Emphasis¶
FaceCue measures the character of the voice, works out which words and phrases carry the weight, and carries that continuously through the head and the brows, so the performance moves with the delivery instead of sitting still underneath it.
A mouth on its own reads as a mouth on its own. This is most of what makes it read as a person talking.
What It Drives¶
The brows lift on the moments being leant on, and settle back between them.
The head moves with the voice. It dips into an emphasised syllable and comes back out, drifts in the pauses, and settles the way a speaker's does across a phrase.
Neither is a gesture placed on a beat. Both are continuous responses to the voice, which is why they read as the character meaning something and not as animation triggered on a keyword.
It Runs Continuously¶
This is worth being explicit about, because it is the difference between FaceCue's emphasis and a gesture system.
Nothing here waits for an event and then plays a clip. The motion is a proportional response to the voice as it goes, so it is always the right size for what the voice is doing, and it degrades gracefully. A flat read produces almost no motion, which is correct. An animated read produces a lot.
Where the Reading Comes From¶
Two things at once, and neither on its own would be enough.
How the line was delivered. Loudness, duration and pitch, each measured as contrast against the words around it, never as an absolute. A word is emphasised because it stands out from its neighbours, not because it crossed a threshold.
How the line is built. Which syllables carry stress in the first place, from the structure of the words themselves.
Take the acoustic half alone and a shout is all emphasis. Take the structural half alone and every sentence is emphasised identically regardless of how it was actually said. Together they find the syllable that was both stressed and delivered as though it mattered.
Why It Is More Precise on a Baked Line
Every drive path includes emphasis, but the baked path places it best, because it knows where the words and syllables are. A live path has the sound but not the structure, so it reads the delivery without the sentence.
That is not a reason to avoid live, it is a reason a baked line looks more deliberate.
Questions and Phrase Endings¶
A question does not end like a statement, and the face knows it.
FaceCue reads the shape of a phrase's ending and poses accordingly: the small lift at the end of a question, the settle at the end of a statement. It commits only when the reading is clear enough, because a face that guesses wrongly at the end of every sentence is worse than one that stays neutral.
Tuning It¶
Three things, and they are the ones worth touching.
How much the brows raise. The most visible, and the easiest to overdo. A brow that moves on every accent reads as surprise, not emphasis.
How far the head moves. Small numbers go a long way here, because the head is large and the eye is sensitive to it.
How quickly it settles. How long the face takes to come back after a beat, which is most of what distinguishes an animated speaker from an agitated one.
What It Does Not Do¶
Worth knowing before you tune.
It is not a gesture library. There are no authored motions, so you cannot ask for a particular head movement on a particular word. What you can do is change how strongly the character responds.
It does not read meaning. It measures how something was said and how it was built, not what it means. Sarcasm delivered flatly produces a flat performance, correctly.
It does not replace emotion. Emphasis is about weight and rhythm. What the character feels is a separate layer, authored instead of measured. See Emotion.