Towards Array-Invariant Speech Enhancement via Geometry-Aware Dynamic Convolution

2026-07-20

AI Summary

This work addresses the limited generalization of existing speech enhancement models trained on fixed microphone arrays to devices with diverse geometric configurations. To overcome this challenge, we propose a Geometry-aware Dynamic Convolution (Geo-DConv) framework, which explicitly incorporates array geometry priors into array-agnostic speech enhancement for the first time. By leveraging a dynamic convolution mechanism that integrates microphone coordinate information directly into the network, Geo-DConv enables adaptive spatial filtering, effectively transforming fixed-array models into array-invariant systems. Extensive experiments on the RealMAN real-world recording dataset demonstrate that Geo-DConv consistently enhances performance across various array topologies when applied to two representative baseline models, significantly improving cross-topology generalization.

Abstract

Multi-channel speech enhancement (SE) systems exhibit superior performance over single-channel methods but are constrained to fixed microphone array configurations. This restricts their real-world deployment across devices with diverse array geometries. While recent array-agnostic SE methods address variable microphone numbers and permutations, they largely fail to exploit explicit array geometry priors when available, missing a crucial cue for optimal spatial filtering. A Geometry-Aware Dynamic Convolution (Geo-DConv) framework is proposed, which explicitly leverages microphone coordinates to transform standard fixed-array SE models into robust array-invariant systems. Experiments are conducted on the recent real-recorded RealMAN multi-channel speech dataset. Results demonstrate that the proposed architecture enables two widely used fixed-array models to adapt to array-invariant settings, with consistent performance improvements across diverse array topologies.

Problem

• array-invariant
• speech enhancement
• microphone array geometry
• spatial filtering
• multi-channel

Innovation

• Geometry-Aware Dynamic Convolution
• Array-Invariant Speech Enhancement
• Microphone Array Geometry
• Multi-channel Speech Enhancement
• Spatial Filtering