I think it's actually kind of sad when people take MBTI types and even Jung's works too seriously, claiming there's something "scientifically" or even "objectively" true about it. They strip it of its historical origins and cultural biases and treat it as some kind of authority on human behavior.
MBTI should be fun, and, at it's most serious application, can be used as a personal tool or a way to appreciate people's differences. What concerns me a lot, especially on forums like this, is when people use it as a way to judge someone on a personal level, giving their personal biases and prejudices more fuel and agency to discriminate as opposed to what typing like this was originally meant for.
Your personal bias is exactly that: your own "preference" for what you like or dislike about a person. MBTI doesn't give those preferences, in a social sense, any more or less legitimacy than what a person is able to dish out or take...
In the late ‘90s I wrote a program that could recognize the vowel formants a, e, I, o, and u. I knew I could synthesize them via three bandpass filters, adjusting the frequencies for each vowel. By doing a real-time low-resolution FFT, I could check the amplitude of the partials, match, and voila. And it worked on any voice.
But full speech? I know nothing about plosives and dental fricatives.
I gotta say, that box was kinda sexy...black slab with VFD...hell yes.