Sentioscope

Speech and voice controls, as each vendor documents them

SceneMixer: a library voice matched to the character

A short matching pass picks a preset voice for a character, a sample can be played before choosing, and once attached the preset is used as reference audio on every shot that character speaks in. The voice lives on the character; the production never handles the clip. From 3 source pages, read 2026-09-12 to 2026-10-13; no maker pays to be listed.

A per-project setting, and no per-character oneThis vendor keeps dialogue language at the project level, so the platform already holds settings that outlast a shot. A per-character voice would fit in the same place, and no document says that it does.Stated to persistNot addressedLanguageSet per project, fifteen namedWhether a voice follows a language changeVoice routesA preset library, or your recordingsHow a selection is storedPerformanceSpoken as the shot rendersWhether the same voice returns nextepisodeWhere it is writtenA compliance checklistAnywhere that explains the mechanismA documentation gap rather than an architectural one
Fig. 1 The sentence that fills the voice field comes from a likeness checklist, which was written to constrain rather than to explain.
SceneMixer on per-character binding, statement by statement. Read from the vendor's languages guide on 2026-09-12.
What the documentation settlesWhat it leaves to a take
A matched preset is used as reference audio on every shot the character speaks inHow many voices the library holds
Without a match, a voice written in words is what the model performs, and can vary slightly per shotHow closely a described voice holds across a long series
Dialogue is set per project to any of fifteen languages, with Cantonese a sixteenthWhether every preset speaks every one of those languages

Inclusion rule. Only statements found on the vendor page named in the caption, and only ones bearing on this one field. Nothing here is filled in from generated output. Order. In the order the statements arrive on the vendor's page.

1Two defaults, and the one that holds

The documentation is candid that the default arrangement is weaker. A voice written in words is performed as described, costs nothing and can land slightly differently from one shot to the next. Matching a preset is what makes a voice hold across an episode, and the guide says so in those terms rather than leaving the reader to infer it.

That is the useful kind of statement for this column: it names which of two routes persists and which does not, instead of describing both as consistent.

2Selection is documented as well as persistence

Most entries that keep a voice on a character say nothing about choosing it. Here every preset has a sample that can be played in the picker, matching and rematching are free, and a change takes effect on the next shot generated while finished segments keep the voice they were shot with.

A recording the user uploads is handled the same way as a matched preset and is not overwritten by later matching. The compliance guide still sets the boundary on whose voice may be uploaded.

3The others where a voice exists before any shot does

Four more entries let a voice outlast a call. Two others keep it on a character object; two create a voice object with an id, which a production has to attach itself.

  • Per-character binding
    A voice matched to a character from the built-in preset library is used as reference audio on every shot that character speaks ina matched preset kept on the characterSceneMixer, character voices guide / recorded 2026-10-13
  • Voice source
    The guide tells users to use a preset voice library or their own recordingstwo routes namedSceneMixer, compliance guide / recorded 2026-09-12
  • Languages
    Dialogue is set per project to any of 15 languages, with Cantonese a 16th for dialogue onlya named list of 15SceneMixer, languages guide / recorded 2026-09-12

4Sources

Read from the languages guide at scenemixer.com on 2026-09-12. The same column across every entry is on per-character binding; everything this vendor publishes about speech is on SceneMixer. What counts as documented is on how read.