...and I think that's exactly the technique the "Monkey Island PC-Speaker coders" used to overlay the percussion over the melody. The IBM PC and its descendants had the advantage of having a relatively fast CPU which could partly make up for the primitive sound system - i.e. it could only play a square wave, but it could be a very complex square wave. Some games even played voice samples - they were extremely distorted, but more or less understandable. Similar to how you can show photos on a monochrome display (like the good old Hercules) - they won't look pretty, but you can do it...
