Task splitting for DNN-based acoustic echo and noise removal

IWAENC |

Organized by IEEE

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.