Hi all, this is my first post at this blog. I decide to make this one short and compact. I hope it would be more understandable this way.

Here we go, human vocal system can be classified into two types of spectrum : voiced, dan unvoiced. Before we go any further, below is an illustrarion of my voice spectrum :

This is my vocal spectrum saying : AlfathAkbar

From this picture, you can see there is some gaps there, makes you think that i stops a while when i say “Al”(stop)”Fath”(stop)”Ak”(stop)”Bar”. But the fact is, i speak from start to end without a single second of silence! So if i speak nonstop, why is there three “band” where my voice magnitude is almost 0 dB? This is why :

Rough analysis on my vocal spectrum

So generally, vocal letters (a,i,u,e,o) is voiced and consonant letters is unvoiced. Here is some depth explanations :

1. Voiced Spectrum

a. Vocal cord (pita suara) vibrates with a pitch frequency

b. Voiced signal can be considered as a periodic signal

c. Signal will change according to the movement of trachea, mouth, and nose.

2. Unvoiced Spectrum

a. Vocal cord doesn’t vibrate

b. The excitation to the vocal tract can be considered as a random noise

c. The energy is relatively small

d. Doesn;t have the characteristic of a periodic signal

One more related fact about this classification, is a decoder of a LPC (Linear Predictive Coding). LPC is a human vocal modelling system. This is the block diagram of LPC decoder :

(on progress)

The noise generator is used to generate the unvoiced spectrum sent by the transmitter. So, we can see that LPC uses this voiced and un-voiced thing as an advantage to model our voice.

That’s it! Thanks for reading.. =)