Hierarchical Modeling of Spatial Cues via Spherical Harmonics for Multi-Channel Speech Enhancement

TL;DR

This paper introduces a hierarchical spherical harmonic transform-based approach for multi-channel speech enhancement, explicitly modeling spatial cues to improve performance with fewer parameters.

Contribution

It presents a novel hierarchical framework that explicitly incorporates spatial modeling using spherical harmonics, enhancing multi-channel speech enhancement.

Findings

Effective recovery of target spatial patterns

Improved performance over baseline models

Fewer parameters and computations needed

Abstract

Multi-channel speech enhancement utilizes spatial information from multiple microphones to extract the target speech. However, most existing methods do not explicitly model spatial cues, instead relying on implicit learning from multi-channel spectra. To better leverage spatial information, we propose explicitly incorporating spatial modeling by applying spherical harmonic transforms (SHT) to the multi-channel input. In detail, a hierarchical framework is introduced whereby lower order harmonics capturing broader spatial patterns are estimated first, then combined with higher orders to recursively predict finer spatial details. Experiments on TIMIT demonstrate the proposed method can effectively recover target spatial patterns and achieve improved performance over baseline models, using fewer parameters and computations. Explicitly modeling spatial information hierarchically enables more effective multi-channel speech enhancement.