Voice Activity Detection (VAD) is a technique for voice pause detection. The VAD technique detects whether or not voice is being transmitted in telephone or mobile networks. The voice pause detection technique is described in ITU RecommendationG.722 and is used in digital telephone and mobile networks to optimize transmission in the voice channel.
The VAD technique is comparable to Digital Circuit Multiplication Equipment( DCME), which uses the pauses in speech for the transmission of other data. Since, from a purely statistical point of view, more than half of the transmission times in telephone calls are not used because of the pauses in speech, these times can be used for more efficient use of the network capacity and to save energy in the cell phones. Among other things, voice pause detection ensures this.
In the GSM standard, VAD technology is used in Discontinuous Transmission( DTX) to perform a shutdown of active circuits during pauses in speech, thereby saving battery power. As soon as a pause in speech occurs during a telephone call, the background noise is abruptly terminated due to the procedure. Since this is extremely annoying, a comfort noise is generated in the VAD process. To ensure that the background noise is as realistic as possible, the DTX module sends silence insertion descriptor( SID) frames to the comfort noise generator( CNG) on the receiving end at regular intervals.
Speech pause detection is also used in Internet telephony. This software application is used to monitor voice traffic. If there is no voice transmission, no data packets are transmitted. The transmission bandwidth thus saved can be used for data traffic. This voice pause detection technique is known as silence suppression.