You're both missing my point, which is that the original poster's "not-Ai music system" would have still needed Ai at its core in order to replicate human voice

You're both missing my point, which is that the original poster's "not-Ai music system" would have still needed Ai at its core in order to replicate human voices with tiny instrument samples. How else would it know what the voices are supposed to sound like?