Speech/audio coding technologies and their applications

Multimedia signal processing by software or firmware can be executed in realtime on the latest high-performance CPU or DSP. Moreover, storage, processing, and communication of speech and audio data will be handled easily using personal computers connected to networks, such as the Internet. This paper describes NTT's speech coding software `DualSpeech' and audio coding technology `TwinVQ' for multimedia services, and their application systems. They feature a low bit-rate, low arithmetic complexity, and high quality of speech/audio. With conventional technologies, these three advantages were incompatible, but `DualSpeech' demonstrates high performance in Internet speech casting, voice mail, Internet tele-conferencing, and network reading systems. `TwinVQ' also demonstrates high performance in network broadcasting systems and network audio players.