Deepfake audio as a data augmentation technique for training automatic speech to text transcription models

TL;DR

This paper proposes using deepfake audio as a data augmentation method to improve speech-to-text models, especially for less-resourced languages, validated through experiments with Indian English datasets.

Contribution

Introduces a novel deepfake audio-based data augmentation framework for training robust speech-to-text models in low-resource language scenarios.

Findings

Augmented data improved speech-to-text model performance.

Framework effective with Indian English dataset.

Validated with existing deepfake and transcription models.

Abstract

To train transcriptor models that produce robust results, a large and diverse labeled dataset is required. Finding such data with the necessary characteristics is a challenging task, especially for languages less popular than English. Moreover, producing such data requires significant effort and often money. Therefore, a strategy to mitigate this problem is the use of data augmentation techniques. In this work, we propose a framework that approaches data augmentation based on deepfake audio. To validate the produced framework, experiments were conducted using existing deepfake and transcription models. A voice cloner and a dataset produced by Indians (in English) were selected, ensuring the presence of a single accent in the dataset. Subsequently, the augmented data was used to train speech to text models in various scenarios.