An Automated End-to-End Open-Source Software for High-Quality Text-to-Speech Dataset Generation

TL;DR

The paper addresses the bottleneck of high-quality TTS data by introducing an end-to-end open-source pipeline that automates text-to-speech dataset generation. It combines language-specific phoneme distribution-aware sample selection, automated and human-in-the-loop quality assurance, automated recording processing, and preprocessing for training pipelines. The approach demonstrates high data-quality and efficiency across six languages, targeting $600000$ words per language and >$30$ hours of audio at $2.75$ words per second, with modular UI/backend and containerized deployment. The work reduces manual effort in dataset creation, enables scalable production of TTS data, and highlights practical considerations and limitations related to ASR-dependent QA and annotator reliability, while outlining concrete future improvements such as improved recording tools and broader language support.

Abstract

Data availability is crucial for advancing artificial intelligence applications, including voice-based technologies. As content creation, particularly in social media, experiences increasing demand, translation and text-to-speech (TTS) technologies have become essential tools. Notably, the performance of these TTS technologies is highly dependent on the quality of the training data, emphasizing the mutual dependence of data availability and technological progress. This paper introduces an end-to-end tool to generate high-quality datasets for text-to-speech (TTS) models to address this critical need for high-quality data. The contributions of this work are manifold and include: the integration of language-specific phoneme distribution into sample selection, automation of the recording process, automated and human-in-the-loop quality assurance of recordings, and processing of recordings to meet specified formats. The proposed application aims to streamline the dataset creation process for TTS models through these features, thereby facilitating advancements in voice-based technologies.

Figures (6)

• Figure 1: Workflow of the System
• Figure 2: System Architecture of the Tool
• Figure 3: Sentence Selection for TTS Recordings
• Figure 4: Screenshot of Admin Dashboard for Recording Upload
• Figure 5: Screenshot of Admin Dashboard for Insights of the annotations done.