Login / Signup

Towards Low-Distortion Multi-Channel Speech Enhancement: The ESPNET-Se Submission to the L3DAS22 Challenge.

Yen-Ju LuSamuele CornellXuankai ChangWangyou ZhangChenda LiZhaoheng NiZhong-Qiu WangShinji Watanabe
Published in: ICASSP (2022)
Keyphrases
  • multi channel
  • single channel
  • speech enhancement
  • noise reduction
  • noisy environments
  • maximum likelihood
  • prior information
  • face recognition
  • feature vectors
  • signal to noise ratio