I'm struggling with EXACTLY the same thing. My advanced action has one more command (that the CC be turned off and on as well), and I've got the same problem. Would love to hear what your solution was!
Mine was to have each button in the toggle pair to be just a pixel or two different in size, and - because my nav is super-simple and uses Webding characters - make use of text alignment (left & center) so that neither the left nor the top edges of the smart shapes were in the same place, but that the play/pause and mute/unmute icons still lined up.