Sentioscope

Speech and voice controls, as each vendor documents them

Entries grouped by the shape of their answer

A route page gathers the entries that answer one column the same way and cross-cuts them against a second. Membership is decided by the vendor wording, so an entry sits on several routes and on no two for the same reason. As of 2026-09-12.

Grouping the register by the shape of the answerA field page reads one column down all sixteen entries. A route page collects the entries that answer a column the same way, so the architecture they share becomes visible and the places they diverge become obvious.Four ways into the same readingThe tableFive columns, sixteen rows, every cell carrying its own readingdate.A field pageOne column read downwards, with the grades of answer itcontains.A route pageThe entries that answer one column alike, cross-cut against asecond.A cell pageOne entry on one column, and what the wording leaves open.Narrowing from the whole register to one statement
Fig. 1 Membership is decided by the vendor's own wording, which is why an entry can sit on several routes and on no two for the same reason.

Reading one column downwards shows the grades of answer a question receives. Reading the entries that share a grade shows what the architecture behind that answer costs, which is the part a schedule is built from. A product handed a recording and a product that invents one are doing different work regardless of how similar the output looks.

Each table here deliberately spans two columns rather than one. The subset of a single column already sits on that field note, and copying it would add a page without adding a sentence. Pairing the grouping column with a second one shows where entries that looked alike stop being alike.

The routes are not exclusive and are not meant to be. An entry making sound in the same pass can also be one that publishes a figure in writing and one that never mentions a mouth, and reading it on all three is how the shape of its documentation becomes visible.

1The routes

  • Sound in the same pass — Eight entries generate audio in the same pass as the picture, which means alignment arrives for free and the words are the part that resists direction.
  • Handed a recording — Two entries take audio as an input and build the picture around it, which puts casting, a session and a delivered file in front of the first frame.
  • Speech from a script — Four entries turn a script into speech as a step of its own, which makes the audio an artefact that can be fetched, approved and reused before rendering.
  • Nothing published — Two entries leave the audio source empty, one because its reference is about pictures and one because its reference is about voices.
  • A voice that is stored — Five entries let a voice be created, named and reused, so a character's voice is wrong once if it is wrong at all rather than on every generation.
  • A voice per call — Five entries establish the voice on the request that uses it, by a clip, an id or a label, and publish nothing that holds the pairing afterwards.
  • Languages named — Two entries answer the language question with names rather than a number: a claim a production can check, instead of one it has to believe.
  • Languages counted — Three entries publish a total for spoken languages, from five to ninety-five or more, and none of them says which languages the total covers.
  • Breadth claimed — Two entries point at language support and give neither a list nor a total, which is the weakest grade this register records as an answer.
  • A figure in writing — Seven entries put a figure in writing: seconds of reference audio, minutes of enrolment, megabytes per file or the length of a single generation.
  • A named driver — Four entries say what the mouth is following rather than that lip-sync exists, which is the difference between planning coverage and guessing at it.
  • Named, not explained — Four entries put lip-sync on the page without saying what moves the mouth, which leaves a reader unable to predict which of three failures to expect.
  • The subject skipped — Seven entries reach the lip-sync field empty, and the pattern in the blanks says more about what API references are for than about the products.

A route is added when enough entries share a grade of answer for the comparison between them to say something a single row cannot.

Other notes: Speech controls, Models, Fields, Side by side, Questions, Terms, Learn, Data. What counts as a documented control is set out on the reading page.

2Voices in a drama pipeline

For the speech side of “Routes”, SceneMixer's guides describe how lines are spoken:

A face becomes specific through things that can be drawn: the shape of the jaw and brow, the set of the eyes, the nose bridge, the lips, skin tone, hair line and cut, age as it shows on the face, and one unmistakable anchor.

SceneMixer, How to Describe a Character’s Face So the Model Draws This Person and Not a Generic Pretty One (read 2026-10-13)

A SceneMixer segment comes out with its sound already in it: each line spoken in that character’s voice, the speaker’s lips moving with the words, and the ambience of the place — rain, traffic, a clock.

SceneMixer, Dialogue and Sound in One Generation (read 2026-10-13)