dots.tts.edit

A unified model for precise text and acoustic editing of speech.

dots.tts.edit follows natural, localized edit instructions while preserving the speaker and the acoustic context outside the edited regions. The same model supports text, emotion, prosody, pause, and compositional speech editing.

Contents

• Overview
• Evaluation
• Text Editing
• Emotion Editing
• Prosody Editing
• Pause Editing
• Compositional Editing