Yeah, I'm kind of skeptical of the specifics, even if in the very broadest sense I think there are blatant traits that can manifest.
It's always easiest to reference myself, plus I've actually consciously done some voice alteration for various reasons (some practical).
Basically if I'm in "normal" mode (i.e., I'm not thinking at all about my voice but merely "thinking out loud"), I'm very monotone, flat, quiet. I would call that my thought process voice. I've had people tell me the timbre is actually reassuring/calming, but to me i just feel kind of flat, no emotive peaks.
Instead, if I'm in creative/fun mode, I get louder and I'm laughing a lot, more volume and more peaks, but still not a ton of nuance.
However, let's say I'm on the phone with someone at work or in a phone meeting. I consciously actually change my voice, it becomes more emotive, it's fishing for a connection, trying to win the other over. It's about as emotive as my voice gets, and much of it, I'm even aware of -- it's a conscious decision. Higher-pitched. Not as flat. In general, any time I'm in a group that isn't people I'm regular friends with, my voice can take on a more emotive tone. But it also drains energy from me, I get tired eventually and the voice quality eventually will go back to flatter presentation if/when I run out of steam or no longer want to be there.... or if we get into a purely informational bent and I stop thinking about my presentation. But I see it clearly as me trying to present a certain way, for certain reasons.
See, as far as MBTI goes, it's not a far cry to label the first as Ti, the second as Ne, and the final one as Fe. I mean, that kind of makes sense... in a broad way.
HOWEVER, you can't know that just by listening to a single clip of me. You have to already have a benchmark FOR me, know the range of what kinds of voice I'm capable of using, and then know what context I am using that voice in. So in terms of the specific and predictive quality, you can't really just listen to a clip and spit out a specific personality assessment. In fact, it's easier to recognize qualities of my voice if you already know my personality, which is kind of the opposite of determining personality by voice. Maybe you can start with a voice and get some ballpark ideas to follow up on, but definitively?
Voice working with body language and how people present themselves (and it's always in a context, then, because you're engaging them in a particular context), can suggest a type, I think you just need other factors to get specific.
