Lip sync becomes noticeable to a sharp observer when latency exceeds 1 frame (33ms). You can get away with a couple of frames for most distribution schemes and only a few people will care. In a large room with IMAG you will have different issues. Since the person is both live onstage and magnified on screen, any quick motion will be clearly out of time on the video side, but since most of the audience will be experiencing audio delay due to time of flight the audio will be partially compensated. You may find you need to have a dedicated mix to video record or broadcast that you can delay to correct lip sync since video systems with lots of processing can easily have 4 or 5 frames of delay (132-165ms) which will be very noticeable. On the PA side you will have to pick the part of the audience that you want to have good lip sync, since you will only be able to keep about 60' of depth in good sync.
Mac