I'll not completely agree with Josh-H : beware that an hyperfocal calculation may well decrease your sharpness, especially at infinity...
Focusing at 1.20m at 17mm f/8 will give you at infinity a blur circle about 30µm on the sensor, ie a blur extending on about 4 or 6 pixels depending on the camera you use. Would you call that "tack sharp"?
Moreover, blurs always add to each other ; in other terms, even if the defocusing blur circle is a bit smaller than the pixel pitch, you may still see some small loss of sharpness relative to the perfect, in-focus zone (this especially if you shoot with a very sharp lens and/or a sensor with weak or inexistant AA filter and print big, so we may count how many angels can sit on the top of a microlens here).
That said, some small "differential sharpness" may be quite efficient to separate a foreground subject from its still-readable background, adding a tad of that much-praised "3d look" to your image.
For me who am affected with a bit of myopia, the contrary (blurred foreground, sharp background) is not as efficient, but your myopia may vary - and I'd say giving the maximal sharpness to the main subject may be the most efficient strategy in most cases.
Bottom lines :
- don't ever expect the result of an hyperfocal calculation to be equally sharp, especially if you take into account the 30µm circle of confusion dating of the middle of last century,
- focusing on a physical subject rather than on a calculated distance may be first easier and second prettier,
- to gauge the effect on a print, carefully inspect the print itself, taking into account the subject matter.
I fully agree with Nicolas.
Most criticism about hyperfocal calculations is not founded, it usually comes down to improper use of the method. Unlike film, digital sensors have a hard limitation of maximum resolution, namely the sensel size. Detail cannot be sharper than the sensel size. When using that sensel size, or rather the sensel pitch, as a COC criterion, then nothing can be sharper, even at the horizon. When the COC is set to the sensel pitch, then the hyperfocal blur circle at infinity will be exactly the size of one sensel.
Nicolas is correct that lens blur and the diffraction blur add (their MTFs multiply), but there is some leeway due to the AA-filter and the Raw conversion process, and the overpowering effect of defocus on the combined system MTF. In my experience a COC of 1.5x the sensel pitch will not produce visible blur, especially after proper sharpening. With advanced (deconvolution) sharpening one might push it a bit further, but my rule of thumb of 1.5x sensel pitch works fine for me with the planning of a shoot. It also tells me, should maximum DOF sharpness be needed, at which distances I need to focus for focus stacking). BTW, there is also a rule of thumb for the aperture number that produces a diffraction blur of 1.5x the sensel pitch. Just take the sensel pitch expressed in microns, and multiply that by a factor of 1.108, and the aperture number is the result. So a 6.4 micron sensel pitch camera will start to suffer from visible loss of microdetail (at the pixel level!) at apertures narrower than f/7.1. How much of that loss of microdetail at the pixel level is still tolerable is a matter of personal choice. Perhaps the image will not be blown up to huge proportions, in which case the overall DOF effect becomes more important than pixel accurate microdetail.
Then focusing at the indicated distance(s), either manually (for focus stacking where the camera shouldn't move) or with AF will work fine, although at larger distances it's still a bit of a guess (unless you use e.g. a laser distance finder).
Having said all that, there can be a clear benefit to just focusing on the main subject, and letting the DOF fall where it does, but that depends on the scene at hand, and the output size.
Cheers,
Bart
