Calculating depth of focus

Most of the optics/maths of this are beyond me, but a D.o.F calculator gives different results for different formats. E.g. Using the one at Cambridge in Colour for a 50mm lens at f/8, focused at 8 metres -

On 35mm (full-frame):
Nearest Acceptable Sharpness: 4.41m
Furthest Acceptable Sharpness: 43.03m
Total Depth of Field: 38.62m

On crop factor 1.5:
Nearest Acceptable Sharpness: 5.21m
Furthest Acceptable Sharpness: 17.24m
Total Depth of Field: 12.03m

So it would appear to me that reading the D.o.F scale of a lens designed for full-frame use would give the wrong D.o.F when it is used on a K-series body.

Philip

But that's no good as a comparison, because the 35mm image is a standard lens and the ALS-C image is a telephoto lens. It's all about magnification, but the complication isn't really necessary in real life.

Take a picture, look at it, magnify the image. If the bits you want are not sharp then use a smaller aperture or adjust where you focus and try again. The very best way to find out what is happening is to shoot some images.

A lens can only focus at one specific plane, anything before or after that is therefore not in focus. How acceptable that nearly in focus zone is depends on the intended use of the image (eg web or 12x15" print) and the viewing distance. At a web size the whole image may look acceptably sharp whereas the large print viewed close up will show some of it as out of focus. But view the 12x15 print from across the room and it too may well now look acceptably sharp. No simple mathematical formula can account for these perceived differences given not all lenses perform the same, some have a flat field of focus, others have a curved field. There are too many variables. As JR says, take a test shot and work from there.

John K

The two Johns,
Your points might be relevant to practical photography but fail to support the original point that was made, quote - "Many of the older prime lenses actually show the depth of field available at each aperture". The usefulness of that statement is challenged by depth of field calculations, produced by science technicians who really understand this property of lenses, showing that for the SAME focal length lens, at the SAME aperture, focused at the SAME distance, the depth of field is DIFFERENT on a different format. Therefore, the depth of field scale printed on the barrel must only apply to the format for which the lens was designed.

Philip

Source of this research, please, and is it peer reviewed? What one reads in a forum isn't always accurate!
Remember that two different 50mm lenses will not be equally sharp, the softer one may give the appearance of having a greater depth of field; but I say again, there is only a single plane where focus occurs.
What you are saying is that one lens placed on a 35mm film camera then on a D-SLR (both producing an APS-C sized image) is giving different depth of fields. I have no research to prove or disprove that but it sounds impossible. The only thing that may affect the DSLR image is post processing to increase the apparent sharpness. It makes no sense to me that the DOF of the lens onto different film stocks or various branded digital sensors is actually going to change. Is the suggestion that the depth of field of a lens at a certain setting on a K-5 (CMOS) is different to that on a K100D (CCD)?

John K

Actually MrB, you're incorrect in your statement and the bit you've missed is that under these circumstances each format will give a different picture.

The DOF of a lens does not magically change just because you've put something different behind it. What is happening is that the camera position needs to be moved to provide the same picture with each format. This changes the magnification, which changes the DOF.

So, to recap, a 15mm on a sub-miniature camera gives immense DOF and is a standard lens. A 35mm lens on APS-C gives much less DOF and is a standard lens. A 50mm lens on a 35mm film camera gives slightly less DOF and is a standard lens. An 80mm lens on 6x6.....etc.

It's nice to understand the relationship, but not essential for everyday photography.

This is my final attempt.

Firstly, as I have stated before, I do not understand the optics or maths of depth of field calculations, so I can only rely on information from those that (presumably) do.

Secondly, the ONLY point I am trying to make is that the depth of field scale on an old lens made for 35mm full frame will indicate the wrong D.o.F. if it is used on a K-series (1.5 crop factor) dslr.

Thirdly, the evidence for that point only is from D.o.F. calculators, e.g. link 1; link 2; link 3; (a Google search will find lots more).

All I can suggest is that you use a D.o.F. calculator and see for yourself that the results (for same f, same A, and same distance) are different for full-frame and 1.5 crop. And if those results are correct, the D.o.F. scale is wrong when the lens is used on a Pentax K dslr.

So finally, if you two want to disagree with that point, you will have to argue not with me, but with the technicians who have produced the algorithms for the D.o.F. calculators, cos I can't explain their results.

Now that's it from me - life's too short.

Philip

Nobody's arguing with you Philip, it's a discussion.

So, if I understand correctly, what we are saying here is that for any single scene shot with a specific lens at a specific aperture, the DOF (the distance within the scene that appears in focus) does not change just because you take another shot with the same lens and aperture on another format.

But because to be able to see exactly the same scene on the different format we would have to move or change focal length, it is those actions that in fact alter the DOF.

Does that cover it?

Sounds good to me Nigel.

You seem to assume the negative meaning, John, but -
argue
to present supporting or opposing reasons or cases in a dispute; to try to prove by presenting reasons; to debate or discuss; to persuade; to give evidence;
(World English Dictionary)
which, as you have stated, is what we have been doing!

Philip

I haven't quite grasped any specific dispute or argument here yet, as in specific opposing views. Its been informative for sure.

So if we seem to be agreeing that the actual lens 'DOF' remains constant, is it the case, (as MrB has said), that lens DOF scales will only be correct when applied to a sensor matching the size the lens was actually designed for?

Is this not a different question to just asking if the DOF itself will change or not?

At best, DOF scales on lenses are notoriously inaccurate. They are good enough for a rough and ready evaluation of the hyperfocal distance, but I would always stop down one stop further than they suggest for the optimum result.

Now stop down mechanisms I find unusable as the ground glass screens are too small to evaluate DOF. But we can see the sharpest point well enough with extreme care. I say that because test have always shown that the eye can be dreadful at judging the sharpest point. The variation in focusing point when manually focusing is surprisingly large.

The marking on many older lenses will be designed for full frame (film) use where the image is often only printed as a 6x4.
These days only a portion of that image is recorded and that is often viewed bigger so it's not surprising the markings are inaccurate for modern use. I used to find them quite reasonable in my film days, though I guess if I'd been shooting slides for projection it might have been different.
The difference is due to changed requirements of the circle of confusion, due to greater enlargement, rather than inherent inaccuracy.

Phillip, I remember a good discussion on this subject a few years back on Digital Photography Review.

To understand DOF you must understand that it originally related to a given print size enlarged from a negative. We can now refer to a print size or screen image, but the relevance is still the same...it is about what the eye can discern when looking at a finished image.

The reason the calculators give a smaller DOF on digital crop sensors compared to full frame for the same lens/aperture/distance, is that the sensor image needs to be enlarged more than the full frame to achieve the same size final image (print). The more you enlarge, the narrower DOF becomes.

So you are correct to assume that the DOF markings on a 35mm film lens are not accurate for digital use. You would need to stop down more than the lens indicated to achieve the same DOF.

.........However.......

As John and others have pointed out, the two different formats do not produce the same image. For a portrait shot for example to be the same size in the frame, you would need to stand further back when using the same lens on a crop sensor. The increase in camera to subject distance will increase the DOF, largely offsetting what will be lost by having to enlarge the sensor image more than with the full frame image.