Amazon Kindle Unlimited Offer: 1-Month For FREE!

D800 vs D800E sample photo at 100%

peterjones 18 5.1k 1 United Kingdom
17 Feb 2012 3:21PM
nothing wrong with a good debate provided the participants have the self control , modesty and courtesy to leave personal attacks out of proceedings; it is very easy to be brave even a legend in your own mind in front of the safety and sanctity of your monitor so perhaps substitute the bike shed with a virtual bike shed where the "combatants" can feel warm and safe.
Steppenwolf 9 1.2k
20 Feb 2012 10:10AM
OK, Strawman, I've run a few tests on a test chart. I used a Minolta 100mm macro, because it has an aperture range of f2.8 to f32, and a 12Mp Sony camera. According to the table I should begin to see image degradation at any aperture smaller than f8 - it says that the diffraction-limited resolution at f11 is 7Mp and at f16 is 3Mp. I must admit the images looked OK at f11, were starting to blur a bit at f16 and looked decidedly mushy at f32. Maybe I need a better printed chart. I certainly couldn't get anything like the dramatic effects of that series of images you linked that showed the image break up completely at f11 - I'm not sure how they did that. Still it was an interesting experiment and I'll think more carefully about using small apertures to get a greater depth of field in future - I've been tending to use smaller apertures than ideal in order to get round my camera's dodgy auto focus.

Anyway, if we take that chart as a rough approximation of where diffraction limits set in, it's interesting to note that, while your figure of a 36Mp FF camera hitting diffraction problems at about f9 may be true, it's also true to say that at f5.6 you could still gain extra image resolution by using up to 60Mp. If you're using f2.8 then the diffraction limit is about 240Mp. So you can still gain extra resolution by using even higher than 36Mp provided you use a large enough aperture. It's by no means unusual to take photos at f2.8 (where the diffraction limit is at about 100Mp on an APS-C camera). So while you're right to point out that diffraction needs to be taken into account, it's not quite the be-all and end-all that you try to imply.
strawman 17 22.2k 16 United Kingdom
20 Feb 2012 1:29PM
Diffraction tends to be the dominant issue limiting lens performance as you stop it down. At the wider open values it tends to be aberrations that limit performance and some of them get better as you stop the lens down. This explains a number of things we notice. Firstly that most lenses are better a stop or two below wide open, and that often the difference in performance between a premium and budget lens tends to turn up in the wide open performance.

Now why is diffraction limitation an issue. Lets take an example of a landscape photographer, and that is a market where a high resolution camera like the D800 will have appeal. A fairly typical example is a lens at 28mm (full frame camera) and you want everything in sharp focus from @ 1M to infinity. F11 gives you 1.2M to infinity, f16 gives you 0.9M to infinity if you hyper focus. So many landscape photographers are going to be shooting in the f11 to f16 range of apertures. Under those conditions diffraction will be limiting the total resolution, and the difference between the photo taken by a 21mp and a 36mp camera could well be quite small.

Now by comparison a crop sensor camera will be in the @ f8 lens setting @ 17mm to get a similar depth of field and a similar angle of view. the full frame cameras should have an advantage in terms of total resolution and other advantages, but the gap between the full frame camera and the crop camera in terms of resolution might not be all that big.

Now go wider and of course the lens needs to be stopped down less to get the desired depth of field so the higher MP camera will have more of an advantage. At 17mm f5.6 will give us @ 1M to infinity in focus is we set the lens correctly.

And remember most lenses work best a couple of stops down, it reduces many of the aberrations that reduce the lens performance. So many f2.8 lenses are at their best f4 to f8, and if its an f4 lens then it is not uncommon to find them best between f7 and f11.

In order to bring it all together, my view is a camera is a photo taking device whose performance varies with the parameters it works under. So in order to decide its ultimate resolution you have to calculate that in the circumstances/settings it will be used. For much of the landscapes I take I would expect a full frame camera to be working in the f8 to f16 region, so I would diffraction to be the major limit in term of resolution for many of them. And I would also expect that for many of the shots the difference in resolution between a 5d MKII and a D800 to be subtle (under those conditions).

Yes I can hear you say for portraits an excellent 85mm lens running about f4 will deliver stunning detail, do your portraits need that??? Or will you frequently be editing the shots to reduce the evidence of imperfections and make them more flattering. A professional friend used a digital MF camera (he hired it for a big project and used it for portraits.) He shot with his Nikon 35mm kit and the MF camera. For landscapes the MF kit had clear advantages (it was for a big wall photo for a company HQ) for the portraits, the existing 35mm stuff had more than enough detail and all the additional information was not wanted Smile

To wrap up, For landscapes if you want big resolution large depth of field images there is still an argument for having medium format or large format cameras. For small sensor systems you can use the shorter focal lengths and use the resulting greater depth of field to allow the use of a larger aperture, reducing the impact of diffraction.

Or will we start to see people shoot 35mm sensor cameras fairly wide open, say f5.6 and use focus stacking to get the depth of field back.

In all of this my message is to get people to think about what they are getting and recognise the point of diminishing returns when it comes. When these new cameras come out there are a couple of forms of hysteria come with it. I am trying to tread a line where the real world benefits are compared. So by doing this I am trying to give a reasoned rational for comparing the different cameras. Its not abut rubbishing types of cameras or manufacturers, but rather about getting people to step back and think, what is the real world advantage.

And can I for the umpteen time pick you up on the be all and end all bit you put in. I have said it represents the point of diminishing returns. The point where you start to say should I spend my money on getting different performance aspects improved. There is more to good image quality than number of pixels.
Steppenwolf 9 1.2k
20 Feb 2012 2:12PM
OK, fair comment. I photograph wildlife and I tend to use long lenses and big apertures (just to get the exposure I need). I usually use a couple of primes (a couple of ancient Minolta G lenses - a 300 f4 and a 300 f2.8) and I honestly can't see any difference between them stopped down or wide open (with regard to aberrations anyway). I've also recently bought (at great expense) a new Sony 70-200 G f2.8 and that certainly does work better when stopped down a bit, which is a bit disappointing. I suppose an APS-C (or smaller) camera that's stuffed with pixels works better for the stuff that I photograph but a bigger format camera (that's got a lower pixel density) works better for landscapes. Horses for courses.
User_Removed 11 4.6k 1 Scotland
20 Feb 2012 3:51PM
You two seem to be getting on better today!
strawman 17 22.2k 16 United Kingdom
22 Feb 2012 9:32AM
An interesting article from Nikon on the D800 , it covers much of what is debated in this thread. They also decided to mention diffraction.

Quote:Stopping aperture down increases depth of field, making the foreground and background sharper. Stop aperture down too far, however, and diffraction will actually cause the image to lose definition. The effects of diffraction are partly influenced by the size of the pixels in the camera image sensor, but with the D800/D800Eís high resolution the effects generally become noticeable around f/11.


Quote:The D800E offers better resolution at apertures where diffraction is not an issue.
This is because it has a less aggressive (or is it no) anti-Alias filter so that also backs the assertion that an anti-alias filter affects the level of detail a camera can record. So having an anti-alias filter will raise the f stop at which diffraction is noticed.

So maths say it happens as you go past f9, but sensors do not offer the perfect resolution and you need to go past diffraction to say you are seeing it so f11 it sounds about correct as the real world point where it is impacting a real camera such as the D800. Boy that is going to upset the pixel peepers over at DPREVIEW demanding canon knock out a 45mp 5D MKII as Nikon have stolen their scone.

So I think the camera makers are aware of the issue and hats off to Nikon for pointing out the real world issues and for making clear that there are issues in just sticking more pixels in a camera. Their points on high res cameras and depth of field control and focus accuracy plus stability are good too. So to get the best from D800, good stability, care on aperture, and precise control of focusing required. Sounds like a well thought out camera.

Me I feel Fly like a G6. Maths, physics and photography combine Grin

Sign In

You must be a member to leave a comment.

ePHOTOzine, the web's friendliest photography community.

Join For Free

Upload photos, chat with photographers, win prizes and much more.