Audio transmission over IP networks opens the way for advanced multichannel audio formats. Such formats can be used for high quality sound reproduction and virtual acoustics beyond today’s stereo and surroundsound formats, e.g., for video-conferencing with significantly improved sound quality. For multichannel formats, coding is a central issue due to the growth in bandwidth for an increase in number of channels. This project studies multichannel audio formats for advanced applications that are suitable for communicationsover IP networks. Application-dependent requirements, channelcorrelation, coding aspects, packet loss handling, psychoacoustic phenomena related to spatial hearing, electro- and room acoustic aspects of end terminals are studied in order to optimize format requirements. A focus is placed on the Ambisonics format which has a layered structure. Furthermore, the multidimensional nature of the perception of realistic sound fields requires new quality metrics. Simulations and real-world perception experiments are used in this project.

Project coordinator: Professor Peter Svensson