Comment by z3t4

Humans are really good at vocals, and it would thus be very hard to replace, eg. making a speech synthesis model sing well. Melodies however are more mathematical and could possibly be generated by a computer.