UPDATE April 2: Hasselblad has confirmed the autofocus issue and is working on it. It involves the way autofocus interacts with the spatial frequency of the subject, which fits perfectly with my field observations of the shooting situations that saw an issue.
I really enjoyed shooting the X1D in the field (Carrizo Plain). I could easily see it being my field camera of choice (assuming that 22mm happens along), even if I had to use manual focus 100% of the time. It works well even on a small tripod and rig.
Update 1: the X1D has/had the latest firmware. The 90mm f/3.2 had firmware 0.5.1, which I have now updated to the latest, version 0.5.3. I have no word as yet from Hasselblad as to whether 0.5.3 vs 0.5.1 could be responsible. I would also emphasize that bugs are bugs, and that most of the time the 90/3.2 focuses properly and that I have seen many a camera have focus errors. However, I deem "most" a non-starter from what I experienced. If a bug as I posit, then any claim by others (no matter how skilled) of “works for me” doesn’t work for me.
Update 2: with lens firmware 0.5.3, I have reproduced the problem and will add a few more examples. Also, I am working with Hasselblad and expect to have an alternate X1D and 90/3.2 in 2 weeks or so to re-test.
Accordingly I’m very disappointed to have to report and document this appalling behavior, but it is totally consistent and matches the front-focus errors I saw back in December 2016 in images from a pre-production camera.
Includes full-resolution examples and crops which can be toggled between razor-sharp manual focus and the disappointment of autofocus.
These failures render autofocus unusable in this sense: if the X1D can fail this badly under brightly-lit high-contrast conditions, then the working assumption must be that autofocus accuracy is suspect at all times, thus requiring manual confirmation of focus via magnified Live View every time, thus making the X1D a manual focus camera, at least with the 90/3.2.
I also experienced various short-term lockup problems requiring popping out the battery.
While I have more material to present from the X1D (already shot), the loaner X1D and the lenses must now go back. My intent is to add more coverage of the X1D when new lenses arrive and/or some other opportunity presents itself.
Howard C writes:
I updated the firmware for my XCD lenses this afternoon to 0.5.3 and went out to test the AF performance of the 30mm, 45mm, and 90mm lenses. Well, you were absolutely right. The AF performance of the lenses is significantly off compared to the MF in magnified live view. Not always, but in many cases.
Interestingly, I had tested the performance of the XCD lenses against the GFX zoom at 32mm and 45mm against the XCD primes last week. Sometimes the XCD primes were clearly superior, and yet there were a few other tests side by side where I could not figure out why the performance of the XCD lenses seemed relatively weak. Now I think I know.
I completely agree that the AF cannot be trusted. I need to manually focus in magnified live view. This was an invaluable observation by you. Hopefully, this can be corrected in a firmware update…soon.
DIGLLOYD: I try very hard when evaluating gear to save people time and money on what to buy (what is best for each person) and to save them the trauma of damaged images caused by camera errors. Let me state strongly that in no way am I singling out Hasselblad here; for examples, the Leica SL had (and still has AFAIK) terrible issues, Nikon had a bad focus fiasco once and has problematic AF with many f/1.4 lenses to this day, the Fujifilm GFX has issues, and so on. The point is to be aware of the issues, and to incorporate the awareness into a decision to buy and/or how to use the gear once purchased.
With lens firmware 0.5.3, I have reproduced the problem several more times (easily, without particular effort), and I have added a few more examples. I concur exactly with Howard that it is “not always, but in many cases”. This is what I saw before, and saw again tonight with updated firmware. I would say that 1 time in 5 the AF will mis-focus (a very, very rough estimate)—at least with some subject matter—and I’m NOT talking about tricky 3D or ambiguous stuff, just stuff that the Sony A7R II would nail every time.
While I have not verified the issue with the 30/3.5 and 45/3.5, I have certain images that I am all but certain were degraded by mis-focus. But lacking MF/AF pairs shot wide open, I can only say that at f/8, a slightly soft image makes no sense at all near where focused—f/8 masks the error enough to fool most users.
I am working with Hasselblad and expect to have an alternate X1D and 90/3.2 in 2 weeks or so to also test, but Howard’s comments seem to rule out a single-camera or single-lens issue and support my original findings in every way.