The GetDPI Photography Forum

Great to see you here. Join our insightful photographic forum today and start tapping into a huge wealth of photographic knowledge. Completing our simple registration process will allow you to gain access to exclusive content, add your own topics and posts, share your work and connect with other members through your own private inbox! And don’t forget to say hi!

Depth of Field on the D800E

glenerrolrd

Workshop Member
One of the hallmark techniques of street shooting has always been using hyper focal distance . On a Leica M you place the infinity mark at the F stop and you can see that everything from your navel to the moon should be in focus . Close inspection would show that this is only true in the most liberal definitions of “in focus “ . With larger sensors (and it appears smaller pixels ) DOF seems to have disappeared .

A recent blog post on Diglloyd (on the free blog) presents a theory ? that DOf should be calculated by using a circle of confusion that mirrors the pixel size .

You could easily see this with MF at wide apertures but I had always considered this a difference in focal lengths . Using an M9 and a S2 the pixel pitch is close and so a 70mm lens has the same DOF on both cameras (but different equivalent FOV)..so all I was seeing was the fact that a 70 on the S2 was a 50 on the M9.

Street shooters are taught to get close and use wide angle lens .....the more you shoot the closer you can get ......but as pixel pitch gets below 5 ... you just don t have adequate DOF to cover subject movement . But the same concept applies to many different subjects .

Add in that optimum aperture is often between f4 and f5.6 and I ve found my limiting factor with the D800E .....
 
Last edited:

donaldt

New member
this is making very little sense to me
DOF should be calculated only based on aperture, focal length, and distance
the reason Medium Format has a shallow DOF is because with the same aperture and focal length, to get the same composition you must get closer and therefore shallower DOF
the only thing that changes with higher pixel counts is that it becomes less forgiving with motion blur and poor focus, and I think sometimes it is being wrongly associated with DOF
 

Guy Mancuso

Administrator, Instructor
Yea and I got motion blur with the 200 f2 at 1/125 on a monopod and I'm a rock I don't move. 36mpx in 35 is better be careful.
 

Oren Grad

Active member
I understand the point Lloyd is getting at, and I more or less agree with it.

Calculated DOF depends, among other parameters, on the assumed threshold at which the circle of confusion becomes too small to discern. But the CoC traditionally assumed for enlarging from film depends heavily on the fact that film grain and its side effects obscure fine distinctions in focus. In practice, digital capture today demands a more stringent CoC criterion. The more pixels you have for a given format, the worse it gets. And if you're heavily into pixel-peeping, it gets worse still.

When I've used an A900 or a 5DII, it almost felt like using a medium format film camera - at least a 645, if not larger - because the apparent depth I was used to having from shooting Tri-X on a 35mm camera wasn't there any more. Shooting 2/3 the focal length on an APS-C format digital camera feels much closer to what it was like shooting 35mm film. I'd probably have a devil of a time trying to do casual walkaround snapshooting with a D800/E, even apart from the bulk and weight of the camera.

You can get a little bit of this effect just by switching from a fast, grainy film to a slower, sharper, less grainy film. But it really hits you between the eyes when you move from film to high-MP digital capture.

Moving from 35mm to a FF DSLR certainly preserves the field of view of your lenses, and there are obvious reasons why that's appealing. But in the end, the experience and the results are still very different. The school of hard knocks has taught me that using the same lenses with a sensor that's the same size most definitely does not preserve the same rendering of the scene.

Repeat after me: "It's a new medium, learn to use it on its own terms..."
 

Jan Brittenson

Senior Subscriber Member
The CoC you want depends on the number of pixels, not pixel size. More pixels means you magnify less at a given print size.

A 70mm on the S2 has less DoF at the same f stop as a 50mm on the M9. It gets the same DoF at one greater f stop than the M9. At the same DoF it diffracts more.

Going bigger is ALWAYS a losing proposition other than for movements - always has been, always will be. The ONLY benefit is more photons, implying less noise. But then 35mm has MFD and the hybrid S2 beat in that regard as well (which suggests photon shot noise is not yet dominant).
 

GrahamWelland

Subscriber & Workshop Member
Welcome to the world of 4.9um pixels ... did you really think that the physics would be different for Nikon 35mm as it is for every medium format camera? That tiny size of pixel will absolutely have a knock on effect in terms of diffraction, apparent sharpness due to motion being rendered, DoF and so on. Luckily the T/S lenses are retro-focus and far enough away from the sensor otherwise I'm sure you'd be having to worry about lens cast colour and light fall off effects too.

If you are pixel peeping at 100% on screen then basically yes I believe that the CoC probably is 1 pixel in size with it's associated effect on what you deem to be acceptable focus for DoF determination. If you break away from the monitor and print then you can be more generous about the CoC you use, depending of course on how far away you pixel peep your prints too. Consensus seems to be to go 1.4x or sqrt(2)x pixel pitch as a better approximation for CoC. Unfortunately when you use these CoC's you rapidly find that your acceptable DoF diminishes away to almost nothing pretty quickly, although again I stress that it really only matters depending upon the size of image that you view and from where. Scale down to the web or even 8x10's and your prints will seem to have plenty of DoF and the effects of diffraction etc will start to disappear also. Print billboards or stare at 100% images on screen and you'll be fighting with diffraction and softer images rendered as apparently smaller acceptable DoF.
 

glenerrolrd

Workshop Member
this is making very little sense to me
DOF should be calculated only based on aperture, focal length, and distance
the reason Medium Format has a shallow DOF is because with the same aperture and focal length, to get the same composition you must get closer and therefore shallower DOF
the only thing that changes with higher pixel counts is that it becomes less forgiving with motion blur and poor focus, and I think sometimes it is being wrongly associated with DOF
Purpose of the post is to discuss diglloyd s POV (supported by Zeiss) that traditional measures of DOF don t factor in pixel size . Your observations appear to be conventional wisdom based on either film or pixel dimensions that are similar .

I will have to look to be sure but I think the D800 pixel size is below 5 and the Leica S is about 6.8 for example . I do not know if this is enough to materially affect DOF . But I do know that DOF with either camera is much much smaller than the old tables provided.

Is pixel size relevant ?

Is DOF a consideration in determining which camera system fits a specific application?
 

glenerrolrd

Workshop Member
Welcome to the world of 4.9um pixels ... did you really think that the physics would be different for Nikon 35mm as it is for every medium format camera? That tiny size of pixel will absolutely have a knock on effect in terms of diffraction, apparent sharpness due to motion being rendered, DoF and so on. Luckily the T/S lenses are retro-focus and far enough away from the sensor otherwise I'm sure you'd be having to worry about lens cast colour and light fall off effects too.

If you are pixel peeping at 100% on screen then basically yes I believe that the CoC probably is 1 pixel in size with it's associated effect on what you deem to be acceptable focus for DoF determination. If you break away from the monitor and print then you can be more generous about the CoC you use, depending of course on how far away you pixel peep your prints too. Consensus seems to be to go 1.4x or sqrt(2)x pixel pitch as a better approximation for CoC. Unfortunately when you use these CoC's you rapidly find that your acceptable DoF diminishes away to almost nothing pretty quickly, although again I stress that it really only matters depending upon the size of image that you view and from where. Scale down to the web or even 8x10's and your prints will seem to have plenty of DoF and the effects of diffraction etc will start to disappear also. Print billboards or stare at 100% images on screen and you'll be fighting with diffraction and softer images rendered as apparently smaller acceptable DoF.
Your examples seem extreme to me .. I shoot with a Leica S2 kit and m9 s except for sports where I use Nikon. So I am familiar with the additional requirements of high resolution and MF . My standard is that if I can see it on a screen at normal size ..then it matters. Of course my files look sharper when printed but much of what I do is shown on a screen . Your requirements may be different ..nothing wrong with that.

lloyd s discussion exactly matches my experience with the D3s and the D3x . Same lens and same technique and the D3S was always easier to nail the focus . Was never certain ..is that just because you can see the plane of focus better in a higher resolution file or because the DOF is actually smaller ?

I think a good assumption for this discussion is that those acquiring D800 s will be seeking the IQ benefits of additional MP s . Haven t seen much discussion on the limitations on DOF .

I think you will start to see complaints of “I can t get anything sharp “ as the camera rolls out to a greater number of photographers . (This camera will be a B%^%%^^ to shoot action with long lenses. ) But I had never seen anything that lays out the relevance of pixel size in determining DOF .

Guy you should be seeing this in spades in doing events . How do you get a group of people in focus ? Maybe no different than with MF ?

Understand that this may be old news to MF shooters but I didn t expect the extent I am seeing. The D800 seems like it has less DOF than my S2 after adjusting for FOV equivalents . How many D800 buyers are expecting such demanding requirements ?
 

glenerrolrd

Workshop Member
The CoC you want depends on the number of pixels, not pixel size. More pixels means you magnify less at a given print size.

A 70mm on the S2 has less DoF at the same f stop as a 50mm on the M9. It gets the same DoF at one greater f stop than the M9. At the same DoF it diffracts more.

Going bigger is ALWAYS a losing proposition other than for movements - always has been, always will be. The ONLY benefit is more photons, implying less noise. But then 35mm has MFD and the hybrid S2 beat in that regard as well (which suggests photon shot noise is not yet dominant).
Then you disagree with LLoyd ? I will let you tell him?

If I shot an 50mm lens on an s2 ,m9 and d800 which would have the least DOF and why? In your hypothesis ..the M9 would have the most because its 18mps , the s2 and the D800 should be similar as they create similar file sizes . Or am I misunderstanding something .
 

Guy Mancuso

Administrator, Instructor
Roger no doubt for me I feel I lost a stop or so of speed with the bigger mpx from shooting similar events with other smaller full frame sensors. Now DOF is a little tougher for me since I'm so used to very little on MF but the D800 is better that I can say , how much I can't pin it totally down against other smaller full frame sensors. I seem to be doing pretty good at F6.3 that's carrying pretty well with the 35 especially. But this is just field feelings so nothing I can pin down in calculation if you know what I mean.
 

Guy Mancuso

Administrator, Instructor
I will say this for sure this cam just takes more care in your technique. You just can't point it at something and expect perfection. So you really do have to pay attention as you would with MF. Now it is easier to shoot than MF and that does help but from a MF shooter that shoots a lot this is pretty similar in watching yourself , you just can't be sloppy. So yes I can see where some people will complain they can't get anything sharp or stuff like that . This is harder to shoot than the D7000 I bought 2 days ago and been using along with the D800 for this gig I'm on. Which I need to be leaving for right about now. I was up late last night my SSD failed on a laptop fall yesterday. The spinning drive is fine but the SSD drive crapped out. Go figure

Remind my to shoot Murphy in the ***. Lol
 

Bob

Administrator
Staff member
Nothing has changed at all and this is not new.
COC is based on the angular resolution of the human eye and the size and viewing distance of that detail.
If you view your images at 100% routinely and at a distance where you can see a pixel, then for sure, the COC approximates the pixel pitch. The 1.4 pixel pitch is related to the fact that square pixels are 1.414 (Sqrt(2)) times the pixel pitch apart on the diagonal.
For optimum crop-ability than sure, keep your COC down to the minimum. If on the other hand, you make only 5x7 prints, than a COC of more traditional dimensions might suit you.
So therefore, COC is literally in the eye of the beholder and the intended use of the image.

Sharpness beyond the resolution limits of the eye are maybe worth bragging rights but not much more.
-bob
 

Jack

Sr. Administrator
Staff member
Just to add a bit more to Bob's cogent response -- folks should understand if they don't already that *the* significant variable in the complete DoF formula is personal choice of CoC constant. With digital and our propensity to view images always at 100% to check for focus, sharpness, detail etc, the defacto CoC has become pixel pitch -- however, this is not a mandate! You are free to use the same CoC you used with your 12MP camera if you want :). Using the D800 pixel pitch for traditional film capture CoC would be roughly equivalent to scrutinizing every negative or transparency you shot under a 70 power microscope. How many of you ever routinely did that?

Finally, if you put a 50mm lens on an M9, a 50 on a D800 and a 70 on an S2 and shoot the same exact framing from the same exact position using f2.8 on the M9 and D800 50's and f4 on the S2's 70, then print each file to the same largish size, say 20x30, you will note the following: 1) ALL THREE will have IDENTICAL DoF in the print; 2) The D800 and S2 prints will look "smoother" (and similar/identical) in tonality than the M9 print; 3) and you may or may not notice certain lens characteristics/rendering differences you normally do see when viewing at 100% on your computer screen. Jack's corollary: It is surprising just how superbly certain sub-par lenses can perform in print when they totally rot at 100% view on a monitor -- and even when used in front of sensors that are sharper everywhere than the lens is.

Flamesuit donned! :D
 

GrahamWelland

Subscriber & Workshop Member
Jack,

I don't think that you need a flamesuit - as predicted the questions are coming about why can't I get 100% sharp images with the same DoF at f/16 or f/22 (or to be honest f/8 -> f5.6) that people used to with 10 or 12mp cameras. Why are identical images looking softer on screen now with 36mp compared to the same shot with the same settings taken on a D3 or D700 etc etc.

Nothing has changed - you need to treat shooting a D800 (as indeed you did the D3x) as if you are shooting with bigger formats such as MFDBs. It's still all about the tiny pixel pitch and physics/optics.
 

Jack

Sr. Administrator
Staff member
Exactly Graham -- it's the same thing we dealt with when moving up in MF backs: Once backs hit full-frame 40MP, the lens faults started being "visible" and only got more visible as pixel counts incrteased.
 

Guy Mancuso

Administrator, Instructor
Kind of my just need to be paying attention comment. My normal stage shooting bumped up some with shutter. Now little secret here I won't be processing these at full res anyway but downsizing them and I will come back with a little less noise and get appearance of detail. It's a little cheat but look what I'm starting with a big mpx cam that's far easier to handle and someone tell me where I can get a 200 f2 in MF. Not happening lol

Btw I can tell you this with certainty very few people will ever look at your images at 100 percent. Need to get over the screen as it is fooling you a little in Print these images are much better. Seriously don't look at them at 100 percent 50 or 67 percent is actually a better judge.
 

Bob

Administrator
Staff member
A decent guideline is to view them at 2-3 times print size on a typical monitor that displays at around 100 pixels per inch.
-bob
 

Jan Brittenson

Senior Subscriber Member
A 645 CoC of 0.045 is equivalent to a 135 of 0.025 when viewed in print.

Go to Online Depth of Field Calculator (which defaults to these CoCs) and check the DoF of a 645 80mm lens at f/2.8 vs 50mm on 135. There's no question 50mm on 135 has better DoF. They match when the 80 on 645 is stopped down to f/4.

Smaller formats have better DoF at the same f-stop. That's a mathematical certainty.
 
Last edited:

GrahamWelland

Subscriber & Workshop Member
Circle of Confusion (CoC)
Definition: "A group of photographers sitting around trying to understand Depth of Field"

Michael Reichmann summed it up perfectly.
 

Shashin

Well-known member
The CoC is a product of the format--how many pixel you have is irrelevant. The Zeiss definition is 1/1500 the format (or crop) diagonal. However, the CoC is also subjective and you can, and some do, create a different value for the CoC to help them visualize DoF.
 
Top