Fujifill GFX 50S — summary

This is the 51st in a series of posts on the Fujifilm GFX-50S. The series starts here.

It’s been more than a month of testing the GFX every day, and now I’m getting ready to actually use the camera for a project, so now is a good time to summarize my thoughts.

Ergonomics, displays, menus, buttons, handling. Excellent. The interface is little strange for a non-Fuji shooter like me, but it’s logical once you get it figured out. Thoughtful controls, nicely customizable. I wrote up a detailed summary here. One thing I didn’t mention in that post was the large color live histogram that’s accessible at the touch of a button (I changed the button). I’ve never seen that before, and it makes exposure selection easy and as accurate as can be given that it’s still not a real raw histogram. The articulated EVF adapter works well and is sturdy (not cheap, though).

Native Lenses. Very good (the 63) to superb (the 120 macro) image quality. I am not a focus-by-wire fan, however, in this or any other camera. The fact that the 63 doesn’t seem to be able to constantly hold the focus point even with the camera continuously on is a worry at best, and a random image-damager at worst. [Added later: I’ve now tested the 110/2 and the 23/4, and they are up to the high standards of the 120/4. The 110/2 is in some important ways a better lens than the Zeiss Otus 85/1.4.]

Adapted lenses.  Few (I’ve only found one, the Zeiss 135/2 Apo Sonnar) full frame lenses cover completely unless they are macro lenses. Several full frame lenses do well at full-height images with 4:5 and 1:1 aspect ratios, though. I have not found a full frame lens with a focal length shorter than 55 mm that gives a really good performance at a broad range of apertures out to even the corners of a 1:1 image. Hasselblad HC lenses are a mixed bag, with some providing mediocre performance at the fine pixel pitch of the GFX, and some doing all right, but none are great. HCD lenses would be better, but I haven’t tested any. Note that I have not tested the following HC lenses: 50, 120 macro version II, 210, or 300. One of the best V-series Hasselblad lenses, the 250/5.6 Superachromat, works well on the GFX, but the ergonomics are not good. Unless you really want to use the corners, you’d be better off with Leica R glass in similar (180mm and 280mm) focal lengths. A lot of the Hasselblad V-series lenses can’t utilize the sharpness that the fine-pitched GFX sensor has available. Long Leica M-mount lenses like the 90/2 Apo ‘cron and the 135/3.4 Apo-Telyt also work fine, except in the extreme corners.

Manual focusing.  Once you realize that the peaking is overactive and that the max magnification is anemic, and you figure out how to get around all that, it’s pretty good. Not great, for sure, but good enough to get consistent focusing that’s nearly optimum (with lenses this sharp, you’ll not get bang on focusing every time no matter how good the focus aids are, IMHO).

Autofocusing. Inconsistent with both the 63 and the 120. However, the problematic f-stops are few, and you can work around them while we wait for the firmware fix that I devoutly hope is coming. The errors are small enough that they won’t affect many kinds of photography. I didn’t do much with autofocus speed, but the mechanism is contrast detection, so don’t expect jack-rabbit performance. There is no inherent reason why CDAF should be inaccurate, though, if you give it enough time. [I’ve found that changing focus modes can ameliorate this to some extent.]

Sensor. Exemplary. 14 bit precision in all modes, including electronic shutter. No gratuitous signal processing that can’t be turned off. Low read noise. Decent full well capacity. Good photographic dynamic range. I was worried that the small microlenses that are part of what makes the sensor so sharp would cause a lot of aliasing; they do cause some, but it’s not much worse than 100% fill factor cameras. I was a little put off by the ¼ second scan time in ES mode, but upon reflection, I’m glad they did it that way and kept the 14-bit precision even when using the fully electronic shutter.

IQ as compared to the a7RII and the D810. With the same top-notch lenses on all three cameras, the GFX is better. It measures better and it looks better if you pixel-peep. However, you’ll have to print large to see a material difference in the finished work. With 15 inch (38 cm) high prints, you won’t see much of a difference. With 30 inch (76 cm) high prints, if your technique is good, you will see a worthwhile difference. But here’s the surprise: at least with the two Fuji native lenses I’ve tested, you see a much more significant difference if you compare all three cameras with their respective manufacturers’ lenses. Whether this quality advantage that Fuji has — with their figurative toe in the lens water — will continue when they have a complete line of lenses is an unanswered question. [Added 10/18. That question is answered in the affirmative, at least so far.] Fuji lenses prices are high by full frame standards, but bargains compared to Hassy H-mount and Phase One. Unless we start seeing some top-notch third-party G-mount lenses, ultimately, the main potential image quality advantage of the GFX rests on the Fuji lenses that are to come. Sure, some may buy the camera to get a skosh more out of their Otus 55s and 85s, but that’s never going to be a big market.

What’s it good for? Landscapes. Still lifes. Studio work. On-location portraits. Fashion. Weddings, but you’ll probably want to bring along something that focuses faster for the candids.

What’s it not good for? Sports. Wildlife.

What’s it overkill for? Travel photography, unless you’ve got a market that really wants image quality. Product shots. Street work.