Busting the FF vs Crop DoF Bokeh Myth

DOF for FF is much shallower for equal FOV.

Thanks for the confirmation.

Gene

therickman wrote:

Cal Dawson wrote:

All the OP said was if all things are the same the end result will be a similar DOF there was no mention of changing lenses/ distance etc that was brought up later by those that failed to see the intent. I don't believe the OP stated that if you change distance (FOV) the DOF stayed the same....

That's exactly it. Jeesh, someone on here understood my OP.

I recall awhile back where Cal was trying to say the same thing as you (but without the photos/experiment), many people wanted to argue and twist his intentions as well, just as they have done with you.

I theorize that dpreview attracts photographers and arguers!

Anyway.... kudos to you for giving evidence to refute a myth that "I" have also read on this forum! Maybe some people understand the math and physics involved but many don't. And keeping it simple in the real world with photos does help some people.

Ohhhh BTW... you unfortunately have now experienced a moment of the dpreview "Mob Mentality"... :-).

-- hide signature --

Insert obligatory quote here...

http://www.picsure.net

Kuivaamo wrote:

This enlargement "hoo-ha" is as much an integral part in determining DoF as viewing distance, aperture, focal length and shooting distance. You need to know all of them in order to calculate DoF. You cannot ignore changing any one of them.

You're wrong. You can, in fact, ignore changing viewing distance within normal use case ranges. Dofmaster confirms this, as do other calculators. Enlargement is the second least important factor, in that you generally have to double or triple the image dimensions to create an appreciable DoF difference at normal viewing distances. Changing the aperture diameter by 10% has a much clearer affect on DoF than going from a 4x6 to an 8x10. The fact that, in a precise mathematical sense, you cannot "ignore changing any of them" is misleading because, in the real world, some of the factors are barely perceptible except at the absolute edges, and among the rest some create difference more quickly than others. Ignoring this is just stupid.

Doing the test as OP did in order to prove Dofmaster calculations wrong is utter fallacy. Just think about it. Dofmaster calculations assume a constant output size. If you don't keep your output size constant, how can you expect your results to agree with Dofmaster?

You should clarify what you mean by "assume a constant output size", because it's not true unless qualified. It assumes a constant ratio between output size and viewing distance , precisely because within normal ranges the difference is negligible.

therickman wrote:

So, you're saying if I enlarge an image taken from an 11MP FF 1Ds to 20X30, and compare that to a 20X30 print taken from a 15MP crop 50D, the 50D's image will be blurrier?

If you take a picture of the same scene from the same distance with the same framing and same f-ratio, and display the entire captured image at the same size, and view from the same distance, then less of the 1Ds image will appear to be in focus compared to the 50D image. That is what people mean when they say "FF has less DOF than crop".

therickman wrote:

Oh yeah, because people refuse to read through things before they jump to conclusions. Such a shame...

You are arguing two entirely different points here:

a) At the same distance , aperture, FL, DoF will be the same between sensor sizes.

b) At the same FoV , aperture, FL, DoF will be narrower on the larger sensor.

Both are correct.

I see the problem being that when someone responds to your a) argument, and you respond with b) counterpoints, and vice-versa. Your original post shows you misunderstood the 'myth' from the get go (in a well done fashion, I might add), and you have only gone on to show that the proper definition of the 'myth' is indeed true (well done, again, really). Then when people agree with your findings, you argue with them.

At least, that's my conclusion after reading through things.

if the FOV is the same then other things are not equal, In your eyes this is not valid because the FOV is not the same. I pretty well crop every picture so the FOV argument does not apply to my shooting. It's all a matter of what you want it to be....

Kjeld Olesen wrote:
If the FOV is not equal, then all things can not be the same.

If all things are not the same, the only sensible comparison - and what the "myth" is all about - is that the use of equivalent FOV and perspective (distance) results in a more shallow DOF on FF when using the same f-stop - not apperture diameter.

This is clearly a myth busting thread that has gone totally ballony, so I'm bailing out no matter what.

Cal Dawson wrote:

that was not the point of the post, When all things are equal except the FOV the resulting DOF will be the same....
--
Cal

Put a Canon to your head, You deserve it....

http://funshots.smugmug.com/

Ok, so do you people agree that most people print at 4X6 or somewhere thereabouts? If so, great. That's the size most of my wedding clients print.

Many people bought the 5DmarkII because the 21MP allows more crop-ability. If you shot a picture with the 50D and printed it at 4X6, you'd be able to get the same exact 4X6 image from the 5DII (actually, from the 5D as well) if you cropped it to an identical FoV (composition) as the 50D shot, using the same lens, same aperture, and same distance from the subject. As some have suggested, because of the inability to change distance from the subject, because of a fnce or scaring away an animal or seating restrictions, etc., the need to crop a larger FF image is quite common. But using the FF over a crop doesn't change the DoF using the same lens at the same aperture at the same distance, especially when all you're going to print are 4X6's. And I've proven that here.

Pretend the following pictures are hard copy prints tacked on a wall.

Here is an exact 4X6 image taken from the crop 50D (on my computer screen, it measures 4X6):

If you have a full frame camera, the same exact image can be extracted with absolutely no change in DoF. Here is the image from the 5D:

Here is the 50D image over the 5D image:

Now, here is an exact 4X6 image cropped from the 5D:

And here is an exact 4X6 image taken from the 50D (no crop):

Gee wizz, they look identical don't they!?

So, to prove my point, there is no difference in DoF with a FF camera or a crop camera at the same distance using the same lens at the same aperture.

Ribber wrote:

therickman wrote:

Oh yeah, because people refuse to read through things before they jump to conclusions. Such a shame...

You are arguing two entirely different points here:

a) At the same distance , aperture, FL, DoF will be the same between sensor sizes.

b) At the same FoV , aperture, FL, DoF will be narrower on the larger sensor.

Both are correct.

I see the problem being that when someone responds to your a) argument, and you respond with b) counterpoints, and vice-versa. Your original post shows you misunderstood the 'myth' from the get go (in a well done fashion, I might add), and you have only gone on to show that the proper definition of the 'myth' is indeed true (well done, again, really). Then when people agree with your findings, you argue with them.

I think there's some truth here. And it emphasizes that not being clear about FoV vs FL and distance is a very bad thing.

There's a truth that matters here, in that there is indeed a myth perpetuated on DPreview forums and elsewhere that FF = less depth of field. It doesn't. It equals the ability to get less depth of field without having to change your composition (much) by changing other factors. This can be a useful thing. For some people it may even be a critical thing. But FF on its own DOES NOT HAVE LESS DEPTH OF FIELD in some inherent way with other factors held constant--it gives you the flexibility to get there, if you need to. People should understand what that means and what the trade-offs in the image are, not just be told "FF has shallower DoF".

climbng_vine wrote:

Kuivaamo wrote:

This enlargement "hoo-ha" is as much an integral part in determining DoF as viewing distance, aperture, focal length and shooting distance. You need to know all of them in order to calculate DoF. You cannot ignore changing any one of them.

You're wrong.

I'm not. Just because the changes are not typically huge doesn't mean that my assertion is incorrect or irrelevant in the context of this discussion (i.e. the OP being wrong). See later.

You can, in fact, ignore changing viewing distance within normal use case ranges. Dofmaster confirms this, as do other calculators. Enlargement is the second least important factor, in that you generally have to double or triple the image dimensions to create an appreciable DoF difference at normal viewing distances. Changing the aperture diameter by 10% has a much clearer affect on DoF than going from a 4x6 to an 8x10. The fact that, in a precise mathematical sense, you cannot "ignore changing any of them" is misleading because, in the real world, some of the factors are barely perceptible except at the absolute edges, and among the rest some create difference more quickly than others. Ignoring this is just stupid.

Be as it may, the OP claimed that the DoF is the same. It is not. Dofmaster gives a DoF of 0.22' for the 50D, 50mm, f/2, 5' case and 0.35' for 5D, 50mm, f/2, 5'. This difference is entirely due to the "enlargement hoo-ha". The reason why the OP's crops look the same is precisely because of the resizing he performed. 0.13' difference in DoF at 5 feet is not a huge difference, but it is significant, and claiming that it doesn't matter or exist would just be stupid.

Doing the test as OP did in order to prove Dofmaster calculations wrong is utter fallacy. Just think about it. Dofmaster calculations assume a constant output size. If you don't keep your output size constant, how can you expect your results to agree with Dofmaster?

You should clarify what you mean by "assume a constant output size", because it's not true unless qualified. It assumes a constant ratio between output size and viewing distance , precisely because within normal ranges the difference is negligible.

That is correct.

springbock wrote:

therickman wrote:

Cal Dawson wrote:

All the OP said was if all things are the same the end result will be a similar DOF there was no mention of changing lenses/ distance etc that was brought up later by those that failed to see the intent. I don't believe the OP stated that if you change distance (FOV) the DOF stayed the same....

That's exactly it. Jeesh, someone on here understood my OP.

I recall awhile back where Cal was trying to say the same thing as you (but without the photos/experiment), many people wanted to argue and twist his intentions as well, just as they have done with you.

I theorize that dpreview attracts photographers and arguers!

Anyway.... kudos to you for giving evidence to refute a myth that "I" have also read on this forum! Maybe some people understand the math and physics involved but many don't. And keeping it simple in the real world with photos does help some people.

Ohhhh BTW... you unfortunately have now experienced a moment of the dpreview "Mob Mentality"... :-).

Don't I know it. It's seriously like little kids who jump into the middle of an argument when they don't know what the original comments were that started the argument, and then take sides with the guy with the larger crowd.

climbng_vine wrote:

You're wrong. You can, in fact, ignore changing viewing distance within normal use case ranges. Dofmaster confirms this, as do other calculators. Enlargement is the second least important factor, in that you generally have to double or triple the image dimensions to create an appreciable DoF difference at normal viewing distances. Changing the aperture diameter by 10% has a much clearer affect on DoF than going from a 4x6 to an 8x10.

I'd appreciate a demonstration of that claim, if you would. For example, according to DOFmaster, http://www.dofmaster.com/dofjs.html , the DOF for an 8x10 image viewed at 25cm on FF from a distance of 10 ft at 50mm f/2 is 0.91 ft.

The aperture diameter at 50mm f/2 is 50mm / 2 = 25mm. So, if we change that aperture diameter by 10% (22.5mm or 27.5mm), we get an f-ratio of 50mm / 22.5mm = f/2.2 or 50mm / 27.5mm = f/1.8. The DOFs for the the same scenario at these f-ratios is 0.81 ft for f/1.8 and 1.03 ft for f/2.2, which represents approximately a 10% difference in DOF.

So, if you would, please show me the math for the DOF of a 4x6 print viewed at 25 cm on FF for a subject distance of 10 ft at 50mm and f/2, and show that it has a much smaller effect on the DOF.

The fact that, in a precise mathematical sense, you cannot "ignore changing any of them" is misleading because, in the real world, some of the factors are barely perceptible except at the absolute edges, and among the rest some create difference more quickly than others. Ignoring this is just stupid.

Please, work the math for me that demonstrates your point above.

You should clarify what you mean by "assume a constant output size", because it's not true unless qualified. It assumes a constant ratio between output size and viewing distance , precisely because within normal ranges the difference is negligible.

If we scale the display dimensions and viewing distance by the same amount, the DOF will not change, of course. For example, an 8x10 image viewed at 25 cm and the same image displayed at 16x20 and viewed from 50 cm would have the same DOF.

However, what you said above is that changing the aperture diameter by 10% has a much stronger effect on the DOF than changing the viewing dimensions from 8x10 to 4x6, so I am assuming you mean at the same viewing distance, since, by changing our viewing distance, we can completely cancel out any change to DOF that results from changing the display dimensions.

In any event, I would really appreciate you working the math for me.

EDIT: Ah! Saved by the bell, I see -- the thread has hit the 150 post limit. Well, please start a new thread, perhaps with a title like "The mathematics of DOF", and post the results, since I really am interested to see the demonstration.