Separate or integrated AV files

Building on the recent timing paper, and acknowledging the limitations wrt audio-visual synchrony, should one expect any difference in variance in synchrony between an experiment that plays a video and accompanying audio as separate files compared to a video with the audio component embedded?