Multi-channel linear prediction-based speech dereverberation with low-rank power spectrogram approximation
Ante Jukić1, Nasser Mohammadiha1, Toon van Waterschoot2, Timo Gerkmann1, Simon Doclo1
1 University of Oldenburg, Dept. of Medical Physics and Acoustics, Oldenburg, Germany
2 KU Leuven, Dept. of Electrical Engineering Leuven, Belgium
RIRs
R_500: REVERB Challenge, 500 ms
R_700: REVERB Challenge, 700 ms
M_450: MARDY, 450 ms
O_750: Oldenburg, 750 ms
Parameters
Sampling frequency fs = 16 kHz
STFT: tight frame based on 64 ms Hamming window, 16 ms shift
MCLP: length of the filters Ls = 25, number of iterations itmax=5
NMF: IS divergence, rank R=40, multiplicative updates, itmax=100, ittol=10-2
M=2 microphones
M=4 microphones