Aperture Series: Fujifilm GF 32-64mm f/4 vs Canon EF 24-70mm f/2.8L II

Disclaimer 1: this article contains mostly results from non-rigorous optical tests, the goal is to compare the sharpness of the mentioned lenses mounted on their respective systems, at equivalent focal lengths.

Disclaimer 2: all the regional crops are examined and compared under 300% magnification to enhance the difference in sharpness. Hence, it is expected if they look a bit soft. In reality we almost never view images beyond 100%, so these crops shouldn’t be used to judge the lenses’ performance under practical situations.

Background

The Canon EF 24-70mm f/2.8L II was introduced back in September 2012. Being a huge step up from the original EF 24-70 in terms of optical performance, it holds its ground well even when pit against the much more recent Sony FE 24-70mm f/2.8 GM (source: the-digital-picture.com). It has been my walk-around lens for the past 5 years and I never travel without it.

The GF 32-64mm f/4 from Fujifilm was among the first 3 lenses initially released with its medium format (I know it only sports a 44×33 sensor but let’s not get caught up in technicalities here) mirrorless system. I bought one right after I received my GFX 50S in January 2017. However, I discovered that the copy I had was quite soft in the center at 44mm (35mm full-frame equivalent) when shooting wide open, softer than the EF 24-70 II on a 5DsR (300% crop from focus point):

Fc-MF-3x-gfx50s-44mm-f4.png

GF 32-64mm (first copy) @ 44mm | 300% @ center frame (point of focus) | f/4

Fc-MF-3x-5dsr-35mm-f2.8.png

EF 24-70mm @ 35mm | 300% @ center frame (point of focus) | f/2.8

Suspecting sample variation, I returned my copy of the GF 32-64. But I know at some point down the road I will need this lens in my arsenal, because the focal range of a standard zoom is just too versatile to pass up.

Six months later, I bought another GF 32-64 for the upcoming trip to Australia. I did a series of test shots and was pleased to find that, not only did this copy produce sharp center wide open at 44mm (which confirmed the sample variation), it also outperformed the EF 24-70 II, from the center all the way to the corner of the frame, at all focal lengths. This was the exact reassurance I needed after buying into the GFX system.

Settings

All photos below were taken with either the GF 32-64mm f/4 lens on a Fujifilm GFX 50S, or the EF 24-70mm f/2.8L II lens on a Canon EOS 5DsR, mounted on a tripod. Accurate focus was manually achieved through magnified live view. The GFX was set to aperture priority mode, with the ISO fixed at 100, and the shutter speed was determined via “Multi” (whole frame) metering mode. The 5DsR was set to manual exposure using the same settings with 7-shot bracketing. Among the bracketed shots, the one with a histogram closest to the one shot on the Fuji was picked for comparison. This wasn’t really necessary for sharpness comparison, so long as I don’t need to make heavy adjustments to the shadows to introduce noise. I just decided to do it this time.

The full-frame equivalent focal length of the GF 32-64 is about 25-50mm. Considering the difficulty to precisely select 25mm on the EF 24-70, I took the test shots at 32/44/64mm on the GF lens and at 24/35/50mm on the EF lens. To minimize the impact from the DOF difference of the medium format and the full-frame sensors, the GFX shots were compared with the 5DsR shots at 1-stop wider apertures. Another obvious difference was the aspect ratio of the two sensors, the GFX being 4:3 while the 5DsR being 3:2. Intuitively speaking, the GFX shots were “taller” and the 5DsR shots were “wider”. This of course didn’t affect the outcome in any way. Also note that the GFX files have the lens profile baked into them – there isn’t really a way to turn it off in Lightroom. To keep things fair, I manually enabled the lens profile correction for the 5DsR files as well.

All crop comparisons will first show the one from GFX + 32-64mm, followed by the one taken with 5DsR + 24-70mm.

Scene 1 @ 32mm (24mm ff.)

scene1-Fc-gfx50s-32mm-f4.jpg

Fujifilm GFX 50S |Fujifilm GF 32-64mm @ 32mm | 1/40s @ f/4

scene1-Fc-5dsr-24mm-f2.8.jpg

Canon EOS 5DsR |Canon EF 24-70mm f/2.8L II @ 24mm | 1/125s @ f/2.8

Let’s first take a look at the 300% crop at the point of focus:

scene1-Fc-C-3x-gfx50s-32mm-f4.png

GF 32-64 @ 32mm | 300% @ center frame (point of focus) | f/4

scene1-Fc-C-3x-5dsr-24mm-f2.8.png

EF 24-70 @ 24mm | 300% @ center frame (point of focus) | f/2.8

The GF 32-64mm has a clear lead in sharpness. Now move on to the bottom left corner (300%):

scene1-Fc-LL-3x-gfx50s-32mm-f4.png

GF 32-64 @ 32mm | 300% @ bottom left | f/4

scene1-Fc-LL-3x-5dsr-24mm-f2.8.png

EF 24-70 @ 24mm | 300% @ bottom left | f/2.8

Fuji still has the lead here. Aside from the degradation of resolution in the extreme corners, chromatic noise in the shadows also contributed to the loss of details for the Canon combo. This difference in sharpness is even more evident in the top right corner (300%):

scene1-Fc-UR-3x-gfx50s-32mm-f4.png

GF 32-64 @ 32mm | 300% @ top right | f/4

scene1-Fc-UR-3x-5dsr-24mm-f2.8.png

EF 24-70 @ 24mm | 300% @ top right | f/2.8

This overwhelming advantage is evident even when the EF 24-70 is stopped down to f/11:

scene1-Fc-UR-3x-5dsr-24mm-f11.png

EF 24-70 @ 24mm | 300% @ top right | f/11

Without stopping down, the GF 32-64 already puts up an convincing performance on the wide end. Now let’s zoom in to 44mm to compare at 35mm full-frame equivalent.

Scene 1 @ 44mm (35mm ff.)

scene1-Fc-gfx50s-44mm-f4.jpg

Fujifilm GFX 50S |Fujifilm GF 32-64mm @ 44mm | 1/45s @ f/4

scene1-Fc-5dsr-35mm-f2.8.jpg

Canon EOS 5DsR |Canon EF 24-70mm f/2.8L II @ 35mm | 1/125s @ f/2.8

Same as before let’s take a look at the 300% crops from center frame (point of focus), bottom left corner and top right corner, respectively:

scene1-Fc-C-3x-gfx50s-44mm-f4.png

GF 32-64 @ 44mm | 300% @ center frame (point of focus) | f/4

scene1-Fc-C-3x-5dsr-35mm-f2.8.png

EF 24-70 @ 35mm | 300% @ center frame (point of focus) | f/2.8

scene1-Fc-LL-3x-gfx50s-44mm-f4.png

GF 32-64 @ 44mm | 300% @ bottom left | f/4

scene1-Fc-LL-3x-5dsr-35mm-f2.8.png

EF 24-70 @ 35mm | 300% @ bottom left | f/2.8

scene1-Fc-UR-3x-gfx50s-44mm-f4.png

GF 32-64 @ 44mm | 300% @ top right | f/4

scene1-Fc-UR-3x-5dsr-35mm-f2.8.png

EF 24-70 @ 35mm | 300% @ top right | f/2.8

At 35mm full-frame equivalent, we can see that the GF 32-64 still maintains a clear advantage over the EF 24-70. How about 50mm?

Scene 1 @ 64mm (50mm ff.)

scene1-Fc-gfx50s-64mm-f4.jpg

Fujifilm GFX 50S |Fujifilm GF 32-64mm @ 64mm | 1/40s @ f/4

scene1-Fc-5dsr-50mm-f2.8.jpg

Canon EOS 5DsR |Canon EF 24-70mm f/2.8L II @ 50mm | 1/125s @ f/2.8

Zoom in to 300% at center (point of focus), bottom left corner and bottom right corner (since there isn’t much in the top right for checking sharpness), respectively:

scene1-Fc-C-3x-gfx50s-64mm-f4.png

GF 32-64 @ 64mm | 300% @ center frame (point of focus) | f/4

scene1-Fc-C-3x-5dsr-50mm-f2.8.png

EF 24-70 @ 50mm | 300% @ center frame (point of focus) | f/2.8

scene1-Fc-LL-3x-gfx50s-64mm-f4.png

GF 32-64 @ 64mm | 300% @ bottom left | f/4

scene1-Fc-LL-3x-5dsr-50mm-f2.8.png

EF 24-70 @ 50mm | 300% @ bottom left | f/2.8

 

scene1-Fc-LR-3x-gfx50s-64mm-f4.png

GF 32-64 @ 64mm | 300% @ bottom right | f/4

scene1-Fc-LR-3x-5dsr-50mm-f2.8.png

EF 24-70 @ 50mm | 300% @ bottom right | f/2.8

Judging from the crops, the GF 32-64 outshines the EF 24-70 yet again at 50mm ff. equivalent.

Scene 2 @ 32mm (24mm ff.)

Next we check out the performance of the two lenses when shooting a plane at close distance, beginning with 24mm ff. equivalent:

scene2-Fc-gfx50s-32mm-f4.jpg

Fujifilm GFX 50S |Fujifilm GF 32-64mm @ 32mm | 1/50s @ f/4

scene2-Fc-5dsr-24mm-f2.8.jpg

Canon EOS 5DsR |Canon EF 24-70mm f/2.8L II @ 24mm | 1/160s @ f/2.8

Center (Focus Point)

scene2-Fc-C-3x-gfx50s-32mm-f4.png

GF 32-64 @ 32mm | 300% @ center frame (point of focus) | f/4

scene2-Fc-C-3x-5dsr-24mm-f2.8.png

EF 24-70 @ 24mm | 300% @ center frame (point of focus) | f/2.8

Bottom Left Corner

scene2-Fc-LL-3x-gfx50s-32mm-f4.png

GF 32-64 @ 32mm | 300% @ bottom left | f/4

scene2-Fc-LL-3x-5dsr-24mm-f2.8.png

EF 24-70 @ 24mm | 300% @ bottom left | f/2.8

Top Right Corner

scene2-Fc-UR-3x-gfx50s-32mm-f4.png

GF 32-64 @ 32mm | 300% @ top right | f/4

scene2-Fc-UR-3x-5dsr-24mm-f2.8.png

EF 24-70 @ 24mm | 300% @ top right | f/2.8

The Canon lens is a hair bit sharper in the bottom left corner, but overall the GF 32-64 wins this round with ease.

Scene 2 @ 44mm (35mm ff.)

scene2-Fc-gfx50s-44mm-f4.jpg

Fujifilm GFX 50S |Fujifilm GF 32-64mm @ 44mm | 1/40s @ f/4

scene2-Fc-5dsr-35mm-f2.8.jpg

Canon EOS 5DsR |Canon EF 24-70mm f/2.8L II @ 35mm | 1/125s @ f/2.8

Center (Focus Point)

scene2-Fc-C-3x-gfx50s-44mm-f4.png

GF 32-64 @ 44mm | 300% @ center frame (point of focus) | f/4

scene2-Fc-C-3x-5dsr-35mm-f2.8.png

EF 24-70 @ 35mm | 300% @ center frame (point of focus) | f/2.8

Bottom Left Corner

scene2-Fc-LL-3x-gfx50s-44mm-f4.png

GF 32-64 @ 44mm | 300% @ bottom left | f/4

scene2-Fc-LL-3x-5dsr-35mm-f2.8.png

EF 24-70 @ 35mm | 300% @ bottom left | f/2.8

The image quality has seen a drastic drop here for the EF 24-70 at 35mm, but do note that this is already in the extreme corner of the frame.

Top Right Corner

scene2-Fc-UR-3x-gfx50s-44mm-f4.png

GF 32-64 @ 44mm | 300% @ top right | f/4

scene2-Fc-UR-3x-5dsr-35mm-f2.8.png

EF 24-70 @ 35mm | 300% @ top right | f/2.8

Comparing the crops, it’s safe to say that the GF outshines the EF at 35mm ff. equivalent.

Scene 2 @ 64mm (50mm ff.)

scene2-Fc-gfx50s-64mm-f4.jpg

Fujifilm GFX 50S |Fujifilm GF 32-64mm @ 64mm | 1/35s @ f/4

scene2-Fc-5dsr-50mm-f2.8.jpg

Canon EOS 5DsR |Canon EF 24-70mm f/2.8L II @ 50mm | 1/100s @ f/2.8

Center (Focus Point)

scene2-Fc-C-3x-gfx50s-64mm-f4.png

GF 32-64 @ 64mm | 300% @ center frame (point of focus) | f/4

scene2-Fc-C-3x-5dsr-50mm-f2.8.png

EF 24-70 @ 50mm | 300% @ center frame (point of focus) | f/2.8

Bottom Left Corner

scene2-Fc-LL-3x-gfx50s-64mm-f4.png

GF 32-64 @ 64mm | 300% @ bottom left | f/4

scene2-Fc-LL-3x-5dsr-50mm-f2.8.png

EF 24-70 @ 50mm | 300% @ bottom left | f/2.8

Top Right Corner

scene2-Fc-UR-3x-gfx50s-64mm-f4.png

GF 32-64 @ 64mm | 300% @ top right | f/4

scene2-Fc-UR-3x-5dsr-50mm-f2.8.png

EF 24-70 @ 50mm | 300% @ top right | f/2.8

So far we can see that the GF 32-64 has an overwhelming advantage over the EF 24-70 (especially in the corners) throughout its 24-50mm full-frame equivalent focal range. For the sake of completeness, let’s take a look at the situation when the subject distance is around 5 ~ 10 meters (16 ~ 32 feet). Since neither of the lenses show strong field curvature, we will be focusing on the details in center frame (focus point), top left corner and top right corner, respectively.

Scene 3 @ 32mm (24mm ff.)

scene3-Fc-gfx50s-32mm-f4.jpg

Fujifilm GFX 50S |Fujifilm GF 32-64mm @ 32mm | 1/125s @ f/4

scene3-Fc-5dsr-24mm-f2.8.jpg

Canon EOS 5DsR |Canon EF 24-70mm f/2.8L II @ 24mm | 1/500s @ f/2.8

Center (Focus Point)

scene3-Fc-C-3x-gfx50s-32mm-f4.png

GF 32-64 @ 32mm | 300% @ center frame (point of focus) | f/4

scene3-Fc-C-3x-5dsr-24mm-f2.8.png

EF 24-70 @ 24mm | 300% @ center frame (point of focus) | f/2.8

Top Left Corner

scene3-Fc-UL-3x-gfx50s-32mm-f4.png

GF 32-64 @ 32mm | 300% @ top left | f/4

scene3-Fc-UL-3x-5dsr-24mm-f2.8.png

EF 24-70 @ 24mm | 300% @ top left | f/2.8

Top Right Corner

scene3-Fc-UR-3x-gfx50s-32mm-f4.png

GF 32-64 @ 32mm | 300% @ top right | f/4

scene3-Fc-UR-3x-5dsr-24mm-f2.8.png

EF 24-70 @ 24mm | 300% @ top right | f/2.8

Scene 3 @ 44mm (35mm ff.)

scene3-Fc-gfx50s-44mm-f4.jpg

Fujifilm GFX 50S |Fujifilm GF 32-64mm @ 44mm | 1/150s @ f/4

scene3-Fc-5dsr-35mm-f2.8.jpg

Canon EOS 5DsR |Canon EF 24-70mm f/2.8L II @ 35mm | 1/500s @ f/2.8

Center (Focus Point)

scene3-Fc-C-3x-gfx50s-44mm-f4.png

GF 32-64 @ 44mm | 300% @ center frame (point of focus) | f/4

scene3-Fc-C-3x-5dsr-35mm-f2.8.png

EF 24-70 @ 35mm | 300% @ center frame (point of focus) | f/2.8

Top Left Corner

scene3-Fc-UL-3x-gfx50s-44mm-f4.png

GF 32-64 @ 44mm | 300% @ top left | f/4

scene3-Fc-UL-3x-5dsr-35mm-f2.8.png

EF 24-70 @ 35mm | 300% @ top left | f/2.8

Top Right Corner

scene3-Fc-UR-3x-gfx50s-44mm-f4.png

GF 32-64 @ 44mm | 300% @ top right | f/4

scene3-Fc-UR-3x-5dsr-35mm-f2.8.png

EF 24-70 @ 35mm | 300% @ top right | f/2.8

Scene 3 @ 64mm (50mm ff.)

scene3-Fc-gfx50s-64mm-f4.jpg

Fujifilm GFX 50S |Fujifilm GF 32-64mm @ 64mm | 1/110s @ f/4

scene3-Fc-5dsr-50mm-f2.8.jpg

Canon EOS 5DsR |Canon EF 24-70mm f/2.8L II @ 50mm | 1/320s @ f/2.8

Center (Focus Point)

scene3-Fc-C-3x-gfx50s-64mm-f4.png

GF 32-64 @ 64mm | 300% @ center frame (point of focus) | f/4

scene3-Fc-C-3x-5dsr-50mm-f2.8.png

EF 24-70 @ 50mm | 300% @ center frame (point of focus) | f/2.8

Top Left Corner

scene3-Fc-UL-3x-gfx50s-64mm-f4.png

GF 32-64 @ 64mm | 300% @ top left | f/4

scene3-Fc-UL-3x-5dsr-50mm-f2.8.png

EF 24-70 @ 50mm | 300% @ top left | f/2.8

Top Right Corner

scene3-Fc-UR-3x-gfx50s-64mm-f4.png

GF 32-64 @ 64mm | 300% @ top right | f/4

scene3-Fc-UR-3x-5dsr-50mm-f2.8.png

EF 24-70 @ 50mm | 300% @ top right | f/2.8

The Canon lens simply could not keep up with the medium format standard zoom in the corners! How about when stopped down?

Take the 50mm top left corner as the point of reference, we found that the EF 24-70 needs to be stopped down to f/11 to rival the sharpness of the GF 32-64 wide open:

scene3-Fc-UL-3x-gfx50s-64mm-f4.png

GF 32-64 @ 64mm | 300% @ top left | f/4

scene3-Fc-UL-3x-5dsr-50mm-f8.png

EF 24-70 @ 50mm | 300% @ top left | f/8

scene3-Fc-UL-3x-5dsr-50mm-f11.png

EF 24-70 @ 50mm | 300% @ top left | f/11

Considering that diffraction starts to kick in beyond f/11 on the 5DsR, this is pretty much the best image quality one can get from the EF lens. Nonetheless, the clarity of the details still comes up short against the GF 32-64 wide open. This is particularly obvious in the center of scene 3:

scene3-Fc-C-3x-gfx50s-64mm-f4.png

GF 32-64 @ 64mm | 300% @ center frame (point of focus) | f/4

scene3-Fc-C-3x-5dsr-50mm-f11.png

EF 24-70 @ 50mm | 300% @ center frame (point of focus) | f/11

Conclusions

Through a series of comparisons at 3 different scenes, we can conclude that the GF 32-64 has an advantage in sharpness over the EF 24-70 at all corresponding focal lengths, when both are wide open. This is especially evident in the extreme corners. Though the Canon lens can close the gap by stopping down to f/11, the overall clarity of details still falls short when examined at 300%.

I want to emphasize again, however, that the EF 24-70 is by no mean an inferior lens. There is little practical value of viewing the images at 300%. It is more of an “extra credit” problem to further differentiate two of the very best lenses. The biggest drawback of the GF lens (imo) is the absence of the 50-70mm full-frame equivalent focal range. I use the telephoto end of the EF 24-70 quite frequently, especially when I am taking only one lens. This might have been a compromise due to possible concerns in cost, optical performance, and/or design complexity, but the shorter zoom does limit the flexibility in composition (sometimes I just wish the field of view could be a little bit narrower). There is not really a standard either, as Pentax, Phase One, Leica S, Hasselblad (HCD and XCD) each has a slightly different take on their respective medium format standard zooms.

Nonetheless, the optical performance of the GF 32-64 far outweighs its downside. The results here have just reassured me that I should use the lens without any hesitation and I will be getting the best image quality possible. So if you own or plan to own a GFX, I highly recommend getting it as your general purpose lens for the system.

Advertisements