[Klaus]

I tested both the Sigma 85 Art and Canon 85/1.4L with multiple samples and camera's (5D3/5D4/1DX2) and the Sigma is more accurate at close distance and faster in low light. At outer focuspoints the Canon starts to hunt, but not the Sigma. So no, the Canon isn't better at focussing than the Sigma.

Btw: there is nothing wrong with the IS of your sample: they all aren't as effective as Canon claims. I could only manage to get around 2.5-3 stops.

See the bokeh fringing test in both reviews - you can clearly see that the Sigma's focus point is moving. The Canon's does not. And it's unlikely to be a sample issue. During the field tests, I had also more outliers on the Sigma - and that was after calibrating the Sigma using the Sigma USB dock. Out of the box, the Sigma was a bit back-focusing (at f/1.4).

I would have been happy with 2.5-3 f-stops efficiency of the IS. The 5Ds R is, of course, picky in pretty much everything. Some of the samples that I've taken at 1/20sec may have been reasonably sharp at 20mp but at 50mp they were clearly blurred (albeit less blurred than without IS). Generally, I'm very surprised about the claims by some people how many f-stops they are getting from IS systems. My baseline is sellable quality (pixel-level sharpness) and not just "acceptable". While I can reach the "max claim" in a few shots, there's no consistency whatsoever - not just at Canon but across the board.