RWR wrote:
Isn’t that true with any lens?
The underlying truth in what he said was this:
If you are using a dSLR/SLR camera, focus is achieved with a different system from the sensor or film. That focus system is aligned at the factory, which is one good reason you are NEVER to touch the reflex viewing mirror, other than with a gentle puff of pure air.
However, despite perfect mirror/focus screen alignment, there is some variability in LENS MOUNTS. That means you will inevitably mount some lens that does not focus exactly where you saw it focus in the viewfinder.
When using an f/0.95, f/1.2, or f/1.4 lens wide open, depth of field is razor thin, especially at close focusing distances. That shallow DOF may be less than the margin of error designed into the focus tolerances for the entire system. In that case, the focus point you GET is not the focus point you SAW in the finder!
You can test for this by placing 25 dominoes upright, one inch apart, on a diagonal line drawn on mat board, so you can see each tile. Label the center one '0'. Clearly label the others from -12 to +12. Put the camera on a tripod, open the lens all the way, meter, and focus MANUALLY — very carefully — on the tile labeled '0'. Make an exposure. View the image at 100% in your post-processing software, and see whether or not your focus is accurate. If the '0' tile is the sharpest, focus is normal.
Note that you have to do this with EACH LENS. The issue arises through a COMBINATION of the tolerance variations with body and lens. Conceptually, if a camera body is off by +3 "units," and the lens is off by -3 of the same units, focus is perfect. But if they're both off by +3, the net result is +6. If the tolerance is +4, you have a problem!
None of this is an issue with mirrorless cameras, because they focus directly, using the sensor itself. Assuming you do your job right, either manually or by positioning the focus point, mirrorless camera focus is accurate every time.