// AI news · 2026-10-23 · Amazon Science
Audible introduces controllable TTS for audiobooks
Audible's controllable TTS system uses natural language descriptions to generate distinct character voices, achieving 97% accuracy in English for character dialogue identification.
Published 2026-10-23 23:20 (Beijing time)Source Amazon ScienceAuto-summarised, fact-checked against source
01Summary
It improves dialogue differentiation by 3.7 points on MUSHRA test, with statistically significant listening gains across eight out of ten voices in both languages. The system leverages LLMs to generate expressive speech annotations and maintains consistent voice profiles throughout books.
Summarised from the original article by an open-source model, then checked automatically: every number and English name must appear in the source. The original article is authoritative.
02Key points
- Achieves 97% character dialogue accuracy in English using natural language descriptions.
- Improves dialogue differentiation by 3.7 points on MUSHRA test across both languages.
- Maintains consistent voice profiles throughout audiobook chapters.
- Leverages LLMs to generate expressive, context-aware speech annotations.
03Source
Better speech synthesis for audiobooks
Amazon Science · published 2026-10-23 23:20 Beijing time
05Related news
From AIMap AI news: collected from 1,100+ official and media sources. Summaries only restate the original article; follow the link for the full story. All AI news · Today