iSEGAN: Improved Speech Enhancement Generative Adversarial Networks

Abstract

Popular neural network-based speech enhancement systems operate on the magnitude spectrogram and ignore the phase mismatch between the noisy and clean speech signals. Conditional generative adversarial networks (cGANs) show promise in addressing the phase mismatch problem by directly mapping the raw noisy speech waveform to the underlying clean speech signal. However, stabilizing and training cGAN systems is difficult and they still fall short of the performance achieved by the spectral enhancement approaches. This paper investigates whether different normalization strategies and one-sided label smoothing can further stabilize the cGAN-based speech enhancement model. In addition, we propose incorporating a Gammatone-based auditory filtering layer and a trainable pre-emphasis layer to further improve the performance of the cGAN framework. Simulation results show that the proposed approaches improve the speech enhancement performance of cGAN systems in addition to yielding improved stability and reduced computational effort.

A short report on improving SEGAN performance

References in corpus (5)

• Conditional Generative Adversarial Nets
• Instance Normalization: The Missing Ingredient for Fast Stylization
• Conditional Generative Adversarial Networks for Speech Enhancement and Noise-Robust Speaker Verification
• Investigating Generative Adversarial Networks based Speech Dereverberation for Robust Speech Recognition
• The Effectiveness of Instance Normalization: a Strong Baseline for Single Image Dehazing

Cited by in corpus (2)

• Frequency Gating: Improved Convolutional Neural Networks for Speech Enhancement in the Time-Frequency Domain
• Adversarial Joint Training with Self-Attention Mechanism for Robust End-to-End Speech Recognition