M Lens DOF question

[jaapv]

Advertisement (gone after registration)

Half a stop tighter than film? May i ask you which cameras and lenses you're using this way?

On a sensor 24x36 DOF is theoretically identical to that on film. However, as a sensor draws more precisely than film, the DOF is more defined, thus the eye perceives a less deep DOF. Thus 1/2 a stop tighter. The camera ore lens is irrelevant in this respect.

[Advertisement]

The Leica Q3 is available for pre-order at B&H Photo Video and Adorama

[UliWer]

 

Anyone else prepared to test a few lenses of the same focal length to sort fact from fiction? Testing at wider apertures would make sense (f/2 or so) but the MATE does not let me do that.

 

The test is harder than I thought it to be. At f:2 and a distance of 1 m it needs only tiny inaccurancies to change the whole image.

 

Make up your minds, if this tells you anything.

 

50 Summar, Summitar, Summicron (1. version) and Summilux pre Asph all taken at f:2, approximately 45 degrees and 1,1 m:

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[lct]

...If you or I find them accurate enough, so be it. They are certainly 'in the ballpark' and good enough that Leica hasn't modified them...

This is the point of course. The DoF markings of our lenses are not less or more accurate with digital than with film. I use them the same way AFAIC, i mean with FF stuff of course. Is it the same for you Howard actually? Tell me yes or no my friend, the rest is litterature.

[lct]

On a sensor 24x36 DOF is theoretically identical to that on film. However, as a sensor draws more precisely than film, the DOF is more defined, thus the eye perceives a less deep DOF. Thus 1/2 a stop tighter...

Not sure if you answer in theory or in practice Jaap. In theory, DoF formulas are the same with film and digital. But who cares really? Only the results count don't they. Hence my questions to you, to Howard and other friends here. Do you have a personal experience with digital FF and, if so, do you use actually your good old Leica, Zuiko or whatever lenses in a different way, as far as DoF scales, with film and digital?

[elansprint72]

The test is harder than I thought it to be. At f:2 and a distance of 1 m it needs only tiny inaccurancies to change the whole image.

 

Make up your minds, if this tells you anything.

 

50 Summar, Summitar, Summicron (1. version) and Summilux pre Asph all taken at f:2, approximately 45 degrees and 1,1 m:

 

Surely you are supposed to shot straight at these charts, not at various sideways angles as shown here?

[jaapv]

Not sure if you answer in theory or in practice Jaap. In theory, DoF formulas are the same with film and digital. But who cares really? Only the results count don't they. Hence my questions to you, to Howard and other friends here. Do you have a personal experience with digital FF and, if so, do you use actually your good old Leica, Zuiko or whatever lenses in a different way, as far as DoF scales, with film and digital?

To be perfectly truthful, I hardly ever even look at DOF marks....

[ho_co]

Advertisement (gone after registration)

This is the point of course. The DoF markings of our lenses are not less or more accurate with digital than with film. I use them the same way AFAIC, i mean with FF stuff of course. Is it the same for you Howard actually? Tell me yes or no my friend, the rest is litterature.

I have full-frame film equipment but have basically shelved it for cropped-frame digital.

 

The lenses of one digital camera don't offer depth-of-field information.

 

With the other I tend to zone-focus without direct reference to depth of field (if that's possible).

 

When I do use the DoF scale, on film I adjusted normally by one stop from what is engraved, and with the M8 I adjust by two to two and a half stops.

 

 

In regard to the specific issue of whether DoF behaves the same with digital and with film, what do you think of the examples to which I've twice posted the link? Poor technique? Bad equipment? Flawed demonstration? Outdated information? How do you explain Ferguson's results?

[ho_co]

Surely you are supposed to shot straight at these charts, not at various sideways angles as shown here?

Pete, these are lined up a lot better than those most people post here!

[ho_co]

The test is harder than I thought it to be. ... Make up your minds, if this tells you anything. ...

Uli, thanks for the post.

 

Others may want to be more exact than I; I've simply eyeballed your web images, not downloaded them and pixel-peeped.

 

To me, your comparison illustrates the general contention that all lenses of a given focal length have the same depth of field when used at a given aperture and distance.

 

The minor differences (the Summitar may have a bit of front-focus) seem irrelevant to this comparison.

[ho_co]

... In theory, DoF formulas are the same with film and digital. But who cares really? Only the results count don't they....

I agree completely, but I'm not satisfied with the answer.

 

"In practice, it doesn't make any difference" is fine, but it didn't hold true with Tim Ashley's Summilux, the lens which first provoked these discussions.

 

In other words, no one needs to know what is discussed here. Tomorrow will come just as surely either way. But I find it extremely interesting.

 

One person claims that the 28 Elmarit and Summicron exhibit completely different depths of field. That flies in the face of theory and practice.

 

Gary Ferguson has shown that depth of field with digital doesn't behave as we would expect. Where do his results come from? Are they valid? I haven't seen them either cited or refuted.

 

How can it be that Stephen has posted pictures that seem at odds with similar ones posted by Michael Reichmann?

 

You, Stephen and Lars have agreed (I think) that various depth-of-field formulas agree pretty closely but not completely.

 

Leica has demonstrated that focus shift is much more noticeable on the M8 because it is a far more sensitive instrument than film (i.e., film and digital are not the same).

 

If I were teaching a photography course, I would say it's time to move on. But if I were teaching a course on camera or lens design, I think we haven't even scratched the surface.

 

The formulas were always approximate, the engraved focusing scales equally so, and digital simply highlights that fact.

 

Depth of field is dried and accepted theory. A century ago, people explained what assumptions they made to develop their formulas. Today, we tend to pull out those formulas and cite them as if nothing has changed; but in my opinion--and I think it's borne out by the fact that we've already got 50 posts on the topic in this thread--digital has changed everything.

[63strat]

63strat (as in Stratocaster?)

 

You are correct -- Fender Stratocaster, my 1963 that I've had for many years.

[SJP]

The test is harder than I thought it to be. At f:2 and a distance of 1 m it needs only tiny inaccurancies to change the whole image.

 

Make up your minds, if this tells you anything.

 

50 Summar, Summitar, Summicron (1. version) and Summilux pre Asph all taken at f:2, approximately 45 degrees and 1,1 m:

Thanks for this. For me these images are close enough to being identical in terms of DOF. Using my own arbitrary reading the DOF is about 24mm in all images. Clearly it is not as appealing as the shark but it does allow a more accurate estimate of the relative performance, and this appears to be more or less equal as predicted.

 

How can it be that Stephen has posted pictures that seem at odds with similar ones posted by Michael Reichmann?

My procedure is different. I made the pictures at a fixed object distance and then cropped the 28mm file to the same frame/perspective as that of the 90mm. If you do that you still see that a wide angle has a larger depth of field than the tele, about a factor 3 as predicted by the approximate equations and the more exact spreadsheet. (This was in reponse to Lars' comment concerning the fact that it is the focus distance that determines the perpective, not the focal length, which is correct. However, he also claimed that the DOF would be the same. It is not. However the difference between a wide angle and a tele does become less pronounced, due to the change in the circle of confusion that results from the cropping.)

 

Michael Reichmann's procedure involves changing the object-camera distance to get the same size of the object in the frame of view, i.e. he does not crop the images. In that case you see the same depth of field (at the same aperture) regardless of focal length. Again this is in full agreement with the approximate equations and the spreadsheet.

 

So as far as I can see the theory is still holding out fine.

 

Gary Ferguson has shown that depth of field with digital doesn't behave as we would expect. Where do his results come from? Are they valid? I haven't seen them either cited or refuted.

The key phrase in his review is "Today’s big print culture means acceptable depth of field has shrunk dramatically. The old notions about hyperfocal focusing and depth of field scales based on a thirty micron circle of confusion, are redundant if you’re aiming at high quality A3 or larger prints."

He is entirely correct in his statement that with the ability to print really large and to examine resolution at the pixel level the digital age has made the old circle of confusion value somewhat inappropriate, unless you print on a standard 'snapshot' size. This explains why people are tending to stop down by 1-2 stops (or more) with digital equipment. The equations as such still work fine, just choose a circle of confusion that is a lot smaller than the historical value.

[ho_co]

Thanks, Stephen!

 

I had read these articles previously and filed them away in memory, but didn't go back and re-read them, and appreciate your doing so.

 

Discrepancies resolved till next week, at least!

[lct]

I have full-frame film equipment but have basically shelved it for cropped-frame digital...

OK Howard so you cannot compare my friend.

Focus shift is only a rangefinder problem and due to a (too) low VF magnification and various QA issues the M8 is probably the worst M body ever made from this viewpoint.

As long as there is no FF RF on the market, one cannot compare pears and pears in person, i mean film RF DoF and digital RF DoF.

The only valid comparisons so far can be made with FF Canons, Nikons or Sonys with the same lenses.

I did this already and i don't see significant difference.

Film and digital OoFs don't look the same for sure but the DoF scales of my early and late R lenses remain usable the same way with my 5D and Leica R4s.

Now i am neither a pixel peeper nor a sharpness maniac so as i am a *very* modest guy, contrary to what you seem to think Howard , i'm still asking other members if they can share the same experience as mine, or why not a similar one with Nikon or Sony/Minolta FF stuff.

[ho_co]

... Focus shift is only a rangefinder problem and due to a (too) low VF magnification and various QA issues the M8 is probably the worst M body ever made from this viewpoint.

Here our opinions differ. We may actually be saying the same thing, but here's how I would express it:

 

"only a rangefinder problem"--yes and no; never a problem in film days, _possibly_ one now; but important to the design and use of the rangefinder camera (I don't see it as a "problem"; I would say "issue" or "factor.")

 

"due to too low a viewfinder magnification"--unrelated to finder magnification; focus shift is a property intentionally built into the lens; although the M8's VF magnification is greater than that of the 0.58x bodies, the "problem" wasn't discovered with film, but only with the increased accuracy of digital

 

"QA issues"--yes, in testing the lens, not in the body

 

"from this viewpoint the M8 is the worst M body ever made"--no, exactly the opposite: From this viewpoint, the M8 is the _best_ M body ever made. Simply because the M8's focus plane is much more exact than that of film bodies, and because Tim Ashley had such difficulty demonstrating to Solms the problems with his 35/1.4: a) many of us became aware of how Leica made use of focus shift in their designs; and Leica discovered that quality control procedures that had worked with film were no longer adequate with digital.

[giordano]

Focus shift is only a rangefinder problem

 

Focus shift is a characteristic of the lens, not of the camera design or focusing mechanism. If you focus a SLR (automatically or manually) with the lens wide open and then shoot at a smaller aperture, any focus shift in the lens will affect the image. (could this be behind some complaints about inaccurate DSLR autofocusing?)

 

In principle a camera/lens system could be designed to compensate for focus shift by adjusting focus as the diaphragm closes. I've never heard of a camera that does this.

[lct]

Focus shift is indeed an RF problem because you don't view through the lens. Not a reflex problem obviously because it is adjusted when you focus.

[lct]

Here our opinions differ...

Yes we cannot agree Howard.

1. Focus shift is indeed an RF problem, not a reflex one obviously. Choosing an M8 to compare film to digital DoF is useless because of this and of course the crop factor. The only valid comparisons are FF ones so far. See my post above.

2. Again, the 0.68x mag. is not enough given the crop factor. The proof is in the pudding i.e. the M8's inability to focus proprely some fast lenses without a magnifier. How many times have we discussed this on this good forum? Leica should have decided (a) to lengthen the base line of the rangefinder a la Zeiss and/or ( to choose a higher VF mag. with possibly a larger VF window a la Epson.

3. Do you really suggest that the M8's RF accuracy is the best of all M bodies? If so, we don't speak the same language Howard so i prefer letting you the last word on that.

[Advertisement]

Hallo,

Look here: Interesting Leica gear

[jaapv]

1. Focus shift is indeed an RF problem, not a reflex one obviously.

 

This often quoted truism is incorrect. A (D)SLR focusses with the lens wide open and exposes with the lens stopped down, so focus shift does occur.

[lct]

I compare DoFs with a 5D and Leica lenses Jaap. Stop down metering of course.