When you speak (make noise with your vocal tract), your vocal folds vibrate and create a rich buzzing sound. Think the sound a kazoo makes. That sound goes through your throat and mouth and nose, and is shaped by them. This is the same process that makes flutes sound like flutes and clarinets sound like clarinets. This sound is composed of lots of different frequencies, which we can plot...
image source
Now onto the more advanced stuff!
Each of these plots is actually more than just loudness (vertical axis) and frequency (horizontal axis). In order to calculate these frequencies, you need to sample sound for a duration of time. The longer the duration is, the more information you gather on the sound. Most of these graphs only look at vowel sounds, so they only sample vowels.
That's where LTASS diverges.
LTASS looks at a lot more time and information. Just like a long-exposure photos get blurry, long-term spectra get blurry. The specific characteristics of the sound segments overlap and lose their defining shapes. What you're left with (after a few minutes of speech), is the entire range of sound spectra Β one voice can make.
image source: This is not what voice recognition looks like. This is a wave form. It can be useful, but it's nowhere near as useful as spectra!
This information can be VERY important in forensic voice recognition. I won't go much into forensic linguistics because I don't know much about it, but most LTASS applications involve speaker identification.
Now the coolest part:
You might be thinking that the shape of the spectrum would depend on what words a person says during that time they're being recorded. After all, the shapes of /a/, /i/, and /u/ are so different!
If you get enough speech, it doesn't matter what they say.
In fact, it doesn't even matter what language they're speaking in!
The LTASS is such a good measure of vocal tract characteristics that a single person speaking in English and Mandarin (two very different languages, with regard to sounds) has more similar shapes within that one speaker than between two speakers of English or two speakers of Mandarin.
Put another way...
It doesn't matter what language someone is speaking, their LTASS could possibly identify them!
(Blue and green are the same group of people in 2 languages, red is a different group.)
Anya is live and ready to show you everything. Watch her strip, dance, and perform exclusive shows just for you. Interact in real-time and make your fantasies come true.
β Live Streamingβ Interactive Chatβ Private Showsβ HD Qualityβ Free Actions
Free to watch β’ No registration required β’ HD streaming
