SP
If it's anything like when they tried something similar on the Six and Ten with two different shots of the presenter (I can't remember the story) there's a slight delay on the picture on the screens with respect to the sound so the lip-sync is out. They use the live shot so you can see the cameras pointlessly moving and then mix to the DVE effect, which doesn't have the delay, as the presenter starts.
I remembered there was quite a discussion about this sometime back, but I've forgotten the technicalities. It seems there's some limitation to having a live capture of the 2 side-by-side, so they have to resort to recording it earlier isn't it?
If it's anything like when they tried something similar on the Six and Ten with two different shots of the presenter (I can't remember the story) there's a slight delay on the picture on the screens with respect to the sound so the lip-sync is out. They use the live shot so you can see the cameras pointlessly moving and then mix to the DVE effect, which doesn't have the delay, as the presenter starts.