Yorùbá speech corpus text with 24,905 verified sentences released
The ÌròyìnSpeech Text Corpus: 24,905 Curated Yorùbá Sentences for Speech and Language Technology
Computation and Language
Summary
It is hard to build speech and language tools for Yorùbá because there are not many accurate texts and recordings. The authors created a collection of 24,905 Yorùbá sentences that are carefully checked for correct tones and easy to speak aloud. These sentences come from news and original writing, covering more topics than past collections. They also fixed text encoding problems that cause errors in Yorùbá letters. This collection can help build technologies like text-to-speech and language analysis for Yorùbá.
What this means in practice
- •For speech technology developers: Use the verified Yorùbá sentences as prompts and references to build and improve text-to-speech front-ends and pronunciation models.
- •For language technology companies: Create products that restore correct Yorùbá diacritics and convert written text to accurate speech using this curated text component.$Commercial implications: Enables commercial Yorùbá TTS and diacritic restoration tools needed for language technologies targeting Yoruba speakers.
Authors
Kola Tubosun, Aanuoluwapo Aremu, Tolulope Ogunremi, Iroro Orife, David Ifeoluwa Adelani
Abstract
ÌròyìnSpeech is a 42-hour, 80-speaker Yorùbá read-speech corpus whose audio has been distributed by ELRA since 2024. This paper describes the release of its text component: 24,905 unique, hand-verified, tone-marked Yorùbá sentences (275,897 tokens; 15,687 types), curated in 2022 as recording prompts. Roughly 11,000 sentences were adapted from openly licensed news material; the remainder were written in-house to broaden coverage beyond the religious translation that dominates existing Yorùbá corpora. Every sentence was checked by hand for tone-mark accuracy, edited for read-aloud clarity and a neutral register, and localised so that non-Yorùbá personal and place names appear in Yorùbá form. Preparing the text for release surfaced systematic Unicode normalisation failures affecting more than 60% of lines (with precomposed and decomposed forms of the same letter co-occurring within single sentences) which we document and correct. The corpus supports diacritic restoration, grapheme-to-phoneme conversion, TTS front-end development and orthographic research, and serves as a validated prompt set for new recording.