Virtual sound source generation: Its various methods and performances

Abstract

There are many means to generate a virtual sound source, or sources in the region of interest. For example, Wave Field Synthesis (WFS) or Higher Order Ambisonics (HOA) are good examples. These methods normally assume that loudspeakers are spatially distributed in the space. The region of interest where the desired sound is generated can have arbitrary shape; enclosed by surrounding loudspeakers or partially enclosed. Therefore, the performance of the method would be affected by the boundary conditions as well as the wave length of desired wave field. In other words, how the waves are distributed in the selected space. In recent work [Choi and Kim, IEEE Trans. Speech Audio Process. 20(7), 1976–1989 (2012)], a new approach was proposed to generate virtual sources in the space that is enclosed by an array of loudspeakers, which have been believed to be problematic with well-known methods. It is proved to be mathematically exact solution. However, “exact solution” does not necessarily mean that it is better than the others. In this paper, performances of these three methods are compared. Theoretical and experimental comparisons have been attempted and observed in this paper.