This is one in a series of posts on the Fujifilm GFX 100S. You should be able to find all the posts about that camera in the Category List on the right sidebar, below the Articles widget. There’s a drop-down menu there that you can use to get to all the posts in this series; just look for “GFX 100S”.
A few days ago, I wrote a post comparing the Fuji 250/4 against the Rodenstock 180/5.6 HR Digaron on a GFX 100S. After that post, several people suggested that I might have a bad copy of the 250/4 GF. Someone on DPR said that he thought his 250/4 was about as sharp in the image field (my characterization) as his other GF lenses. He posted some Siemens star images, which I found intriguing, but inconclusive. Yesterday he sent me the results of some slanted edge testing that he had performed on his GF lenses (he used MTF Mapper for the analysis), which seemed to bear out his assertion about the relative sharpness of the 250/4. I asked for the raw file for the 250/4 shot, and he obliged.
There were some possibly material differences from my test conditions:
- I used a low contrast target; he used a high contrast one.
- I used a GFX 100S; he used a GFX 50R.
- I used natural diffused sky light; I don’t know what he used.
- I used a matte target; I don’t know what kind of substrate he used.
- He used 15 meters for the target distance; I used about 25 meters.
If he used an F-surface target, those last two might cancel each other out.
Here’s an Imatest STF analysis of his test shot:
And here’s a similar analysis of my shot:
There is quite a bit of overshoot on his edge profile. That shouldn’t happen with sharpening turned off. Lightroom has performed some behind-the-scenes sharpening for some cameras in the past; is it doing that for the GFX 50R?
A big difference is the pixel pitch. We can correct for that mathematically by converting to cycles per picture height.
There is a 12% difference in MTf50 and a 5% difference in MTF 30. The microlenses on the two cameras are about the same size, so the loss in sharpness due to pixel aperture should be about the same.
As a way to track down the overshoot, I extracted the G1 raw channel from each raw file, and analyzed that.
With the GFX 50R:
The overshoot is gone! Shame on you, Lightroom.
And with the GFX 100S:
Because 1/4 of the pixels are used, the Nyquist frequency and cycles/pixel are twice what they would be for the full image. I will correct for that mathematically.
Now the MTF50 values in cy/ph are essentially on top of each other, but the MTF 30 values are a bit further apart. I did notice when I was comparing the microlens sizes of the GFX 50S and GFX 100 that The GFX 50S did tend to be a bit sharper at high spatial frequencies.
I think that, based on this test, there is little difference between our two copies of the Fuji 250 mm f/4 GF lens.