VOICES & NARRATION

A trigger volume starts the conversation. Charlotte's line plays from an attached component, and her OnAudioFinished event hands off to the main character's response, spawned on the character's mesh through the same class override node that anchors the player's voice to the center channel on surround configurations. Two linked response lines run inside one MetaSound, with the second wave player started from a trigger delay so the timing between them is authored, not edited into the file.

While the response plays, the character's head drives the sound. The trigger enables its tick only for the length of the line: each frame, a function library call takes the head socket's facing direction and dots it against the direction to the camera, and the result feeds a Facing parameter on the MetaSound. When the last line finishes, OnAudioFinished disables the tick again, so the cost exists only while someone is speaking.

Inside the graph, Facing maps to the cutoff of a one pole low pass, so the voice is open when the character faces the listener and darkens as the head turns away. It is the same directivity idea as a source dipole, driven from the animation instead of a plugin.