Footsteps worked pretty well when I tried that on the other hand. I wonder if lot of it has to do with how well the model understands what the english description of the sound should sound like...

i do think that’s the case. i tried a few different ways to write x and got meaningfully varied results