It's unclear what difference that makes in the context of my post. The fact remains that the modulation and signaling protocols developed for use in modems were designed within the constraints of electrical transmission across miles of copper wire, not through sounds waves in an unknown acoustical environment. The constraints for the two situations are entirely different. Acoustic couplers are designed to minimize any acoustical impact of their use to avoid the exact acoustical problems you'd have transmitting across a room.
Yes. Also, the upper limit on the dynamic range of the telephone network is somewhere around 3400 Hz, well within the range of human hearing. This protocol is designed to send data in a frequency range that most people can't hear, or at least can't hear very well.