Spatial sound reproduction

Certain software techniques and associated hardware devices enable spatialized sound reproduction for virtual environments. The spatialization of sound, i.e. the simulation of our hearing's spatial localization cues for virtual sound sources, is one of the most active research themes in acoustics today. Consumer applications are legion: video games, home cinema, music, whether at home or on nomadic devices. In virtual reality applications, the addition of spatialized sound also enhances the sensation of immersion [1]. These technologies are all based on intensive fundamental research into the spatial perception of sound in humans [2]. Reference works such as [3] [4] can also be consulted for further details.

Generally speaking, spatial sound reproduction systems are concerned with reproducing for the listener's ears the appropriate localization cues for a virtual sound source according to its desired position. These systems can be categorized into two families. The first series, which can be described as perceptive approaches, is based on the fact that a simple model of sound level difference or arrival time at the two ears is sufficient to create the illusion of positioning a "phantom" source between the physical positions of the loudspeakers. This principle is the basis of classical stereophonic reproduction in particular, but can be extended to more complex loudspeaker arrays. The second series of approaches aims to reconstruct the acoustic field in the vicinity of the listener with physical precision. In this category, we distinguish between binaural approaches, which reconstruct the sound field at two points (the listener's ears), and more global approaches, which reproduce the correct sound field in an entire zone of space. Each of these approaches has its advantages and disadvantages in terms of computational resources, device complexity and number of simultaneous listeners. For further information, please refer to [3] and [5]. We also present some examples of the integration of spatial sound restitution devices into immersive virtual reality systems. In particular, we draw up a comparison of the different approaches and their optimal conditions of use. Finally, we devote the last section to a brief overview of a few existing software libraries for implementing such techniques, some of which benefit from hardware acceleration.