# Voice — Process
**Role in the pipeline:** one credited human actor per character, recorded once as a voice bible, then used for authored monologues and (under contract) as the synthetic voice for live answers.
## Steps
1. **Casting** with the historian and dialect coach (period, region, age; Quaker plain speech where relevant).
2. **Contract**: face/voice use on every surface, synthetic voice creation, future retraining, credit, residuals or buyout, and a clause on who may change the voice later.
3. **Voice bible** (30–60 min): calm exposition, sharp, grief, humor, public address, period vocabulary list, names and places pronounced correctly.
4. **Monologues** recorded with the director, synced to motion capture.
5. **Synthetic voice** built from the actor's recordings on a low-latency TTS platform; locked version; watermark if the platform supports it.
6. **QA**: pronunciation list (Nantucket names), loudness, caption accuracy.
## Deliverables
Authored monologue audio · locked TTS voice ID · pronunciation lexicon · caption files · credits text
**Owners:** voice director, actor, sound designer. Tutorials: [[Voice — Tutorials]]