Speech Dereverberation Using Non-Negative Convolutive Transfer Function and Spectro-Temporal Modeling

By Nasser Mohammadiha, Simon Doclo. 2016

33 citations recorded by the source corpus.

This paper presents two single-channel speech dereverberation methods to enhance the quality of speech signals that have been recorded in an enclosed space. For both methods, the room acoustics are modeled using a non-negative approximation of the convolutive transfer function (N-CTF), and to additionally exploit the spectral properties of the speech signal, such as the low-rank nature of the speech spectrogram, the speech spectrogram is modeled using non-negative matrix factorization (NMF). Two methods are described to combine the N-CTF and NMF models. In the first method, referred to as the integrated method, a cost function is constructed by directly integrating the speech NMF model into the N-CTF model, while in the second method, referred to as the weighted method, the N-CTF and NMF based cost functions are weighted and summed. Efficient update rules are derived to solve both optimization problems. In addition, an extension of the integrated method is presented, which exploits the temporal dependencies of the speech signal. Several experiments are performed on reverberant speech signals with and without background noise, where the integrated method yields a considerably higher speech quality than the baseline N-CTF method and a state-of-the-art spectral enhancement method. Moreover, the experimental results indicate that the weighted method can even lead to a better performance in terms of instrumental quality measures, but that the optimal weighting parameter depends on the room acoustics and the utilized NMF model. Modeling the temporal dependencies in the integrated method was found to be useful only for highly reverberant conditions.

Subfields: Biomedical Engineering, Computational Mechanics, Signal Processing

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

In Biomedical Engineering

1. Donald S. Williamson
2. Deliang Wang

In Computational Mechanics

1. Donald S. Williamson
2. Deliang Wang
3. Lei Xie
4. Sining Sun
5. Ke Tan
6. Yuxuan Wang
7. Zhong-Qiu Wang
8. Zhuo Chen

In Signal Processing

1. Donald S. Williamson
2. Deliang Wang
3. Lei Xie
4. Yujun Wang
5. Sining Sun
6. Fei Xiang
7. Junbo Zhang
8. Ke Wang

Citations in the corpus

• Time-Frequency Masking in the Complex Domain for Speech Dereverberation and Denoising (2017, neutral): “Another approach uses the RIR magnitude response and nonnegative matrix factorization (NMF) to remove reverberation [27].”
• Speech dereverberation and denoising using complex ratio masks (2017, methodology): “Other spectral-magnitude based approaches employ inverse filtering [7] or non-negative matrix factorization (NMF) [8].”
• Investigating Generative Adversarial Networks based Speech Dereverberation for Robust Speech Recognition (2018, background): “Single-microphone speech dereverberation has been intensively studied in the signal processing community and a variety of approaches have been proposed [5, 8, 9, 10, 11].”