Ultra Low Delay Audio Source Separation Using Zeroth-Order Optimization

Abstract

In this paper, we introduce the "Random Directions" probabilistic optimization method, demonstrating its efficacy in real-time, low-latency signal processing applications. Applied to an ultra-low delay, time-domain, multichannel source separation system, our "Random Directions" is compared with gradient-based method "Trinicon" and frequency domain methods like AuxIVA and FastMNMF. Results indicate that our approach often outperforms Trinicon in terms of the Signal to Interference Ratio (SIR) and presents the least non-linear distortions among all methods, as measured by the Signal to Artifacts Ratio (SAR). This study suggests that probabilistic optimization methods, traditionally perceived as slow, can indeed be effective for real-time applications.