View Full Version : Depth of Field Equations, Lens Design Assumptions and Soft Focus Lenses

1-Mar-2004, 11:57
What are the lens design assumptions behind DOF equations? I ask because I have just acquired a lens to which the standard equations apparently don't fully apply, if at all. The manual for the lens, a Wollensak f6 10" Veritar, states as follows:

"With a conventional, anastigmat type lens it is customary to focus on the catch lights of the eyes, because there is a range of sharpness both in front and behind the plane focused on. This range of sharpness or depth is small at large apertures."

"With the Portrait Veritar there is no usable depth of field in front of the plane focused on. However, because of the particular design of the lens, there is considerable depth behind the plane focused on. Within this region there is a considerable amount of recognizable detail."

Does this mean that standard DOF equations don't apply at all? Do they apply but only for far depth of field? Do they apply for both near and far depth of field as the lens is stopped down? Stroebel, in the 5th edition of his book, talks about soft focus lenses and specifically about designs in which there is no depth of field in front of the plane of focus, but he does not address the calculation of depth of field for these lenses.

Leonard Evens
1-Mar-2004, 12:55
I've never seen anything like that. I just read the entire section on soft focus lenses in Cox's Photographic Optics, and while he describes many methods to produce the soft focus effect, he doesn't say much about depth of field. He does make two remarks which might be relevant. Some lenses try to increase apparent depth of field by moving one element of the lens during exposure. He says such lenses are used in cinemetography. If that happens, it is possible that DOF would be extended in one direction and reduced in the other. The most common optical trick used in soft focus lenses is to increase the amount of spherical aberration. Such lenses will usually have a focus shift when you stop down. The effect might then be perceived as less DOF in front and more in back.

I hope someone else can give a more pertinent answer, as this seems quite interesting.

I suspect, though, that DOF formulas are not too relevant. Such formulas are derived on the assumption that the lens functions essential as a point and strict geometric optics apply. As soon as you depart from those assumptions, figuring out just what will happen is going to be quite complex.

David A. Goldfarb
1-Mar-2004, 14:09
Soft focus is pretty subjective, so trying to calculate DOF with a soft focus lens might be a purely academic exercise (and I say that as a professional academic).

There's another thread around here from way back that references an article where several leading portrait photographers were asked to focus a camera with a soft focus lens, and they each chose a different point of "optimal" focus. Likewise, Julia Margaret Cameron responded to her critics who accused her of being unable to focus the camera, by asserting that it was up to her to decide where the focus should be and not them. Linhof won't cam an Imagon for the same reason--there is no objective standard of "best focus" aside from issues of focus shift as the lens is stopped down.

1-Mar-2004, 15:54

Thanks, but all judgments about depth of field are subjective. Also, I would think that portrait photographers would have varying views on optimal focus regardless of whether they are using a regular or soft focus lens. My question isn't about "best focus", it's about whether dof, given a set of assumptions, can or cannot be calculated. If not, why not? If it can, how?

1-Mar-2004, 16:16
In other words, is the behaviour of this lens and others of similar design (I get the sense that not all soft focus lenses are designed the same way) predictable as it relates to depth of field? If so, presumably that behaviour can be expressed mathematically. If it is not predictable, why not? If it is, what is the mathematical expression of that behaviour? Given that it does not behave in accordance with standard equations at large apertures, does it behave in accordance with those equations on the far side of depth of field at those apertures or at both far and near dept of field at smaller apertures? Or is there a completely different equation at work?

Ernest Purdum
1-Mar-2004, 17:44
The basis for depth of field calculations is the "circle of confusion" which is the size of of the blurred circle that a lens makes when imaging a point. How large a circle of confusion will be acceptable depends on how the image is to be viewed, but tables are often based on a circle which is equal to the resolution of the normal eye. This can result in the formula d (diameter of the circle of confusion = Focal length of lens divided by 2500.

There is a difference between soft focus and out of focus. A point which is out of focus just has a larger circle of confusion. A point imaged with a soft focus lens has a sharp center surrounded by a sort of halo.

The Portrait Veritar is an unusual design. There is a positive meniscus lens in front, and a biconcave and biconvex achromat at the rear, the two groups being widely separated. Many other soft focus designs are either a single achromat like the Imagon, or, closely similar to a normal anastigmat. The latter types may have an adjustable element to allow setting the amount of diffusion desired. The design difference is presumably responsible for the unusual depth of field characteristics of the Veritar. (Some very early soft focus types were not achromatized. These had large amounts of chromatic aberration which contributed to the softness.)

Trying to calculate depth of field for a soft focus lens seems to me to be an excessively difficult exercise, since there are too many factors regarding what a particular photographer would consider acceptable. Doing tests would seem a more practical approach. Until a lot of experience with the particular lens is gained, perhaps bracketing the aperture settings for important shots would be advisable.

Cesar Barreto
1-Mar-2004, 17:48

I suppose optical geometry still works fine. At least, if you consider ideal glasses, ideal lenses, ideal polishing, and so on... Those portrait lenses plays around the rules to create beautiful images, not correct images. So I'm in doubt if equations of any kind are useful here. Maybe you could find one for a specific lens, set at specific apertures, focusing at specific points. But I never heard of portrait photographers worried about equations. So I guess the trouble of finding real Dof with these lens remains with the user's taste. I hope you enjoy this piece of glass. Wish I had one...

1-Mar-2004, 18:16
Ernest and Cesar,

I think that depth of field is probably of less concern to portrait head and shoulders photographers precisely because one is working close up within a narrow aperture range (in this case, between f6 and f8). I'm interetested in using the lens for photographs of people that include more of the boy and landscapes at longer distances and smaller apertures. That is the context of my question. Ernest, I'd like to thank you for your essay on this site about soft focus lenses. By the way, I looked at the Cooke Optical site this afternoon and saw that there were depth of field tables for their lenses with the exception of their new, rather pricey, soft focus lens. I sent them an e-mail asking whether depth of field tables for that lens exist and whether, like the Veritar, there is no depth of field in front of the plane of focus.

3-Mar-2004, 18:00
I received a response from Jon Maxwell of Cooke Optics. Cooke comes across as a very classy company, and that I am indebted to Mr. Maxwell for responding to my questions despite the fact that I told him that I have acquired a Veritar and am not in the immediate market for his company's lens. This is a company whose products I'd like to buy when the need arises.

If I understood him correctly, the Cooke soft focus lens responds, in terms of depth of field, pretty much like a normal lens except that it is soft at large apertures. Indeed, Mr. Maxwell indicated that they plan to put depth of field charts for their lens on their web site when they have time to do so.

That suggests that the Cooke lens operates differently from the Veritar, and so the question about the functioning of depth of field for the Veritar remains open.