Autofocus: Soccer with the Sony a9 and Nikon D5

Sony FE 70-200mm F2.8 GM @ 133mm | ISO 25600 | 1/1000 sec | F2.8
Image processed to taste in Adobe Camera Raw
Photo by Carey Rose

Shooting this soccer match, with the field's dim and aging lighting system, proved a much bigger challenge for the a9 - but it still performed remarkably well. We're still using Lock-On AF: Expand Flexible Spot most of the time, and when we were shooting with the D5, we used 3D Tracking. Here's how the two break down.

Rather obviously, the a9 at 20 fps will net you more images. This makes a big difference if you start shooting just a fraction of a second before a player heads the ball, for example - with the a9, you're far more likely to get just the right moment you're looking for than with 12 fps on the D5, and when you couple this with the incredible AF coverage of the a9, you're likely to have some more interesting compositional options, or at least have to crop less later on.

With 20 fps, there's no doubt that it's easier to get just the right moment in crucial situations.
Processed to taste in Adobe Camera Raw
Sony FE 70-200mm F2.8 GM @ 200mm | ISO 25600 | 1/1000 sec | F2.8

On the other hand, the a9 images during bursts would fall in and out of crisp focus (as we will see with some frisbee and cycling bursts coming up in the review), whereas the D5's seemed to be more consistently spot-on in this regard. The D5 was also less prone to jump to a different subject after an obstacle blocked the initial subject (though it's not perfect, as you can see in sample bursts provided below). We should note that the 'blocked shot AF' response on the D5 and the 'AF Tracking Sensitivity' on the a9 were both set to the default setting of '3,' and the D5's Subject Motion parameter was set to 'erratic,' while the a9 has no provision for adjusting this.

Lastly, while the D5's JPEGs are less sophisticated in terms of noise reduction and sharpening than the a9, the color still looks more pleasing out of camera, and the auto white balance was far more effective than the a9, which I ended up having to manually white balance.

Setting white balance manually without a gray card isn't ideal (I used a white line on the ground), but it's an improvement over the auto white balance performance.
Sony FE 70-200mm F2.8 GM @ 200mm | ISO 25600 | 1/1000 sec | F2.8

Perhaps the biggest difference between the two, though, was that the D5 still just feels more immediate. I found that I still prefer the D5's optical viewfinder, though shutter blackout does put it at an objective disadvantage compared to the a9. Also, AF acquisition speed on the D5 is just absolutely instantaneous, whereas the a9 still appears to have a bit of a hiccup before it begins tracking with Lock-On. It was fast enough for these subjects, but the D5 is still just that smidgen faster.

Sample bursts for download are below. Nikon D5 images are all out-of-camera JPEGs, re-saved at quality '8' for size considerations. The first Sony burst was processed from Raw with default settings; the second two are out-of-camera JPEGs with slight color adjustments, and all are saved with at quality '8' as well.