# Documentary Pronunciation Workflow for AI Narration

> Build and test a pronunciation sheet for names, places, languages, dates, abbreviations, and contested terms before visual production and final narration approval.

Updated: 2026-07-11
Audience: Documentary writers, voice directors, editors, and AI narration teams
Canonical: https://onira.studio/guides/documentary-pronunciation-workflow

## Key takeaways

- Extract difficult terms before recording or synthesis.
- Record source, dialect, phonetic note, stress, and approved audio.
- Approve pronunciation in full sentences, not isolated words only.

## Extract the pronunciation inventory

Scan the accepted script for personal names, place names, languages, titles, abbreviations, numerals, units, scientific terms, historical spellings, borrowed words, and terms whose pronunciation signals a disputed identity or period.

Add each term once with every script occurrence or timestamp. Mark risk by prominence, repetition, sensitivity, and whether an error would undermine factual trust.

- Written form and script location.
- Language, region, period, or dialect.
- Risk and review owner.

## Research a defensible spoken form

Prefer a relevant speaker, institution, archival recording, official naming authority, or specialist source where available. Dictionaries and phonetic notation can help, but a transcription should preserve the intended language and dialect rather than force every term into one accent.

Record the source URL or file, access date, phonetic transcription or plain-language cue, stressed syllable, acceptable variants, and the editorial choice. Preserve uncertainty when no single pronunciation is authoritative.

- Source and speaker context.
- IPA or readable phonetic cue.
- Chosen form and acceptable alternatives.

## Test the term inside performance

Generate or record the full sentence and neighboring lines. A correct isolated word can fail when speech synthesis changes stress, elides sounds, or rushes a transition. Review pacing, emphasis, emotional register, and consistency across repeated uses.

When a system cannot produce the approved form reliably, rewrite the sentence, use a tested phonetic spelling in the synthesis input, split the segment, or replace the voice path. Do not let a pronunciation workaround leak into visible captions.

- Full-sentence audio sample.
- Repeatability across occurrences.
- Caption text remains orthographically correct.

## Lock audio before final visual timing

Approve the pronunciation sheet and measured narration before final shot planning. Correcting a repeated name after scene timing is built can change pauses, subtitle timing, music, and clip lengths throughout the film.

Archive the approved samples with the script and final narration. Reuse the record for recurring people and places, but review it when the language, narrator, historical context, or channel style changes.

- Named audio approver.
- Measured accepted narration.
- Reusable series pronunciation library.

## Publication checklist

- Names, places, languages, numerals, and specialist terms are extracted.
- Every material term has a source and chosen spoken form.
- Dialect, period, and acceptable variation are documented.
- Terms are approved in full sentences and repeated contexts.
- Caption spelling remains correct after synthesis workarounds.
- Pronunciation is locked before final scene timing.

## Sources

- [International Phonetic Association chart](https://www.internationalphoneticassociation.org/IPAcharts/IPA_chart_orig/IPA_charts_E.html)
- [Library of Congress primary-source guide](https://www.loc.gov/programs/teachers/getting-started-with-primary-sources/)

## Questions

### Should a pronunciation sheet use IPA?

IPA can be precise for trained readers, but also include a source recording or plain-language cue that the production team and voice system can use consistently.

### Which accent should a documentary narrator use?

Choose an accent and register appropriate to the audience, subject, and series. Do not treat one accent as neutral, and preserve names and terms respectfully within the chosen performance.

### Why approve pronunciation before visuals?

Corrected pronunciation can change segment duration, pauses, subtitle timing, music cues, and shot lengths. Audio-first approval avoids expensive downstream rework.

## Product boundary

- Onira delivers a final MP4; it does not upload or schedule posts on YouTube or social platforms.
- Onira provides a reviewable production workflow; creators remain responsible for approving the story, facts, rights, disclosure, and final publication.
- Director chat is limited to regenerating one selected PREVIEW timeline video clip; other available Studio controls are separate direct actions.
- Creators must review facts, sources, rights, realistic-synthetic-media disclosure, and platform policy before publishing.
- Onira does not guarantee YouTube monetization, reach, factual accuracy, or legal clearance.