Generative adversarial networks for speech processing: A review

Generative adversarial networks (GANs) have seen remarkable progress in recent years. They are used as generative models for all kinds of data such as text, images, audio, music, videos, and animations. This paper presents a comprehensive review of the novel and emerging GAN-based speech frameworks and algorithms that have revolutionized speech processing. We have categorized speech GANs based on application areas: speech synthesis, speech enhancement & conversion, and data augmentation in automatic speech recognition and emotion speech recognition systems. This review also includes a summary of the data sets and evaluation metrics commonly used in speech GANs. We also suggest some interesting research directions for future work and highlight the issues faced by current state-of-the-art speech GANs.