a microscope; the bigger the print the greater the viewing distance (for everyone apart from photographers, who's viewing distance is limited only by the length of their nose!) - and the greater the viewing distance, the less detail our eyes can resolve. If you want to see if your lounge carpet needs vacuuming do you stand at the door and look at the whole carpet, or do you get down on your hands and knees with a magnifying glass?

My normal print size is 44 x 22" - and I stop down to whatever I have to to get the shot I want; if that means F32 (well into diffraction territory) then "so be it" - and people tell me that my prints look just fine. In reality diffraction just isn't an issue.

[chicagonature]

Hi Mike,

I think photographers worry far far far far too much about diffraction;

- Many worry about diffraction, and yet do a poor job sharpening an image - and yet correct multi-pass sharpening will have a far greater effect on the image.

- Many worry about diffraction, and ruin the image with sub-optimal depth of field due to not stopping down sufficiently (ie they "ruin the image trying to save the pixels")

- In the real world, nobody examines a large print with a microscope; the bigger the print the greater the viewing distance (for everyone apart from photographers, who's viewing distance is limited only by the length of their nose!) - and the greater the viewing distance, the less detail our eyes can resolve. If you want to see if your lounge carpet needs vacuuming do you stand at the door and look at the whole carpet, or do you get down on your hands and knees with a magnifying glass?

My normal print size is 44 x 22" - and I stop down to whatever I have to to get the shot I want; if that means F32 (well into diffraction territory) then "so be it" - and people tell me that my prints look just fine. In reality diffraction just isn't an issue.

Colin,

That's true about viewing distance. Photographers do get a bit close.

I guess it's the resolution you would lose quite a bit of at f/22. By looking at the difference between f/8 and f/22 (with digital) it sure seems like a huge difference. So, I wonder if there might be a trade-off that can be made between choosing, say, f/11 with digital and get the same result as f/22 with film. I'm just wondering.

[Black Pearl]

Colin is right that photographers worry far to much about diffraction - the only problem in the case of the D800E is it seems to be aimed specifically at photographers who worry too much.

[Colin Southern]

Colin,

That's true about viewing distance. Photographers do get a bit close.

I guess it's the resolution you would lose quite a bit of at f/22. By looking at the difference between f/8 and f/22 (with digital) it sure seems like a huge difference. So, I wonder if there might be a trade-off that can be made between choosing, say, f/11 with digital and get the same result as f/22 with film. I'm just wondering.

Hi Mike,

Here's what I think might be a good example. Take a look at these three images (be sure to click on them to see them at the highest resolution I've made available) (which equates to about only 5% of the information I've captured by the way). What do you see?

3 Stunningly attractive ladies - Nice lighting - nice images - very saleable - professional quality?

3 images seriously degraded by horrible diffraction? (yep - they're all shot at either F20 or F22 - well into diffraction territory).








(and to add insult to injury, the last two were using a zoom lens (and we all know how they're not as sharp as primes!), and all 3 had a UV filter on the lens thus (obviously ) further severely reducing the image quality

Point I'm trying to make is that I think far too many folks get caught up in the "basic theory" (because it's something they can understand), but fail to appreciate that in "real world" shooting there are other far more significant factors that they're not taking into consideration, thus the good-sounding theories of "primes giving better IQ than zooms" - "UV Filters degrading IQ" - "Diffraction degrading IQ" etc just don't hold a lot of water in real-world situations. So we see someone shooting at ISO 400 instead of ISO 3200 to avoid noise, and end up with motion blur due to the shutterspeed being to low - shooting at F11 instead of F32 to avoid diffraction, and end up with far less pleasing effects due to a far higher shutterspeed (think andscape) - or rather than zoom a zoom lens they try to "zoom with their feet" (because everyone knows primes are sharper) - take 3 steps back and fall off a cliff (OK, I exaggerate ... it was only 2 steps back before falling off the cliff ).

[Colin Southern]

Colin is right that photographers worry far to much about diffraction - the only problem in the case of the D800E is it seems to be aimed specifically at photographers who worry too much.

It's a commercial reality, unfortunately (says Colin, waitng for his 1Dx and 4x 600EX-RTs + ST-E3-RT!). One thing you'll NEVER hear a manufacturer say is "we're well past the point of diminishing returns - this is a lot more than anyone actually needs".

[herbert]

If you were to bottom line it, are you saying I'd be slightly better off with the D800 than the D800E if I'm going to use it at smaller apertures?

Hi Mike,

As Colin has pointed out, a good sharpening workflow will really sort out your sharpness. I've had a think about this with respect to the D800 and it may be the case that the camera with the anti-alias filter will respond better to a standard sharpening workflow: capture sharpen, content sharpen and then output sharpen.

If you are using capture sharpening to correct for the anti-alias filter then the D800E will not need the capture sharpen step until the diffraction limited aperture. (The D800 will need some to overcome the anti-alias filter.) After that both will require capture sharpening but the D800E will require less.

So the D800 will look better in raw straight out of the camera, but there may be little difference after correct capture sharpening. Especially when you step back and do not examine at 100% pixel level.

I am also wondering about sharpness and how softness from diffraction differs from depth of field softness.

In my experience depth of field sharpness is more of a problem than diffraction. The hyperfocal distance is calculated using borderline sharpness requirements. If you have high standards then regions within the hyperfocal distance will still be soft. I believe it is better to focus on what is important as discussed by Keith Cooper in this article:

http://www.northlight-images.co.uk/a..._distance.html

Personally I am a big fan of telephoto lenses and shoot a lot of landscapes at over 100mm. This avoids any foreground and allows shooting close to wide open. When I do use a wide angle lens then it is because the foreground is essential to the atmosphere of the scene. In this case I am not worried by diffraction and use f22. I also take a few bracketed aperture shots so that I have options when I get home to post process the images. This is the advantage of digital over film. The cost is the time and effort to get there. Once you have arrived then you can shoot a huge stack of images just to be sure.

Alex

[herbert]

I wonder if I can shoot at a larger aperture and still achieve the same perceived sharpness. For example, say I shot a landscape at f/10 and focus at the hyperfocal distance instead of setting the lens to f/20 and using that hyperfocal distance, the subject in the middle will definitely be sharper because of the vastly less diffraction. BUT, how will the out-of-focus area of the f/10 shot compare to the the same area in the f/20 shot impaired by diffraction? Could it be possible for the f/10 shot to appear sharper than the f/20 picture even at the nearest and farthest ends.

Hi Mike,

I realised I had not addressed this question in my last post. This will depend on the blur size of a single point coming from the part of the scene that is the focus point. This is the circle of confusion. When the spread reaches the defined limit of resolution the human eye can see then that is the specified distance limit for the depth of field. You can read more about circle of confusion here:

http://en.wikipedia.org/wiki/Circle_of_confusion

So to compare how objects away from the focus point will look at f10 and f20 you need the circle of confusion of the point for the the two apertures. The equation is buried in the wikipedia article. I reproduce it here:



c = Circle of Confusion
S1 = Distance to object in focus
S2 = Distance to object out of focus
f = Focal length
N = F-number

Note: This formula is exact for a simple paraxial thin lens or a symmetrical lens. A real SLR lens will be more complex and may not be easy to calculate. So see this as an approximation.

What can be seen is that the CoC is linearly related to aperture number (N). So the same object will be twice as blurred at f10 as it will at f20. This will have more of an effect on your final image than the lack of an anti-alias filter.

Alex

[chicagonature]

Hi Mike,

Here's what I think might be a good example. Take a look at these three images (be sure to click on them to see them at the highest resolution I've made available) (which equates to about only 5% of the information I've captured by the way). What do you see?

3 Stunningly attractive ladies - Nice lighting - nice images - very saleable - professional quality?

3 images seriously degraded by horrible diffraction? (yep - they're all shot at either F20 or F22 - well into diffraction territory).

(and to add insult to injury, the last two were using a zoom lens (and we all know how they're not as sharp as primes!), and all 3 had a UV filter on the lens thus (obviously ) further severely reducing the image quality

Point I'm trying to make is that I think far too many folks get caught up in the "basic theory" (because it's something they can understand), but fail to appreciate that in "real world" shooting there are other far more significant factors that they're not taking into consideration, thus the good-sounding theories of "primes giving better IQ than zooms" - "UV Filters degrading IQ" - "Diffraction degrading IQ" etc just don't hold a lot of water in real-world situations. So we see someone shooting at ISO 400 instead of ISO 3200 to avoid noise, and end up with motion blur due to the shutterspeed being to low - shooting at F11 instead of F32 to avoid diffraction, and end up with far less pleasing effects due to a far higher shutterspeed (think andscape) - or rather than zoom a zoom lens they try to "zoom with their feet" (because everyone knows primes are sharper) - take 3 steps back and fall off a cliff (OK, I exaggerate ... it was only 2 steps back before falling off the cliff ).

Hi Colin,

Thanks for the examples and your time.

Now, with the Nikon D800/D800E Technical Guide, they show pictures that show dramatic drops in resolution between f/8 (very sharp) and f/22 (very soft). I'm shooting landscapes with lots of detail (Outdoor Photographer Magazine - November, 2011), not people with skin tones that benefit from some softness and not too much detail, so there may be a difference here. Still, I get your point and I know that the real world is what really matters. Post processing has always helped me a lot. So, thanks!

[chicagonature]

Hi Mike,

In my experience depth of field sharpness is more of a problem than diffraction. The hyperfocal distance is calculated using borderline sharpness requirements. If you have high standards then regions within the hyperfocal distance will still be soft. I believe it is better to focus on what is important as discussed by Keith Cooper in this article:

http://www.northlight-images.co.uk/a..._distance.html

Alex

Alex, thanks for the comprehensive response. I have a couple more days to wait before I decide on the D800 or D800E.

By the way, I have had good success with my own sharpening regimen (with medium format film, at least) that uses Shadow/Highlight Detail followed by "Local Contrast" and, at the very end, FocalBlade for sharpening. "Local Contrast," simply put, is setting Photoshops USM to approximately 20, 50, 0 (strength, radius, threshold) and this dramatically restores what I describe as "fidelity," "sparkle," are "life" to the picture. It increases micro-contrast and helps to create separation and, very importantly, restores the "3rd dimension." FocalBlade is a great focusing program/plug-in that allows for independent sharpening of both surfaces and edges, among other things. So, I'm probably already on my way. I've also been hearing about Deconvolution Sharpening as well, that is supposed to be better than Unsharp Mask because it tries to reverse the defocusing or diffraction process to help recover important details and edges.

Thanks again!

Mike

[herbert]

Hi Mike,

Deconvolution requires that you know how a point source of light is spread out after it passes through your optical system. This is the Point Spread Function or PSF. If you have a perfect PSF for your optics then it is very easy to deconvolve an image (with some Fourier domain transforms). However the PSF is rarely known.

Deconvolution is very useful in wide field microscopy where out of focus light can come from in front of and behind the subject (since they are transparent). This can be combined with Z-stacking to gain extra information to help the deconvolution. Also the PSF can be measured using tiny beads that act as a perfect point source of light. In addition the PSF covers a wide spread of pixels with a relatively high fraction of spread light. In this case deconvolution is almost magical in the effect it has on the image.

However I would think that the PSF of a standard camera will be very narrow. If the PSF only covers a few pixels then the deconvolution will not have much additional information to use and so the effect will be less pronounced. This may be why we do not hear very much about it in photography. It probably is not much better than using Unsharp Mask.

Interestingly the new Photoshop CS 6 contains a deblurring algorithm that appears to magically get rid of camera shake blur. This works using an estimated PSF. In the case of camera shake you have smeared the correct pixel location over maybe a dozen pixels and so you have more information in your PSF to reconstruct the image. Have a look if you have not seen it already:

http://prodesigntools.com/photoshop-...adobe-max.html

I am looking forward to actually trying the system out because I have a few images that might be resurrected from the low resolution dungeon.

Alex

[Colin Southern]

If you are using capture sharpening to correct for the anti-alias filter then the D800E will not need the capture sharpen step until the diffraction limited aperture.

Hi Alex,

Don't forget that capture sharpening isn't just to counter the softening of the AA filter though - it also counters softening introduced by digitization & demosaicing processes.

[Colin Southern]

Now, with the Nikon D800/D800E Technical Guide, they show pictures that show dramatic drops in resolution between f/8 (very sharp) and f/22 (very soft). I'm shooting landscapes with lots of detail (Outdoor Photographer Magazine - November, 2011), not people with skin tones that benefit from some softness and not too much detail, so there may be a difference here.

Hi Mike,

Landscape is my "forte" - and the same rules apply; you'll notice that they always have to show 100% crops to illustrate diffraction, and that becomes the issue.

Consider the following examples ...

1. You want to display an image online. Assuming that the camera captures something along the longest dimension in excess of 5000 pixels - and you want to display online at around 1000 pixels then by the time you down-sample, you're only using a paltry 4% of your original data ... and any diffraction has long since been sampled out. So it just isn't a real-world issue.

2. You want to make a small print (say 12 x 8") - a 5000px image will give you 416ppi print which equates to up to 270 tone changes per square millimetre. At these resolutions out eyes are totally incapable of resolving that kind of detail (including the diffraction) - so again, it just isn't a real-world issue.

3. You want to make a big print (say one of my 88" long monsters). 5000 pixels will give you 56ppi; although many will cringe at that one has to remember that a print that large ISN'T viewed up-close ... it's viewed from several meters away, and although the diffraction would be visible if viewed up close, that's just NOT how the general public views large prints (and our eyes can't resolve the diffraction from normal viewing distances) - so once again, it's just not a real world issue.

In fact, probably the ONLY time these diffraction boffins have ever seen the effects of diffraction is when looking at images in Photoshop at 100% (where you'd need a monitor 5 feet wide and 3 feet high to be able to see the whole image).

Here's an F32 "diffraction degraded" 1200px image ...

[chicagonature]

Hi Mike,

Landscape is my "forte" - and the same rules apply; you'll notice that they always have to show 100% crops to illustrate diffraction, and that becomes the issue.

Consider the following examples ...

1. You want to display an image online. Assuming that the camera captures something along the longest dimension in excess of 5000 pixels - and you want to display online at around 1000 pixels then by the time you down-sample, you're only using a paltry 4% of your original data ... and any diffraction has long since been sampled out. So it just isn't a real-world issue.

2. You want to make a small print (say 12 x 8") - a 5000px image will give you 416ppi print which equates to up to 270 tone changes per square millimetre. At these resolutions out eyes are totally incapable of resolving that kind of detail (including the diffraction) - so again, it just isn't a real-world issue.

3. You want to make a big print (say one of my 88" long monsters). 5000 pixels will give you 56ppi; although many will cringe at that one has to remember that a print that large ISN'T viewed up-close ... it's viewed from several meters away, and although the diffraction would be visible if viewed up close, that's just NOT how the general public views large prints (and our eyes can't resolve the diffraction from normal viewing distances) - so once again, it's just not a real world issue.

In fact, probably the ONLY time these diffraction boffins have ever seen the effects of diffraction is when looking at images in Photoshop at 100% (where you'd need a monitor 5 feet wide and 3 feet high to be able to see the whole image).

Here's an F32 "diffraction degraded" 1200px image ...

So, Colin, what would you get if most of everything you shoot is landscapes and, say, some macro (both requiring very small apertures)?:
- D800E with some resolution gain when shot a f/8 or smaller, but diffraction will affect a little more above f/8 OR
- D800 that protects against moire', is a tad softer when shot at f/8 or smaller aperture, but is a little better at apertures smaller.

Thanks!
Mike

[Colin Southern]

So, Colin, what would you get if most of everything you shoot is landscapes and, say, some macro (both requiring very small apertures)?:
- D800E with some resolution gain when shot a f/8 or smaller, but diffraction will affect a little more above f/8 OR
- D800 that protects against moire', is a tad softer when shot at f/8 or smaller aperture, but is a little better at apertures smaller.

Thanks!
Mike

Hi Mike,

For landscape the best camera is the one with the greatest dynamic range -- end of story! I personally wouldn't be bothered by "moire potential" - I imagine that it could be processed out with a low radius blue quite easily on the odd occasions that it's an issue.

In terms of diffraction - IMO - it's a non issue on any camera.

[herbert]

Hi Alex,

Don't forget that capture sharpening isn't just to counter the softening of the AA filter though - it also counters softening introduced by digitization & demosaicing processes.

Oops. Thanks for the reminder. I was thinking sigma foveon sensors at the time which have no anti alias filter and no Bayer filter.

Alex