Global IP Sound released a new set of software modules for enhancing the audio quality of VoIP calls. The modules include functionality to add comfort noise during silence and automatically adjust audio levels.
Global IP Sound's Voice Activity Detector (VAD) classifies every frame as active speech frame or non-active speech frame. Non-active frames are passed to the comfort noise encoder that encodes the parameters of the background noise. The Discontinuous Transmission (DTX) defines the packet size and packet interval for the background noise parameters to obtain a decrease in bit rate. On the receiving side, the Comfort Noise Generator (CNG) decoder analyzes the background noise parameters and generates a comfort noise signal. Collectively the modules makes it possible to maintain a high perceived sound quality while reducing the average transmitted data rate during silent periods of speech. The VAD can operate in several modes depending on the usage scenario. http://www.globalipsound.com
- Global IP Sound's sound processing software is designed to provide better than PSTN quality sound for calls over packet networks. Instead of using buffers, which can add to latency, Global IP Sound's approach is to encode input speech into a series of sound descriptions that are separately transmitted to build a “sound profile�? that aids in lost packet compensation. The speech quality processing is thereby provided at the edge of the network, rather than at the core. The software can be implemented in IP phones, softphones, media gateways, and wireless VoIP on PDAs. The company claims better than PSTN quality VoIP even in adverse network conditions of up to 30% packet loss.
