Fish Audio releases Drama 3 preview, supporting localized regeneration of voice content
BIBIBI
AT A GLANCE
Fish Audio releases Drama 3 preview, supporting mid-sentence voice changes, multi-character dialogue, and word-level regeneration.
Article
Fish Audio releases Drama 3: turning TTS into editable audio—if word is wrong, regenerate only that word
Beating Insights AI Breaking News:AI voice company Fish Audio has released the Drama 3 preview of its voice model, featuring more granular voice control. Users can directly describe tone, pacing, and characters in natural language, without having to combine a host of emotion or audio tags themselves.
Drama 3 can change voices and speaking styles mid-sentence, and can also generate dialogue for multiple characters at once. If any part is unsatisfactory, users can regenerate just word rather than the entire segment.
Fish Audio's previous-generation S2 already supported natural-language voice control. Drama 3 further refines this control by incorporating mid-sentence voice changes, multi-character generation, and localized regeneration into a single model. This will make post-production revisions much more convenient for game dialogue, spoken-word content, and AI characters.
Drama 3 is currently still in the Preview stage, and developers can apply for API testing access.
Original link https://m.theblockbeats.info/flash/368821
Key points
01
Fish Audio releases the Drama 3 preview of its voice model.
02
Drama 3 Supports using natural language to describe tone, pacing, and characters.
03
Drama 3 Can change voices and speaking styles mid-sentence.
04
Drama 3 Supports generating dialogue for multiple characters at once.
05
Users can regenerate just word without having to regenerate the entire segment.
06
Drama 3 It is currently in the Preview stage, and developers can apply for API testing access.
AI-assisted interpretation
The following is analysis, separate from reported facts. Verify important claims independently.
Drama 3 is a voice-generation model. Users can use natural language to specify the desired tone and switch voices within the same sentence; if they are dissatisfied with only word, they can modify just that word.
Why it matters to readers
Mid-sentence voice changes, multi-character generation, and word-level regeneration may make post-production revisions more convenient for game dialogue, spoken-word content, and AI characters.
It demonstrates the direction of voice-generation tools evolving from regenerating entire segments toward more granular editing.
Risks and unknowns
It is currently still a Preview version, and no formal launch date has been specified.
API The scope of eligibility for testing was not specified.
The report did not specify the actual reliability of partial regeneration, mid-sentence voice switching, and multi-speaker generation.
The report did not provide pricing information.
Related Developments
Loading event timeline…
Related concepts
TTS
This term is not in the glossary yet. Browse related concepts in the glossary.
