The problem is multiplied by an information-theory phenomenon: like a mathematics textbook, each line of dialog these days is distilled to contain maximum information. There is little redundancy. If you miss a line, it really matters.
Combine that with this trend of filming long, brooding pauses punctuated by only occasional dialog, and you get a lot of resentful watchers after they've waited for 30 seconds for the actor to finally say something, and then it trails off unintelligibly, leading to a 90-second adventure in rewinding not far enough, then too far, then waiting for buffering...