Cinematography Mailing List - CML

Perceived Sharpness

Hoping that CMLers with big screen experience can help with this one. Is there an apparent *increase* in depth of field as one uses a softer lens or adds filtration? Does it follow that the sharper the lens the more defined the depth of field is to the viewer?

So the sharper the lens, the less depth of field, as perceived by the audience. If you were shooting in a small room, with low lighting level, lots of action and a hand held camera would you steer clear of sharper lenses? I may be lacking some (fuzzy) logic on this question.

Mike Brennan

Using "professional" lenses one expects them to be sharp: I don't know of "soft" lenses. If a lens appears 'soft', one of the elements is out of wack and should be repaired or it is a substandard cheap lens. On 16mm, using an Angenieux 25 mm, F.95 lens at one ftcandle wide open, you will obtain a hair sharp picture on one chosen field, but with no depth of focus. On 35mm an equivalent 50mm .95 will be twice, no, four times as critical. There is no solution for this, it is just an optical law. On certain projects, if needed one can make the scene appear softer by adding filters and what not. Maybe you should wait until that mythical 10x faster emulsion is available. How one can see at F 8 through a viewfinder with 1 candle burning is naturally the next problem.

Robert Rouveroy csc

The Hague, Holland

There are many DP's who believe that older, less sharp lenses create the illusion of more depth because the fall-off in focus is less abrupt-looking. For example, a number of DP's shooting in anamorphic prefer the older, less sharp (and more distorted) C-Series lenses from Panavision -- since depth-of-field is more of a problem with anamorphic, they feel that the C-Series lenses softness seems to make it less obvious (I believe that Steven Burum made comments like this in the A.C. article on "Mission Impossible." Adrian Biddle made similar remarks regarding his use of C-Series lenses for "The World Is Not Enough.")

Since depth-of-field tables are calculated around a chosen "circle-of-confusion" for the film format, it seems to make sense that as film stocks and lenses get sharper, then the circle of confusion number should be adjusted. It's similar to the argument I have with some AC's who want to constantly split the focus between two subjects -- if I feel that the lens is very sharp, then I also believe that it is more clearly only in-focus at one point and that the charts telling you how much area you can hold in focus were designed with older, softer lenses in mind. Personally, I don't think that for moments where you want more depth-of-field (like handheld shots) that you should switch to a softer lens or use more diffusion, because then you'll have a matching problem. I still think that stopping down is the only real way to increase depth-of-field. But there are a lot of artistic reasons for shooting a movie on older, softer lenses.

Besides the more gradual apparent fall-off in focus, and perhaps slightly more depth-of-field (debatable), there is a gentler look to the image without having to resort to filters. I've shot a couple of features on Panavision's old Ultra-Speeds (due to the budget) and while I wasn't happy about the flare problems, they did seem to produce more complimentary close-ups of the actors because they were less sharp. (But in general, I prefer to get the sharpest lenses I can because there will be enough softening down the road somewhere...)

David Mullen Cinematographer / L.A.

Since depth-of-field tables are calculated around a chosen "circle-of-confusion" for the film format, it seems to make sense that as film stocks and lenses get sharper, then the circle of confusion number should be adjusted. My understanding is that the "circle of confusion" is based on perception as opposed to the sharpness of the lens. That is to say that if the cone is at or smaller than X percent (1/1000th of an inch in 35 & 1/500th in 16??) the viewer will perceive that point as "in focus". If this assumption is in error, please correct me. But - if my assumption is correct, then the circle of confusion number won't necessarily change, but maybe DOF for a particular lens may? (Or is DOF independent of lens sharpness and strictly an immutable law of physics?)

Brent Reynolds

Absolutely, but what alters perception if not the sharpness of the lens and film stock. How 'sharp' is 'sharp' and what is its relative'soft'? Perception is all about one thing relative to another, in this case sharp to soft.

Anders Uhl

© copyright CML all rights reserved