All three columns filled, against the longest list and a blank

SceneMixer is the only vendor in this register with something published in all three columns. Synthesia publishes the longest named language list here and nothing at all about the language of the documents its product generates. As of 2026-09-22.

Inclusion rule. Two vendors chosen because one fills every column and the other fills the longest single cell. The last row records each vendor's largest figure. Order. Register columns in fixed order, then the headline figures.

1Breadth and depth are different achievements

One vendor answers all three questions with sets of fifteen. The other answers one of them with a hundred and forty-three named rows and leaves the middle column blank. Neither is a better disclosure; they are answers to different readers.

A studio that needs its working documents in Japanese cannot use the longer list at all, because it says nothing about documents. A company dubbing an existing catalogue into Zulu cannot use the shorter one, because Zulu is not in it.

2The empty middle cell is the whole reason the column exists

A vendor naming 143 languages for translation says nothing about the language of a generated script. That is not an oversight in the disclosure; it reflects a product that translates finished material rather than generating production documents.

Recording it as empty rather than inferring from the voice list is the discipline this register is built on. Seventeen of the rows here are empty in that column, and the one that is not is the comparison this page exists to make.

3Two figures that cannot be added

Fifteen and 143 are in different units. One is a set covering interface, documents and dialogue together; the other counts languages and dialects in one table for a single translation feature.

Any table that put both in a column headed languages supported would be producing a figure neither vendor published. The register prints the unit beside the number instead, which is less tidy and is the only way the two rows can sit next to each other honestly.

4Sources

Read from the SceneMixer languages guide and the Synthesia supported languages table. Both rows are kept in the register under one rule, set out on how counted; the column is described on document language.

5Both of them against the rest of the column

Inclusion rule. Tools with a public English site selling AI video generation or AI video dubbing. Site languages count hreflang alternates served from the host being read; x-default and alternates pointing at another host are not counted, for every vendor alike. Order. Alphabetical by tool name.

Other pairings that turn on the same kind of difference: Rask AI and Vidnoz, D-ID and Rask AI. The full set of pairings is here; single cells are indexed at entries.

• Output languages
  Script analysis, episode outlines, storyboard scripts and character descriptions follow the interface language, across the same 15tied to the interface settingSceneMixer, languages guide / recorded 2026-09-12
• Voice languages
  Spoken dialogue is set per project to any of the 15, with Cantonese a 16th option for dialogue onlyset per projectSceneMixer, languages guide / recorded 2026-09-12
• Voice languages
  A table headed “All languages and dialects supported for Synthesia translation feature”, running 143 named rows from Afrikaans to Zulunamed list publishedSynthesia Docs, supported languages / recorded 2026-09-22
• Output languages
  No public list (as of 2026-09-12)no list publishedsynthesia.io / recorded 2026-09-22