Inworld TTS-2 A/B — listen test

Same script across models. Listen in order:

1. TTS-1.5-Max (current), Damon, English narration.
2. TTS-2 STABLE, Damon, same narration.
3. TTS-2 BALANCED, Damon, same narration.
4. TTS-2 STABLE, Pär clone (v1), short quote.
5. TTS-2 STABLE, Damon, Swedish (language=sv-SE).
6. TTS-1.5-Max, Damon, same Swedish (no language tag) — contrast.

Raw model output: no lexicon SSML, no wobble, speakingRate 1.0, applyTextNormalization ON. Deciding whether to switch the platform default.