jwr1995 / WD-TCN

Utterance Weighted Multi-Dilation Temporal Convolutional Network (WD-TCN)

This respository contains the WD-TCN mask estimation network code from our paper "Utterance Weighted Multi-Dilation Temporal Convolutional Networks for Monaural Speech Dereverberation". The code is designed to replace the TCN MaskNet from the SpeechBrain project which documentation is provided for at: https://speechbrain.readthedocs.io/en/latest/API/speechbrain.lobes.models.conv_tasnet.html#speechbrain.lobes.models.conv_tasnet.MaskNet

Please cite the following paper

@misc{wdtcn,
  title = {Utterance Weighted Multi-Dilation Temporal Convolutional Networks for Monaural Speech Dereverberation},
  author = {Ravenscroft, William and Goetze, Stefan and Hain, Thomas},
  booktitle={2020 17th International Workshop on Acoustic Signal Enhancement (IWAENC)},  
  month = {sep},  
  year = {2022},
  doi = {10.48550/ARXIV.2205.08455},
  url = {https://arxiv.org/abs/2205.08455},
}