Currently video conference software is designed on the assumption that all participants must see everything in the same sequence, synchronized to within milliseconds (or at least tens of milliseconds). Being able to relax these constraints by two orders of magnitude, to allow items out of sync by two seconds, would make the systems a lot easier to design.
This may also make it easier to design software which allows a smooth transition between synchronous and asynchronous modes (Worthington, 2013).
Reference:Fairhall, S. L., Albi, A., & Melcher, D. (2014). Temporal Integration Windows for Naturalistic Visual Sequences. PloS one, 9(7), e102248. http://journals.plos.org/plosone/article?id=10.1371/journal.pone.0102248#pone-0102248-g003
Worthington, T. (2013). Synchronizing Asynchronous Learning: Combining Synchronous and Asynchronous Techniques. In Proceedings of 2013 8th International Conference on Computer Science & Education (ICCSE), 26 Apr - 28 Apr 2013 , Sri Lanka. URL: http://dx.doi.org/10.1109/ICCSE.2013.6553983