Product

FaceCue for Unity Turns Dialogue Into Facial Animation: What Is Claimed and What Has Not Been Tested

ยท RenderBob team

Digital Production covered FaceCue on 17 November, a Unity tool that drives a character's face from dialogue audio. The publication says it did not test the claims.

A stylized clay fox bust sits in a facial-animation studio. An audio waveform feeds three pose frames of the same fox, with small mouth and eyebrow changes, while rig curves trace the bust and an empty validation monitor waits beside calibration markers.

Digital Production covered FaceCue on 17 November, a Unity tool by Christos Nasioutzikis of Firecrest Roost that drives a character's face from dialogue audio. The publication says the workflows are developer-claimed and that it did not test rig compatibility, language quality or runtime cost.

What it claims to do

  • Inputs. Dialogue audio, plus optional transcript text, drives lip sync, eye movement, expression, head motion and breathing.
  • Lip sync. Phoneme-based, with co-articulation. Ten languages are individually tuned, including German, and 50 or more are supported. A transcript can switch language mid-line, and a less precise universal audio route exists for other languages.
  • Rig mapping. A setup wizard recognises MetaHuman, Character Creator, Daz and VRoid rigs and builds visemes from FACS shapes. A Face Rig Designer adjusts the mapping, and a Clip Composer previews baked lines without entering Play Mode.
  • Runtime. Baked cues drive the rig at runtime instead of storing keyframes.

Practical details

  • Price. 129 euros, currently discounted to 65 euros, with regional prices and licence terms on the Unity Asset Store.
  • Versions. Unity 2021.3 through Unity 6.6.
  • Platforms. Baking runs in the Windows editor. Live visemes are unavailable on Linux and WebGL, and Apple-silicon Macs use a fallback.
  • Not stated. Whether it needs an internet connection.
  • Transcript caveat. Transcript text improves alignment but is not a complete lip-sync input on its own.

Baked cues, and an untested result

A tool that bakes cues, rather than processing the face live, can fit a normal build. Whether the result looks good is the question Digital Production does not answer.

A short test plan

  • Use one line of your own dialogue in your main language.
  • Try your real character rig, not the sample.
  • Check non-English lines if you ship in more than one language.
  • Test on your target platform, given the platform limits above.
  • Judge with the sound on, since timing is what viewers notice.

What to watch

Independent hands-on reviews, and clarity on offline use.

RenderBob 3 and this story

RenderBob 3 keeps references and prompt logs with the studio, so a character's source material and the generated shots built from it stay together as a project grows.

More from the blog

All posts