Yeah. 6 seconds for you to receive their message, have the video captured, composited, encoded, buffered, uploaded to your service's intake server, processed, redistributed across the backbone to a local sub-server, buffered again, downloaded by them, decoded and displayed?
That's pretty damn fast. Tell them to manage their expectations.
Some freaking people just don't understand exactly how ridiculously, insanely fast 6 seconds is for all of that work.