Factors Affecting Depth of Field

spraynpray

"Don't believe everything you read on the Internet.
--Abraham Lincoln"

That IS funny!

FlowtographyBerlin

Wow, what a mess. Some of you guys are really confusing many things at once, hence leading to correct conclusions that appear wrong and vice versa. One is that you're varying sensor size now, without taking the different crop factors into account in the sense of what they do for the image. It's like saying there's a different DOF for the same image (whatever sensor format) full/uncropped and after you cropped it. But anyway.

More important:

What's new here is the claim that there is a noticeable perceived difference of DoF between a D3200 and D7100 in otherwise identical situation.

Without example pictures from a D3200 and a D7100 demonstrating this difference so we can all see what's actually going on, I feel we're just rehashing the old thread again.

Exactly. So what about @DaveyJ posting some images to illustrate what he's referring to?

Golf007sd

A few videos that might help those that don't fully understand what Circle Of Confusion (CoC) is, as well as, DoF clarification.

brownie314

From the thread "How to make the Nikon D3200 really perform!" this topic emerged.

DOF is a function of lens only. Now whatever sensor you put behind the lens is your business - you can take all of the image circle, or only a tiny part of it - but that doesn't change the optics of the lens. DOF is set by the parameters of the lens.

spraynpray

Funny how I keep saying apparent and few people are even trying to see where I'm coming from. OF COURSE I understand the basic facts about DoF, but I am also opening my mind to other possible apparent influences on the DoF.

brownie314

It is pretty simple. DOF is a function of focal length, aperture and distance to subject (how far away you are focusing). Take a 50mm lens as an example. If I put it on a FF camera, and take a picture of a small child at a 10 feet away, at f/1.8 - that combination of FL, distance to subject, and aperture will result in a certain depth of field. Now take the same lens and put it on a crop sensor. Now I will have to take a few steps backwards to get the same field of view I had on my FF camera. So the distance to my subject has changed! Now instead of focusing on something 10ft away, I am focussing 20 ft away. So I have the same FL, same aperture, but different (greater) distance to subject. And since distance to subject is one of those things that effect DOF - my DOF will be different (in this case larger DOF). There, all of the apparent this, sensor size that stuff explained.

FlowtographyBerlin

@spraynpray: I hear you. Now. Somewhat. I think. No actually, I really do.

It's also funny how everyone keeps repeating the same stuff. :-)

Ironheart

Sigh, some of you folks are just not getting what this discussion is really about. What we are discussing is that the CoC is not a fixed value based solely on sensor size. If you change the CoC value from .030 to .025 in your favorite DoF calculator, you will notice the difference in the "calculated" DoF. There are only four inputs into a DoF formula and we are holding three of them constant: Focal length, f/stop and distance to subject are all the same. The only thing that is changing is which camera/sensor is used to make the photo, but in this case the sensor size is not changing (DX). If you want to talk about DX vs. FX and apparent DoF changes, go back to the other thread where this has been hashed to death:
http://forum.nikonrumors.com/discussion/510/dof-fx-vs-dx/p1

I'm going to post some photos of exactly what we are discussing here with two different DX sensor cameras. It would be great if DaveyJ could also, since he sparked this discussion back on the D7100 thread :-)

SquamishPhoto

If this was something legitimately observable then don't you think it would already have been discussed at large in a variety of forums and webpages? No articles from anyone professing that the lack of AA on the 7100 changing the perceptual nature of DOF, at least that I could find, just speculation from you guys from one post from DaveyJ - add to that he didn't included photos or even the most basic shooting information. Knowing that, does anyone here really think that they've stumbled onto to something this significant before the rest of the photographic world? I don't want to poo poo on this, but if you guys have to rely on "perceptual" references to carry your argument then don't you dare ever criticize anyone who credits DXO with reliability. :]

brownie314

Sigh, some of you folks are just not getting what this discussion is really about. What we are discussing is that the CoC is not a fixed value based solely on sensor size. If you change the CoC value from .030 to .025 in your favorite DoF calculator, you will notice the difference in the "calculated" DoF. There are only four inputs into a DoF formula and we are holding three of them constant: Focal length, f/stop and distance to subject are all the same. The only thing that is changing is which camera/sensor is used to Ithe photo, but in this case the sensor size is not changing (DX). If you want to talk about DX vs. FX and apparent DoF changes, go back to the other thread where this has been hashed to death:
http://forum.nikonrumors.com/discussion/510/dof-fx-vs-dx/p1

I'm going to post some photos of exactly what we are discussing here with two different DX sensor cameras. It would be great if DaveyJ could also, since he sparked this discussion back on the D7100 thread :-)

IH,
Yes, you are right. I assumed this was another one of those forums threads where everyone is confused (or has a COC about...hahaha) between FF and crop DOF. I have seen so many, I assumed this was just another one. But yes, this is something different. My apologies.

Since I am here, my 2cents, for what its worth - you are splitting extreme hairs. But that is what we are all here for right. Let me get my popcorn and sit back and watch.

spraynpray

"Since I am here, my 2cents, for what its worth - you are splitting extreme hairs. But that is what we are all here for right. Let me get my popcorn and sit back and watch."

@brownie314: We are indeed, but then the DoF is so small at macro distances that hairs need to be split.

For the non-believers:

OK, I'll spell out what I'm thinking: The factors I keep talking about affect the CoC - if you double the sensor resolution or improve the lpmm of the lens, then the CoC is affected (improved) so the distance from the focal plane before the CoC gets too large is greater = larger depth of field. Y'all are not considering the whole system working.

Please lay out your contrary argument similarly clearly so we can consider it.

Correlli

The factors I keep talking about affect the CoC - if you double the sensor resolution or improve the lpmm of the lens, then the CoC is affected (improved) so the distance from the focal plane before the CoC gets too large is greater = larger depth of field. Y'all are not considering the whole system working.

Please lay out your contrary argument similarly clearly so we can consider it.

Either I misunderstand you or I don't agree.

Since we are splitting hairs: the factors that affect the CoC are the magnification of the image and the viewing distance. These determine what the CoC needs to be. Resolution of the lens or sensor might be limiting factors (that prevent the system to achieve the required CoC) but do not affect the CoC. But then again: English is not my native language so I might just be misinterpreting words...

Msmoto

Splitting hairs...yes.....but for some of us this is an interesting topic. Practically speaking I am afraid I just don't do this, but shoot with an overall image quality/impression in mind.

brownie314

Splitting hairs...yes.....but for some of us this is an interesting topic. Practically speaking I am afraid I just don't do this, but shoot with an overall image quality/impression in mind.

Agreed - this is interesting to me too, but like you, not so practical. I can only see this coming into play if someone were doing extreme crops at extremely thin DOF. Because if you are not doing the extreme crops, then I don't see any of this making any difference because uncropped, it is hard to see the difference between, for example, the D7000 and the D7100.

spraynpray

Hmm, by extension I would have to argue that doing extreme crops reduces resolution therefore DoF....

roombarobot

Hmm, by extension I would have to argue that doing extreme crops reduces resolution therefore DoF....

Likewise, a crop sensor does the same thing, it essentially just crops the center of a full-frame image.

In the very simplest case, if sensor size actually changed the inherent depth of field, then when I switch my D800E from FX into DX mode, the depth of field should change. Of course it doesn't!

Ade

"Extreme crops" does affect CoC (and DoF), not because it reduces resolution but because it ultimately affects magnification if printed / enlarged to the same size of an uncropped image.

The CoC value is an approximation where a photographic "dot" is already so small that the average human eye cannot distinguish anything smaller. It is, in fact, the "lppm" of the human eye -- customarily chosen to be 5 lppm at a distance of 10 inches. Increasing resolution beyond this point is "meaningless" -- we're already at the limit of the human eye.

In the FX era, we've tripled resolution from the D700 (12mp) to the D800 (36mp) yet the perceived DoF remains the same. The D800E without AA filter also did not cause a change in DoF when compared to a D700.

In comparison, both the D3200 and the D7100 have DX-sized sensors of around 24mp. I'm as curious as anyone to see any difference in DoF.

SquamishPhoto

In the FX era, we've tripled resolution from the D700 (12mp) to the D800 (36mp) yet the perceived DoF remains the same. The D800E without AA filter also did not cause a change in DoF when compared to a D700.

I think Ade nailed it. If this concept were real then it would've clearly bore itself out in this newest batch of FX DSLR's, and it very clearly did not. And I think that snp wants very badly to believe the logic in his head is sound, but I don't see science helping him in any way.

Paperman

@roombarobot

Likewise, a crop sensor does the same thing, it essentially just crops the center of a full-frame image.

In the very simplest case, if sensor size actually changed the inherent depth of field, then when I switch my D800E from FX into DX mode, the depth of field should change. Of course it doesn't! -

It does .... Just as I showed it in the DOF calculator before. .

The image is no longer the original image - it is magnified. The object has to cover same area in frame for DoFs to be same.

Correlli

Just being curious: does anyone really use those dof calculators in real life? I mean, does anyone really measure the dof he/she wants as well as the subject distance and enters those values into one of those calculators to set the f-stop according to the result?

Paperman

I doubt it especially if one is after large DOF shooting wide angle . Everything is in focus from 1.5-2m to infinity anyway even at @ f5.6-f8.

I downloaded an Apps for it 2 years ago but never used it.

Ade

Many landscape photographers use DoF calculators to set the hyperfocal distance -- an old practice which I don't subscribe to.

I admit I've consulted the DoF calculator a few times "after the fact" -- usually after I screw up a bunch of pictures by having the DoF way too thin!

spraynpray

OK, now some of us are starting to open our minds.

I am talking about a real world camera system here not a theoretical calculation done using data on paper which is not based on a system.

Let me go from a different angle: If we accept that a 1Mp Bayer array type sensor on a full frame camera would not give enough resolution to be able to give you a CoC of a current accepted diameter unless you stood back a long way to view the image, at what point in Mp's does it occur that we have enough Mp's to attain that CoC diameter? If this question is valid, I also ask the same question about lens resolution. If we can see that there is a normal viewing distance and that using that distance the standard CoC (that cannot be attained in the above example), are we saying then that improving the system resolution would attain the resolution required to see the CoC? Of course we are - and then if we take lens and sensor performance beyond the ability to see the standard CoC does not bring any improvement to the DoF? In my mind it may.

heartyfisher

If the theory that the Sensor resolution effects Perceived DOF then we really dont need the D3200 vs D7100 pictures to evaluate (though it would be good) any other sensors with different resolutions will do .. eg the D700 vs the D800e. Can some one provide test images from these 2 cameras something simple like shooting a ruler slanted at 45 degrees using the same lense and aperture and distance. in my mind its pretty obvious that we will see a difference in DOF.

roombarobot

@roombarobot

Likewise, a crop sensor does the same thing, it essentially just crops the center of a full-frame image.

In the very simplest case, if sensor size actually changed the inherent depth of field, then when I switch my D800E from FX into DX mode, the depth of field should change. Of course it doesn't! -

It does .... Just as I showed it in the DOF calculator before. .

The image is no longer the original image - it is magnified. The object has to cover same area in frame for DoFs to be same.

@paperman, so, you are really claiming that if I take a picture with my D800E in FX, then switch it to DX-mode and keep everything else the same, that my DoF will change?

That would be equivalent to saying that every time one crops a photo in post-production it changes the DoF. Of course it does not.