Doug Kerr
Well-known member
In a recent thread here there was a discussion about two ways to, from a certain vantage point, more closely frame a "distant" object:
• Use a lens with a "greater" focal length.
• Crop the taken image to produce the delivered image.
I mentioned the matter of, among other considerations, the effects of the two approaches on depth of field (and its cousin, out-of-focus blur performance, this being a factor in the properties of the "bokeh" created from out-of-focus objects).
As always, we need to remember that "depth of field" is an artificial, and rather arbitrary, construct, not a basic optical property. And a parameter of our calculation of this situation is our adoption of a value of the criterion that defines "negligible", or "acceptable" blurring: the circle of confusion diameter limit (COCDL).
Consider now two cases. In each, the camera position is the same, the camera is focused on a subject at the same distance, and the aperture (as an f-number) is the same. Suppose that we start with a 50 mm lens, and conclude that the framing of the desired scene is "not tight enough". Consider these two cases:
Case A: We replace the 50 mm lens with a 100 mm lens, and use the full frame for our delivered image.
Case B: We leave the 50 mm lens in place, but crop the taken image to a portion having half the linear dimensions of the full taken frame.
We note that:
• The resolution of the system for the subject, in terms of "lines per subject height", is half as much in case B as in case A. (This is not an announced topic of this note, but is significant nevertheless.)
• As always, the depth of field depends on which philosophy of choosing a COCDL is used.
If we choose a COCDL based on the amount of blurring that would be noticeable by a typical human observer, and assume that both images would be viewed at the same delivered size and from the same distance, then:
•• The depth of field for case B would be about twice that for case A.
If we choose a COCDL based on the amount of blurring that would "noticeably degrade" the resolution potential of the system, then:
•• The depth of field for case B would be about four times that for case A.
Now, as to out of focus blur performance: The two cases are as described earlier. We will assume the camera, in each case, to be focused at the same distance, and a consistent aperture, as an f-number. We will consider the size of the blur circle created in the image from a "background" point source at the same distance in each case.
We will compare the diameter of the blur figures in the two cases compared to the delivered image size (we can also say, compared to the size of the principal subject as seen on the delivered image).
Then:
• The (relative) diameter of the blur figure in case B would be about 1/2 the diameter of the blur figure in case A.
Interesting.
Best regards,
Doug
• Use a lens with a "greater" focal length.
• Crop the taken image to produce the delivered image.
I mentioned the matter of, among other considerations, the effects of the two approaches on depth of field (and its cousin, out-of-focus blur performance, this being a factor in the properties of the "bokeh" created from out-of-focus objects).
As always, we need to remember that "depth of field" is an artificial, and rather arbitrary, construct, not a basic optical property. And a parameter of our calculation of this situation is our adoption of a value of the criterion that defines "negligible", or "acceptable" blurring: the circle of confusion diameter limit (COCDL).
Consider now two cases. In each, the camera position is the same, the camera is focused on a subject at the same distance, and the aperture (as an f-number) is the same. Suppose that we start with a 50 mm lens, and conclude that the framing of the desired scene is "not tight enough". Consider these two cases:
Case A: We replace the 50 mm lens with a 100 mm lens, and use the full frame for our delivered image.
Case B: We leave the 50 mm lens in place, but crop the taken image to a portion having half the linear dimensions of the full taken frame.
We note that:
• The resolution of the system for the subject, in terms of "lines per subject height", is half as much in case B as in case A. (This is not an announced topic of this note, but is significant nevertheless.)
• As always, the depth of field depends on which philosophy of choosing a COCDL is used.
If we choose a COCDL based on the amount of blurring that would be noticeable by a typical human observer, and assume that both images would be viewed at the same delivered size and from the same distance, then:
•• The depth of field for case B would be about twice that for case A.
If we choose a COCDL based on the amount of blurring that would "noticeably degrade" the resolution potential of the system, then:
•• The depth of field for case B would be about four times that for case A.
Now, as to out of focus blur performance: The two cases are as described earlier. We will assume the camera, in each case, to be focused at the same distance, and a consistent aperture, as an f-number. We will consider the size of the blur circle created in the image from a "background" point source at the same distance in each case.
We will compare the diameter of the blur figures in the two cases compared to the delivered image size (we can also say, compared to the size of the principal subject as seen on the delivered image).
Then:
• The (relative) diameter of the blur figure in case B would be about 1/2 the diameter of the blur figure in case A.
Interesting.
Best regards,
Doug