ISCA Archive Interspeech 2023
ISCA Archive Interspeech 2023

SCP-GAN: Self-Correcting Discriminator Optimization for Training Consistency Preserving Metric GAN on Speech Enhancement Tasks

Vasily Zadorozhnyy, Qiang Ye, Kazuhito Koishida

In recent years, Generative Adversarial Networks (GANs) have produced significantly improved speech enhancement (SE) task results. However, they are challenging to train. In this work, we introduce several improvements to GAN training schemes, which can be applied to most GAN-based SE models. We propose using consistency loss functions, which target the inconsistency in time and time-frequency domains caused by Fourier and Inverse Fourier Transforms. We also present self-correcting optimization for training a GAN discriminator on SE tasks which helps avoid "harmful" training directions for parts of the discriminator loss function. We have tested our proposed methods on several state-of-the-art GAN-based SE models and obtained consistent improvements, including new state-of-the-art results for the Voice Bank+DEMAND dataset.