Sensitivity Analysis of Generative Spatial Audio Metrics: A Study on Responsiveness, Smoothness, and Symmetry
For researchers evaluating generative spatial audio, this provides a first step toward understanding metric sensitivity, but is incremental as it only analyzes existing metrics.
The paper proposes a framework to analyze sensitivity of generative spatial audio metrics along continuous spatial trajectories, defining desiderata of Responsiveness, Smoothness, and Symmetry. It finds that FAD with localization-specific embeddings and acoustic maps perform well, while intensity vectors degrade with complexity.
Evaluating generative spatial audio for First-Order Ambisonics (FOA) remains challenging due to a limited understanding of how metrics respond to changes in spatial parameters such as azimuth and elevation. We propose a framework to analyze metric sensitivity along continuous spatial trajectories, drawing on principles of sensitivity analysis in parametric sound synthesis. Using controlled FOA scenes with increasing scene complexity, we define three desiderata for metric behavior: Responsiveness, Smoothness, and Symmetry. We assess standard distribution-based and sample-based metrics, including Fréchet Audio Distance (FAD), intensity vectors, and acoustic maps. Our findings show that FAD using localization-specific embeddings and acoustic maps yield high Responsiveness and robust Smoothness and Symmetry across conditions, while intensity vectors degrade with increasing scene complexity. This is the first step towards investigating the sensitivity of metrics for generative spatial audio.