Skip to content

Text-to-speech

Poneglyph generates placeholder audio from text using a TTS engine. Useful for previewing line lengths, blocking out cinematics, and giving designers something to listen to before VO is recorded.

  • Drafting a scene where you want to hear pacing.
  • Producing a non-final mix for a marketing test.
  • Any time recording a real actor would be premature.

File → Export → Export .wav generates TTS for the selected lines and languages. The job runs in the job center. The resulting .wav files land in your Wwise Originals folder by default; per-table overrides (the DivertTTS config) can route specific tables elsewhere.

Each character can have a different TTS voice; the mapping lives in the Characters table. Languages without a configured voice fall back to the default TTS voice for that language.

Translations rendered from TTS are flagged as such (RecordingState = TTS or similar). When real recordings come in via audio import, the flag flips to Recorded automatically.