Imagemaster wrote:
It may be an issue, but it is such a small issue in comparison to the other factors that it does not even require any consideration. To think that it is a necessity to worry about sensor size for DOF, whether planning on making an 8x10 print or a 30x40 print, is a joke.
The sensor size is not a joke. If a 22.5mm x 15mm 8 MegaPixel image is enlarged to 8x10 and compared to a 36mm x 24mm 8 MegaPixel image enlarged to 8x10, it makes a huge difference to how much the circle of confusion is enlarged. Joke's on you.
We'll keep on saying it because it is true: perceived depth of field depends on linked factors: 1) Chosen acceptable circle of confusion size in the print, which is related to 2) print viewing distance, 3) print size, 4) degree of enlargement of image to print size, 5) focal length of the lens, and 6) distance to subject focus plane. The sensor size directly interacts with the focal length of the lens, the distance to the subject and the desired angle of view.
Note that circle of confusion at the sensor is not the same as circle of confusion at the print. Most DoF calculators pick a CoC at the sensor based on assumptions about final print size and viewing distance and acceptable print CoC that are often hidden and not stated. Properly done, DoF calculations start at the print and work backwards or include all the linked factors in an interactive application where some are declared fixed and others are declared variable.
Some of the factors are interconnected because you can hold one constant and calculate a couple other in terms of each other or in other combinations of constant and variable. This is why some people will say sensor size doesn't matter because they can make it a dependent variable by holding other factors constant or alternatively they can adjust focal length and distance to compensate for sensor size.
Because of the inherent variability of so many factors, if DoF is critical to a shot, I bracket for DoF by using several apertures.
Do the math (geometry). Math rules the universe.