If you’ve been following the online forums about the D800/D800E you may have come across a few message threads about a “Left AF point” issue. Or maybe a few hundred threads about it. In today’s blog I’m going to discuss the issue, and then present a video demonstration of the sadness on a recently-acquired D800 that is afflicted by the disease.
First, is the problem real? Yes. I’ve had two D800 bodies so far, the first I acquired soon after they became available in the states and a newer one I bought within the past few weeks. The first body had perfect AF – its left and right AF points were highly accurate and consistent across all my lenses, including the ones most sensitive to the Left AF issue, the 24G f/1.4 and 14-24G f/2.8. The second body? Not so much. The good news is the second body has highly consistent autofocus, just like the first. The bad news is it’s consistently bad, at least when focusing using its leftmost AF point. After finetuning this second body with my 24G (-5 AF tune for you geeks), the center and right AF points produce perfectly focused images at f/1.4, over a wide range of focusing distances. The left AF point? It’s a little off starting at MFD (minimum focusing distance), then gets progressively worse as the focusing distance increases. At around 4 feet and beyond the focusing error becomes unusable wide-open on my 24G.
Those few to a few hundred online threads about this issue have spanned the spectrum from “you’re imagining things” to “you’re testing it wrong” to “mine sucks too, let’s go have a beer and commiserate”. To my knowledge the first person to discover and disclose this issue was Ming Thein, a photographer who maintains a blog Here. He even went so far as to visit a Nikon service center to demonstrate the problem in front of Nikon technicians, where they cycled through several samples of the 24G and D800 bodies, not to mention other full-frame Nikon bodies like the D3. More recently Thom Hogan, a respected Nikon shooter, blogger, and author of must-have guides about Nikon bodies wrote about his experience with the issue, disclosing that he’s personally observed it on 2 bodies, out of a sample size of 12 bodies. Well I guess it’s my turn to chime in, late as always 😉
What follows is my video demonstration of the issue, as recorded from a live session with my second D800 body. I set up three siemens star charts perfectly parallel to the camera’s focusing plane, one centered on the rightmost AF point, one on the center AF point, and finally, one on the boogeyman leftmost AF point. I mounted my D800 on my sturdy-as-all-getout Benro J-3 ballhead and C-358m8 legs. The charts were illuminated with a 5000K LED lamp to produce an exposure of f/1.4 ISO 100 1/50. Not the most light possible but certainly well above the darkness threshold where the D800 starts to hunt like a blind pig. To isolate clarity issues to the AF system alone, I configured the body in 2-second exposure delay mode, which combines the mirror vibration-avoiding MLU mode with the human vibration-avoiding timer-release. This level of precision of setup isn’t really necessary; you can simply point your camera at something of interest and see if it takes a sharp photo. Expecting any less from a $3,000 body seems rather silly. But in the interest of repeatability, comparability, and most of all, flak invulnerability, I went through the trouble of setting up the most precise test I could think of.
The methodology is simple. For each of the three AF points I first enter Live View and focus on the siemens star using the ultra-precise Contrast-Detect autofocus and then snap a photo, which becomes my reference “sharp” photo. I then exit Live View and take a sequence of three photos using the camera’s Phase-Detect autofocus, racking focus in between each photo so that the camera has to work to acquire focus. After each photo I zoom to 100% in playback and then compare the photo to the reference shot focused in Live View. For my test the charts were mounted 4 1/2 feet away from the sensor plane indicator on the body. As mentioned previously, at closer focusing distances the problems are less severe, at further distances forget about it. Watch the video below for the results.
So what’s Nikon’s position on this? Nobody knows. Of the D800/D800E owners who have sent their bodies in, the early adopters reported frustration and lack of success, whereas the most recent service attempts seem to have been much more successful. I’d like to send my body in as well, except I haven’t been able to convince Nikon USA to send me a prepaid printing label that will spare me the $50+ insured shipping expense. That’s Nikon USA’s standard warranty policy/procedure. And unless and until this issue is recognized by Nikon as something beyond the normal low-percentage defect rate type of issue I suspect that policy will remain in place. When will Nikon acknowledge the issue publicly? Will they issue a recall, even a silent one?
You can download the Siemens Star chart I used in this video at Chart