BibleTTS: a large, high-fidelity, multilingual, and uniquely African speech corpus

TL;DR

BibleTTS provides a large, high-quality, multilingual speech dataset for ten African languages, enabling the development of advanced text-to-speech models with studio-quality recordings.

Contribution

This work introduces a new, open, high-fidelity speech corpus for ten African languages, with aligned recordings and baseline TTS results, filling a gap in multilingual speech resources.

Findings

High-quality, aligned speech data for ten African languages.

Baseline TTS models demonstrate the dataset's utility.

Open license facilitates commercial and research use.

Abstract

BibleTTS is a large, high-quality, open speech dataset for ten languages spoken in Sub-Saharan Africa. The corpus contains up to 86 hours of aligned, studio quality 48kHz single speaker recordings per language, enabling the development of high-quality text-to-speech models. The ten languages represented are: Akuapem Twi, Asante Twi, Chichewa, Ewe, Hausa, Kikuyu, Lingala, Luganda, Luo, and Yoruba. This corpus is a derivative work of Bible recordings made and released by the Open.Bible project from Biblica. We have aligned, cleaned, and filtered the original recordings, and additionally hand-checked a subset of the alignments for each language. We present results for text-to-speech models with Coqui TTS. The data is released under a commercial-friendly CC-BY-SA license.