My one observation is that the systems with the largest sweet spots and most stable image have very wide and even dispersion/power response. Perceived imaging is enhanced by diffusion between the speakers and even to the sides. It's not L/R crosstalk that I've heard mess up imaging so much as poorly managed in room reflections and gnarly off-axis response.
Also, digital has nearly perfect L/R isolation vs. analog systems and yet they don’t really image much better IMHO. Some studies have suggested ~65 dB separation to be more than adequate.
I guess my point is, while interaural crosstalk is a real thing it’s not really the problem we might think it is. The bigger issue is probably frequency cancellation in the center, making center channel speakers really useful for movies and 3 channel recordings as well as mono.