The simplest option that comes to mind is to create two actors, one for the visual and one for the replication.