Posted on

The Fujifilm GFX-50s experience

It’s only natural to compare yourself against ‘better’, whatever that means, and when it comes to photography gear, its ‘medium format’ that’s had me wondering.  So I’ve headed out into the landscape for a couple of days with a Fujifilm GFX-50s in one hand, and my Nikon D850 in the other, to find out if Fujifilm’s medium format system is ‘better’.  Then back in the studio I’ve pushed, prodded and printed some images to see how they compare.

Captured with the GFX 50S

Differences in the field

The GFX and D850 are actually quite different cameras which intersect right where it matters for landscape photography.  Both sit at around 50 megapixels, both claim similar bag real estate, and both have the controls and functions needed for making high end photographs.  The GFX costs about twice as much and is of course medium format, with a sensor size around 70% larger than the Nikon’s full frame.

Left: Nikon D850   Right: Fujifilm GFX-50S

When it comes to handling, you could say they’re similar but different, and either can be adapted to quickly enough.  The GFX is a little larger and heavier most of the time depending on the lens being used.  I set out with three GFX lens’s and four DSLR primes, and both systems came to about the same weight.

In use, I found the most noticeable difference was simply between DSLR and mirrorless designs.  Alternating between an optical and electronic viewfinder shows the world in different ways, both are perfectly usable but have different advantages.  The GFX’s electronic viewfinder is pretty large which can help when arranging compositions, and it’s great being able to display a histogram or shift focus points to the edge of the frame.  But I prefer the D850’s optical viewfinder when it comes to seeing fine details, grasses for example, it’s a minor difference but the EVF on the GFX just seems a little course to me coming from a DSLR mindset.  I also prefer how an optical viewfinder displays the scene more purely, which I find serves as a more reliable memory point upon later reflection.  I’d be happy with either system but what matters most is mastering the one you have.

The GFX like other medium format cameras doesn’t use the 3:2 image ratio, instead it’s a 4:3 image which to my way of thinking is simply more pleasing and efficient.  I more often feel constrained by too narrow an image than too wide an image and I’d like to see some ‘full frame’ cameras follow this path, but I won’t hold my breath.

Print tests and early thoughts

To me it’s the print that really brings an image to life, so I made several full size prints.  Prints give an opportunity to see the image as a whole, revealing its character, flaws and  intricacies.  They also provide a chance for others to see and respond in their own way.

When I presented the prints to a group of photographers and several others without explaining why, their responses gave some perspective.  It turns out the differences I’d already found were largely buried when viewed in context of the whole, it’s an effect that can’t be duplicated on a screen because even the highest resolution monitors can only show a portion of the full sized image at a time.

The overall impression was noticed first, perhaps the light had changed between captures or the blacks were a little deeper, perhaps one image was a little cooler or warmer than the other.  Then some areas of interest were pointed out, areas which attracted attention.  It took time for viewers to scan the remainder of the image where flaws and differences could be noticed, and they weren’t jumping out.

There’s a lot competing for our attention in an image and we don’t want nitty gritty issues pulling our attention away from the scene.  These prints showed that the differences between the GFX and D850 won’t always be noticeable, but I have a hunch the next generation of medium format bodies will stretch this gap.

Print comparison between the Fujifilm GFX-50s +23mm and Nikon D850 +20mm
Print comparison between the Fujifilm GFX-50s +120mm and Nikon D850 +85mm

 

Noise

Let’s start with noise, given the GFX’s larger sensor could be expected to produce cleaner images. The D850 uses a minimum ISO of 64 compared to the GFX’s 100, and since I’d typically shoot landscapes at the lowest ISO, that’s how the comparison was made.  The GFX does a little better pulling detail from deep shadows and I’ll always welcome cleaner shadows, but in well exposed areas I can’t see a meaningful difference, both are decent.

Original scene, four stops underexposed
Nikon D850, four stops under exposed, wth exposure corrected (ISO 64)
Fujifilm GFX-50s, four stops under exposed, wth exposure corrected (ISO 100) (actually it was 3.7 stops, I stuffed up)

A note on processing

Adobe automatically applies lens corrections to the GFX raw files, which is not the case for Nikon DSLR files (but is becoming the norm for mirrorless).  Adobe applies these corrections by default to most raw files that contain correction data, but it’s unfortunate this is a forced correction.  To get around this you’ll need to use non-Adobe software to view the GFX files without lens corrections applied.

I was first a little stunned by the GFX’s images in Lightroom, and I wondered how two 50 megapixel images could show such a difference in micro detail.  The D850 appeared a little soft, while the GFX went the other way and almost seemed too crispy.  So I processed some files with Iridient developer for a second opinion and the difference was narrowed, but still in favour of the GFX.

ISO

An ISO comparison would have been nice to add, but since I don’t typically use high ISO’s in the landscape they won’t be compared in this review.

Lens comparisons

These lens comparisons may present an incomplete and shallow view of the world but were the best I could do at the time, they’re mostly concerned with detail & image quality, without too much thought given to the aesthetics of blurs and backgrounds or anything else.

Three GFX lenses were compared with the D850

GFX

D850

23mm F4 R LM WR (19mm equivalent) 20mm Nikkor F1.8G
32-64mm @ 32mm F4 R LM WR  (26mm equivalent) 24mm Nikkor F1.8G
32-64mm @ 64mm F4 R LM WR  (53mm equivalent) 50mm Sigma Art
120mm F4 R LM OIS WR Macro (98mm equivalent) 85mm Sigma Art

Equivalents are taken across the frame width, which makes the GFX image a little taller at the equivalent focal length.

Wide prime

23mm F4 R LM WR (19mm equivalent) vs 20mm Nikkor F1.8G

Of the four lens comparisons, the differences here at the wide end were most noticeable.

The Fujifilm 23mm proved a great performer on the GFX, even wide open at F4 it shows strong detail into the corners, while the Nikkor 20mm shows a little softness by comparison and has a tendency towards purple fringing.  By F8 the Nikon is resolving almost as much detail but is still suffering the purple fringing at high contrast edges.

I’m not sure there’s a DSLR wide angle alternative that can match the GFX with the 23mm lens, it makes for a great landscape combination if you don’t mind the additional weight and bulk, and it accepts filters too, unlike some wide lenses.

Left: D850 + 20mm    Right: Fujifilm GFX-50S + 23mm.   Wide angle comparison

When downsized the GFX and D850 images look very similar, which shouldn’t be surprising given they were both processed using Adobe standard settings.  But let’s look closer….

Left: Nikon D850 + 20mm     Right: Fujifilm GFX + 23mm   100% crops at corners

This is an extreme close up of the corners.  The Nikkor 20mm can produce purple fringing towards the edges, and while it’s normally correctable with processing, it can be severe enough to leave strange ‘scars’ if not corrected well.  In this image, Adobes fringing correction turned the purple blotches into grey blotches, which made them less noticeable but not gone.

However, the reality is nowhere near what you’re probably imagining, and when I displayed these images as one meter prints the uncorrected fringing was barely noticed.  On most scenes the fringing isn’t this extreme.  See the end of the article for another example image.

Overall, it’s the 23mm on the GFX which produces the better images at this focal length, the Nikon 20mm’s fringing was a little disappointing, even if it barely noticed by my test group.

Wide Zoom

32-64mm F4 R LM WR @ 32mm (26mm equivalent) vs 24mm Nikkor F1.8G

It seemed a little unfair to compare the GFX’s 32-64mm zoom to a prime on the D850, but my challenge was to see if the D850 could match medium format, so the Nikkor 24mm F1.8G was a reasonable candidate, and I own it so it will have to do :).

Overall the 32-64mm was well behaved and was enjoyable to shoot with, it may be a ‘general purpose’ lens, but from what I can see it’s at least equal to the best full frame primes.  At F4 the GFX was marginally sharper towards the corners than Nikon’s 24mm, at F8 both hold detail well and are a near match.  Both are subject to a little purple fringing towards the edges of high contrast scenes, the 24mm perhaps more so.  The 32-64mm may be just in front for this round, but it’s surprisingly close.

Lake mountain bend

 

Left: D850 + 24mm, Right: GFX + 32-64mm @ 32mm,  F8 center of scene.

At scene center, there’s not much difference between these two, the GFX has a little more definition,  partly from the longer focal length.

Left: D850 + 24mm, Right: GFX + 32-64mm @ 32mm,  F4 edge of scene

At the edges, both are prone to a little chromatic aberration, and again there’s not a lot of difference between the two, the GFX may be a touch in front.

The gap is narrowed here, and I don’t think you’ll see a meaningful difference between these two cameras/lenses most of the time.

Mid zoom

32-64mm F4 R LM WR @ 64mm (53mm equivalent) vs 50mm Sigma Art

Here again it’s a close call between the GFX’s 32-64mm and the D850 with Sigma’s 50mm Art lens attached, the Sigma is one of the sharpest around.  At F4 the GFX does a little better towards the edges, but by F5.6 and F8 it’s almost too close to call.  I’m sure a test chart could nail the differences, but I’m not seeing much between them in the real world.  Even in terms of overall camera weight there’s little difference here.  The GFX with the zoom lens gives more versatility, while the Nikon with 50mm Art provides a wider equivalent aperture.  For most landscape shooting the GFX and zoom would be my preference for its versatility, and no loss in quality.

Balliang Scene at 50mm.  I’ll spare the fine details – both cameras performed similarly

Short telephoto

120mm F4 R LM OIS WR Macro (98mm equivalent) vs 85mm Sigma Art

The GFX & D850 at 120mm & 85mm are both well performing combinations, sharp and free of distortion.  The Nikon+Sigma is a little heavier, and the GFX +120mm a little bulkier.  I’ve been happy with the Sigma 85mm Art in the past and have used it successfully for art reproductions which can be quite demanding, it’s strong to the edges and produces very pleasing bokeh when used for portraits too.

Unfortunately I didn’t capture the same framing with each lens so I’m using a little judgement here.  At F4 the 120mm is already sharp and the 85mm still very good, at F5.6 and F8 both are sharp across the frame.

Left: D850 + 85mm Sigma Art   Right: GFX + 120mm.  F8 center of frame.  The D850 has a slight disadvantage because of the shorter focal length.

The GFX seems ahead, but again I’m surprised there isn’t a greater difference.  I see more difference stepping down a sensor size than stepping up to medium format.  Perhaps more thorough testing would reveal more, I don’t know.  I’m certainly happy with the results from either of these, and would lean towards the GFX for more technical photography where cleaner data may become noticeable.  Aside from sharpness, I found a more pleasing background blur in some scenes with the 85mm, but that could have been specific to those scenes.

Lens Conclusions

Not only is the D850 competing against the larger sensor of the GFX but also it’s excellent glass.  The differences between the two are magnified by the quality of the glass, Nikon zooms would fall noticeably short against the GFX, while the GFX’s 32-64mm zoom held up against some decent full frame primes.  With both cameras at around 50MP, the GFX does provide more detailed images and finer micro contrast, but it’s no clean sweep, in real world shooting the differences can be hard to distinguish at times.  A well processed and captured D850 image may well turn out better than a sloppily processed GFX image, so this isn’t only about the camera’s and lenses, technique comes into play throughout the whole process.

Autofocus

The GFX autofocus received a bad rap when it was first released with inconsistent focusing (focus shift), but I found it reliable and accurate.  I believe a firmware fix was released some time ago to change the focus strategy.  Even under a rainforest canopy with polariser attached I found no focus issues.  That’s not to say the don’t exist, but they didn’t for me in my short time with it.

Electronic shutter

Beware of the electronic shutter!!  The GFX shutter can operate mechanically, electronically, or as a mixed electronic front and mechanical rear curtain.  Less mechanical vibrations equals sharper images.  The D850 can also use the mixed shutter option, which is my default setting.

However the fully electronic shutter on the GFX can cause some strange effects!  So don’t leave it on ‘just in case’, it’s only useful for static subjects.  Because the electronic shutter takes longer to move across the frame, a whole quarter second from bottom to top according to Jim Kasson’s blog, it can cause some weird sloping effects on anything moving through the scene, it’s the same effect as in the famous 1913 photo by Jacques Henri Lartigue, Le Grand Prix A.C.F.

Here’s an example:

People don’t walk like this.   The sloping shutter effect

 

The other GFX

There’s another GFX I haven’t mentioned yet, the GFX-50R.  The GFX-50S was originally intended as a studio camera, so when it received a huge interest from DSLR upgraders, Fujifilm answered with the 50R, the same sensor and same image quality packaged into a smaller more ruggedised body.

Personally I’m not sure this was the right answer, it’s lighter and arguably smaller, depending on which way you look at it, but it lacks a deep grip and doesn’t feel too steady in the hand.  It also uses a smaller viewfinder and has no option to mount the tilting EVF, instead it has a tilting rear LCD.  The 50R may appeal to many, but in many ways the 50S seems a better platform for landscape photography.  Hopefully the next ’50s’ could retain the best of both these models and shrink a little, just have a look at the Hasselblad X1D-50C.

Left: Fujifilm GFX-50R       Right GFX-50S

Hasselblad X1D-50C

It would seem incomplete to leave out the Hasselblad X1D-50C, the GFX’s main competitor.  By comparison it’s a work of art in concept and design, keeping clean, simple and small.  I’ve only had a brief opportunity to play with it so I won’t go jumping to conclusions.  Here it is next to my D850.

Left: Nikon D850       Right: Hasselblad X1D-50C

Final conclusions

It’s a little unfair to compare the D850 and GFX directly, they’re similar in pixel count but made for different purposes.  The D850 is well refined and made for anything.  The GFX-50S comes with medium format ‘prestige’, it has the studio shooter in mind and wasn’t intended to be thrown in bags and carted around the world, and it’s certainly no action camera.  For the landscape photographer it’s image quality that brings these two together, but weight, robustness and function may be just as important.

Neither of these fit the description of a perfect landscape camera on all counts, if weight is the overriding priority the latest round of full frame mirrorless cameras may be a better fit, the best of which, perhaps the Z7 or A7RIII should be capable of similar image quality to the D850 if paired with the right glass.  And it’s only a matter of time before Canon produces a worthwhile body too, they’re already producing some excellent glass, and let’s not forget what’s likely from Panasonic and Sigma.  Camera tech is changing rapidly but it’s the ‘right glass’ at the ‘right weight’ with the ‘right quality’ and the ‘right functions’ that need to come together for the perfect landscape combination (in my mind).

When it comes to image quality the D850 can come close to and at times can match the GFX depending on the lenses used.  The GFX does extract greater detail than the D850 given the chance, but whether it’s enough to make a difference in all but the most technical or well controlled applications, I’m not yet convinced.  I’m just wishing the GFX’s ‘more’ was a little bit ‘MORE’.  I’d say the next generation will see this gap widen as the medium format sensors improve and increase in resolution, while full frame starts to push up against a wall.

Every system has its compromises and I’m more than happy to photograph with the realisation that excellent gear is excellent gear.  The differences between the GFX and the D850 can seem insignificant within the bigger picture.  We’re beings of emotion, and as my large prints showed we’re easily lead, we see what takes our interest and imperfections only need to remain un-noticed within that context.

In the future, perhaps in the next few years, ‘excellent gear’ will come in smaller, lighter and more convenient packages, ultimate pixel quality will be doubled, and doubled again, and the real challenges in photography will remain beyond our gear.  I’ll always welcome ‘better’ and I’m unlikely to shoot landscapes with ‘poor gear’, but for now as much as I like the GFX with it’s 3:4 aspect ratio, better data, and a zoom option that matches the quality of my primes, it’s still a little bulky and there’s just not enough difference between these two to get me excited most of the time, especially given the price difference and what we’re likely to see within the next few years.

More reference images

Wide prime

23mm F4 R LM WR (19mm equivalent) vs 20mm Nikkor F1.8G

Here’s another wide angle image…

Nissan Hut scene as captured with the GFX
Center of scene  F8.   Top:  D850 + 20mm     Bottom: GFX + 23mm.

Both look pretty close in the center, how about the corners…

Top:  D850 + 20mm     Bottom: GFX + 23mm.   Corner of scene, F8

Again, in the corners the GFX is resolving more cleanly than Nikon, and the Nikon is showing some purple fringing.  It’s the 23mm on the GFX which produces the better images at this focal length.