Rocky,
Lenses do not change just because the sensor is smaller. The focused image circle and perspective will not change unless of course one changes position. Now why would one do that? For a 28 mm wide-angle lens, one would expect to get the whole castle, for example, in one shot from experience say with a regular 35mm Canon Eos 3 film camera. Now if that same lens were on a 5D or 1Ds series, and also full frame, the whole castle would still fall on the CMOS silicon sensor. So apart from a different "look" the depth of field and perspective would be identical.
When, however one grabs a digital Rebel, where the crop factor is 1.6, one has to retreat until on again gets the whole image of the castle to fit on the smaller screen. Since one is further away, the perspective is that of the new distance from the castle (as perspective is only dependant on viewing position, not focal length.
Depth of field is influenced by focal length, distance from the subject and the crop factor of the imaging sensor. But first there's a factor to deal with. What does one consider to be "in focus"? Obviously there is a plane of very sharp focus where the most detail can be recognized. However, to the front and back of this details are not resolved. At some point, we say, that's so poor that we say it's out of focus. The concept is then that a point source is clearly resolved at the plain of focus but that becomes increasingly blurred as one goes away from this plane backwards or forwards. The point at which we cannot recognize the detail, a point for example, because that little circle is too blurred we call the circle of confusion.
Obviously different people's eyes have not the same in ability to discriminate fine detail. So we state that an average person from a distance of 10” could recognize a detail of 0.01 inches on an 8x10 sheet of paper. (If you have 20/20 vision, you will do better than that as you are better than average). If all we change is film or senor size, then a
larger film/digital chip size means a relatively
larger depth of field in which focus is going to be acceptable looking at the print from 10". Increasing the area of film/digital chip capturing light, more detail can be resolved. That is obvious, right? So, in fact, it's really intuitive that DOF might in fact behave in the way it does.
In general, a way of increasing depth of field is to simple decrease the size of the window, or aperture by which light enters the camera. Lens with wide open large apertues, say f 1.2, have shallow depth of field but can give beautiful portrait images. The DOF is narrow so other distracting objects nearby are blurred magically and we are drawn to the eyes and face, as should be. The same aperture for a picture of a car would hardly work. As we want to get a deepr scene we are pushed to coose progressively small apertures. This sounds good. Geometrically it works. However not without limits. There are a complicating factors. The first is diffraction. That is that as the lens aperture becomes smaller, the edges of the metal blades interfere with the
waves of light. This causes a ripple effect and blurring. As the hole becomes smaller, this effect dominates and so degrades the sharpness one would expect, even though from all other aspects, a smaller aperture should lead to a larger depth of field.
Now lets add on another consideration that will alter the sharpeness according to aperture size. With pixel based photon capture in sort of little wells or "sensels" we now look at
light as particles. It turns out that as sensel wells get smaller, the additional degradation occurs as the aperture induced blurring has to be resolved with tiny sensor, but that one point may now be spread over 2-3 adjacent sensels, so how do we now know it's a point and not a smudge?
To deal with these factors you either learn some rough rules of the road or else use a calculator.
Here's one for example
for DOF calculation but this is already simplified.
A simple approach is to know that if one goes tinier in aperture than f8 then the effects of the aperture and or sensel pitch come into play and soften the image somewhat. The easiest thing to so is to try. If you open up an image in you image editing program, you'll find access to the speed, ISO setting and aperture for each shot. The distance to the subject is all one has to record. Then sitting at the computer, one can look at the images from using different apertures and print some out. That will tell you about your camera and your lens and your own values of what's satisfactory and sharp enough, better than anything.
And lest we get lost in all this "pixel peeping", remember that sharp may not necessarily be a good thing!
Asher
