Bandwidth-Scalable Fully Mask-Based Deep FCRN Acoustic Echo Cancellation and Postfiltering

TL;DR

This paper introduces a scalable deep neural network architecture for acoustic echo cancellation that supports multiple bandwidths, including fullband, with modular training and a bandwidth extension stage, achieving robust performance in challenging scenarios.

Contribution

It presents a fully mask-based FCRN model with bandwidth scalability and a flexible bandwidth extension stage, enabling effective AEC across various bandwidths.

Findings

Competitive performance on ICASSP 2022 challenge

Robust noise and echo suppression at high bandwidths

Effective handling of delayed echo and dynamic changes

Abstract

Although today's speech communication systems support various bandwidths from narrowband to super-wideband and beyond, state-of-the art DNN methods for acoustic echo cancellation (AEC) are lacking modularity and bandwidth scalability. Our proposed DNN model builds upon a fully convolutional recurrent network (FCRN) and introduces scalability over various bandwidths up to a fullband (FB) system (48 kHz sampling rate). This modular approach allows joint wideband (WB) pre-training of mask-based AEC and postfilter stages with dedicated losses, followed by a separate training of them on FB data. A third lightweight blind bandwidth extension stage is separately trained on FB data, flexibly allowing to extend the WB postfilter output towards higher bandwidths until reaching FB. Thereby, higher frequency noise and echo are reliably suppressed. On the ICASSP 2022 Acoustic Echo Cancellation Challenge blind test set we report a competitive performance, showing robustness even under highly delayed echo and dynamic echo path changes.