Highlights
This release collects the changes introduced from v5.4.0 through v5.4.4.
The main additions are:
- complete Fish Audio S2 Pro integration
- combined audio-reference and vector emotions for IndexTTS-2
- a faster, more visual Multiline TTS Tag Editor workflow
20260711_16fps_1.5x_cuts.mp4
Fish Audio S2 Pro
Fish Audio S2 Pro joins the suite with two particularly useful strengths:
- native multi-speaker dialogue generation
- free-form inline speech instructions for controlling delivery
It also integrates with the suite’s character switching, long-form text, and SRT workflows.
IndexTTS-2 emotion blending
IndexTTS-2 emotion sources are no longer mutually exclusive.
You can now combine:
- an audio emotion reference
- a direct emotion vector
- vector adjustments, named emotions, or text-derived vectors
The audio reference preserves the reference performance while additional vector controls can reshape or adjust its emotion inline.
This also integrates with the Multiline TTS Tag Editor, allowing emotion controls to be composed directly with character and audio-reference tags.
Faster Multiline TTS Tag Editor
The tag editor now supports faster direct editing of inline controls:
- click a character, language, reference, preset, parameter, or supported inline tag to swap it
- click and hold, move to an option, and release for faster gesture-based selection
- get engine-aware choices for IndexTTS-2, Higgs Audio v3, Step Audio EditX, CosyVoice3, and OmniVoice
- edit IndexTTS-2 vectors through the visual emotion radar
- create and manage emotion presets
- compose emotion controls directly with character and audio-reference tags
- switch individual characters inside compound tags
- wrap long inline tags within the editor
- keep quoted emotion descriptions directly editable