Overview of Methods for Data Augmentation for Speech-to-Text and Text-to-Speech

Abstract

In the field of machine learning and deep learning, data augmentation is a widely used technique to expand the amount of training data available. This involves altering existing data instances or generating new synthetic data, with the aim of enhancing the quantity and variability of the training set.It has shown to be especially useful when working with low resource languages and domains, where datasets are limited. This paper provides an overview of the data augmentation methods used for speech-related tasks, specifically for speech to-text and text-to-speech applications.The goal of this paper is to provide researchers and practitioners with a comprehensive understanding of the data augmentation methods available for speech-related tasks, their strengths and potential applications.