[JimmyDranox]

Of course, for the same field of view, the DOF of a FF camera is shallow. And for the same lens, is deeper. Only that I don't remember someone to mention the same field of view before.

Anyway, the attitude, and the language used in this thread makes me to forsake any participation in this kind of discussions. I thought that we can talk in a civilized manner, without charges of misunderstanding and harsh words.

[Uluru]

Sorry, but you seem to have a fundamental misunderstanding of how to go about DOF calculations. Ether you or someone you have read is trying to reinvent the wheel, and not making a very good job of it.

I think it is other way round.

When one uses DoF calculator, it shows results when using same lens and a chosen aperture on a certain camera.
I'm not writing anything about "equivalent lenses" and "adjusting apertures". Nor DoF calculator cares about it — it simply shows what you get when you have a certain camera + lens combo. When you change the camera model / format, it recalculates the DoF for that same lens and aperture setting.

So read my post again, please. And particularly observe the paragraph where I mention comparison between a Pentax camera with a 35mm camera. In such a "test" reviewers are commonly using same lenses by a third party supplier (say Sigma 50/1.4 available for Pentax and Nikon mounts) to "neutrally" assess the AF performance of the camera. And then they miss the point.

[clackers]

To emphasize my point, APS-C offers greater depth of field than FF under the following conditions:

1. Same lens aperture used in both formats.

2. Same field of view used in both formats.

Not this again!


You've cheated by moving the camera, John - you're really talking about FoV, not DoF, and changing two variables, not one.


The DoF isn't the relative percentage of the frame that's in/out of focus, it's the absolute amount of the subject.


Check out this guy who took two shots focused on the 20 inch mark of a ruler, same position, just different sensor.


Full Frame vs Crop Sensor Depth of field myth


You'll notice the 5D doesn't have a shallower DoF, and if anything, in the 30D the '1' in '19' and '21' is less sharp.


I'm beginning to think this is a myth exposed, as Stevebrot has also highlighted the similar 'teles compress images' received wisdom!

[John Poirier]

Not this again!


You've cheated by moving the camera, John - you're really talking about FoV, not DoF, and changing two variables, not one.


The DoF isn't the relative percentage of the frame that's in/out of focus, it's the absolute amount of the subject.


Check out this guy who focused on the 20 inch mark of a ruler, same position, just different sensor.


Full Frame vs Crop Sensor Depth of field myth


You'll notice the 5D doesn't have a shallower DoF, and if anything, in the 30D the '1' in '19' and '21' is less sharp.


I'm beginning to think this is a myth exposed, as Stevebrot has also highlighted the similar 'teles compress images' received wisdom!

You have misread/misunderstood my post.

The camera position does not change. You use two different lenses.You use two different bodies, in different formats.

Therefore, the FOV remains the same.

The aperture remains the same.


The examples in the article that you linked to show two different fields of view. In other words, field of view is the major variable. It is the wrong variable to use when comparing depth of field between formats. The article is breathtakingly dumb. Frankly, I think all the fuss about APS-C vs FF DOF is dumb. Ya work with what ya got. The only reason I'm commenting is that oft-repeated misinformation tends to breed in dark corners.

The norm in comparing the DOF of different formats is to remove field of view as a variable. This is done by keeping the camera position and aperture the same, changing the camera body, and changing lens focal length. That has been the standard in photography for many, many years.

Why is that the standard? Because the goal of adjusting DOF is to produce a desired degree of apparent sharpness (a useful concept, that) in images of three-dimensional objects for output viewed at what is usually termed "normal viewing distance"- in the real world. It's not to produce exactly 8.3337 millimeters of offiicially sanctioned, pixie-dust enhanced DOF no matter what is actually in the picture.

You might find that info in less pungent form if you read the fine print associated with online DOF calculators. You might not. It does appear in lots of those nasty old printed books.

(I'll add that fully realized apparent sharpness is basically an optical illusion. It is a combination of a number of factors, including but not limited to the subject itself, lens sharpness, DOF, microcontrast, and inteplay of colours. And that's even before you start sharpening and messing with curves/clarity.)

Please read my comments carefully, and do some homework by (a) reading some books and (b) taking some pictures yourself. If you don't have access to digital FF, borrow a 35mm film body. Better still, borrow a 4x5 camera and see how much DOF you get with that. If you are right, 4x5 will have giganormously more DOF than APS-C. (Not!).

You don't have to take my word for it. So here's what you do:

1. Get an APS-C body and a FF body (or an APS-C body and a smaller format body.)

2. Choose a suitable 3-dimensional test subject and mount the cameras on a tripod at an appropriate distance. Take the usual precautions to eliminate vibration. Focus manually, using Live View if you have problems with optical viewfinders. Don't bother posting results if you don't use a tripod.

3. Select prime lenses for the two formats that have matching FOV, or use zoom lenses if you must. Either way, make the framing the same for both formats.

4. Take pictures at identical apertures, focusing on the same subject plane. While you're at it, run through the full aperture range of both lenses.
(Are you one of those folks who believe that lenses are "diffraction limited" because someone on the Web says so, and never stop down beyond 5.6 or whatever? Now is your chance to begin to learn the truth about the benefits to apparent sharpness for 3-dimensional subjects when stopping down beyond "diffraction limits".)

5. Make uncropped, identically sized prints from both formats. I recommend somewhere in the 8x10 to 12x18 range. 12x18 or larger is better if your lenses are good.

6. Place the prints side by side and view them from a reasonable distance.

7. Decide for yourself which format has the appearance of greater DOF. (greater apparent sharpness front to back from the focus plane, viewed at normal viewing distance.)

If you don't want to make prints, you can try viewing on a monitor- although monitor resolution may not adequate to fully tell the story with really sharp lenses. Just make sure you back off from the monitor a reasonable distance.

[clackers]

The camera position does not change. You use two different lenses.You use two different bodies, in different formats.

With all due respect, John - you're doing it again.

Tricky accounting.

You're changing two variables - FL and sensor size, when any valid test requires one.

FoV is *not* the issue being discussed.

[DSims]

OK, after a little bit of research, I finally figured out what is going on here:

It would be best to say: Compared to crop cameras, 35mm cameras allow for more DoF, therefore more focusing leeway, and shallower look of the OoF.

It's very simple. The DOF calculators either have you input the Circle of Confusion, or transparently insert the value for you when you select the camera model.

The numbers used for Circle of Confusion inherently factor in the size of the sensor or film plane, and they assume the same physical size for the viewed output. In other words, they incorporate the FOV in the numbers. This is why, for example, Nikon FX has a CoC of 0.030mm while Nikon DX has a CoC of 0.020mm - exactly matching the 1.5x crop factor.

But the actual measured physical depth (or distance) of field in the original subject is exactly the same, as common sense would tell you:

The DoF isn't the relative percentage of the frame that's in/out of focus, it's the absolute amount of the subject.

Check out this guy who took two shots focused on the 20 inch mark of a ruler, same position, just different sensor.

You'll notice the 5D doesn't have a shallower DoF, and if anything, in the 30D the '1' in '19' and '21' is less sharp.

Clackers' link is helpful in demonstrating what we already know should be true, if we think about it (same prime lens, same everything simply going to a different sized sensor - each of which simply crops a different amount of the lens' projected image): Full Frame vs Crop Sensor Depth of field myth



So whether you're using an FX or DX camera, the measured DOF is exactly the same for both (regardless of what any charts say), while the perceived DOF relative to the size of the entire frame (and the final viewing size) may differ noticeably. And if someone can come up with an example of where the FX image actually appears to have more DOF (although I have trouble imagining this can be done) then fine, but it will never measure as more DOF.


So, technically, sensor size alone should have no affect on the accuracy or precision of the Auto Focus, although many other, related (and possibly even required) factors certainly could affect it.

[Rondec]

With all due respect, John - you're doing it again.

Tricky accounting.

You're changing two variables - FL and sensor size, when any valid test requires one.

FoV is *not* the issue being discussed.

I do think the assumption is that you have a certain framing you want on both cameras. Let's say you are taking a portrait with a 35mm camera and a 50mm lens. If you would take the lens of the full frame camera and put it on a crop camera, you will suddenly have poor framing. You would have two options to keep your framing the same either (a) you back up or (b) you switch to a wider angle lens (a 30-ish mm lens). Either one of these things will give you more depth of field. That's not always bad, but it is true.

[DSims]

Now check it with a camera, and you might understand why I recommend people ignore DoF tables. If you're using Nikon, take the lens off your D800 and put it on your D7000, that's approximately the same pixel pitch. Now shoot with the D800 in DX mode or measure your DoF in the middle of the frame, you should have the same DoF, no matter what the DoF calculator says.

Normhead is right. The common DOF calculators are just plain wrong. As amazing as it seems, these incorrect calculators keep getting perpetrated. (But I suppose it wouldn't be the first time in our post-Renaissance era that an insistent majority of widely accepted experts has been wrong!)


Can anyone here please explain (or guess) why this has happened? All it takes is two photos of a ruler to prove that the numbers for min and max in-focus distance shouldn't change if the only variable is the DSLR body (Norm's D800 and D7000 being perfect examples here).

Why? Why, why, why???



I looked at some other forums too, and we keep giving (and getting) these idiotic answers anytime a "newbie" correctly identifies that the actual numbers in these calculators are wrong - at least as soon as you change the CoC or camera model. We've somehow bought into the nonsense ourselves, and say the calculator is theoretical and therefore not completely accurate, rather than acknowledging that one or more parameters in the theory itself (that is, the calculation) is flatly wrong, not just "a little off because real lenses and cameras are imperfect!"



And if no one can answer this honestly - and do better than the other forums - then I'm going to force all of you to join my newly formed PF Flat Earth Society chapter.

[John Poirier]

With all due respect, John - you're doing it again.

Tricky accounting.

You're changing two variables - FL and sensor size, when any valid test requires one.

FoV is *not* the issue being discussed.

No, FOV is not the issue. Therefore it should be a constant rather than a variable when testing. The use of focal length and sensor size combined to maintain fixed field of view constitutes a single constant in this case, assuming that pixel pitch is constant.

Do you seriously believe that a print made from a larger format sensor will appear to have more DOF than a same-size, same aperture, identically composed print from a smaller format sensor at the same viewing distance? I invite you to produce proof if that is what you think.

I laid out a test method to see for yourself which format produces greater apparent DOF (or perceived DOF as DSims called it).

Are you going to test this for yourself?

---------- Post added 05-05-2014 at 08:45 AM ----------

So whether you're using an FX or DX camera, the measured DOF is exactly the same for both (regardless of what any charts say), while the perceived DOF relative to the size of the entire frame (and the final viewing size) may differ noticeably. And if someone can come up with an example of where the FX image actually appears to have more DOF (although I have trouble imagining this can be done) then fine, but it will never measure as more DOF.

Correct.

[jsherman999]

Normhead is right. The common DOF calculators are just plain wrong..

I've honestly lost track of that sub-discussion - can you tell me why you think DOF calculators are wrong?

---------- Post added 05-05-14 at 11:51 AM ----------

Do you seriously believe that a print made from a larger format sensor will appear to have more DOF than a same-size, same aperture, identically composed print from a smaller format sensor at the same viewing distance? I invite you to produce proof if that is what you think.

There is no proof, because it's an impossibility

John is correct in what he says, but I don't know if someone else is actually talking about something different because I can't be bothered to parse the previous argument right now. But John is correct above.

For the same FOV, distance to subject, and F-stop, same display size, a larger sensor will have less DOF.

Using the same lens on two formats, same F-stop, same distance, same display size, the larger sensor will have more DOF. (but the two images will be radically different because they will be framed differently - so DOF comparison between the two is hardly even significant.)

Regardless, AF accuracy isn't tied to DOF as much as some people are suggesting here, although it can be a factor.

.

[Rondec]

I think keeping focal length the same on both formats is a false comparison. If you take a photo with a 50mm lens at f2.8 on a full frame camera (say, a D800), you can crop it down to APS-C size and yes, it will appear that this cropped photo has less depth of field than the full frame photo. The only place where I can see someone doing that consistently is with telephotos, where your lens might not be long enough and you are cropping to APS-C or less.

It just doesn't make sense that you wouldn't try to frame on either format as close as you can, even if it means stepping in or moving back a little.

[John Poirier]

I've honestly lost track of that sub-discussion - can you tell me why you think DOF calculators are wrong?

---------- Post added 05-05-14 at 11:51 AM ----------



There is no proof, because it's an impossibility

John is correct in what he says, but I don't know if someone else is actually talking about something different because I can't be bothered to parse the previous argument right now. But John is correct above.

For the same FOV, distance to subject, and F-stop, same display size, a larger sensor will have less DOF.

Using the same lens on two formats, same F-stop, same distance, same display size, the larger sensor will have more DOF. (but the two images will be radically different because they will be framed differently - so DOF comparison between the two is hardly even significant.)

Regardless, AF accuracy isn't tied to DOF as much as some people are suggesting here, although it can be a factor.

.

You have summed things up nicely.

From my perspective in attempting to parse discussions, I see people failing to adequately distinguish between the two scenarios you have laid out. This leads to a good deal of confusion

This is highlighted in the blanket assertion that the greater DOF of smaller sensors is a "myth". It is a myth when viewed under some conditions., and a fact when viewed under other conditions. The conditions under which the greater DOF of smaller sensors is a fact happen to coincide with what you will see hanging on a wall when focusing is done accurately. The conditions under which DOF is less for smaller sensors have some relevance when discussing AF performance, but become moot for real-world output when accurate focus is a given. I think in general the former is a more useful basis for comparison of formats.

[clackers]

Are you going to test this for yourself?

What, that bizarre proposal of yours to change both lenses and camera position, to arrive at some proxy quantity you're now calling 'apparent DoF'?



Look at the link I gave, and somebody has done it for - what shall we call it? *Real* DoF, with the same lenses, same subject and same distance from the subject - only the sensor size was different.


You haven't given a satisfactory reconciliation of their findings with your 'large format equals small DoF' assertion. And you haven't given a theoretical disproving of DoF/hyperfocal calculators or tables of the kind Uluru has pointed to (and if you think these instruments are invalid, the burden of proof is on you).


So, you'll pardon us for being unconvinced.


As Jsherman says:


"Using the same lens on two formats, same F-stop, same distance, same display size, the larger sensor will have more DOF."

That's all, John. I'm not sure everyone understands that. I didn't - I had heard differently.


As I said, FoV/perspective is a different matter entirely - not one that was being discussed, either!

[DSims]

So why are these DOF calculators wrong?

Can someone help here?


I can see there's a flaw somewhere. I think it's either in 1) the move to digital, because there's an assumption that all film will have relatively the same grain in smaller sizes, unlike sensors which can have much higher pixel densities if needed, or more likely 2) has something to do with the standard of how it will look in an 8x10 (or 8x12) print size, and that this is inherently assuming the FOV must be made comparable on different sensor/film sized (or something similar/related to this).

So what is it?

The DOF calculators state the DOF in absolute units of measurement, which change when only the crop (e.g. FX vs DX) has been changed, when we know (and can see) that in practice the physical distance that's in-focus from the actual environment being photographed is identical on both with the same lens at the same settings. But the calculator insists on reporting exactly a 1.5x difference (due to the CoC), which is wrong (in reality there is no difference in the absolute measurement).

[John Poirier]

What, that bizarre proposal of yours to change both lenses and camera position, to arrive at some proxy quantity you're now calling 'apparent DoF'?


As Jsherman says:


"Using the same lens on two formats, same F-stop, same distance, same display size, the larger sensor will have more DOF."

That's all, John. I'm not sure everyone understands that. I didn't - I had heard differently.


As I said, FoV/perspective is a different matter entirely - not one that was being discussed, either!

You seem not to have read my comments properly. I have clearly stated that camera position remains the same.

It is difficult to respond to your comments when that level of confusion exists. One of the problems with the Web is that confusion can arise in ways that could be avoided in face-to-face conversations. I do not have the patience to try to unravel this one. I think what we have is a circle of confusion.

If you agree with the part of JSherman's comment that you quoted, I presume you will agree with the other part of his comment in which he states the following: "For the same FOV, distance to subject, and F-stop, same display size, a larger sensor will have less DOF"

If you agree with both parts of JSherman's analysis, we may well be on the same page. If not- too bad, so sad. With that I consider the subject closed.