A scripting convention for directing AI text-to-speech performance purely through punctuation and capitalization in the script text, without re-recording. Three dots (ellipses) create a natural pause; ALL CAPS on a word adds emphasis; a trailing question mark changes how the sentence's intonation resolves. First observed on ElevenLabs' 11v3 model, which reads pacing and emotional cues directly from the text it's given — the technique is a property of how the voice model parses text, so it should transfer to any platform running 11v3 or a similarly text-sensitive TTS model, not just the one it was demonstrated on.
Apply: Insert ... for a pause, wrap a word in capitals for emphasis, or end a line with ? to shift its landing intonation, then regenerate the line from the edited script rather than trying to fix delivery after the fact. Pairs with Emotion Tagging for Per-Character AI Performance Direction for per-line performance direction and with Video-Analysis-to-Voiceover Script Pipeline when scripting narration against existing footage.