As a physicist, it's not at all clear to me that tighter tolerances would lead to higher efficiency or less noise. I assume it shakes out in the CFD simulations, but I would be curious to know the explanation.

I thought the primary gain in efficiency came from the large blades, with the blade shape the next most important factor. Gaps between the blade and housing feels like a single-digit percent effect.