All fleeting speech cues are verbalized using a slightly softer voice to set them apart from actual document content.