1 ABSTRACT

Packetized speech transmission systems implemented with Voice over IP are gaining momentum against the traditional circuit switched systems despite the fact that packet switched VoIP is two to three times less efficient than its circuit switched counterpart. At the same time, it only supports a rather bad “toll” quality. We believe that it is time for a new architecture developed from scratch – an architecture that includes an Internet enabled speech codec and its transport system. This architecture manages the perceptual service quality while using the available transmission resources to its best. The transmission of speech is managed and controlled with respect to its speech quality, mouth-to-ear delay, bit-rate, frame-rate, and loss robustness. Beside the architecture, we describe the requirements for the Internet speech codec and its transport protocol and present an interface between the speech codec and the transport protocol.