Busting the FF vs Crop DoF Bokeh Myth

boudro_ wrote:

I think you wasted your time proving something that you could have figured out using a pencil, paper and 30 seconds.

It seems others on here prefer the pencil and paper scenario, and their arguments are incorrect. I actually took the time to show solid photographic evidence that people who claim a full frame camera produces thinner DoF than a crop camera are wrong.

i think because nobody considers what you actually tried to bust as what the FF vs. APS-C DOF myth is

it was your title, it was all about busting the big myth and then you didn't touch what the vast majority consider to be the actual myth

anyway it should at least bust whatever you want to call this for those who thought there would be a difference even presented this way'

(If I read someone post that these samples disprove my own observation, I'm seriously going to blow a stack. Please read my original post before you comment further. I never, I repeat NEVER claimed similar images from a 5D and 50D will have the same depth-of-field. I disproved the myth that depth-of-field will still be different using the same lens at the same distance from subject at the same aperture.)

Why do I get the feeling people are still going to completely ignore the evidence - and my original post - and still try to make me out to be a moron?

therickman wrote:

boudro_ wrote:

I think you wasted your time proving something that you could have figured out using a pencil, paper and 30 seconds.

It seems others on here prefer the pencil and paper scenario, and their arguments are incorrect. I actually took the time to show solid photographic evidence that people who claim a full frame camera produces thinner DoF than a crop camera are wrong.

the only problem is i don't think very many consider the scenario you shot when they refer to the myth

i think either you read some stuff by a few who were unusually confused or read too much literally into what most mean when they toss around the FF DOF myth

Gene L. wrote:

Unfortunately, (correct me on this if I am mistaken) there is no standard the circle of confusion relative to a sensor's size, and without a standard, this is a subjective matter.

Really nice post, Gene! I thought I'd add to the discussion about a "standard CoC". Wikipedia has a nice discussion on that point, including a standard:

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

Most online DOF calculators use 0.03mm or 0.033mm for FF and then scale accordingly for other sensor sizes.

However, the choice of the CoC is subjective, standard or not. What is absolute is that the CoC scales linearly according to sensor size. In other words, you can choose whatever CoC you want for FF, but you need to divide that value by the crop factor to get meaningful results for other sensor sizes, assuming, of course, that we are comparing images with the same perspective, framing, and display dimensions and viewing these images from the same distance. Failure to state these assumptions is what causes so much confusion, such as in threads like this.

For example, in these days of pixel peeping, where people often compare 100% crops from images with different pixel counts, the effect is the same as changing the display dimensions of the image, so that's another important factor to consider.

Gene L. wrote:

therickman wrote:

Duh! That's not what I'm trying to prove here.

Please don't patronize me. I am a mature adult, and while I may not have earned your respect, I certainly have not earned disrespect.

I didn't mean to offend. My apologies.

Many people have argued that the DoF and bokeh using the same lens on a full frame and 1.6x crop camera at the same aperture and the same distance from the subject will yield different DoF and different bokeh... And they're flat out wrong!

While I don't recall seeing such a post making that claim, it certainly doesn't surprise me that you have come across it more than once. There is a lot of misinformation on the Internet and this stuff takes some time and a bit of thinking to absorb the underlying principles.

It's all over the place on here. People always say a FF camera has a "narrower DoF than a crop camera." That statement on it's face is untrue, and my OP has disproven this very myth. What people don't explain is that at the same field of view or composition, the full frame has a narrower DoF. That is true because of the farther distance needed by the crop camera for the same FoV (composition).

therickman wrote:

It seems others on here prefer the pencil and paper scenario, and their arguments are incorrect. I actually took the time to show solid photographic evidence that people who claim a full frame camera produces thinner DoF than a crop camera are wrong.

It is because these people assume the same perspective, framing, and display dimensions for the captured image . Under these conditions, FF will have a smaller DOF for the same f-ratio.

As I've said numerous times in this thread, why would people compare the DOF between systems by framing the scene differently and then cropping to the same FOV? Is that how people usually use a camera? Do they usually crop away 61% of the image? Or is it not that people with FF usually use a longer lens to achieve the same framing?

Does it make sense that a FF user would use the same focal lengths on FF as they would on 1.6x and then crop to the same framing? But, sure, if they did, then your test shows exactly what is to be expected -- the DOFs would be the same for the same perspective, focal length, f-ratio, and display dimensions for the FF image cropped to the same FOV as the 1.6x image. I've never seen anyone contest that point.

therickman wrote:

It's all over the place on here. People always say a FF camera has a "narrower DoF than a crop camera."

What they mean is that for the same perspective, framing, f-ratio, and display dimensions, FF has a narrower DOF than crop .

That statement on it's face is untrue, and my OP has disproven this very myth.

It is because you have misinterpreted the implied conditions.

What people don't explain is that at the same field of view or composition, the full frame has a narrower DoF. That is true because of the farther distance needed by the crop camera for the same FoV (composition).

True, they should spell out the conditions. But the implication of the same perspective and framing is, in my opinion, more natural than your test (same perspective, focal length, and cropping to the same framing), or your statement above (different perspective, same focal length, same framing).

In any event, you've not "disproved" any "myths" -- you've merely highlighted the importance of explaining the conditions of the comparison. However, you must also understand that when the conditions are so "obvious" to people, they often leave them out. For example, in almost all my posts here, I've left out the conditions of same viewing distance and same visual acuity, as these conditions are "obvious" to me.

Cheburashka wrote:

therickman wrote:

It seems others on here prefer the pencil and paper scenario, and their arguments are incorrect. I actually took the time to show solid photographic evidence that people who claim a full frame camera produces thinner DoF than a crop camera are wrong.

It is because these people assume the same perspective, framing, and display dimensions for the captured image . Under these conditions, FF will have a smaller DOF for the same f-ratio.

As I've said numerous times in this thread, why would people compare the DOF between systems by framing the scene differently and then cropping to the same FOV? Is that how people usually use a camera? Do they usually crop away 61% of the image? Or is it not that people with FF usually use a longer lens to achieve the same framing?

Does it make sense that a FF user would use the same focal lengths on FF as they would on 1.6x and then crop to the same framing? But, sure, if they did, then your test shows exactly what is to be expected -- the DOFs would be the same for the same perspective, focal length, f-ratio, and display dimensions for the FF image cropped to the same FOV as the 1.6x image. I've never seen anyone contest that point.

When people say "a full frame camera has a narrower DoF than a crop camera," that statement is false. It's a myth. I've disproven it. This statement confuses a lot of newbies, and I needed to clarify that FoV is the most important factor.

therickman wrote:

When people say "a full frame camera has a narrower DoF than a crop camera," that statement is false. It's a myth. I've disproven it. This statement confuses a lot of newbies, and I needed to clarify that FoV is the most important factor.

No, it is not a "myth", you've not disproven anything, and FOV is not "the most important factor". It is but one of the factors. As I've said more than once throughout this thread, all you've demonstrated is the importance of stating the conditions of the statement. Quoting myself from above:

It is because these people assume the same perspective, framing, and display dimensions for the captured image . Under these conditions, FF will have a smaller DOF for the same f-ratio.

And the reason that people usually fail to state these conditions, is because of this paragraph that I also wrote in my post above:

As I've said numerous times in this thread, why would people compare the DOF between systems by framing the scene differently and then cropping to the same FOV? Is that how people usually use a camera? Do they usually crop away 61% of the image? Or is it not that people with FF usually use a longer lens to achieve the same framing?

So, what you've demonstrated is that you misunderstood the premise that was implied in other peoples' statement and have made inaccurate statements yourself:

http://forums.dpreview.com/forums/read.asp?forum=1029&message=32516595

Resizing an image doesn't change the depth-of-field. That's preposterous.

and from your post above:

I needed to clarify that FoV is the most important factor.

You seem so amped up to tell people that they're wrong, rather than to understand what is actually going on and why people are saying what they are saying.

Like others here, I don't recall ever reading any discussions on these boards about what you have proved. The debates I've seen on these forums regarding the different DoFs on FF and 1.6x is usually in relation to using the same framing.

therickman wrote:

There's been so much discussion, argument, and name-calling because of the disagreements (misunderstandings) over the depth-of-field and bokeh of full frame versus 1.6x crop cameras, I decided to finally do a controlled test with my 50D and 5D.

Cheburashka wrote:

You seem so amped up to tell people that they're wrong, rather than to understand what is actually going on and why people are saying what they are saying.

So, please explain to me why DoFMaster's calculations disagree with reality. It's not taking into consideration "pixel density," or whatever else. One drop-down is for 50D/40D/30D/20D/20Da/10D, and the other for 5D (Mark II). Obviously the 50D has a higher pixel density than the 20D. But the calculator has all of them in one option, the figures are the same. Here's what it records:

5D
50mm at f/2
Subject distance 5 ft
Depth of field
Near limit 4.83 ft
Far limit 5.18 ft
Total 0.35 ft

50D, 40D, 30D, 20D, 20Da, 10D
50mm at f/2
Subject distance 5 ft
Depth of field
Near limit 4.89 ft
Far limit 5.11 ft
Total 0.22 ft

5D
85mm at f/2
Subject distance 5 ft
Depth of field
Near limit 4.94 ft
Far limit 5.06 ft
Total 0.12 ft

50D, 40D, 30D, 20D, 20Da, 10D
85mm at f/2
Subject distance 5 ft
Depth of field
Near limit 4.96 ft
Far limit 5.04 ft
Total 0.08 ft

5D
85mm at f/4
Subject distance 5 ft
Depth of field
Near limit 4.88 ft
Far limit 5.12 ft
Total 0.24 ft

50D, 40D, 30D, 20D, 20Da, 10D
85mm at f/4
Subject distance 5 ft
Depth of field
Near limit 4.93 ft
Far limit 5.08 ft
Total 0.15 ft

These numbers imply that the OOF areas would be different despite being at the same distance and same aperture. And it shows the DoF for a crop is actually narrower than a full frame at the same aperture and distance. Again, pixel density is not taken into consideration since the 50D thru 10D are all one option.

http://www.dofmaster.com/dofjs.html

therickman wrote:

So, please explain to me why DoFMaster's calculations disagree with reality.

It shows the DoF for a crop is actually narrower than a full frame at the same aperture and distance.

DOFMaster's numbers are based on the assumption that the 5D will not be cropped down to the same size as the 1.6X, nor that it will be printed 1.6X larger. Your conclusion, on the other hand, requires one of those assumptions.

therickman wrote:
So, please explain to me why DoFMaster's calculations disagree with reality.

Glad to lend a hand! Note that your conception of "reality" differs from mine. My conception of "reality" is that the most "natural" DOF comparison is with the same perspective and framing, whereas your test was with the same perspective and focal length, and then cropping to the same framing. As will be demonstrated, the online DOF calculator you cited (and all others I know of), use my version of "reality" for their computations.

It's not taking into consideration "pixel density," or whatever else. One drop-down is for 50D/40D/30D/20D/20Da/10D, and the other for 5D (Mark II). Obviously the 50D has a higher pixel density than the 20D. But the calculator has all of them in one option, the figures are the same.

As I noted above:

http://forums.dpreview.com/forums/read.asp?forum=1029&message=32518147

pixel density has nothing to do with DOF as long as you display the image at the same size (coldbivy noted that more pixels "encourages" one to print larger, and that does affect DOF).

Here's what it records:

5D
50mm at f/2
Subject distance 5 ft
Depth of field
Near limit 4.83 ft
Far limit 5.18 ft
Total 0.35 ft

50D, 40D, 30D, 20D, 20Da, 10D
50mm at f/2
Subject distance 5 ft
Depth of field
Near limit 4.89 ft
Far limit 5.11 ft
Total 0.22 ft

So far, so good. Online DOF calculators assume that the whole of the captured image is being displayed at the same size and viewed from the same distance with the same visual acuity, so we would expect a more shallow DOF from the 1.6x DSLRs as their framing is 1.6 tighter and thus magnifies the OOF areas 1.6x as much (note that 0.22 ft x 1.6 = 0.35 ft). The other comparisons all show the same thing for the same reason.

Now, using the same DOF calculator:

http://www.dofmaster.com/dofjs.html

Compare the 5D (or any other FF DSLR) at 5 ft 80mm f/3.2 to the 50D (or any other 1.6x DSLR) at 5 ft 50mm f/2 and tell me what you notice about the DOFs. Recall, once again, that the online DOF calculators are assuming that the whole of the image is being displayed and viewed from the same distance with the same visual acuity. Note also that 50mm x 1.6 = 80mm for the same AOV, and that f/2 x 1.6 = f/3.2 for the same aperture.

you test two cameras not even a (5dmkII), in only a few different environments, and think you have accomplished anything?

very disappointed

therickman wrote:

There's been so much discussion, argument, and name-calling because of the disagreements (misunderstandings) over the depth-of-field and bokeh of full frame versus 1.6x crop cameras, I decided to finally do a controlled test with my 50D and 5D.

My conclusion:

At the same distance from subject, using the same exact lens, there is absolutely no difference in DoF and bokeh between images produced by a full frame and crop camera.

Test setup consists of tripod with camera exactly 5 feet from the Ansel Adams book with center focus on the "L". John Freeman's "Photography" book is exactly 1 foot behind, "Lighthouses" book 2 feet behind. Images are JPG straight out of camera with no PP applied. Sharpening in-camera at 7. Contrast in-camera at +1.

Here is the setup shot at 5 feet with the 5D with 50mm f/1.4 lens

Here is the setup shot with the 50D with 50mm f/1.4 lens

And here are my findings

Here is the setup shot with the 5D with 85mm f/1.8 lens

Here is the setup shot with the 50D with 85mm f/1.8 lens

And here are my findings

Here's a link to my full-size test images:
http://www.flickr.com/photos/picsurephoto/sets/72157621866502388/

GTW
http://www.flickr.com/genotypewriter

It's about how you use your gear and not what gear you use,
It's not what you can afford to get but what you can afford to lose

I like your test very much, thank you! I haven't a FF camera and a cropped one to do myself comparison, and even if I know clearly the theory behind your results is interesting to see what happen in true comparison shots!! Don't misunderstand yours tests, and all you'll see is a good comparison, thank you! Roberto!

this is correct. circle of confusion is defined for a particular enlargement, which is where, if you artificially hold fl and perspective constant (artificial because that isn't how real pictures are made, where framing is a paramount concern), the op's example runs into trouble--just as other previous threads on this same topic have become similarly tangled.

the samples the op's provided are a useful exercise. they are similar to ones available all over the web, and which used to be a pretty common photography class exercise. but by failing to hold the final enlargement size constant, the op has mis-applied the lesson, and potentially misled a lot of people.

you have to make the comparison with the same FOV...

My conclusion:

At the same distance from subject, using the same exact lens, there is absolutely no difference in DoF and bokeh between images produced by a full frame and crop camera.

And that is where your test fails. When using a crop versus a full frame, you would not use the same focal length AND the same distance from the subject. One or the other would change because you would (presumably) want similar framing. This in turn would affect depth of field.

Look at it this way: With a large-format camera, take a photo of your wife in front of the Eiffel tower. Now switch to a tiny-sensor pocket digicam, keep the same focal length and subject distance. Now take a photo of your wife's bellybutton (in front of the Eiffel tower, but you can't see that because it is outside the frame). In real life, you would use a shorter focal length or increase subject distance to get the Eiffel tower in the photo.