A GPU Compute Framework for Real-Time Ray-Traced Acoustic Simulation in Unity - An analysis in Performance, Accuracy, and Validation

Sammanfattning

This thesis evaluates the viability of GPU compute shader audio ray tracing forreal-time spatial audio, comparing its performance and physical accuracy against ahigh-fidelity wave-based reference (FDTD). We present a custom compute shaderframework built within the Unity engine that utilizes backward Monte Carlo raytracing and dynamic Linear Bounding Volume Hierarchies (LBVH) to generate RoomImpulse Responses (RIR) for each sound source present in a scene. The generatedRIR is processed by a custom native C++ DSP plugin to handle real-time earlyreflections via convolution and late reverberation derived by the Schroeder equation.Testing in a complex architectural scene showed stable real-time performance ataround 80-100 FPS while casting 16,000 rays per frame on an RTX 2060 hardwarebaseline. Accuracy evaluations showed that while the system can successfully modelearly reflections and directionality, it fails to physically sustain late reverberationtails. The acoustic energy decays faster than the wave-based ground truth due toray budget constraints and compute shader control-flow divergence, which limit theamount of computation that can fit within the allowed time frame. We concludethat while GPU audio ray tracing is highly effective, it should be used primarily as asecondary tool to enhance the overall perceptual quality and spatial accuracy of theacoustic environment.