On the relative bias of void tracers in the Dark Energy Survey