YouTube auto dubbing languages: which get expressive speech

YouTube's February 4, 2026 post says auto dubbing covers 27 languages and that Expressive Speech is available in eight: English, French, German, Hindi, Indonesian, Italian, Portuguese and Spanish. The help page marks expressive languages with an asterisk and is the place to check the current list, since YouTube may have added languages since.

Which eight languages get Expressive Speech?

These are the languages YouTube's blog names. YouTube's help page says expressive speech replicates the original audio's pitch and intonation to produce more natural-sounding dubs.

Why do the language counts differ between pages?

The blog post is dated February 4, 2026, and the help page is edited over time, so the two can give different counts. The help page also separates source languages from target languages, with English described as the most widely available target. When you need to know whether a specific pair works, use the help page and test one upload rather than trusting a count from a news post.

How do I find out if my language pair is supported?

Open YouTube's automatic dubbing help page and look for your source language first, then for the target language you want. An asterisk on a language means expressive speech is supported for it. Then upload one short video, not your whole catalog, and inspect the result before deciding what to do with the rest. The help page also lists reasons a video is not dubbed at all, such as no speech, only music, copyrighted content or an undetectable source language.

Keep in mind that the asterisk is the only expressive marker this post relies on; for any language without one, read the help page's wording rather than assuming.

Is lip sync one of the expressive languages?

No, it is a separate feature. The blog describes Lip Sync as a pilot only, and the help page says it alters the speaker's lip movement to align with the dubbed audio. See YouTube lip sync dubbing for what that means for your own videos.

What if I dub the audio myself?

Sume does not publish a language list for its text to speech either. Its schema has a language field that takes a BCP-47 or ISO-639 code, and an optional emotion string in generation_config; neither promises YouTube-style pitch matching. For a language you need, render a test sentence and listen. See multilingual text to speech API.