Actually, it would require rephrasing sound representations.
Currently it works like this:
You’ve got sound and direction. If direction is “here”, nothing is added. Otherwise “From direction you hear” is added to the message. Notice how it “sounds” when you smash an object you’re standing on.
You’d either have to deal with sound messages as informative as “A pumping bass.” (that’s the entire message you’d get) or make a copypasted set of functions for differently phrased sound messages.
It isn’t hard, but it isn’t a trivial, obvious fix.
