Originally Posted by Sheiky
Both, actually. Primarily, phase detection AF is the one that's used - that's the camera's AF system. In bodies with Live View, you have the option of using contrast detection AF (with or without face detection) or the 'quick mode' where the mirror flips down briefly so the phase detection system can be used.
Inaccurate but precise focusing with a given lens (consistent front-focusing, for example) is due to a misalignment of the lens optics, which project the focal plane of the lens either in front of or behind the plane of the AF and image sensors (the reflex mirror reflects most light up to the VF, some light down to the AF sensor, which should be in the same optical plane as the image sensor, and if it's not, that's what you'd use the [Adjust all by same amount] in the AF Microadjustment C.Fn to correct). Since the AF calculations are based on this same image, I'd expect this problem to affect focus confirmation as well. AF microadjustment will correct this inaccuracy, and thus if you AF microadjust a MF lens (yes, that sounds like an oxymoron), focus confirmation should be corrected as well.
So, if a lens has a simple and consistent inaccurate AF, my guessis that both AF and manual focus confirmation would be affected, and that performing an AF microadjustment would correct both.
On the other hand, if a lens has randomly imprecise AF (sometimes FF, sometimes BF), that's a mechanical issue with the lens and nothing is going to correct that, but MF with a focusing screen will be fine.
Then there's the additional case of a lens with inconsistently inaccurate AF - meaning that at some subject distances it consistently front-focuses and others it consistently back-focuses. I hadn't considered this possibility until I ran across it inRoger's (lensrentals.com)description of the Sigma 50mm f/1.4 lens- hecalls this 'schizophrenic AF' ("Closer than 5 feet it will front focus, further than 20 feet it will backfocus."). There's no way to adjust for that, it must be lived with. If he's correct, it may well account for the varying reports of AF issues or non-issues with this lens - more obvious on FF than crop due to the greater DoF with crop, and highly dependent on what you shoot. (Also highly dependent on how you test - for example, the LensAlign Pro instructions suggest testing a 50mm lens at 4 feet from the target, where the Sigma would apparently front-focus, but Canon's guidance would suggest testing it at 8 feet, where it would apparently be fine.)




Reply With Quote