Is the APS-C DOF penalty all that important?

I have been discussing with folks on dpreview and dyxum forums various topics regarding availability of different lens types as well as the issue of DOF penalty in the APS-C format vs. FF. “Earthquake” over at dyxum provided me some rules of thumb for the relative f-stop penalty on APS-C required to obtain DOF as shallow as with FF (a little over 1 stop). I realized though that I didn’t have any sense of what this penalty translated to in terms of distance, so I decided to run the numbers.

First I found a plethora of DOF calculators on the internet via Google search. The one I used is

http://www.dofmaster.com/dofjs.html.

There is no limit on the conditions that can be specified to compare, but I wanted to look first at one classic example. It seems to me that the portrait situation is as relevant as any, more than most. Portraiture is universally recognized to benefit from shallow depth of field as the means of emphasizing the subject and separating it from the background.

For the FF lens on a FF camera I chose a portrait classic, the Minolta AF 100mm f/2.0. While most of you will not be familiar with this lens from the perennial number three camera company, I can assure you it represents the ultimate in quality for 35mm format portraits.

For a Sony APS-C camera the mathematically correct lens comparison would be a 66.7mm f/2.0. Uh, that is a bit of a problem, since no such lens exists. No matter, this isn’t a real photo shoot experiment, just a paper exercise. Why not say that is the lens we are comparing? It will certainly demonstrate the relative issues between the two formats. Concerning the maximum apertures, the f/2.0 is what the Minolta lens actually has, and it is also the most likely choice for a reasonably priced high quality offering in the hypothetical 66.7mm lens. Besides, the idea is to compare lenses with the same maximum aperture to see how bad the DOF penalty is without resorting to exotic and expensive solutons.

The only other assumptions to make are the subject-to-camera distance, subject-to-background distance, and f-stop. I chose 7ft for subject-to-camera since that seemed to be about right for an adult bust shot. For subject-to-background I fell back on the traditional recommendation of 6 feet. Just on a whim I chose to stop down the lenses two stops to f/4. Seems like a normal way to operate for those hoping for high IQ.

So what are the results? Under the “test” conditions the 100mm lens has a DOF of 4.1 inches while the 66.7mm lens would have a DOF of 6.2 inches. This difference is essentially erased if you open up the shorter lens to f/2.8. Yes the penalty in this common shooting situation is real, but is it significant? Barely 2 inches total, about 1 inch each front and back. Compared to the 6 feet between the subject and the background it is completely negligible. The effect of the difference on the appearance of the subjects clothing, props close to the subject, etc. would be greater, but not terribly much.

It seems to me this calls into question the DOF penalty as a reason for dissatisfaction with APS-C and as a driver for upgrading to FF. Sure there may be other reasons such as IQ for wanting FF, but I can’t see how DOF is all that important. Are there other shooting situations where the DOF difference is more meaningful?

Robert