Image quality versus full-frame

The Fujifilm GFX 50R uses the same medium format chip as its 50S predecessor, pictured above.

The greater size of the GFX 50R's sensor relative to full-frame competitors means any given pixel count is less demanding on the lens (which has to resolve fewer pixels per mm). But the main advantage of that extra silicon is the light it can capture: that's the primary source of medium format's image quality advantage.

Talk of better gradation, tonality or 'magic' of medium format are essentially just ways of expressing that larger sensors can capture more light, which then gives every tone in the image a better signal-to-noise ratio (assuming the sensor technology isn't significantly outdated).

Key takeaways:

  • GFX 50R offers greater IQ potential thanks to larger sensor and less resolution demand on lenses
  • The best full-frame sensors can offer very similar image quality, thanks to a similar number of pixels and lower base ISO allowing comparable levels of light capture
  • Transitional areas from in-focus to out-of-focus are going to come down largely to optics, and we plan a shootout in the future to investigate this further

Despite the size difference, the full-frame sensors in the Nikon D850 and Z7 offer very similar resolution to the Fujifilm and have the ability to shoot at ISO 64, which allows them to capture 2/3EV (56%) more light per unit area than the 50R. Lenses will make a difference, of course, but, in theory, the Nikons should be very close in terms of underlying image quality. So we shot them both side-by-side.


Side-by-side comparison

Here we've shot the Fujifilm GFX 50R directly alongside the Nikon Z7. We've used the 45mm F2.8 lens on the Fujifilm and the Nikkor S 35mm F1.8 Z on the Nikon to give the same angle of view.

We've set the Fujifilm to F10 and the Nikon to F8, meaning both are shooting with a roughly 4.5mm aperture diameter, giving the same depth-of-field. This 2/3 of a stop difference in F-number lets us use the same exposure time for both, since it's equal to the difference in ISO ratings.

When compared at their native pixel counts, the two images look similarly noisy. Scale the Z7's shot up and correct for the rotation and the Nikon Z7 looks a tiny fraction noisier, and a touch softer, despite more sharpening being applied.

But, despite the Nikon being put at a disadvantage by the additional processing, there's still very little to differentiate between the two cameras. Thanks to its lens (and its microlens design), the Fujifilm appears a fraction sharper overall, but the differences are difficult to recognize unless compared side-by-side at pixel-level.

Given that these images cover dynamic range that extends across at least 12EV, this should give an idea of the difference the way the cameras handle tone and gradation across most of the range that their sensors are capable of.

Overall, this comparison isn't supposed to show which format is better (Fujifilm has already announced the development of a 100MP GFX camera with a BSI sensor, which it'll be hard for the smaller format to match), but it does give an idea of how the GFX 50R compares with the best of its current peers.


Studio portrait

Back when we reviewed Fujifilm's GFX 50S, we set up a controlled comparison between that camera (which shares the exact same internals as the GFX 50R) and the Nikon D810 and EOS 5DS R. And because the GFX 50R will perform identically in such a test, we decided it was prudent to show those results here as well. Any potential advantage of the newer backside-illuminated sensor in the Nikon D850 will, if anything, further close any perceived image quality gap seen here.

The images were shot with the same lighting setup and focused as precisely as was possible. Exposure on the Fujifilm was set to ISO 100 and F8, the same as the Canon EOS 5DS R. The Nikon D810 was set down to ISO 64 and F6.3, to match total light capture and depth-of-field (the higher exposure giving more light per unit area in proportion to how much smaller its sensor is, which it can tolerate thanks to a higher pixel well capacity).

The idea here is to allow an assessment of both detail and tonality. Raw files are available to download for your own inspection in the Raw processor of your choice.

We can see a slightly noisier image from the Canon, evident in the backdrop resulting from a lower total signal:noise ratio thanks to its smaller sensor (and that's before you consider the camera's lower dynamic range due its high read noise - it's an older sensor at this point). Although subtle in some instances, this reduces the file's flexibility in post. Notably, the Nikon D810 does not suffer from this problem, because its signal:noise ratio is essentially the same as the Fujifilm thanks to its capturing the same total light at ISO 64 as the GFX at ISO 100. This means the D810 can offer base ISO dynamic range similar to the GFX, and so its files are likely to be just as malleable in post-processing, given you have the extra light ISO 64 requires.

Although we've seen much discussion on the virtues of the GFX system's larger sensor format in terms of the transition from in-focus to out-of-focus areas, that will be due to the optical characteristics of a respective system's lenses, not sensor quality. We plan on doing another shootout in the future between Fujifilm's GF 110mm F2 lens and other systems' high-end 85mm primes to examine this in further detail.