Global Rotation Equivariant Phase Modeling for Speech Enhancement with Deep Magnitude-Phase Interaction

TL;DR

This paper introduces a novel deep learning framework that models speech phase with global rotation equivariance, improving various speech enhancement tasks by respecting the circular nature of phase.

Contribution

It proposes a magnitude-phase dual-stream architecture with GRE-preserving modules, advancing phase modeling in speech enhancement.

Findings

Reduces Phase Distance by over 20% in phase retrieval.

Improves PESQ by more than 0.1 in zero-shot denoising.

Demonstrates superiority across multiple speech enhancement tasks.

Abstract

While deep learning has advanced speech enhancement (SE), effective phase modeling remains challenging, as conventional networks typically operate within a flat Euclidean feature space, which is not easy to model the underlying circular topology of the phase. To address this, we propose a magnitude-phase dual-stream framework that aligns the phase stream with its intrinsic circular geometry by enforcing Global Rotation Equivariance (GRE) characteristic. Specifically, we introduce a Magnitude-Phase Interactive Convolutional Module (MPICM) for modulus-based information exchange and a Hybrid-Attention Dual Feed-Forward Network (HADF) bottleneck for unified feature fusion, both of which are designed to preserve GRE in the phase stream. Comprehensive evaluations are conducted across phase retrieval, denoising, dereverberation, and bandwidth extension tasks to validate the superiority of the proposed method over multiple advanced baselines. Notably, the proposed architecture reduces Phase Distance by over 20\% in the phase retrieval task and improves PESQ by more than 0.1 in zero-shot cross-corpus denoising evaluations. The overall superiority is also established in universal SE tasks involving mixed distortions. Qualitative analysis further reveals that the learned phase features exhibit distinct periodic patterns, which are consistent with the intrinsic circular nature of the phase. The source code is available at https://github.com/wangchengzhong/GRE-Net.