We rent chamber time a few times a year, and it’s rarely to check how something ‘sounds’ — a fully anechoic room is actually a bad listening environment, disorienting in a way most people find unpleasant after a few minutes.
What it’s good for is isolating a single variable: does this HRTF convolution actually produce the elevation cue we designed it to, with zero room reflections to contaminate the result. It’s a measurement tool, not a mixing room, and we try to get engineers out of it again within the hour.