Efficient large-scale multichannel audio coding

Abstract

Systems employing multichannel audio are becoming increasingly common. Multichannel audio is used in domestic surround sound systems, multitrack audio recording and new applications are emerging. Digital multichannel audio systems often employ non-standard formats that complicate the exchange of audio data between different systems. Typical multi-track recording environments consist of a large number of channels. Each channel has its own acoustic configuration with respect to equalisation, levels and effects. Another characteristic of multitrack recording is that only a small subset of channels are non-silent at any one time. A family of lossless, compact and general multichannel audio formats are proposed that support a large number of audio channels and the embedding of non-audio information. Non-silent tracks are combined using Reed Solomon codes. The strategy can therefore be implemented efficiently using one of the many existing Reed Solomon encoder/decoder hardware designs, and several configurations are illustrated.