Bilkent Samsurya

Independent Researcher

Abstract

Accurate acoustic simulation is crucial for immersion, yet many industry pipelines overlook wave phenomena, especially at low frequencies. We present a 2D finite-difference time-domain (FDTD) framework integrated with Unreal Engine that models sound as a wave process and embeds occlusion, diffraction, reflection, and interference. Scenes are top-down projected to a 2D grid to derive obstacle masks and boundary conditions. A Python FDTD solver drives a sine sweep at the source, while virtual quadraphonic microphone arrays capture pressure responses at listener positions. Deconvolution produces multi-channel, direction-aware impulse responses that are injected into Unreal Engine’s audio pipeline for dynamic playback. Benchmarks align with analytical expectations, and we outline hybrid extensions for practical deployment. The approach offers a tractable, wave-based alternative for interactive applications.

10.64560/58989191

Keywords