On Tue, 14 Oct 2003, Dario Bonazza 2 wrote: > > > Good point. The 24mm should become a 36mm, while depth-of field must be > > > conidered one stop less, hence pictures taken with the 24mm f/4.5 are > like > > > those taken at 36mm f/3.5, while 24mm f/11 is like 36mm f/8. > > > However, I was expecting some more sharpness there (not more unsharp > mask!). > > > > Can you explain your logic here? In my experience the DOF is based > > on the lens focal length, not the 35mm equivelent focal length. > > > > The 24mm on the *ist D gives you the field of view of a 36mm lens but > > the depth of field of a 24mm lens (because that is what it is). A > > 36mm lens at f3.5 would have much less depth of field. > > Glad you noticed that. I'll try to explain this concept (all but mine). > > The depth of field is based on the concept of confusion circle: > 1) Your eyes see as pinpoint each spot size below their resolution. > 2) Your eyes can appreciate dimensions of each spot size above their > resolution. > Images look unsharp (out of focus) when the size of dots forming them are > above eye resolution and look sharp (in focus) when dots forming them are > below that limit. > > The confusion circle is a parameter in optical design, and depth of field as > indicated in DOF scales is related to it. > > Since *ist D CCD sensor diagonal is 1.5 times smaller than that of 35mm > film, using a 24mm designed for film on such a sensor not only gives you an > angle of view equivalent to that of a 36mm lens, but in order to do that it > only uses (enlarging it) a central portion of its possible image field. > So, to get a print (or file as seen on PC monitor) the same size of that > taken with a "true" 35mm lens on a 24x36mm sensor, you have to enlarge the > image of such a 1.5 factor. > Think of doing that with film: should you want to get a 35mm perspective > print out from a 24mm slide or negative, you have to enlarge it 1.5 times > more and then crop the print to the same size of that made with a true 35mm > lens. The only difference is that the *ist D crops during shooting. > You'll enlarge the image taken with the 24mm more than that shot with the > true 35mm on a larger sensor (35mm film format), hence you'll push image > resolution of the 24mm to a 1.5x higher extent. In other words, some dots > which would stay below eye resolution when sooting on 35mm format will jump > out, and a lesser part of the image will look in focus. > The 1.5x ratio between the two sizes (35mm film and CCD) roughly correspond > to 1 stop wider. > I dindn't invent such theory. It's explained (not so well) in *ist D > instruction manual (see bottom of page 137).
I looked this up in the manual. There is a big difference between what you said and what they said, and to me what they said makes more sense. You computed DOF by taking the focal length of the lens, multiplying it by 1.5, then removing a stop from the f-stop. Pentax only recommends doing the last operation. This results in very different results. Using my depth of field calculator for a 50mm lens at f5.6 I get a DOF of 1.75m to 2.34m. Using your method this would change to 1.91 to 2.10m. Using Pentax's method I get 1.81m to 2.23m. What suggests to me that your method is wrong is that it would give me a narrower depth of field than 35mm (using an equivelent lens). I know from practice (using a prosumer digital camera with a 1 2/3" sensor) that this isn't the case. alex
