Visemes and Expression: How a Face Comes Alive

What visemes are

Visemes are mouth shapes that correspond to the sounds of speech. Putting them in the right order lets the avatar "speak" the words with its lips.

Syncing with the voice

Timings are extracted from the spoken audio, and the lip movement matches the voice exactly. A mismatch instantly gives away a fake, so synchronization matters.

Emotions

Beyond the lips, the face conveys the emotion of the answer — warm, sad, thoughtful. This makes the interaction feel alive rather than mechanical.

Why it matters

The brain is sensitive to the slightest mismatch between lips and sound — it immediately reads as "not real." Precise synchronization and lively expression remove that effect. That's why the image comes across as warm rather than eerie.

• Visemes — mouth shapes for the sounds of speech.
• Timings from the voice → synchronized lips.
• Expression conveys the emotion of the answer.
• Synchronization removes the "fake" effect.