Accuracy of depth-of-field scale of lenses on M8

[wparsonsgisnet]

Advertisement (gone after registration)

Thanks for the info, Steve. I'll start my testing that way.

[sean_reid]

I agree with Sean of course that circumstances and subjects and intentions are important. I don't agree however that anything but aperture and reproduction ratio (a combined product of focal length and subject distance) decides d.o.f. For the same c.o.c. criteria, any 35 mm lens whatever at, say ten meters, will produce the same d.o.f. I could go even further: Any lens, irrespective of focal length, and at a given aperture and ratio of reproduction on the sensor, will produce the same d.o.f.

 

That's only true in theory. In reality different lenses (of the same nominal and/or actual focal lengths) can show different depth of field in final files. Do careful experiments and you'll see what I mean. I've come to this conclusion after several years of paying very close attention to how lenses draw. It wouldn't be the first time I've come to conclusions that go against convention. One must know a lot about the behavior of a specific lens to see what its *actual* DOF will be (once one has defined for him or herself what "acceptable" focus is). See my list above - all those aspect can play a role in the final illusion of focus at distances where one did not focus. And that list is not exhaustive.

 

The formulas for DOF are too simplistic, they don't account for all the variables. I again recommend ignoring the formulas and working from direct experimentation. I stand by this based on many, many hours of looking closely at how lenses behave.

 

A day or two outside experimenting can teach one a lot about the DOF of his or her specific cameras and lenses.

 

Cheers,

 

Sean

[Advertisement]

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

[sean_reid]

 

Per Sean's prescription, I will test this hypothesis. It's supposed to rain this weekend, anyway.

 

There you go. It should be interesting. Notice how the perceived DOF (and DOF depends upon perception) can be different between objects near the center of the frame vs. those near the edges. Then try a lens that is sharp across the frame against one that is sharp on center but softer in the outer zones ...etc. Its educational and the more one does it the more he or she realizes how limited these formulas are in their ability to describe a complex phenomenon.

 

Cheers,

 

Sean

[sean_reid]

Even in a glass lens, low resolution and contrast may mean that the transition from (not very) sharp to unsharp seems smoother and maybe more pleasant than in a crisper lens, where it will be seen as more abrupt. This may bear on Sean's argument. Soft focus lenses or 'soft filters' (diffusers) may subjectively seem to create more depth of field, for this reason.

 

Now that's getting closer to the jugular. There's much more to consider where that came from as well. Depth of field, as we perceive such in a final print, is both subjective and relative. There is only one exact plane of distance that is truly in focus, the rest relates to perception.

 

Cheers,

 

Sean

[jaapv]

Now that's getting closer to the jugular. There's much more to consider where that came from as well.

 

Cheers,

 

Sean

 

Ah- let's get to the level of correction in the unsharp zones vs the correctionlevel in the plane of focus, Sean

[sean_reid]

Ah- let's get to the level of correction in the unsharp zones vs the correctionlevel in the plane of focus, Sean

 

And so on....and what we find in the end is that no formula is going to nail this. Fortunately we have direct experience which trumps formulas. What one has to do is to take camera A and lens B and see how the two together behave. Make pictures of subjects at different distances and at different apertures. Then process in one's usual way and make prints at the size one usually makes. Then, trust the eyeballs. The eyeballs always trump the math.

 

Do the Canon 28/2.8 and the Leica 28/2.8 Asph show the same DOF for objects near the outer edges of the frame? - absolutely not. A quick test will make that evident.

 

Cheers,

 

Sean

[ampguy]

Advertisement (gone after registration)

Excellent thread. Lots of great info.

 

My own experiences have been that while both FF film, and the M8 benefit from either infinity settings, or inwards of 2 stops from lens barrel settings, I have also noticed:

 

a) no differences between M8 and FF DOF (w/o relation to the different images, but same lens, same distance to subject), and

 

that different lenses of same focal length give different DOF results, even after corrected for collimation and shift, their dof "windows" are different between say a Summicron 35/2 and Hexanon 35/2.

 

Furthermore, if you do the math, dofmaster is built on several constants that a vendor may be using, or may be using a variant of, dofmaster doesn't account for this. The output of a variable such as coc, can't also be used as the input for the equation, as it is in dofmaster. This is a circular reference.

 

Leica lenses are designed for .023mm coc with film, and the DOF tables are given in the lens brochures. Trust this more than dofmaster, but for a fun experiment, do that math that yields the results provided for in the Leica tables (equations are on 2 sites I can link to later), and understand it for yourself.

[01af]

Given the 1.33 of the M, all my lenses are longer—and therefore have narrower DOFs.

No, your lenses are not any longer, so that cannot be the reason for the reduced depth-of-field. The reason actually is the smaller circle-of-confusion.

 

 

... is the infinity mark on the lens still correct? If so, why wouldn't the hyperfocal distance be correct when getting inside the corresponding aperture mark? That is, if I'm shooting at f8 and I set the infinity mark to 5.6, won't I have everything in focus from infinity to somewhere about the distance on the lens at the other 5.6 mark?

Yes. It would. Yes.

 

 

I don't agree however that anything but aperture and reproduction ratio (a combined product of focal length and subject distance) decides d.o.f. For the same c.o.c. criteria, any 35 mm lens whatever at, say ten meters, will produce the same d.o.f. I could go even further: Any lens, irrespective of focal length, and at a given aperture and ratio of reproduction on the sensor, will produce the same d.o.f.

That's wrong.

 

 

That's only true in theory.

Even in theory it's not true. The angle-of-view that captures your field-of-view does matter for depth-of-field. So for a given image format, calculated DOF does not depend on just aperture and reproduction ratio but on aperture, focal length, and distance (well, and on pupil magnification, too, but let's not get any deeper into this ...).

 

I said "calculated DOF" to discriminate it from actually perceived DOF which additionally depends on several factors which already have been mentioned ... field curvature, residual spherical aberration, out-of-focus rendering, absolute sharpness at the plane-of-focus, and so on. Calculated DOF implies an ideal thin lens and also disregards the effects of diffraction.

Edited June 12, 2010 by 01af

[mjh]

Furthermore, if you do the math, dofmaster is built on several constants that a vendor may be using, or may be using a variant of, dofmaster doesn't account for this. The output of a variable such as coc, can't also be used as the input for the equation, as it is in dofmaster. This is a circular reference.

There are no “constants that a vendor may be using” except the focal length. The circle of confusion isn’t part of the lens design and there is no way a lens could be built with a certain circle of confusion. The CoC is a measure of acceptable unsharpness, i.e. a circle that is is equal or smaller than the maximum CoC is deemed to be indistinguishable from a point. And of course, acceptable unsharpness is in the eye of the beholder; the vendor has no say in this. For historical reasons, the image diagonal divided by 1500 is often used as the CoC, but it can be argued that this somewhat arbitrary figure was chosen to match the resolving power of lenses and film 80 or 90 years ago; since our eyes would still perceive this as slightly unsharp we should use a still smaller CoC. But in any case the CoC is a fraction of the image diagonal and thus it is smaller for a camera with a smaller sensor such as the M8.

[okram]

Sean, nice to hear from you again.

I hope you are in good health.

[ampguy]

Hi MJH,

 

The constraints I am referring to are the values that make up the CoC. The CoC actually is used when designing a lens. Leica states this and uses .023mm [ref. the Gunther Osterloh Advanced M Manual].

 

In any case, the image diag/1500 is one of the ways, the "Zeiss method" of calculating a CoC. The 1500 as you note, is a constant, but does vary, and has also been known as the "z-constant" - where Zeiss may use 1830 or 1500, Canon 1480, and Leica 1800.

 

Another way to derive the CoC during (target design) or after lens design (measuring) could be lp/mm x diagonal of target print / diagonal of medium. For example, vendor a could be establishing their CoC by targeting 5lp/mm x (325mm - diag of an 8x10 / 43.3mm - diag of ff 35 medium).

 

The best way for me to focus is to focus on the subject, or if the subject is out past 15 or 20 meters, to simply focus on infinity. I don't do much hyperfocal shooting except with small animals that are jumping to and from me faster than I can focus.

 

With the tests I've done between film ff (I don't have an M9) and the M8, I now know that the DOF is the same using the lens barrels as I have for 4 decades.

 

Thanks!

 

There are no “constants that a vendor may be using” except the focal length. The circle of confusion isn’t part of the lens design and there is no way a lens could be built with a certain circle of confusion. The CoC is a measure of acceptable unsharpness, i.e. a circle that is is equal or smaller than the maximum CoC is deemed to be indistinguishable from a point. And of course, acceptable unsharpness is in the eye of the beholder; the vendor has no say in this. For historical reasons, the image diagonal divided by 1500 is often used as the CoC, but it can be argued that this somewhat arbitrary figure was chosen to match the resolving power of lenses and film 80 or 90 years ago; since our eyes would still perceive this as slightly unsharp we should use a still smaller CoC. But in any case the CoC is a fraction of the image diagonal and thus it is smaller for a camera with a smaller sensor such as the M8.

[stunsworth]

The best way for me to focus is to focus on the subject, or if the subject is out past 15 or 20 meters, to simply focus on infinity.

 

I don't want to sound funny, but have you actually done this in practice and found the results acceptable? At f2? With a lens of 50mm or longer? I only mention this because you don't mention focal length or aperture, both of which have an effect on perceived DOF.

[pgk]

DofF scales on physically small lenses should be considered as 'guidance' only IMHO. When shooting, hyperfocal settings can be used (i use them) but are often more dependent on experience, the subject matter and common sense than precise mathematical computation and precision setting on the lens. If in doubt try it - it cost virtually nothing these days:).

[adan]

Just for the record, I would note that spherical aberration can have the effect of adding apparent DoF because it smears out the image to be "almost sharp" over a wider range, a rough equivalent to stopping the lens down a bit.

 

That means that it is quite possible, with imperfect lenses in the real world, for, E.G., two different 35 lens designs to show slightly different DoF, even though in theory (which deals in "ideal" lenses) that shouldn't happen.

 

In this diagram, the bright white areas, which are what will be captured as a "sharp point", cover a wider distance. Nothing will be perfectly sharp (light rays converging to a dimensionless point), but more points at different distances will be equally almost-sharp if the lens has SA (top and bottom examples).

 

File:Spherical-aberration-slice.jpg - Wikipedia, the free encyclopedia

 

This can make it easier to "focus", say, a 90 f/2 pre-APO compared to the APO/ASPH version. The APO is sharper at perfect focus - the pre-APO is "as sharp as it's going to get" over a slightly wider range of subject-lens (and thus lens-sensor/film) distances.

 

That is pretty much a restatement of Sean and Lars' points from 2007.

Edited June 13, 2010 by adan

[Peter Branch]

........Do the Canon 28/2.8 and the Leica 28/2.8 Asph show the same DOF for objects near the outer edges of the frame? - absolutely not. A quick test will make that evident.

 

D.o.F. is a very complex subject but just taking the example above one would actually expect on the basis of calculation alone that the D.o.F. would be different. If they were not then something would almost certainly be very wrong.

 

In simple terms it is often assumed that lenses of a given focal length when used at the same aperture will give the same D.o.F. This is only the case if the "Pupil Magnification" of the lenses is the same. It is almost certain that in the example quoted they are not. Therefore the result will be different. "Pupil magnification" is a complex property of the optical design but can be appreciated by looking at a lens from the front and the back and seeing the difference in the apparent diameter of the diaphragm. With retro focus lenses as used on SLR cameras the apparent size is often significantly different.

 

For this reason Leica publish D.o.F. tables for each of their lenses and a close study shows that the data is unique to each design.

 

This topic has come up before and I have attempted to see if there is any useful laymen’s guide to the optical theory behind the effect but a Google search failed to find anything.

[jaapv]

 

This topic has come up before and I have attempted to see if there is any useful laymen’s guide to the optical theory behind the effect but a Google search failed to find anything.

 

 

Have a look here:

 

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

 

 

You'll need a working knowledge of algebra, I fear

[Guest mc_k]

...

Even in theory it's not true. The angle-of-view that captures your field-of-view does matter for depth-of-field. So for a given image format, calculated DOF does not depend on just aperture and reproduction ratio but on aperture, focal length, and distance (well, and on pupil magnification, too, but let's not get any deeper into this ...

 

...there are two different kinds of derivations, depending on whether the viewing conditions are known or not. These are sometimes given side by side. The resulting formulas for depth of field will be in terms of different variables.

[Peter Branch]

Have a look here:

 

Depth of field - Wikipedia, the free encyclopedia

 

 

You'll need a working knowledge of algebra, I fear

 

Thanks for the link - I was not aware of this source. Having gone through it I think I can claim that my suggestion that it is not easy to find a laymen's guide is vindicated.

 

Whilst it is certainly true that as the magnification decreases, i.e. the lens is focused on greater distances, the effect on DoF of the pupil magnification decreases it is a matter of opinion as to whether it becomes, as seemingly claimed, negligible. Like many things in this subject area the effect may be negligible in practice – because of the many other factors involved – but if one sets up tests designed to see it, and is sufficiently careful, then it is not that difficult to find. Indeed I would suggest that the advent of digital sensors, particularly full frame ones, has made such things much easier to observe – or curse!

[jaapv]

Some otherarticles:

 

Understanding Depth of Field in Photography

 

 

Depth of Field and the Small-Sensor Digital Cameras - photo.net

 

And a simple one:

 

Depth of Field

[Advertisement]

Hallo,

Look here: Leica M

[DWatts]

I am gearing myself mentally up for an extended trip where I will be using my M8 gear (body plus 5 lenses) as my my main camera setup. How accurate is the depth-of-field scale of Leica lenses on the M8? Is the scale on the lenses as accurate as on film M bodies? Any feedback is appreciated.

-------

-----------------------------------------------

 

 

 

 

 

 

 

No matter if you use an analogue M camera, a Leica M8 with a slightly smaller sensor or the new Leica M9 with full size sensor, the reproduction scale remains the same, if the same focal length is used. That is why the depth of field engravings do not need to be changed, also with your Leica M8, they are still accurate.

 

 

 

Thank you very much for your understanding and have a nice time with your Leica M8.

 

 

 

 

 

Mit freundlichen Gruessen / kind regards

 

 

 

Martin Seeliger

 

 

 

Leica Camera AG

 

Informationsservice

 

Gewerbepark 8 / D-35606 Solms / Germany

 

Leica Camera AG / email address removed by moderator

 

 

 

 

-----Ursprüngliche Nachricht-----

Von: Deiss, Ulli

Gesendet: Donnerstag, 11. März 2010 16:54

An: Seeliger, Martin

Betreff: WG: Depth of field.

 

 

 

 

 

 

 

-----Ursprüngliche Nachricht-----

 

 

Gesendet: Donnerstag, 11. März 2010 16:43

 

An: Deiss, Ulli

 

Betreff: Depth of field.

 

 

 

As I am now using a Leica M8, will the depth of field engravings on my older lenses, designed for use on film Leica's, still be accurate ?

 

 

 

Thanking you in advance for your reply.

 

 

 

 

 

 

 

 

 

 

 

 

 

 

Aus Leidenschaft zum perfekten Bild / Passion for perfect picture.

html.de Forum - HTML für Anfänger & Fortgeschrittene

----------------------------------------------------------------------------------

Leica Camera AG / Oskar-Barnack-Strasse 11 / D-35606 Solms

AG mit dem Sitz in Solms / AG Wetzlar HRB 966

Vorsitzender des Aufsichtsrats: Franz Jung

Vorstand: Rudolf Spiller (Vorstandsvorsitzender)

Dr. Martin Picherer (stellvertretender Vorsitzender),

Andreas Lobejäger

----------------------------------------------------------------------------------

Privileged/Confidential Information may be contained in this

message.

If you are not the addressee indicated in this message (or

responsible

for delivery of the message to the intended recipient), please be

aware

that any dissemination or copying of this communication is strictly

 

prohibited. If you have received this communication in error,

please

immediately notify us by replying to the message and deleting it

from

your computer. In such case, you should destroy this message and

kindly

notify the sender by reply email. Please advise immediately if you

or your

employer does not consent to email or messages of this kind.

Opinions,

conclusions and other information in this message that do not

relate to the

official business of Leica Camera AG shall be understood as neither

given

nor endorsed by it.

Edited June 28, 2010 by andybarton