I strongly feel the audio sensor and entity should be implemented as the other ones. I.e., the audio sensor should be a spot that serves as a listener. And an audio aspect should have a radius.
I think this better resembles the real world. How do you guys think?