Normalized Features for Improving the Generalization of DNN Based Speech Enhancement

09/07/2017
by   Robert Rehr, et al.
0

Enhancing noisy speech is an important task to restore its quality and to improve its intelligibility. In traditional non-machine-learning (ML) based approaches the parameters required for noise reduction are estimated blindly from the noisy observation while the actual filter functions are derived analytically based on statistical assumptions. Even though such approaches generalize well to many different acoustic conditions, the noise suppression capability in transient noises is low. To amend this shortcoming, machine-learning (ML) methods such as deep learning have been employed for speech enhancement. However, due to their data-driven nature, the generalization of ML based approaches to unknown noise types is still discussed. To improve the generalization of ML based algorithms and to enhance the noise suppression of non-ML based methods, we propose a combination of both approaches. For this, we employ the a priori signal-to-noise ratio (SNR) and the a posteriori SNR estimated as input features in a deep neural network (DNN) based enhancement scheme. We show that this approach allows ML based speech estimators to generalize quickly to unknown noise types even if only few noise conditions have been seen during training. Further, the proposed features outperform a competing approach where an estimate of the noise power spectral density is appended to the noisy spectra. Instrumental measures such as Perceptual Evaluation of Speech Quality (PESQ) and short-time objective intelligibility (STOI) indicate strong improvements in unseen conditions when the proposed features are used. Listening experiments confirm the improved generalization of our proposed combination.

READ FULL TEXT

page 1

page 2

page 3

page 4

research
09/07/2017

Improving the Generalizability of Deep Neural Network Based Speech Enhancement

Enhancing noisy speech is an important task to restore its quality and t...
research
04/07/2020

SNR-Based Features and Diverse Training Data for Robust DNN-Based Speech Enhancement

This paper analyzes the generalization of speech enhancement algorithms ...
research
07/22/2019

ML Estimation and CRBs for Reverberation, Speech and Noise PSDs in Rank-Deficient Noise-Field

Speech communication systems are prone to performance degradation in rev...
research
06/01/2018

Machines hear better when they have ears

Deep-neural-network (DNN) based noise suppression systems yield signific...
research
10/11/2017

PROSE: Perceptual Risk Optimization for Speech Enhancement

The goal in speech enhancement is to obtain an estimate of clean speech ...
research
02/10/2020

On Cross-Corpus Generalization of Deep Learning Based Speech Enhancement

In recent years, supervised approaches using deep neural networks (DNNs)...
research
09/12/2023

Assessing the Generalization Gap of Learning-Based Speech Enhancement Systems in Noisy and Reverberant Environments

The acoustic variability of noisy and reverberant speech mixtures is inf...

Please sign up or login with your details

Forgot password? Click here to reset