Skip to content

jyhan03/icassp22-dataset

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

19 Commits
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

Overview

This repository provides the dataset and its simulation code used in our ICASSP22 paper titled "DPCCN: Densely-Connected Pyramid Complex Convolutional Network for Robust Speech Separation And Extraction".

The simulation code of in-domain and source domain dataset can be found in Yi Luo and Libri2Mix, respectively.

Dataset Generation

For constructing the target domain dataset, Aishell2Mix, please refers to the scripts:

  • data_select.py, randomly select the clean speech from Aishell-1 to generate the mixed speech pairs.
  • data_simu.py, generate Aishell2Mix mixture using the data_select.py results.
  • run_simu.sh, a simple shell running script for data_simu.py.
  • flac2wav, transform the original LibriSpeech to .wav mode and 8k sample rate.
  • get_aux_aishell.py, generate enrollment speech for Aishell2Mix.
  • get_aux_libri.py, generate enrollment speech for Libri2Mix.

Simulated Results

The simulated results can be found in lst, where *_cfg.lst is the config file generated by data_simu.py and *_aux.lst is enrollment speech list generated by get_aux_aishell.py and get_aux_libri.py.

To obtain same experimental results, we recommend commenting the config file generation code in data_simu.py and use our provided *_cfg.lst directly.

Contact

Email: jyhan03@163.com

About

Dataset simulation for DPCCN.

Resources

License

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published

Languages