An Experimental Study on Speech Enhancement Based on Deep Neural Networks

TLDR

The study proposes a regression‑based speech enhancement framework employing a multi‑layer deep neural network architecture. The framework trains a deep neural network on a large simulated speech dataset to learn a nonlinear mapping from noisy to clean speech. Experiments show the DNN approach yields cleaner speech without musical artifacts, outperforms log‑MMSE in objective metrics, and is preferred by listeners in 76.35% of cases.

Abstract

This letter presents a regression-based speech enhancement framework using deep neural networks (DNNs) with a multiple-layer deep architecture. In the DNN learning process, a large training set ensures a powerful modeling capability to estimate the complicated nonlinear mapping from observed noisy speech to desired clean signals. Acoustic context was found to improve the continuity of speech to be separated from the background noises successfully without the annoying musical artifact commonly observed in conventional speech enhancement algorithms. A series of pilot experiments were conducted under multi-condition training with more than 100 hours of simulated speech data, resulting in a good generalization capability even in mismatched testing conditions. When compared with the logarithmic minimum mean square error approach, the proposed DNN-based algorithm tends to achieve significant improvements in terms of various objective quality measures. Furthermore, in a subjective preference evaluation with 10 listeners, 76.35% of the subjects were found to prefer DNN-based enhanced speech to that obtained with other conventional technique.