Bar-Bet: Does DOF change with format?

[rdenney]

Let's not forget that a 90 mm lens for 35 mm will be a telephoto, and pupillary magnification comes into DOF too ... :-)

I have four 50-ish lenses:...

Your pupils must magnify better than mine.

In response to a similar argument (more on that in a minute) along these lines on Photography on the Net a few years ago, I did actually perform the experiment, with the following lenses:

- 47mm Super Angulon
- 50mm Flektogon
- 50mm Canon (for both 35mm and APS-C)

I didn't have a Pentax 110 (hanging his head in shame). But I did have an APS-C Canon DSLR.

The exposures were made on 6x9, 6x6, 35mm, and APS-C, respectively. The principle argument in that case was that perspective was not a function of focal length with respect to format, but rather a function of camera position. That's not likely to be an item of debate on this forum. But in addition to illustrating that, I also was able to show that a 24x36mm frame from the middle of all three images were basically identical in appearance, including their depth of field. (I will add that some lenses show smaller airy disks off-focus than others, all else equal, and this must be a function of their design. It's not surprising, given that some lenses show different bokeh effects than others, too.)

Back to the argument. What argument? The only argument I've seen so far is how one should prove the point. Nobody seems to disagree with that point, though there are lots of arguments about the terms used. It would appear to me that Mr. Bedo now has the means at his disposal to turn a bar bet into a full-fledged brawl, complete with airborne chairs and broken crockery.

Rick "expecting a report" Denney

[rdenney]

CoC is not a fixed value, but defined by "agreement" about what looks acceptable sharp. So in 8x10 you have much larger CoCs than in 35mm, becouse you need less enlargment.

[raises his hand...]In the telescope world, the fuzzy spot produced by all optics, and generally constrained by diffraction in good optics, is called the airy disk. I have always considered this the term appropriate to the fuzzy-spot effect, but maybe it does not apply to spots made fuzzy because of being out of focus. The circle of confusion seems to me the standard by which the dimension of the airy disk is evaluated. At some point, it is small enough so that the detail it portrays is no longer confused. But that is an assumption on my part. Right? Wrong?

Rick "who divides all DOFMaster CofC values by four to convert their 8x10 print assumption to 16x20" Denney

[Struan Gray]

Your pupils must magnify better than mine.

In theory at least.

Any differences will show up at large magnifications, 1:1 and above. In principle, you should see the change in DOF clearly by reversing the lens, but you can find yourself fighting the asymmetry another way as one of the object/film conjugates moves inside the front element. Aberrations complicate the issue more than somewhat too.

The Pentax 110 50 is an oddball. The pupillary magnification of the seperate lens is around 2, a little more if you put an aperture where it was supposed to be in the original system, which had an combined aperture and shutter inside the camera body. The right way round, it makes a great loupe (or wide-field telescope eyepiece), but I'm too lazy to expose fine-grain film with it to settle someone else's bar bet. I trust the maths.

At least I know to capitalise "Airy" :-)

Struan

[Mike Tobias]

Don't think anyone's posted this yet so here it is. To date the best examination of this issue I know of:

http://www.naturfotograf.com/D3/D3_rev06.html

Short answer (if you don't want to read it), maybe there's a difference, but it's so small as to be almost unmeasurable.

Mike

[Jeff Conrad]

A bit off the original topic, but ...

Rick, the Airy disk is a different animal from the defocus blur spot; it has different causes as well as different appearance. The Airy pattern arises from diffraction, and consists of light and dark concentric rings. The defocus blur spot arises from, well, defocus, and is more uniform in appearance. Defocus and diffraction combine to reduce sharpness, but for what we've been talking about here, diffraction (and the Airy disk) shouldn't be much of an issue, save perhaps each lens's smallest aperture. Or at closeup distances, where diffraction does become significant, and where the simple model breaks down anyway. Or unless it is essential to test some even more obscure knowledge (or perhaps technobabble one's way out of a lost bet. Probably works better with pints than with decaf.).

Just out of curiosity, why do you multiply DOFMaster CoCs by four for 16x20? Two would seem to do (unless you always pull out your loupe).

[Darin Boville]

Don't think anyone's posted this yet so here it is. To date the best examination of this issue I know of:

http://www.naturfotograf.com/D3/D3_rev06.html

Short answer (if you don't want to read it), maybe there's a difference, but it's so small as to be almost unmeasurable.

Mike

No good, Mike. Your post contains a link to photographs. We're only doing equations on this thread.

--Darin

[Jeff Conrad]

Short answer (if you don't want to read it), maybe there's a difference, but it's so small as to be almost unmeasurable.

Ummm ... the difference is night and day when you compare the images at f/8 in the second example. For the FX-format image at f/8 and the DX-format image at f/5.6, the DoFs are about the same, as also would be expected. Yet again, it depends on the conditions under which the comparison is made. The usual assumptions for the "same picture" case (which he examines in the "same FOV" example) again are

• Same subject distance (i.e., object conjugate) for both formats.
• Subject distance significantly less than hyperfocal for both format.
• Magnification significantly less than unity.
• Same final-image size for both formats
• Same f-number for both formats.

At magnifications of 0.13 and 0.2, the third condition is violated. Even so, with the final assumption (comparing the results at f/8), there's clearly much greater DoF for the DX-format image.

I think this also illustrates some problems with strictly photographic comparisons:

• The subject distance is too small, so what's being compared isn't within the model's valid range.
• Though it's obvious that f/8 DX image has considerably greater DoF than the f/8 FX image, it's not immediately apparent precisely how they differ (it's not exactly 1.5 because the magnification in the numerator isn't negligible).

It seems to me that the results clearly demonstrate the validity of the DOF models we've been discussing, and which have been well understood for years. It boggles the mind that people continue to attempt to rediscover this stuff.

[Paul Kierstead]

What baffles me is that it is absolutely, 100%, beat you over the head obvious that, in practice, the DoF when using LF is vastly less then for 35mm. Anyone who has shot any number of pictures at all will easily observe this. I suppose one could argue that LF has the same DoF, it just requires a lot more light. Now that is in practice. One can mathematically masturbate till the cows to come home to show all sorts of other conclusions (some of them even interesting), but it is all just that; masturbation. If you are concerned about outcomes, then it is clear.

Of course, when it comes to bar-bets, it pretty much depends on wankery, or else it wouldn't be a bet would it?

[Jeff Conrad]

What baffles me is that it is absolutely, 100%, beat you over the head obvious that, in practice, the DoF when using LF is vastly less then for 35mm.

So one would think. But it seems pretty obvious that the FX format has a lot less DoF than the DX format in the cited article, yet the author reaches the opposite conclusion ...

[Dan Fromm]

Paul, my flower pictures don't agree with you. But then, I went up in format to be able to get more of the main subject's setting in the frame, not to shoot at a higher magnification than with 35 mm.