Monday, December 17, 2007

Depth Of Field

When people talk of depth of field, they generally mean the area from near to far that falls within the focus of the lens, implying that everything is sharp within this area.

This is not entirely correct. What in fact happens is that at smaller apertures, the degree to which the image looks out of focus gets less, eventually being sufficiently small so as not to be noticed on your print.

Here's some points to consider:

1) what looks sharp on a small print, may well look significantly fuzzy on a large one - depth of field is inversely proportional to the size of the print.

2) Depth of field most certainly won't work when you inspect your image at 100%. This is particularly true of higher pixel count cameras in which parts of the image may look significantly out of focus at 100% but look just fine in a normal size print.

3) You may want to forget about things like hyperfocal distance (at which everything is supposed to be sharp from infinity to half the hyperfocal distance) and focus instead on what's really important in the image. If two areas are of equal importance and one is significantly further away than the other, you might even want to blend two exposures. As there is some image magnification as you focus, blending isn't straight forward, but usually is achievable, even without things like Helicon Focus.

If you do want to rely on depth of field tables or settings and plan fairly large prints, you may want to "pretend" that you are using an f stop two stops or twice as wide - eg. your camera is at f16 but you rely on the table or lens setting for f8.

4) given a rock or a person or a bush in the foreground and mountains in the background, my experience suggests that the focus should be on the foreground and stop down enough so that the background, if not absolutely sharp, is adequately recorded.

5) don't forget that there is a law of dimminishing returns - as you get to quite small apertures, diffraction robs you of any gain made in depth of field by blurring the sharpest parts of your image. Eventually you find that the fuzzy bits don't get sharper and the sharp bits get fuzzier - clearly a losing proposition. For me, on a full size sensor 1Ds2, that cutoff is f16. At f16, I can sharpen the image sufficiently to compensate for the loss by diffraction, but by f22 I cannot. So, that means there is never ever a reason to go beyond f16, unless in total desperation you really need a slow shutter speed to blur water and even then you should have had a neutral density filter with you.

1 comment:

Jakub said...

Great post! It should be a preamble of every "Depth Of Field" discussion. Nothing but facts, although not so obvious until one learn that DoF is a mathematic formula depending on final print size.

One important feature that I use to tell a good camera from a bad one is DoF preview button.