Nikon 24mm f/1.4G autofocus hugely erratic
Thanks to all the readers who completed the autofocus accuracy survey on the Nikon AF-S 24mm f/1.4G. So far 8/9 respondents reported focusing problems of various kinds.
After shooting some controlled tests today, I can say unequivocally that autofocus with the Nikon 24mm f/1.4G is the worst I’ve ever encountered with any lens of any focal length of any brand. As I looked over both my controlled trials as well as ad-hoc shots, the phrase came to mind: “you ain’t gonna believe this shit”. And not in the positive sense that phrase can have. But you will believe, once I post my report (soon).
On the plus side, I’m more impressed than ever with the imaging quality at f/1.4, which if anything I’ve under-rated. Problem is, that quality appears only with manual focus (Live View), and in all my ad-hoc test shots, nearly all far focus shots were misfocused.
Manual focus (left), autofocus (right) on Nikon D3s
There are autofocus problems on both the D3x and the D3s, and I observed egregious misfocus at a subject distance of about 20 feet with the D3s, shown above— a high contrast target in sunlight, no easier target than that. I observed problems with distance focus as well So it’s not just a distance issue.
Sometimes autofocus is OK, because even a blind squirrel finds an acorn once in a while.
Fine focus adjust? Forget it. Neither +20 nor -20 had any noticeable effect, maybe ±500 might have some effect, but ±20 is the maximum that the D3s/D3x offers.
I’ll be documenting my findings soon in my review of the Nikon AF-S 24mm f/1.4G ED in DAP. I had planned for some careful testing, but the problems are so severe that the 24/1.4G should be avoided until Nikon fixes the issue. The tried and true manual focus Zeiss ZF.2 21mm f/2.8 Distagon comes to mind, always having satisfied.
Bottom line: even the best optical design is worthless if it cannot be focused reliably. Professionals are taking unacceptable risk if they trust a job to this new Nikon 24mm— don’t even think about using it for a job unless you are certain that your sample focuses accurately and consistently on your camera body or bodies.
The example below is less bad, but still way off, focusing on the grass in the foreground instead of the Golden Gate Bridge, turning a 12 megapixel Nikon D3s into a 2 megapixel camera. The effect is obvious— observe the 4-5 pixel blur circles on highlights. Other than that, I love the “look” of the image.
