Video: D850 versus D5 subject tracking

Photographer Matt Granger's YouTube channel is chock-full of videos that run the gamut from fairly technical testing to fun challenges and prize giveaways. But we took particular notice when, last weekend, he pitted his new Nikon D850 against his Nikon D5 in terms of autofocus tracking. (What can we say; we're nerds.)

But why even bother? After all, the D850 and D5 purportedly have the exact same autofocus hardware—performance should be virtually identical. In fact, in our conversations with Nikon, we were told that the D850 includes an additional processor to handle autofocus calculations, just like the D5; they also claimed this processor was omitted from the D500.

More importantly, we were told flat out to expect 'D5 levels of performance' from the D850.

So why is it that, during an early shoot with the D850, Granger said, "I felt like I was getting more [images] well-tracked with the D5 than I did with the D850." Cue the tests in the video below:

Granger performed both low-light and daylight tests. In the former, his subject moved primarily in the X axis (read: across the frame); in the latter, his subject moved in all three dimensions. Both cameras were set to shoot at 7fps (the D850's max), and shared all other settings as well.

In both instances, Granger concludes that the D850 is simply not as good at tracking moving subjects as the D5.

In the first low-light test, he notes some hesitation: the AF point occasionally lagged behind the subject before catching up (though at 3:00, where he says the D5 does not exhibit this behavior, it looks like the D5 also hesitates a bit, but perhaps not as severely).

We've been told to expect 'D5 levels of performance' from the D850's autofocus system.

In the second test, the D850 really struggled to track the subject as she moved between shade and direct sun, seemingly performing better if the subject was initially acquired in bright light.

Of course, Granger concludes that this disparity isn't really a reason to ignore the D850 or cancel that pre-order. In fact, there have been similar claims across the interwebz of autofocus performance disparities between previous Nikon cameras that are supposed to share the same AF system (the D810 and D4S for example—although we haven't dug into those). But it's definitely something worth testing further.

So, given our recent coverage and in-progress full-review, what does the DPReview staff make of all this?

Our experience so far

For white water kayaking, we found the D850's 3D Tracking to perform very well indeed.

We've been shooting a lot with the Nikon D850 over the past couple of weeks, and as with just about any recent high-end Nikon camera, we find subject tracking to work very well. Even though the D850 comes with that excellent AF selection joystick, sometimes we find it preferable to let the camera do some of the work for us.

In particular, when shooting white water kayaking in Oregon, 3D Tracking performed admirably. That shouldn't be too much of a surprise, though, considering these are fairly distinct subjects, particularly in terms of color, when compared with their surroundings.

As with Matt Granger, though, we've been surprised by some of our own experiences while shooting with the D850. With both the D5 and even the D500, we've become accustomed to being able to initiate 3D Tracking on a subject's eye for a tight headshot, and have the camera track it remarkably well as either the camera or subject moves.

This sort of candid kid shot is a situation in which accurate 3D Tracking or Sony's Eye AF both come in very handy.

With the D850 though, we've noticed it is more apt to be 'jumpy,' in that it will jump from our subject's eye to another portion of their face. Further complicating our assessment of this sort of behavior is that it seems to be inconsistent—sometimes the D850 will track perfectly, other times it gets distracted by something else in the scene.

We also noticed some inconsistency in a situation where we've historically been able to rely solely on 3D Tracking, namely: motocross. With the D850, we were able to get a ton of keepers using 3D Tracking, but sometimes, when we looked through our bursts and found that one perfect moment, it would be slightly out of focus because the AF point had jumped from one part of the rider or motorcycle to another.

When we switched to a type of zone focusing called D25, we came away with a better hit rate, so long as we kept the zone over our subject (Note: you would expect this result from just about any camera with advanced autofocus).

3D Tracking on the left nailed focus in this instance, but D25 on the right nailed focus more reliably, more often.

So what's next?

More photos. Lots more photos.

We'll be doing our standard autofocus treatment on the D850, including our bicycle tests and our close-range, low-light autofocus evaluation. We'll also be seeking out sporting events in the Seattle area to see if we can draw some parallels between our usual evaluations and real-world shooting situation performance.

We should also note that, in both Granger's testing and our own experience, we haven't yet been able to use Nikon's new battery grip for the D850. For it to be as even a comparison as possible, we'll be looking to add the grip and the much more powerful EN-EL18a battery if at all possible.

Only more time and more testing will tell if the D850 can truly stand up to the D5's autofocus performance, though we should reinforce that D5-level subject tracking is an awfully lofty benchmark to reach for. In our experience, it's a benchmark that has remained out of reach of any camera to date, and whether or not the D850 measures up will be revealed in our full review.