jerryc41 wrote:
The general rule of thumb I have heard for years is two stops down from maximum - just a "Rule of Thumb," though.
That is in regard to the sharpness at the point of focus, which is not what the OP is interesting in.
He is asking about sharpness across a range of Depth of Field. Or, if one needs good sharpness from this distance to that distance... do you just use a DOF table to match the total distance, or is there more to it? And indeed there is a lot more to it!
The old rule of thumb about stopping down two stops relates to where the quality of the lens surface (how accurately it is shaped during manufacture) becomes the major cause of poor focus (i.e., at a wide open aperture) compared to the point when further stopped down where diffraction becomes the primary cause of blur. Somewhere between those two apertures will be the point where the focused distance is the sharpest the lens can produce.
Of course that is at some fixed DOF and may not be the DOF needed for any give image.
For an idea of what happens at different DOF distances, particularly in respect to larger distances where diffraction becomes more significant, there is a really useful calculator at
http://www.tawbaware.com/maxlyons/calc.htm which can show a graph that includes diffraction. You have to go down just below the graph and click on the button for "Add diffraction blur to focus blur" before calculating DOF.
A good example to play with is the use of an 85mm lens, focused at 10 feet, mounted on a camera with an APS-C body. A traditional DOF chart suggests that using f/2.8 will provide a narrow DOF of about 5 inches, while f/24 will increase the DOF to 3' 10". It ain't really so! Add in the diffraction and certainly the DOF for f/2.8 is about 5 inches, and it is sharp at the point of focus where the blur diameter is about 0.003mm. But at f/24 we don't get 3' 10" of DOF where the CoC stays below 0.019. Instead we get only a distance of maybe 1 foot. But horror of horrors the blur size at the point of focus is 0.015, which is nearly (at it's best) as large as the worst case blur of the edges of calculated DOF for that sensor! (F/24 was selected because it hit right at 0.015mm blur on the graph. At f/30 the blur from diffraction is equal to the desired CoC maximum of 0.019mm, and DOF is non existent!)
At that particular distance/focal length/sensor combination, f/16 might work for a portrait, where narrow DOF is okay and a sharp image is not the point. If wouldn't work well for a landscape shot!
For portraits the way to get sharper images is the use of more powerful strobes! For landscapes the answer is wider angle lenses.