ISCA Archive Interspeech 2021
ISCA Archive Interspeech 2021

Dual-Path Filter Network: Speaker-Aware Modeling for Speech Separation

Fan-Lin Wang, Yu-Huai Peng, Hung-Shin Lee, Hsin-Min Wang

Speech separation has been extensively studied to deal with the cocktail party problem in recent years. All related approaches can be divided into two categories: time-frequency domain methods and time domain methods. In addition, some methods try to generate speaker vectors to support source separation. In this study, we propose a new model called dual-path filter network (DPFN). Our model focuses on the post-processing of speech separation to improve speech separation performance. DPFN is composed of two parts: the speaker module and the separation module. First, the speaker module infers the identities of the speakers. Then, the separation module uses the speakers’ information to extract the voices of individual speakers from the mixture. DPFN constructed based on DPRNN-TasNet is not only superior to DPRNN-TasNet, but also avoids the problem of permutation-invariant training (PIT).


doi: 10.21437/Interspeech.2021-858

Cite as: Wang, F.-L., Peng, Y.-H., Lee, H.-S., Wang, H.-M. (2021) Dual-Path Filter Network: Speaker-Aware Modeling for Speech Separation. Proc. Interspeech 2021, 3061-3065, doi: 10.21437/Interspeech.2021-858

@inproceedings{wang21x_interspeech,
  author={Fan-Lin Wang and Yu-Huai Peng and Hung-Shin Lee and Hsin-Min Wang},
  title={{Dual-Path Filter Network: Speaker-Aware Modeling for Speech Separation}},
  year=2021,
  booktitle={Proc. Interspeech 2021},
  pages={3061--3065},
  doi={10.21437/Interspeech.2021-858}
}