Multichannel Audio Coding for Multimedia Services in Intelligent Environments

Abstract

Audio is an integral component of multimedia services in intelligent environments. Use of multiple channels in audio capturing and rendering offers the advantage of recreating arbitrary acoustic environments, immersing the listener into the acoustic scene. On the other hand, multichannel audio contains a large degree of information which is highly demanding to transmit, especially for real-time applications. For this reason, a variety of compression methods have been developed for multichannel audio content. In this chapter, we initially describe the currently popular methods for multichannel audio compression. Low-bitrate encoding methods for multichannel audio have also been recently starting to attract interest, mostly towards extending MP3 audio coding to multichannel audio recordings, and these methods are also examined here. For synthesizing a truly immersive intelligent audio environment, interactivity between the user(s) and the audio environment is essential. Towards this goal, we present recently proposed multichannel-audio-specific models, namely the source/filter and the sinusoidal models, which allow for flexible manipulation and high-quality low-bitrate encoding, tailored for applications such as remote mixing and distributed immersive performances.