My thought is that human speech contains a huge amount of delicate, not conciously noticable quailites that we are innately attuned to pick up on but find hard

My thought is that human speech contains a huge amount of delicate, not conciously noticable quailites that we are innately attuned to pick up on but find hard to explain explicitly. This is also true of visual facial expressions, but deep fakes know to avoid close ups on the face otherwise the deceit would be apparent, and we are accustomed to seeing faces from afar. Audio on the other hand, we just dont accept perceiving it badly: We'll ask the person to speak louder and clearer, turn up the volume, come closer etc.