PDA

View Full Version : Interesting info on DOF etc



ADG
2-Dec-2004, 05:35
Found some great info and comment on depth of field etc here:
http://www.reasonableexpectations.com (http://www.reasonableexpectations.com)
I am no expert or mathemetician so cannot comment on the accuracy of the findings, but it makes great reading and looks credible to me.

Leonard Evens
2-Dec-2004, 08:10
I read through it quickly. It seems generally correct, but there were a couple of places where I think it may be misleading. These seem to result from too much dependence on the relation of depth of field to subject magnification.

Let me give one example. In the first article, "What is depth of field and how is it determined", he shows two pictures of his daughter. Both were taken with the same camera and same f-stop, but one with a 105 mm focal length setting on a zoom lens and the other with a 300 mm focal length setting on another zoom lens. But the subject distances were adjusted so that the image magnification in the plane of exact focus was the same. In such circumstances, if the subject is close enough, the depth of field about the exact plane of focus is essentially the same and not dependent on focal length.

But this can be misleading, since it depends strongly on just what the distance to the plane of exact focus is, and how far is close enough will depend generally on the focal length. For example, it is easy to arrange that the magnifications in the exact subject plane are the same but that for the shorter focal length lens, the distance to the exact subject plane is greater than the hyperfocal distance for that focal length/ fstop combination and the same distance is well less than the hyperfocal distance for the longer lens/fstop combination. In that case, the back depth of field for the shorter focal length lens would be infinite while the back depth of field for the longer focal length lens would be relatively small.

This is illustrated in what he then says about the background. The background in the image from the 300 mm lens is clearly more out of focus that the background in the image from the 105 mm lens, as anyone can plainly see by looking. But he then says this is a subjective illusion. To justify that he shows that the backgrounds look the same if the latter background is enlarged three times. Well that is obvious nonsense. If you enlarge three times, you are not viewing both images the same way, and of course everything changes, including depth of field and blurriness of the background.

Personally, I think the emphasis that many people put on the relation to image magnification is not too relevant in most large format application. As noted above, it is limited to certain situations. If you are doing portraiture and you expect to get more depth of field by using a shorter focal length lens and moving closer, you will be disappointed. But in many other applications, you want very large back depth of field, and you are outside the range where the argument applies. Also, for most applications, including portraiture, the point of view is more important that the depth of field, and that is determined by the camera position, so you won't want to treat all subject distances the same. You want to move somewhat back and use a longer lens to avoid the usual kinds of "distortions", the sources of which are well understood. In addition, magnification is a function of the ratio of subject distance to focal length. But that ratio will be different for different parts of the scene which are still within the depth of field region, i.e., the region where the circle of confusion is small enough that you consider everything in focus. So if an image has any significant three dimensional extent, even if you arrange the magnification so that it is the same at the plane of exact focus, it may differ significantly within the dof range depending on focal length. This is perfectly clear when actually taking pictures.

I always refer people who are confused about depth of field to Bob Atkins's article at www.photo.net. As far as I can tell, it gets everything right and the explanations are clear. If you want to get into the technical details and mathematics, Jeff Conrad's article on this website is the place to go.