Blind Signal Dereverberation for Machine Speech Recognition

TL;DR

This paper introduces a method for dereverberating speech signals by leveraging long-window Fourier transforms and spectral normalization, improving speech recognition in reverberant environments.

Contribution

It proposes a novel dereverberation technique that uses training data and spectral normalization in the log spectral domain to enhance speech recognition accuracy.

Findings

Effective dereverberation in reverberant environments

Improved speech recognition performance

Utilizes spectral normalization for noise reduction

Abstract

We present a method to remove unknown convolutive noise introduced to speech by reverberations of recording environments, utilizing some amount of training speech data from the reverberant environment, and any available non-reverberant speech data. Using Fourier transform computed over long temporal windows, which ideally cover the entire room impulse response, we convert room induced convolution to additions in the log spectral domain. Next, we compute a spectral normalization vector from statistics gathered over reverberated as well as over clean speech in the log spectral domain. During operation, this normalization vectors are used to alleviate reverberations from complex speech spectra recorded under the same reverberant conditions . Such dereverberated complex speech spectra are used to compute complex FDLP-spectrograms for use in automatic speech recognition.