Echo Cancelation and Noise Suppression by Training a Dual-Stream Recurrent Network with a Mixture of Training Targets

Fatemeh Alishahi, Yin Cao, Youngkoen Kim, Asif Mohammad

Research output: Chapter in Book or Report/Conference proceedingConference Proceedingpeer-review

Abstract

Nonlinear echo in presence of background noise can degrade the performance of digital signal processing algorithms. Deep neural networks with their ability to model complex nonlinear functions can potentially address this issue. In this paper, a deep and causal neural network based on dual streaming of the near-end microphone and far-end speech signals is employed to leverage the real-time nonlinear echo cancellation and noise suppression. The extracted features of two streams are coupled into a shared neural network for joint echo and noise cancellation. The training target is a mixture of spectral mapping and masking-based targets which are gated through a feedforward neural network. The model is evaluated in terms of both signal-level and perception-level metrics for different scenarios with a range of SI-SDR as low as -25 dB. Furthermore, the effect of mixing of training targets is assessed by evaluating different models.

Original languageEnglish
Title of host publicationInternational Workshop on Acoustic Signal Enhancement, IWAENC 2022 - Proceedings
PublisherInstitute of Electrical and Electronics Engineers Inc.
ISBN (Electronic)9781665468671
DOIs
Publication statusPublished - 2022
Externally publishedYes
Event17th International Workshop on Acoustic Signal Enhancement, IWAENC 2022 - Bamberg, Germany
Duration: 5 Sept 20228 Sept 2022

Publication series

NameInternational Workshop on Acoustic Signal Enhancement, IWAENC 2022 - Proceedings

Conference

Conference17th International Workshop on Acoustic Signal Enhancement, IWAENC 2022
Country/TerritoryGermany
CityBamberg
Period5/09/228/09/22

Keywords

  • Supervised speech enhancement
  • deep neural network
  • recurrent neural networks
  • training targets

Fingerprint

Dive into the research topics of 'Echo Cancelation and Noise Suppression by Training a Dual-Stream Recurrent Network with a Mixture of Training Targets'. Together they form a unique fingerprint.

Cite this

Alishahi, F., Cao, Y., Kim, Y., & Mohammad, A. (2022). Echo Cancelation and Noise Suppression by Training a Dual-Stream Recurrent Network with a Mixture of Training Targets. In International Workshop on Acoustic Signal Enhancement, IWAENC 2022 - Proceedings (International Workshop on Acoustic Signal Enhancement, IWAENC 2022 - Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/IWAENC53105.2022.9914701