What tonal transcription actually means
In Yorùbá, tone is grammatically and semantically load-bearing. The same sequence of consonants and vowels can produce words with completely different — sometimes opposite — meanings depending on whether each syllable is spoken at high (́), mid (unmarked), or low ( ̀ ) pitch.
Transcribing Yorùbá without tones is not Yorùbá. It is approximated phonetics that a native speaker cannot use and that carries no cultural, liturgical, or scholarly validity. Yet this is exactly what every general-purpose AI transcription tool produces.
Minimal pairs — where the stakes are obvious
| Correctly marked | Meaning | Without marks | Problem |
|---|---|---|---|
| àṣẹ | Spiritual authority / “so be it” | ase | Undefined — not a Yorùbá word |
| Ọrúnmìlà | Divinity of wisdom and Ifá | Orunmila | Looks like a proper noun but tone is irretrievably lost |
| ọkọ | Husband | okọ / oko / ọko | Could mean husband, farm, vehicle, or arrow |
| igba vs igbá | 200 / calabash | igba | Two completely different cosmological objects in Ifá |
| Ègúngún | Ancestral masquerade | Egungun | Without marks, the sacred name becomes a romanized approximation |
In ceremonial contexts, these are not minor errors. They are the difference between a correctly transmitted text and a corrupted one.
How Ifa Scribe handles tonal accuracy
Yorùbá-optimized AI
The transcription engine is built on top of Gemini AI with Yorùbá-specific prompting and phonological context — not a generic multilingual model treating Yorùbá as a secondary language.
Àmi ohùn output by default
Output always includes tonal diacritics on every marked syllable — not stripped to ASCII. You receive proper Yorùbá orthography compatible with academic and liturgical standards.
Confidence flagging
Segments where the audio is unclear or the tonal interpretation is ambiguous are flagged for human review — rather than silently generating plausible-looking errors.
E-Orator verification loop
After transcription, paste the result into E-Orator to hear the AI read it back. Your ear can catch tonal errors that visual review misses.
Tonal transcription for different learner needs
- Students: Capturing class recordings with tonal marks means you study the real written form of the language — not placeholder ASCII. Your notes are immediately usable for vocabulary review.
- Practitioners: Liturgical texts must retain tonal accuracy to function correctly in practice. Ifa Scribe output is reviewed against instructor standard and flagged for uncertainty — not guessed.
- Scholars: Citable diacritical Yorùbá transcriptions from oral sources — with export to standard word processing formats used in academic publication.
The E-Orator connection
Ifa Scribe’s E-Orator (Pillar II) is the direct complement to tonal transcription. Where E-Scribe converts audio to marked text, E-Orator converts marked text back to tonal speech — creating a verification loop that is unique to this platform. No other transcription tool offers native Yorùbá speech synthesis as part of the same subscription.
See all three Ifa Scribe pillars →Access Ifa Scribe — $5/month
E-Scribe, E-Orator, and E-Companion are included together. Register through the Yorùbá Language Program.