New audio applications for multimedia and MPEG-4 : Complexity and hardware

Abstract

This paper presents in its first part a brief overview of the new enhanced audio features which are expected to be relevant for the outcoming multimedia applications, moving from the already existing multichannel equipments towards interactive and immersive simulators, videogames, educational applications and many others; most of the needed features (namely 3-D rendering, synthetic audio generation and reverberation, physical modeling of sources and environments) have already been investigated by research activities throughout the last decades, but only nowadays the processing power potential made available by technology can probably be considered enough to solve these problems at a high-quality / acceptable-cost level. Some present and recent past architectures will be shortly introduced in the second half of this paper, and compared with what could be considered as the ideal properties of an enhanced audio system, able to support the future tools and standards for Multimedia and Virtual Reality, and naturally MPEG-4 above them all.

The main goal of this brief presentation is to analyze in which direction these attractive features of the new audio applications are pushing the computational requirements, to detect how the DSPs and multimedia processors are evolving, and finally to conclude with some considerations on the state of the art architectures.