There is some amount of the "Focus follows brain" problem here. What we want is for things to do what we meant, all the time, and in this case it's very possible that the visitor wanted to hear the music. It is not practical (without yet to invented technology) for that to work so we have a substitute - there's a switch and you should remember to press it.

"Focus follows brain" is how everybody wants windowed UIs to work. When I type on the keyboard the letters go where my brain thought they should go - duh, but of course that's unimplementable, so the Windows UI provides "Click to focus" - if I click on a Window the typing goes there until I click another window, meanwhile some Unix systems do "Focus follows Mouse" - if I move the mouse over a Window then my typing goes there even without clicking. Neither is what we actually wanted, both are trying to approximate.