For a mercifully brief period of my young life, I earned some money doing telemarketing. That all came to an end when the bosses told me that to improve my percentages, I should stop trying to sell to anyone who sounded "dark." There's more to that story for another time.
Almost anyone who has sold things over the phone can tell you that smiling when you talk is a proven sales tactic (so is mimicking the regional accent of the area you're calling, if you can pull it off). I was reminded of this when I heard a song* in which one line was sung with an obvious, audible smile.
Consider the significance of that. Your hearing is so finely tuned that it can detect the facial position of a total stranger by voice alone. That facial arrangement has no inherent significance, except that it is itself an encoded representation of an emotion. So, when you hear that voice, you're decoding a face from sound, and decoding an underlying emotion from the position of a face you can't see.
* Okay, I'll admit it, the song was "Can't Stop the Feeling" by Justin TImberlake, and the line is "Just imagine..." which is sung just before the chorus.