Ben asks: In your experience is it necessary to test and micro adjust m.Zuiko lenses?
Short quick answer: For general shooting, there is no reason to AF adjust native lenses in the m4/3 system.
Long answer: (if you care for the tech stuff and reasons why, and have time to read it!)
Before we dive in, for those that don’t know… Some cameras have a feature that let you fine tune your Auto-focus points. On the E-M1 mkII you can find it from Custom Menu A2 AF Focus Adjust option. That is what is in question here.
Let me explain my background on this first, and then what I have found with current lenses.
I am a big rangefinder fan. Leica’s taught me how to spot focus errors very quickly. Leicas, as long as your eye works, and all the machinery is calibrated, are perfect focus tools. The catch was having the equipment aligned right. If my camera was serviced, and the image didn’t line up with what my eye saw in terms of focus, I knew my lens was wrong. This always worked out exactly. Many rangefinder lenses front and back focus due to focus shift- an inherent tendency when being stopped down. Not all lenses do this though. Or noticeably at least. A good rangefinder shooter learns to see how much focus shifts based on image review and is able to compensate focus accordingly.
When shooting SLRs, I shot canons for over 10 years, I would feel that a lens was potentially not focusing right. I could usually tell within one shooting session if a lens was off, or if there was a potential autofocus issue. Once we got auto-focus adjust on DSLRs, a whole new world opened up. I would use a LensAlign ( http://michaeltapesdesign.
So now we come to mirrorless. When I first started using mirrorless cameras I was amazed that I never felt like my focus was wrong. Nor did I feel compelled due to the images to check my lenses. It all just seemed right. Contrast detect focus works. However, we have cameras like the E-M1 series that also have AF adjustments built in. The E-M1 was touted as the successor to 4/3 SLRs and it is compatible with the excellent lens range. If you are using 4/3 lenses with an adapter, I would absolutely be using a lensalign to adjust focus. I would do so as a whole for all the points together, and not for individual points. No reason to go through all that work. Unless you have some exotic reason that one point needs to front or back focus intentionally. Very rare cases. 4/3 lenses when mounted with the adapter use the phase detect focus exclusively.
Using m.zuiko lenses, or other native m4/3 lenses like panasonic, sigma, etc… there is conventionally no need to use this adjustment. Contrast detect focus is working from the sensor data directly for focus. (most basic explanation) This inherently makes it much more accurate when focus is achieved, and I have never felt or seen a lens not be right if the lens was in proper order. (if you drop a lens and there is a misaligned element, that is totally different) This is also why we can focus these cameras while shooting in infrared and not need to make adjustments. They are shooting based on the live sensor data.
When it comes to on-chip phase detection and contrast af, Olympus provides no data that I am aware of that says how much is one used over the other. Being that it is on-sensor phase detection, it makes me wonder if this form of phase detection is more accurate since it is using the image on the sensor as well I would think. I feel Olympus is using the phase to quickly get to 90% and then fine focus based on contrast. This would eliminate the need to use the adjustments as well. During continuous shooting, I have heard that only phase is used. In this case an adjustment might be merited, but the problem is how would you use a lensalign with continuous autofocus? If the camera switches to contrast when in single autofocus, we have no hard way to test and adjust for pure phase on a native lens. It would be great if we had more insight as to how the autofocus system works.
I might have to give my lens align a try with continuous autofocus and since neither camera nor target move during testing, I wonder if continuous focus would freeze enough to get a proper reading.