Separate names with a comma.
Discussion in 'Adapted Lenses' started by Phocal, May 10, 2015.
My understanding is that focus tuning is needed on cameras that use phase detect sensors that are physically separated from the image sensor.
These would be seeing the image via a secondary mirror when the main mirror in a DSLR is down. The focus tuning compensates for the inevitable difference in distance of the phase detect sensors vs. image sensor due to tolerances in the mirror assembly.
The EM-1, which is what you seem to be using for this test has the phase detect pixels on the image sensor itself, so there would be no variations.
So, I don't understand what focus tuning would buy you in this setup...
phase detect works by using the differences in phases of light to calculate how far the lens needs to move it's elements to adjust focus. You can have issues where the AF sensor off the mirror is out of perfect alignment with the sensor on an SLR. That is not an issue with mirrorless. But the other issue you can have is a lens that is not starting at "zero" for lack of a better description. The AF sensor says move to position 25 to achieve focus. Lens thinks it is on zero and moves 25 positions. However the lens was miscalibrated and it was actually starting from position -4, meaning it is only at position 21 when all is said and done and you have poor focus. Phase detect is not like Contrast detect where it sits and samples focus continually until it is perfected. It just calculates where focus will be and goes there (which is why it is so great for tracking and speed). If the lens doesn't go where it was told, focus will be off.
If it was just a matter of the body being off due to the mirror path, then every lens would need the same correction and it would be a simple matter. Sadly, that is not the case.
At a glance I can't think of any reason why the distance would be different however in practice I've found I needed +5 on the tele end and -3 on the wide on one of my lenses, using the EC-14 it needed -5 on both (35-100m f2.0 at 100mm f2.0 this is a difference of around 5mm at distance 3 meters). The interesting part for me was without the focus adjust on the EC-14 at the wide end it would actually focus slightly past infinity (at closer focusing distances it was very hard to tell any difference at all).
One reason I can think of is that the polarizer (or PD part of the sensor stack) may be a different thickness than the bayer filter of the rest of the sensor and there is a slight difference depending on the focal length between where it thinks is in focus vs reality. Shooting wide open at closer distances body movement is going to make most of this moot, however it's best to start with the most accurate so that by the time the picture is taken the deviations are less likely to add up to out of focus.
To expand on that slightly, when it's hunting the camera tells the lens to go to the end of range -> the lens starts moving (towards infinity or closest, whichever the camera specified) -> The body tells it to stop and the lens stops focusing and announces its current position and waits for instructions -> the camera tells the lens to move to X position (based on what it's last position was). The lens itself can only really say "moving", "stopped at X due to interrupt" and "at the position you requested", the body has all the smarts (it can say other things related to aperture/IS/etc, and there's a handshake when it's first attached which defines all of the ranges it can operate at naturally).
If it really runs open loop then this seems like a poor implementation on the EM-1.
Since the camera does contrast detect for native u4/3 there is no reason why it could not self calibrate the phase detect.
Move the 4/3 lens using the phase detect sensor to where it thinks it should go and then do fine adjusts until the contrast detect says it's perfect.
Store the difference as an adjustment in the lens and be done. Sounds like a firmware update worth asking for...
My recently acquired 50-200 SWD lens benefits from some adjustment (at the long end) on the E-M1 as I think the lens may vary.
Following along as best I can. This interests me because I have a 50-200 SWD that seems fine native, but degrades slightly with the ec-14, and poor with the ec-20. So I'd like to try and adjust focus. How necessary is the software? Targets are cheap.
I finally got around to focus tuning the 50-200 with the EC-20 TC. Focus with it was too soft. I don't think I'd ever had an in focus shot with it.
The process is easy, but tedious. I estimated the final adjustment. One day I'll go back to do the fine, one-step tune.
Results were very good. Getting a shot in perfect focus with a hand held 800mm equivalent is still tough for me, but at least I can get them now. Surprisingly, focusing is much faster than it was. It still hunts once in awhile, but with a little ore-focus, it was quite snappy most of the time. I didn't expect such a difference, but there was.
Still best resting on something solid, but I suspect I will be just as good and likely better with it than a MF lens. So I scratched a list of wanna have adapted long telephotos off my list.
Picking a fast-moving bird or insect out of the air is still a dream, but the situation is much improved. This lens is so versatile. As much as I want the 40-150 Pro, it's hard to justify.
Anyway, here is my sharpest shot with the EC-20. It's not a great picture of anything, just a dead flower head, but shows the best I could get with the camera supported. What do you think? Is it pretty sharp or am I a victim of wishful thinking?
Autofocus on the dead flower head.