SceneMixer: a library exists, and persistence does not
Two voice routes are named and neither is described as attaching to anybody. A library exists, recordings are accepted, and how a selection survives from one shot to the next is not addressed anywhere public. As of 2026-09-12.
| What the documentation settles | What it leaves to a take |
|---|---|
| Users are told to use a preset voice library or their own recordings | How a selection from that library is recorded against a character |
| Voice routes are described; persistence between shots is not | Whether a project holds one voice per character or one per shot |
| Dialogue is set per project to any of fifteen languages, with Cantonese a sixteenth | Whether a chosen voice follows a project that changes language |
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.
1A per-project language setting makes the blank stranger
This vendor is unusually specific about language, and language is set at the project level rather than per call. A platform that already keeps a per-project setting has somewhere obvious to keep a per-character voice, and the documentation does not say that it does.
So the blank reads as a documentation gap rather than an architectural one. That still matters for this register, because a production choosing between tools has only the pages, and the pages are what a comparison can be built from.
2The compliance framing cuts both ways
The sentence that fills the voice field here comes from a checklist about likeness, which means it was written to constrain rather than to explain. Checklists say what may be used; they do not usually say how a choice is stored.
Read that way, the absence is unsurprising and uninformative about the product. The cell records what is published and the page says plainly which document it came from, so a reader can weigh it accordingly.
3Others whose documentation never addresses persistence
Seven more entries reach this column empty. Most of them say much less about voices overall, which is what makes this one worth reading: two routes named, and no pairing described.
- Hedra — not documented by the vendor.
- Luma Ray — not documented by the vendor.
- sync-3 — not documented by the vendor.
- Synthesia — not documented by the vendor.
- Veo — not documented by the vendor.
- Vidu — not documented by the vendor.
- Wan2.2-S2V — not documented by the vendor.
- Per-character bindingNot documented by the vendor (as of 2026-09-12)no public description
- Voice sourceThe guide tells users to use a preset voice library or their own recordingstwo routes named
- LanguagesDialogue is set per project to any of 15 languages, with Cantonese a 16th for dialogue onlya named list of 15
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.