Diffuse Sound Field Synthesis

Franz Zotter,Stefan Riedel,Lukas Gölles,Matthias Frank

2024-02-21

Abstract:Can uncorrelated surrounding sound sources be used to generate extended diffuse sound fields? By definition, targets are a constant sound pressure level, a vanishing average sound intensity, uncorrelated sound waves arriving isotropically from all directions. Does this require specific sources and geometries for surrounding 2D and 3D source layouts?

Audio and Speech Processing,Sound

What problem does this paper attempt to address?

This paper mainly discusses how to generate an extended diffuse sound field using uncorrelated ambient sound sources. A diffuse sound field refers to uncorrelated plane waves from all directions, with constant sound pressure level, zero average sound intensity, and isotropic arrival characteristics. The study analyzes different layouts of sound sources with different shapes (such as circular and spherical) and decay exponents through numerical simulations and a series of calculations to determine if an ideal diffuse sound field can be achieved.

The authors found that for a circular layout, an exponent of 1/2 is required to generate an ideal diffuse sound field, while for a spherical layout, an exponent of 1 is needed. Although no layout can perfectly maintain a constant expected sound pressure level, the flatness is acceptable. The spherical design is used to describe the optimal sound source layout with high diffusivity, and non-spherical or non-convex layouts can be improved through the restoration of isotropy or mode matching for diffuse synthesis.

The paper also discusses how to synthesize a diffuse sound field using a surround loudspeaker array, which is important for acoustic measurements and psychoacoustic research in laboratory environments (such as building sound insulation, absorption, and microphone array response). In spatial audio reproduction, the perception of envelopment in a diffuse sound field is crucial, and the use of surround loudspeakers can achieve this goal. Although there is spatial aliasing, the localization effects based on psychoacoustics suggest that these methods perform well in larger listening areas.

In addition, the study proposes the use of modified radial decay 1/r^β and hypergeometric functions, Gegenbauer polynomials, and circular and spherical harmonics for representation to gain a deeper understanding of the physical principles of diffuse sound fields. For non-circular/non-spherical layouts, the synthesis of diffuse sound fields can be optimized through mode matching techniques, but it may affect isotropy.

Overall, this paper aims to establish a systematic theory to better understand and process the synthesis of diffuse sound fields, and provide technical support for applications in related fields.