Speaker: Maya Tia Kanani.
Abstract: Real-time acoustic simulation remains a significant challenge due to the high computational cost of accurately modelling sound propagation across the full audible frequency range. Building on previous work on perceptually optimised hybrid acoustic simulation, this project extends the framework by investigating frequency-dependent modelling techniques that exploit the strengths of different acoustic propagation methods. The proposed approach combines theoretical wave-based models to accurately capture low-frequency phenomena, where diffraction and interference are dominant, with geometric acoustics and stochastic scattering methods to efficiently simulate higher-frequency sound propagation. The resulting hybrid framework aims to provide physically plausible room acoustics while maintaining the computational performance required for interactive and virtual environments. The implementation is evaluated through numerical simulations and comparison with existing acoustic modelling approaches, considering both computational efficiency and the perceptual quality of the generated sound field. The project demonstrates how integrating complementary modelling techniques across different frequency ranges can improve the realism of real-time room acoustic simulation while preserving practical performance for interactive applications.