Voice over IP continues to collect more and more support from companies. It allows you to reduce operating costs and investments in infrastructure and also allows an increase in productivity by adding new features.
The importance of codecs
Encoding and decoding is performed by special programs, called codecs. There are different and the type used for voice is that G.729 compresses any sound in blocks lasting 10 milliseconds each. Is primarily used for VoIP conversations and encode sound at a rate of 8 Kbits per second.
This is a frequency coding is not suitable to play music or even to send faxes, but sufficient to transfer voice conversations. We remind here that the human voice occupies only a fraction of the “spectrum” (together) that the frequency ‘ ear can hear. More precisely, the human voice is from 300 to 3,400 hertz (cycles per second) while the set of audible sounds is from 20 to 20,000 Hz
It is considered therefore that the digitization of voice is sufficient to have a channel from 4 kHz to which you combine two sidebands kept clear to prevent interference. Accounts with debt then you arrive at a sampling rate of 8 KHz, which is the system that converts the original sound into digital pulses takes 8,000 samples per second which are then processed by the G.729 codec which produces packs of 10 compounds from 20 milliseconds bytes each.
The protocols used
The packets are then grouped in pairs and routed on the Internet using one of three typical protocols: IP, UDP (User Datagram Protocol) or RTP (Real-time Transport Protocol) that add another 40 bytes per packet. When you must leave the line ADSL packages also need to encapsulate PPP (Point to Point Protocol).
It is an encapsulation additional need to manage a communication between two fixed points (the connection between the router and the provider). This addition brings another 7 +4 bytes per packet. Summing all subsequent encapsulations we see that a channel for conveying voice quality normally takes 34.4 Kbits per second.