Sample images here: Fujinon GF 20-35mm f/4 R WR Sample Images
During my recent trip to Victoria, where I tested the Fujinon XF 8mm f/3.5 R WR, I also had an opportunity to shoot with a Fujinon GF 20-35mm f/4 R WR medium format zoom lens on a borrowed GFX 100S body. The GF 20-35mm is not a new lens, but one that has been shipping since mid October of 2022, however it is not a lens that Beau Photo has on demo, nor in rentals unfortunately, so I’d never had a chance to test one out previously.
Similarly to the XF 8mm, I was somehow not expecting to be blown away with this lens. Why? Well first off, Fujinon GF prime lenses are incredibly good, and I have shot a fair bit with the Fujinon GF 23mm f/4 R WR, which is stunning and one of my “must-have” rated lenses for a GFX. For an even wider angle zoom lens to be as compact as the GF 20-35mm is, and only a slight bit larger than the GF 23mm prime, I was expecting some compromises. Strictly speaking, I suppose there is one main compromise: the lens is not fully corrected optically for distortion. That is not at all surprising to me, given the size and focal length range of this zoom, but what did surprise me, similar to the XF 8mm, was how little the software correction impacted image quality.
Spoiler: I am also rating this GF 20-35mm to be a “must-have” lens for a GFX system, especially if you are doing landscape, architecture or anything where you want a very wide field of view and/or a deep depth-of-field, which can be a challenge on medium format. If you open the sample gallery (linked to above) and download some of the large images, you’ll see what I mean: stunningly sharp from corner to corner! The samples are only 24 MP and not 100 MP, but trust me… they look just as good at full resolution.
Example of distortion auto-corrected and uncorrected with the Fujinon GF 20-35mm at 20mm
Let’s get the distortion comment dealt with first: yes this lens suffers from barrel distortion at the 20mm end (with a somewhat more complex characteristic than the XF 8mm, as you can see in the above animation), and that distortion transitions to pincushioning at 35mm. However as mentioned, I honestly couldn’t see any obvious impact to sharpness, even in the extreme corners. Fine branches, brick detail in buildings, everything I looked at closely was rendered beautifully. Now I’m not sure if software distortion correction has just gotten a lot better in recent years, but whatever the reason, the sharpness, resolution and overall look of the image, even in the extreme corners, is near flawless. I should note that out-of-camera jpeg images are perfectly corrected, and most raw converters (Capture One, Lightroom, Photoshop etc.) will default to correcting this distortion automatically, so you will likely never see what the animation is showing unless you use a program that allows you to disable lens corrections, like Capture One. The pincushion distortion at 35mm, the long end of the zoom, is less overall and has a smoother and less complex characteristic than the barreling at 20mm. Sorry, it turns out that I did not have a photo at 35mm with nice geometric detail, where the distortion character was as easy to visualize as in the above shot.
While in Victoria, I was under the impression I would have the lens for one more day, but as it turned out, Fujifilm needed it back for another scheduled demo, so I did not test it quite as thoroughly as I would have liked. For example, I didn’t do any flare tests with it, although looking at some shots where the sun was only just barely outside the frame, I don’t see any issues. Given Fujifilm’s excellent track record of making lenses that are extremely flare resistant, I expect it will perform well, but still… apologies… I can’t actually officially report on its flare resistance for this review. As far as chromatic aberration, this lens basically doesn’t have any. When looking very closely at the extreme corners and toggling the CA correction on and off in Capture One, it actually gets worse when it’s toggled on! With it off, there is basically no CA. When I was checking it previously, I just looked at the images and toggled the CA correction on and off, incorrectly thinking the slight CA was for the uncorrected version. It would seem the lens profile in Capture One needs to be tweaked. Most impressive performance for an ultra-wide zoom!
Mechanically, the lens is seems typically well built for a Fujinon. The zoom and focus rings are smooth, the aperture ring has precise feeling detents and overall, the build quality feels great. It is not a very large lens either, being noticeably smaller and lighter than the Fujinon GF 32-64mm f/4 R LM WR for example. The GF 20-35mm can focus really close as well, down to 35cm which offers 0.14x magnification, and that is great for such a wide angle lens. Vignetting is also minimal and even wide open, I’d say that for most shots, you wouldn’t really need to correct for it.
If the lens has a weakness, then the only one I can really come up with is that at medium distances and wide apertures, it can exhibit unpleasantly nervous bokeh. If you download the wider photo of the rose bed, shot at 35mm and wide open at f/4, and look around carefully, some areas will exhibit what almost looks like a double image where the background blur doesn’t look very nice at all. However, who buys an ultra-wide zoom to shoot wide open to try and get background blur? I think that usually, the 20-35mm will be stopped down to get the best corner quality and deepest depth of field… at least that is likely how I would use it.
Speaking of stopping down, as far as edge and corner sharpness, there is hardly any improvement when shooting at f/8 compared to f/4 at 20mm. The corners are already very sharp even wide open, and only going into the absolute extreme corner is there any hint of softness. Due to my limited time with the lens, I also did not perform my usual corner sharpness tests at longer focal lengths, but I am fairly confident it will perform similarly well. I just shot a quick test series at 20mm while waiting for my fiancée to finish at the conference she was attending, but that is all I had time for.
One big advantage of a lens as wide as the 20-35mm, is the ability to actually take shots where everything is sharp, from a distant background to close foreground detail. The vertical shot of the Empress Hotel, with the flower bed extending to the bottom edge of the frame (only a few feet from the camera) is a good example. It was shot at f/16, and everything is nice and sharp. The last few images in the gallery, those only at f/11, also show very good sharpness from the closest foreground detail, to the distant background. To achieve hyperfocal shots like that on medium format, without resorting to a tilt-shift lens, you need an optic as wide as the 20-35mm.
So in conclusion, the GF 20-35mm f/4 R WR is a very impressive lens. Even though it requires some software correction for distortion, image quality holds up well when compared to what you’d get from the GF 23mm f/4 R WR for example, even though the 23mm prime lens is perfectly corrected optically for distortion. As much as I love prime lenses, if I were buying into a GFX system, the GF 20-35mm f/4 R WR would likely be my ultra-wide of choice and not the GF 23mm. As much as I love both the GF 23mm and GF 30mm f/3.5 primes, I think a kit with the 20-35mm, the upcoming (roadmapped) GF 55mm f/1.7 and the GF 80mm f/1.7 would be my ideal three lens kit. I could easily do 95% or more of what I normally shoot with a kit like that, and that is very compelling indeed!
Sample images here: Fujinon GF 20-35mm f/4 R WR Sample Images
∞
Note that the photo of the 20-35mm zoom sitting on the wood stump in the headline image of this post was supplied by Fujifilm Canada and is credited to Adam Gibbs.
