Non-negative source-filter dynamical system for speech enhancement

TR Image
Illustration of the proposed model. The power spectrum S is decomposed as a product of a filter part V^r, an excitation part V^e, and gains g. The smooth overlapping filter dictionary W^r implicitly restricts V^r to capture the smooth envelope of the spectrum. W^e captures the spectral shapes of the excitation modes.

Model-based speech enhancement methods, which rely on separately modeling the speech and the noise, have been shown to be powerful in many different problem settings. When the structure of the noise can be arbitrary, which is often the case in practice, model- based methods have to focus on developing good speech models, whose quality will be key to their performance. In this study, we propose a novel probabilistic model for speech enhancement which precisely models the speech by taking into account the underlying speech production process as well as its dynamics. The proposed model follows a source-filter approach where the excitation and filter parts are modeled as non-negative dynamical systems. We present convergence-guaranteed update rules for each latent factor. In order to assess performance, we evaluate our model on a challenging speech enhancement task where the speech is observed under non-stationary noises recorded in a car. We show that our model outperforms state-of-the-art methods in terms of objective measures.