Online neural fusion of distortionless differential beamformers for robust speech enhancement

TL;DR

This paper introduces a neural fusion framework for combining multiple fixed beamformers in real-time, significantly improving speech enhancement robustness in dynamic acoustic environments by effectively adapting to rapid changes.

Contribution

It proposes a novel frame-online neural fusion method that estimates combination weights via neural networks, outperforming traditional adaptive convex combination in non-stationary scenarios.

Findings

Enhanced interference suppression in dynamic environments

More effective adaptation to rapid acoustic changes

Maintains distortionless speech signals

Abstract

Fixed beamforming is widely used in practice since it does not depend on the estimation of noise statistics and provides relatively stable performance. However, a single beamformer cannot adapt to varying acoustic conditions, which limits its interference suppression capability. To address this, adaptive convex combination (ACC) algorithms have been introduced, where the outputs of multiple fixed beamformers are linearly combined to improve robustness. Nevertheless, ACC often fails in highly non-stationary scenarios, such as rapidly moving interference, since its adaptive updates cannot reliably track rapid changes. To overcome this limitation, we propose a frame-online neural fusion framework for multiple distortionless differential beamformers, which estimates the combination weights through a neural network. Compared with conventional ACC, the proposed method adapts more effectively to dynamic acoustic environments, achieving stronger interference suppression while maintaining the distortionless constraint.