/Adaptive-Dereverberation-Algorithm

Weighted RLS based adaptive dereverberation algorithm

Primary LanguageMATLABMIT LicenseMIT

Adaptive-Dereverberation-Algorithm

MATLAB implemention of Weighted RLS based adaptive dereverberation algorithm.

Layout

 ./
 +-- lib/
 |   +-- +util/                       utility functions
 |   |   |-- stftanalysis.m
 |   |   |-- stftsynthesis.m
 |   |-- demo_fdndlp.m
 |   |-- fdndlp.m
 |   |-- config.m
 +-- wav_sample/                      reverberant audio samples
 |   |-- sample_1ch.wav
 |   |-- sample_4ch.wav
 |   |-- sample_8ch.wav
 +-- wav_out                          dereverberant autio samples
 |   |--  drv_sample_1ch.wav
 |   |--  drv_sample_4ch.wav
 |   |--  drv_sample_8ch.wav
 |-- rls_ada.m                        Weighted RLS based adaptive dereverberation algorithm
 |-- config.m                         Configurations
 |-- demo_rls_ada.m                   Demo for WRLS-ADA
 |-- README.md

Run the Demo

  • Just run the script file demo_rls_ada.m and the audio sample in wav_sample will be used.
  • To use your own data, change the filepath and sample_name in demo_rls_ada.m.
  • The configrations are gathered in config.m. Be careful to change the settings.

Reference

[1] Yoshioka T, Tachibana H, Nakatani T, et al. Adaptive dereverberation of speech signals with speaker-position change detection[C]. 2009 IEEE International Conference on Acoustics, Speech and Signal Processing. IEEE, 2009: 3733-3736.