I voted receive full information but not process in real-time, but the story is more complicated.
I receive massive amounts of information from non-verbal cues that are taken physically and not processed in the social-emotional direction in real-time. I see all the facial features in miniscule detail down to the nosehairs, tooth grooves, and iris specks, and I see all the facial features flitter-flutter, and I can replay this video from memory later, but I almost never think about what the people with the facial features are thinking in their minds when they are sending these non-verbal cues to me, especially not the complicated subtle social stuff about me and them and others we know and all our relations with each other. If you gimme the RMET with the multiple-choice pre-selected states of mind and the wonderfully grainy pictures for to refer, then I have no problem doing the analyses and matchingks, and I got 34/36 on that test. I missed the questions with answers "distrustful" and "nervous". If I had to come up with the fill-in-the-blank states of mind, then I would score poorly. After real-time interactions, I can analyze the non-verbal cues in my video, but I won't be able to come up with most of the complex social-emotional meanings, because my social knowledge is eggstremely limited due to not knowing that the noises and movements that people were making in my direction were a form of mind-to-mind communication until the 8-10 age range, a rather late start for learning communication and social cognition. Also, I receive equally massive amounts of information from non-social sources, like all the stuff in the room - books, couch wrinkles, carpet fibers, dust specks on the blinds, wood grains on the tables - no filtering of physical information and no emphasis on social meanings. My automatic interest in dust specks and carpet fibers likely causes me to miss some hoooman non-verbal cues if I am spending lots of time automatically looking at these things instead of hooomans.