I second the idea of training the player. Something to provide audio context, even if there ends up being no visual component.
