wickerprints
Offline
• • • •
Upload & Sell: Off
|
 p.2 #5 · Depth of field revelation. | |
Sami Ruusunen wrote:
I'm interested the "larger front pupil gives different dof" statement, does that mean that medium format 100/4 lens gives different dof than 100/4 35mm lens when both used on same 35mm camera?
Indeed, the assertion is not correct.
From the point of view of the image plane, a lens of a given f-number (ignoring transmission and aberrations) projects the same light. A microscopic creature living on the surface of the sensor can only see the exit pupil; moreover, it can't even tell how far the exit pupil appears. It could be 10mm in diameter and 50mm away, or it could be 20mm in diameter and 100mm away from the image plane--that is to say, a point on the image plane cannot tell you the focal length. To the creature, both scenarios look the same; the light strikes the sensor at the same angles, with the same collective intensity.
Consequently, there is no way for the creature to infer the diameter of the entrance pupil, since it doesn't even know the focal length.
That said, however, different lenses may produce different DOF and/or extents of OOF blur, due to the various reasons you cited. The single most common reason why they vary is because these things are typically measured when DOF is small or OOF blur is large, and this happens at small subject distances, where the lens is no longer at a focal length corresponding to infinity focus, and the effective f-number is no longer equal to the relative f-number--thus the magnification and pupil ratio *do* affect the DOF, but only in the sense that the imaginary creature sees the incoming light differently because the lens has refocused in a (design-specific) way that no longer makes the exit pupil look the same. It still can't tell what the focal length is or even the subject distance. It just sees an exit pupil that is f/5.6 instead of f/2.8, for example.
As for the other characteristics of the lens, such as aberrations, these too play a significant role. The extent of DOF that is observable may be diminished if there are, for example, spherical aberrations; similarly, the perceived extent of background blur may be greater in lenses with undercorrected spherical aberration when shooting a sharp foreground subject--though the blur disk may actually not be larger, its boundary may be less well-defined, thus increasing the perception of blur.
Ultimately, modeling DOF and background blur is only approximate unless you can ray-trace the actual optical formula for the lens; and even then, one needs to account for properties such as dispersion and diffraction. And even if you could do *that*, remember that real lenses vary in their manufacturing precision. In the end, it's just not that important to go around comparing millimeters of DOF or OOF blur. It makes for an interesting academic exercise, but in terms of appreciating the artistry of photography, such technical minutiae are at most a minor footnote. Even if you could empirically show that one 40mm f/2 lens has twice the background blur as another 40/2 lens, what good does it do to try to attribute it to some aspect of its design? It's not as if you can take that knowledge and apply it in another case; each lens is different. Just use the lens you want to use and be done with it.
|
more of these please