Task splitting for DNN-based acoustic echo and noise removal

2022/05/13 by Sebastian Braun, Braun, Sebastian, María Luis Valero +1 · 1 citation
Computer Science · Engineering · #Speech and Audio Processing #Advanced Adaptive Filtering Techniques #Ultrasonics and Acoustic Wave Propagation

paper · pdf · doi:10.48550/arxiv.2205.06931

Abstract

Neural networks have led to tremendous performance gains for single-task speech enhancement, such as noise suppression and acoustic echo cancellation (AEC). In this work, we evaluate whether it is more useful to use a single joint or separate modules to tackle these problems. We describe different possible implementations and give insights into their performance and efficiency. We show that using a separate echo cancellation module and a module for noise and residual echo removal results in less near-end speech distortion and better performance during double-talk at same complexity.

Cited by

• Model as Loss: A Self-Consistent Training Paradigm

Related

• Low-Complexity, Real-Time Joint Neural Echo Control and Speech\n Enhancement Based On PercepNet
• Deep Residual Echo Suppression and Noise Reduction: A Multi-Input FCRN Approach in a Hybrid Speech Enhancement System
• Personalized Acoustic Echo Cancellation for Full-duplex Communications
• Efficient High-Performance Bark-Scale Neural Network for Residual Echo and Noise Suppression
• AEC in a NetShell: On Target and Topology Choices for FCRN Acoustic Echo\n Cancellation