Mamiya depth of field discrepancies

For awhile now I have been having problems achieving front to back focus with

some of my Mamiya lenses, specifically the 55mm f/2.8. When setting the lens

to the hyperfocal distance markings on the lens barrel for a given aperture and

subject distance, the front to back sharpness just isn't there like it should

be. I decided to punch in the formulas for my Mamiya lenses on the Depth of

Field calculator site and sure enough there are discrepancies. Apparently

Mamiya has an overly optimistic view on what constitutes acceptable sharpness.

Perhaps this is old news. Most of my lenses appear to be off somewhat but the

older versions (non-N) are the worst. At any rate, maybe this could be useful

to others like myself who took Mamiya's lens markings for granted. You may

want to do the calculations yourself. For some situations the differences are

not noticeable but for critical focusing I've found that Mamiya's distance

scale just doesn't cut it.

How many lenses have you tested? I've not seen the problems you are describing with any of my lenses...

 

- Randy

Daniel, this issue seems to be growing, and not just amongst Mamiya users. I guess it's due to the current trend for printing bigger, allied to negative scanning or digital capture when the image is scrutinised at 100% on the monitor.

 

The fact is that all depth of field scales assume fairly modest enlargements, indeed when depth of field scales were first laid down in the 1930's contact printing was still fairly common and a 10x8 print was considered a large photograph.

 

Most medium format lenses have their depth of field scales computed with about a 50 micron "circle of confusion" at the limits, that means that a point in the subject would be recorded on the negative as a blurry patch 50 microns in diameter. To put that in context the average human hair is about 70 or 80 microns thick. So if you enlarge that negative by a factor of 10 then at the depth of field limits the point now becomes 500 microns in diameter, or half a millimetre. And to most experienced photographers a half a millimetre blurred patch looks distinctly soft.

 

If you close the aperture down by two stops you'll halve the circle of confusion, I appreciate that's not always practical, but it illustrates the difficulty in achieving front to back sharpness in large prints without using lens tilts.

The best thing to do is to calculate by yourself your own depth of field tables.<br>

You can choose the circle of confusion that you need, and have the DOF automatically calculated for each of your lenses.<br>

It is easy to do, and you can find on the web many pages where a DOF calculation software is available. Search for "Depth of field calculation" and you'll find them.<br>

I use a Mamiya 7II and have been stopping down 2 more stops from the markings on the lens, but calculating my own depth-of-field table sounds intriguing.

 

What size circle-of-confusion is a good starting point assuming a 16x20 print from a 6x7 negative?

Well, Randall, all what you say (circle of confusion, print size, negative format) can be chosen on many DOF calculators.

Thanks for your thoughts guys. Randy, the lenses in question are :80mm f/2.8 N, 55mm f/2.8 and 150 f/4. None of them were completely accurate when compared with the guidelines for the Depth of Field master calculator. In all cases, Mamiyas scales assumed greater depth of field for a given aperture than actually existed.

For example:

 

According to Mamiya's scale on the 55mm, maximum depth of field is attained at f/22 by setting the distance to 7 feet which they claim should give good sharpness from just over 3.5 feet to infininty. According to the online depth of field table however, setting the focus to 7 feet only gives aceptable sharpness from 4.12 ft to 23.3 ft with a circle of confusion of 0.045mm which is recommended for the 645 format. Maximum depth of field is actually acheived by focusing at 10 ft. in which case all objects from 4.98 ft to infinity are sharp.

 

The shortcomings of Mamiya's calculations are rather obvious. I see obvious loss of sharpness in distant objects in an 11x14 print which is why I decided to investigate the issue further. I'm glad I did and at the very least I can put my mind at ease lol! Anyway thank you for all of your thoughts.

As indicated this is a common problem -certainly my Mamiya 7 has the same issue- but it isn't a new one. On many medium format lenses use a circle of confusion in their dof markings that is just bigger than a critical user demands for enlargement or even projection. Interestingly my Bronica lenses' barrel markings are much less optimistic and the dof scale is broadly usable as is. If you calculate your own dof tables and use that to work out whether you need to adjust the dof marked by one,two or three stops, I suggest you use a COC of 0.03 mm which is used often for 35mm. If you are producing massive prints and viewing them from an unusually close distance, then even that would be insufficient.

 

Be aware however that if all this pushes you to use very small apertures, then diffraction will reduce sharpness anyway and you'll be swapping one type of unsharpness for another. There will be a happy medium in there.

Good point Dave. I was having problems because I was trying to avoid excessively small apertures to avoid the diffraction you mention. I would choose the widest aperture that would still give me adequate front to back focus in accordance with Mamiya's distance scale. Unfortunately when focusing was critical and there wasn't much margin for error the scale proved to be erroneus and I was often dissapointed with the visible lack of sharpness in ojects I wanted to appear sharp. The idea to use a much smaller circle of confusion is a good one when you want to be absolutley sure that critical subject matter appears in focus over a given distance range.

I do not trust Mamiya dof scales either: I calculated my own based on my own assessment of the critical sharpness I need in large prints.

 

Then, when looking at the dof scales on my Voigtlander lenses I realised also that the circle of confusion they assume for the 15mm Heliar is different to most other Voigtlander leness: it seems the Heliars dof markings are rather more generous than for other lenses.

 

I can understand why different c-o-cs may be used for very long and very wide lenses (or in "special" conditions), and various cases are discussed in several definitive books on photography (Focal Encyclopedia of Photography, etc), but I have not previously noticed significant discrepancies with Nikon or Leica lenses, where a consistent dof formula appears to be used across the manufacturers lens range. Interesting!

 

It makes me wonder if some manufacturers do indeed deliberately and significantly "tweak" their dof markings according to focal length, and what their particular reasons for doing so may be.

 

Interesting!

I never had a lens that would have pessimistics DOF markings. I generally use one or two stops lesser markings.

I think you want to find a hyperfocal distance calculator. This is the best site I found this winter when researching the issue:

 

http://www.dofmaster.com/

 

I hope this helps, Regards, Phillip

That's the one I used too Phillip. A good site indeed.

The most important is to choose your CoC (Circle of Confusion) which can vary, depending on the format of your prints.<p>

Many data are not exact ones, but is it important ?<br>

- The focal length, which should be the real distance from the optical center of the lens to the focal point (film), also called flange focal distance, and not the "commercial" focal length.<br>

- Same for the film format : the height of 120/220 films is actually 56mm, not 60mm and the film width called 6x6, 6x7, 6x9... depends on the real size of the exposure window of the camera. For the same format, it can vary by more that one half centimeter.

On Toyo 6x7 backs the format is 56x67mm, on Horseman backs 56x68mm, on Mamiya RB 56x69,5mm, on Pentax 67 55x70mm, and on Linhof Super Rollex 56x72mm.<br>

On some 645 cameras, the size of the window is 40,5 x 56mm.<br>

But for DOF calculations, using the "official" focal length and film format is probably sufficient. Otherwise, it's fly.... or, as I recently wrote "enlarging the anal sphincter of dipterians". ;>)<br>