ElevenLabs Dubbing v2 keeps emotion intact
ElevenLabs says Dubbing v2 carries over the original performance, not just the transcript, across 90+ languages. The pitch is sync-aware phrasing and delivery that sounds acted, not machine-translated, for creators, marketers, and production teams.
The meaningful change here is less about language coverage and more about moving dubbing from “translated audio” toward “performance transfer.” If it works on real content, that is a real step up for localization workflows that currently lose tone, pacing, and emotional shape.
- –Conditioning on the original performance should reduce the flat, over-literal feel that makes a lot of AI dubbing unusable
- –Sync-aware translation is a practical win because timing constraints matter as much as lexical accuracy in video and broadcast
- –ElevenLabs is packaging this across ElevenCreative and ElevenProductions, which splits self-serve creator use from managed studio workflows
- –The 90+ language claim is strong, but the hard test is whether quality holds on fast, emotional, multi-speaker content
- –The enterprise restrictions in the model terms suggest ElevenLabs is treating Dubbing v2 as a serious commercial audio product, not a casual feature
DISCOVERED
1h ago
2026-05-28
PUBLISHED
1h ago
2026-05-28
RELEVANCE
AUTHOR
ElevenLabs