A Grapheme-to-Audio Mapping and Phonological Rule System for Yorùbá Concatenative Speech Synthesis

This document specifies the grapheme-to-audio mapping rules and phonological environment conditions for the TTSYoruba concatenative speech synthesis system, deployed at TTSYoruba.com and integrated into the YorubaName.com multimedia name dictionary. The system operates on a diphone inventory of five file types (_l, _m, _h, _f, _r), corresponding to low, mid, high, falling, and rising tonal realisations of each Yorùbá syllable. The specification covers: (1) the full consonant diphone inventory for all Yorùbá consonants across seven oral vowels and their nasal equivalents, including open vowels ẹ and ọ; (2) the phonological environments that condition contextual tone assignment, distinguishing lexical from contextual rising and falling contours; (3) nasal disambiguation rules for the three functional environments of the grapheme -n- (oral, nasal, and syllabic); (4) word-initial vowel inventory; and (5) extended mappings for contour tone diacritics: i.e. the caron (ˇ) for rising contour and circumflex (ˆ) for falling contour, as proposed in Túbọ̀sún (forthcoming). The specification was originally compiled in 2017 and substantially extended in 2026. It is deposited here to establish prior art and to support reproducibility of the phonological rule architecture described in the companion paper: Túbọ̀sún, K. et al. (forthcoming), A Situational Speech Synthesizer for Yorùbá.