Apologies if this is in the wrong sub-forum but I thought this might be of interest to the MF users as well.
I've been thinking for a while about how the P645 DA 28-45mm zoom might work on the A7r with shifts, and now with the release of the GFX if there is potentially any available movement on that camera.
Now this is far from a perfect test. The equipment I used consisted of a Kipon P645-EF adapter mated to a Metabones, with the A7r attached.
The Kipon adapter is fairly well made, but the bayonet fits have some rotational play, so take from that what you will! It's also prone to internal flare. I haven't quite cracked it yet with the flocking, so I did notice a bit with this lens even on a very overcast day.
I had to fiddle with the adapter so I could set the lens at about F11. This isn't precise as there is no aperture ring on the SDM lenses, or lever to set aperture on the adapter, so I had to use the 120 macro to set the pin push and tape it down in place.
The lens is a rental from Calumet, so no clue on its history and no chance to cherry pick (which I try to do when buying lenses)
The weather this weekend was just solid cloud, so the light was really flat and grey. The max shift on the adapter is 15mm.
So what did I find?
@28mm there is nearly 12mm of usable shift along the long edge, with only a slight amount of the end of the IC in the corners. By 15mm the IC is clearly cutting off. See the example below which shows shifts of 0-5-10-12-15mm
Along the short edge there is no cut off of the IC till 15mm, and then it is only just in the corners and could be corrected fairly easily dependent on subject.
@35mm there appears to be a clean 12mm of shift along the long edge. At 15mm it appears the adapter is now cutting off the image, noticeable by the heavy darkening at the top, but note the straight shape to it. Clearly different to the end of the IC @28mm
The Kipon adapter does have a fairly narrow opening, and looking at it, the rear lens element is further away from this opening than the other P645 glass I have, especially between 35&45mm, so I think the beam angle is getting cut off.
@45mm the results are identical to @35mm. It looks to me that if the opening was bigger, or this lens was on a Cambo/Arca, there could be a cleanly usable 15mm of shift along the long edge, if not more.
From roughly working it out, I'd say there is about 64mm of image circle projected @28mm, hard to say at the other focal lengths because of the adapter cut off, internal baffling may come into play at certain settings.
Overall image quality is OK. I didn't do any head to heads with my 24mm TS-E II but I'd say it's as sharp in the center, if not a little better than mine. Towards the edges it drops off but its gradual, there is no sudden drop off or smearing of detail.
If just shooting for web/small print sizes, I'd probably be happy with it. It's much more even across the frame than my TS-e, which seems to be worse off to one side. Sharpness seems best between 32-40mm, there is a drop off at 45mm.
At larger print sizes it may not hold up, but I'd have to try it to see for sure.
There is noticeable barrel distortion @28mm, but seems easy enough to correct. 32-45mm its pretty much gone.
CA's are kinda bad though, noticeably worse than my TS-e (mine is pretty much free of them). They mostly vanish with LR's auto setting, but there are some still at the edges around 10-12mm. Mostly noticeable when pixel peeping, I can't see them at web sizes.
I mentioned flare earlier, and there was some quite noticeable on a few images. It was a soft and contrast reducing lighter area in the frame. As I sometimes notice it with my other P645 glass I'm putting it down an internal reflection off part of the adapter I didn't think to flock.
Non of the images did I use a hood. The hood is unusable for shifts. It cuts into the image around 6mm of vertical shift. Notable here is that when the lens is mounted in the Kipon, to get vertical shift, it is 90degs to how it would be on the Pentax body. So the shorter petals of the hood are on the top/bottom!!
If you mounted the lens to a Cambo/Arca then the longer petals will be at the top/bottom, so they will cut in even earlier. I don't have a compendium hood at the moment so on a few shots I took I shaded with my hand or filter pocket.
This version may be a bit of a dud, I have read great reviews for the lens online, but also comments that samples vary quite a bit. All of my, much, older P645 glass (55mm/75mm/120mm A) all perform great, even with CA's
What I do find interesting to me, is that on the GFX this would be a 22-35 equivalent!? Now that sort of range is more useful, especially with potential movements on something like the Actus (or a dedicated adapter by someone like Kipon/Hcam!)
Thinking about it you wouldn't have to use as much vertical shift either, due to the squarer frame of the GFX. Since I saw the prototype I've been curious about the GFX as an upgrade/move to MF later in the year.
When the GFX is available for hire locally, I'll try and put something together to test the combo out, even if i have to buy an adapter.
Below are some examples of vertical shift in portrait orientation @28&35mm (0-5-10-12-15mm shifts) and some vertical shift in landscape orientation.
They have been played with in LR, using my usual starting points for archi images, but I have a load of Raw files on my dropbox, if anyone would like a link to download and have a look at them on their own machines. Just PM me.