Four voices at the altars.
The four selected voices now speak all 172 god lines, with Signature Presence baked into every recording.
The auditions have become the cast. Shaun chose Velvet temptation for Vessara, Roots beneath stone for Ilveth, Crystal intellect for Veyra and The fractured host for Vorr, all with Signature Presence. This separate Mac build carries those choices beyond their opening lines.
The four voices
Each god now has 43 recorded lines. The four approved greetings are reused from their original dry masters. Another 168 lines were generated using voice references extracted from those approved synthetic performances, then checked and treated locally.
- Vessara — Velvet temptation: intimate, coaxing and seductive.
- Ilveth — Roots beneath stone: slow, patient and ancient, with pitch-preserving pacing.
- Veyra — Crystal intellect: airy, precise and otherworldly.
- Vorr — The fractured host: unstable, broken and theatrical.
Presence without a live generator
Signature Presence is baked into stereo recordings using the Voice Chamber’s treatment graph and preset values: a shaped lead, delayed room, quiet detuned doubles and filtered echoes. The offline renderer uses the same settings; sample-for-sample equality with browser audio engines is not claimed.
The game plays ordinary recorded assets. There are no generation requests during play and no added live effect graph. Existing voice volume controls still apply. The remaining 87 narrator and wizard recordings are preserved.
Keep the words intact
Every final dry god recording passed a normalized transcript comparison against its exact script. Missing word endings, repeated phrases and awkward joins prompted replacement takes; the original results and rejected takes remain in the local evidence. The final 172 stereo masters also pass format, duration, headroom and effect-tail checks.
The separate Shipping app passes 435 dialogue assertions, 1,300 campaign rules checks and 88 native-input checks. Dialogue tests cover all four gods, natural completion, skipping, choices and wizard replies. Current main was incorporated before building, including the hooded guide and recent movement repairs.
Listen in context
The selected auditions are approved; the full campaign performances still need a human listening pass. Voice-reference synthesis helps retain identity across lines, but acting and emotional continuity are judgments to make in the story. This release supplies the complete cast for that review, with previous playtests retained.
This is an audio iteration. It does not establish a frame-rate improvement or a Windows release.