There's zero shake here -- a surprise for me -- and her eyes are tack sharp. Wide open at f2.8.

(Note: The JPEG blocking artifacts you see on the bridge of her nose and cheeks are not in the original image.)

Shot with the 70-200 f2.8 G Master lens at 186 mm. This is the full version of the crop on the next slide.

Well, this crop was supposed to show how well the detail is preserved at midrange ISO sensitivities, but I doubt you can see it. So see the original. Since her hair's in motion it suffers from a little softness and color contamination, but it's not bad.

This and the next slide are 100 percent crops from sequential photos (sort of -- there's one shot between them, but it looks similar enough that I stand by my point). It shows how the focus system has trouble with the helmet blocking his face.

You can see that the autofocus and metering systems jumped from his helmet to his face (this, of course, is the focus I wanted). This is a particularly difficult situation for every camera, and my point is not that the A9 fails, but that there are no miracles here.

Here's the original. Another example of the lack of shake, and an example of what the camera can do with still subjects. It's hard to judge the skin tones, since he's wearing makeup.

Here's the original. It's not a great shot, but it gives you a sense of how much highlight tonal range is captured. Note that I was shooting in sRGB and I can't yet view the raws, so it probably gets better.

This shows how close to the edges of the frame the autofocus system can lock on. That dark spot in the lower right corner isn't vignetting -- it's a human getting in my way.

This is one of the few shots I took with the 24-70 mm lens. You should be able to see the range between the details in the black camera he's holding up and the reflections in the ice.

Another example of tonal range; in the original, you can see the texture on the puck as well as details on the ice shavings.

We really need tracking metering systems to keep up with tracking autofocus; these are two sequential shots -- the underexposed shot came first -- and they've barely moved, but the metering system took about two shots to adjust. You could probably fix the exposure, but not outdoors if the subject is moving through sunlight and shade.