I wonder if high fidelity is even needed, as long as the speakers and mic both respond somewhat to ultrasound. Perhaps a high data rate isn’t needed for the kinds of info being transmitted. Say you used wide-bandwidth pulses of ultrasound at high volume, at, say, 1 pulse/second. Perhaps the data would be transferred even though the speaker would struggle to emit it, and the mic would struggle to accept it. With heavy error correction, the data could get through, and it wouldn’t matter much what the frequency response was of the speaker and mic. It’s not like you’re transferring music.
I apologize if I’ve given anyone any ideas!
