|
|
|
Before voice can be transmitted over an IP network, it must be digitized. Analog to digital conversion usually takes place at a rate of 8000 samples per second. The digital stream is then encoded by A-law, mu-law, or bit-rate reduction methods and finally grouped into packets for transmission. Echo-cancellation to eliminate acoustic or electronic network reflection effects also takes place upon reception. To reduce packet transmission rates, silence suppression detects when there are periods of silence and, during those times, does not transmit packets.
When the packets are received, they must be put in proper order and converted back into an analog voice signal. In addition, jitter must be removed and the effects of packet loss mitigated. Various algorithms are used to deal with jitter and packet loss. In addition, silence suppression is eliminated by adding artificial samples, often in the form of comfort noise, a random, low-level signal that gives the impression that the connection is still alive during periods of silence.
VoIP is a specific set of protocols for the transmission and reception of digital packets. In addition, the User Datagram Protocol (UDP) provides for packet error detection; although the UDP does not guarantee that packets arrive in the order they are sent. As a result, it is possible to get misordered packets which may affect voice quality. The Real-Time Protocol (RTP) numbers the packets in sequence so that they can be put in proper sequence on the receiving side and played out from the jitter buffer at the rate at which they were originally transmitted.
|
|
|