MTF for Fujifilm GFX GF Lenses
See my Fujifilm GFX wish list.
I was reading the Fujifilm GFX pdf and had to laugh at the MTF (Modulation Transfer Function) charts, which at best show just a general “shape” of what performance might be like, a sort of vague hint. I just cannot understand the point of even including nearly meaningless graphs. Do it right, or remove them, Fujifilm:
- unspecified aperture
- unspecified spectral weighting
- unspecified line pairs per millimeter
- unspecified distance (infinity, 1:20, macro range, whatever).
- Whether MTF as computed includes software correction.
All these graphs claim to outperform (by far) every lens Zeiss has ever made (Zeiss publishes MTF measured from real lenses, using the K8 MTF tester). Including vastly outperforming every Zeiss Otus, the Coastal 60/4, etc.
The charts shown by Fujifilm are obviously computed (not measured) as well as diffraction-free graphs—fantasy MTF. Why not just give the lenses a number rating on a 1 to 10 scale, and just say “all our lenses are an unbeatable perfect 10!”.
Image circle radius = 27.4mm = sqrt( 43.8^2 + 32.9^2 )/2
I look forward to seeing what real lenses on a real camera actually deliver.
Fujifilm GF 63mm f/2.8 R WR
Assuming a reasonable aperture like f/5.6, there looks to be somewhat-weak corner performance, possibly with a tiny bit of lateral color. Field flatness (field curvature) looks better than any Zeiss Otus, which is quite an accomplishment given the far larger sensor area—meaning I am deeply skeptical of this chart.
Fujifilm GF 120mm f/4 R LM OIS WR
A perfect performance impossible to criticize. But at least it seems that the 120/4 ought to have a perfectly flat field at some distance and some aperture.
In a testament to the power of meaningless MTF charts, here we see the world’s best zoom lens (or prime lens at the tele end). Heck, who needs a prime lens with a zoom that beats every Zeiss DSLR lens ever made, and on a much larger sensor. Fantasy computed diffraction-free MTF is a con game. The only things to be said here is that the long end may be a bit better than the wide end, and that the wide end has a little astigmatism and possibly a little lateral color.