Inplace Gated Convolutional Recurrent Neural Network For Dual-channel Speech Enhancement

TL;DR

This paper introduces a compact inplace Gated Convolutional Recurrent Neural Network (GCRN) for dual-channel speech enhancement, effectively preserving spatial cues and improving speech quality through a novel spectrum recovery method.

Contribution

It presents a novel inplace GCRN architecture that combines inplace convolution with recurrent neural networks for efficient multi-channel speech enhancement.

Findings

Effective preservation of spatial cues in dual-channel speech enhancement.

Improved speech quality through a new spectrum recovery method.

Compact model suitable for real-time applications.

Abstract

For dual-channel speech enhancement, it is a promising idea to design an end-to-end model based on the traditional array signal processing guideline and the manifold space of multi-channel signals. We found that the idea above can be effectively implemented by the classical convolutional recurrent neural networks (CRN) architecture. We propose a very compact in place gated convolutional recurrent neural network (inplace GCRN) for end-to-end multi-channel speech enhancement, which utilizes inplace-convolution for frequency pattern extraction and reconstruction. The inplace characteristics efficiently preserve spatial cues in each frequency bin for channel-wise long short-term memory neural networks (LSTM) tracing the spatial source. In addition, we come up with a new spectrum recovery method by predict amplitude mask, mapping, and phase, which effectively improves the speech quality.