Originally posted by Topsy I'd be surprised if this method would change focus confirmation for manual lenses?
It's funny, I was playing around with it just yesterday, annoyed by having to look at the display all the time when shooting. (one thing that made me enjoy film more recently)
Messing around with it seemed to improve the confirmation somewhat but I put that down to placebo/a wide range of what the camera thinks is "in focus" and luck with getting the sharper end of it a few times in a row.
Nearly started a topic but thought better of it last night!
But I may not entirely understand how it works with MF lenses to be honest!
I’ve performed focus confirmation calibration for some of my M series lenses which has improved my results with manual lenses. Firstly, as there is a range where the focus confirmation light will come on, I consistently focus my manual lenses from a setting short of what will be the correct distance out to the distance where the confirmation light comes on. This eliminates the variation that comes from alternating which direction I approach the focus point from. I then fine tune the focus confirmation using the Apply All sub menu, as Apply One only works with chipped lenses (F and later). Apply One (where set for an autofocus lens) I believe overrides the Apply All setting so this method doesn’t upset your autofocus lenses calibrations. I don’t think it matters what method you use for calibration - whatever you are comfortable with/find quickest.
Why do manual focus lenses need focus adjustment? Firstly the relationship between the distance from flange to focus sensor and flange to the imaging sensor needs to very precise (I would presume equal?). Bearing in mind the path to the focus sensor passes thru the main mirror and bounces off a secondary mirror behind that, it’s not hard to see where some camera to camera variation will come from. Presumably some calibration is done in factory, but where the mirror settles on its stops after some use might take a while to become consistent and that in itself influences the distance to the focus sensor. That would be in addition to any manufacturing tolerance variability of which there will always be some. Secondly, aberrations of wide open lenses seem to impact focus confirmation accuracy/reliability (not all colours will be focussed at the same point) so you are also attempting to account for this. And we know most older manual focus lenses have a fair few colour aberrations wide open.
I’ve only check a few of my short tele M lens this way and the same Apply All setting seems close enough for those settings. Once done I was much more successful focussing my M lenses with consistent results at wide apertures.
---------- Post added 13th Jul 2021 at 12:35 AM ----------
Originally posted by kinkindoll My method to do the AF adjustment is setting up a tripod, use liveview to take a picture first, then go back to OVF to take several pics with different AF adjustment, then use LR to compare the sharpness and accurate vs the pics from liveview. All use center point focus.
That’s the way I’ve undertaken focus fine tuning also, but it can be slow to do, needing sometimes to go back to the target and retest and re-review. The dot method allows all the tuning to be done just using the camera and target, and by its nature requires repetitive focussing activity which will tend to average away the little bit of variability between each focus attempt (greatly lowering the sampling noise). That slight inconsistency from focus attempt to focus attempt can sometimes make the ‘take it back to the computer’ approach a bit annoying slow if you only take one or two shots at each adjustment as you may not catch the shot to shot variability on the first attempt.
You won’t notice this variability between focus attempts while shooting the target with a newer lens without a focus scale but if you have an older lens with a distance scale and mark the live view focus point with a pen on masking tape applied to the moving focus ring, the variability between each focus attempt becomes quite apparent.
The dot approach by its nature leads to dozens of focus lock attempts (done quite quickly) and that higher number of samples will improve the signal to noise (statistically speaking) of the testing process and improve confidence in the result. I quite like it’s approach for that reason and will try it out on my DFA 28-105 as I’m currently torn between +3 and +4 for that lens. My testing suggested +4 but in the field I lean towards +3. The dot method is a different approach and I will be interested to see what that suggests is the optimal adjustment. It’s also a very quick way to assess the adjustment at different focal lengths.