A physics-based low-order filter approximation for scattering from a rigid sphere.

Area of Science:

• Acoustics
• Digital Signal Processing
• Computational Physics

Background:

• Efficient simulation of object scattering is vital for real-time virtual and room acoustics.
• The analytical solution for rigid sphere scattering is computationally intensive for interactive applications.
• Rigid spheres serve as simplified models for complex 3D object scattering.

Purpose of the Study:

• To develop a physically informed, low-order digital filter approximation for rigid sphere scattering.
• To enable efficient, real-time simulation of sound wave diffraction for arbitrary source and receiver positions.
• To provide a computationally feasible alternative to the complex analytical solution for interactive audio rendering.

Main Methods:

• Approximation of low-frequency scattering using combined first-order low- and high-pass filters based on spherical harmonics.
• Modeling high-frequency scattering by approximating shortest and longest paths around the sphere.
• Combining low- and high-frequency approximations with a blending function for wideband results.

Main Results:

• The low-frequency approximation shows good agreement with the analytical solution, with mean RMS errors below 0.5 dB for distances > 2x sphere radius.
• Errors increase to approximately 1.5 dB at smaller distances for the low-frequency approximation.
• The high-frequency approximation yields a mean magnitude error of about 2 dB.

Conclusions:

• The proposed digital filter approximation offers an efficient method for simulating rigid sphere scattering in real-time.
• This approach effectively balances accuracy and computational cost for interactive acoustic simulations.
• The wideband approximation provides a viable solution for virtual acoustics applications requiring real-time performance.