Another thing you could do is use bits 1-3 of the video data (the VFD only uses one in every 4 pixels from the video controller) and easily hook up another 3 displays that could show different things

Easily is relative, if I tried surface mount soldering I'd probably have a brick, but that would be cool. Imagine one display for gpsapp, one for whatever the next cool hack is, and one for the player visualizations. Seems like overkill to move the lyrics away from the player visualization but you could do that too.