To understand how VoIP works, you will be taken through the 
process of voice transmission from one end to the other. 
The process starts with a person talking into the 
mouthpiece on one end of a VoIP call.

This analog voice signal must first be sampled and 
digitized. Voice sampling is usually done 8,000 times per 
second (8KHz). In order to reduce bandwidth, a voice CODEC 
is used. A voice CODEC is a compression/decompression 
algorithm that is optimized for the voice frequency range. 
The bit stream uncompressed is 64Kbps. By using an 
available CODEC, the bit stream can be reduced to 8Kbps or 
less.

In order for the compressed voice data to be sent over the 
Internet, it must go through a process called 
packetization. This is a packet consisting of a small 
sample of the voice data (usually 10-30 milliseconds).

While being routed through the Internet, these packets can 
get delayed or even lost. This can cause degradation in 
voice quality. Simply put, there are various mechanisms in 
place to compensate for these problems and help smooth out 
the audio.

Once all the packets arrive on the listening end of the 
call, they must be reassembled to their original state. The 
packets are decompressed and converted from a digital to 
analog voice signal.