On 27 Dec 1998 02:37:30 GMT, "Michael K. Davis"
[most of a really excellent post has been deleted for
space reasons, and a couple of comments are added to
parts of the original post - go back and read the
entire original....]
>The diameter that lens manufacturers use varies from one manufacturer to
>the next, but most run between 1/100th inch (very generous -- likely to
>dissappoint you even at viewing distances equal to the print diagonal) to
>1/300th inch (pretty conservative -- definitely yielding an illusion
>of sharpness in images viewed at distances equal to or greater than the
>print diagonal.) The fact remains, however, that for a given focal length
>and format diagonal (after cropping!) a print of given magnification,
>viewed at a distance equal to its diagonal, will have exactly the same
>depth of field for a lens made by one manufacturer as that of another.
>The math rules -- depth of field can not be engineered into a lens.
Ummm, what about the poor lens vs. the really superb...? ;-)
Upon close examination of the images, the sharper lens may
appear to have less DOF, since there could be a greater
visible difference in sharpness between the in-focus image,
and the limits of DOF - as determined by the selected disk
diameter, assuming that the disk diameter is not unusually
small... (I know you have specified the relative viewing
distance - but that is often violated...! ;-)
>There's a web page that I read once, where the author strongly advocates
>that for landscapes, you should set your focus at infinity and then select
>an aperture that will give you an acceptable near sharp. He has even
>posted an image of a Newfoundland village taken from a hill where there's
>a cannon in the foreground, a village near infinity and hills and ocean at
>infinity where the near sharp is well beyond the foreground cannon, but he
>argues that for his tastes, this is allowable (the cannon is at 30 feet
>from his 90mm lens!) because everything else, out to infinity, is razor
>sharp and that is much more important to him. He says he is tired of
>using DoF calculations that encourage focusing at the hyperfocal distance
>because this has so often resulted in less than crystal clear backgrounds.
>
>I hear him loud and clear but I honestly believe he is throwing away
>useful depth of field by focusing at infinity whenever infinity is part of
>his subject. There is twice as much depth of field between the point of
>best focus and the far sharp as there is between the near sharp and the
>point of focus. So he's throwing away 2/3's of his depth of field because
>he never again wants to see infinity out of focus.
>
>I want to encourage him to come back to the math and trust it. All he
>needs is control over the one variable that MOST depth of field
>calculators deprive you of using -- the ability to specify the maximum
>permissible diameter for circles of confusion.
Yes - using most DOF guides, infinity images look
insufficiently sharp (I counter this by "fudging"
focus toward infinity some, and using a somewhat
smaller lens stop). The web page author does throw
away a lot of useful DOF by focusing at infinity,
but I do find that equal sharpness at the near and
far limits does generally look worse than unequal
sharpness favoring the distant focus somewhat
(even if the near focus sharpness is compromised).
This may be due to the differences in the scale
of the detail...
I prefer very distant subjects either to be very
sharply rendered, or to be very unsharply rendered
(they look very bad to me if they look "almost in
focus, but not quite"...;-).
Thanks for a nice post!