2.35:1, exactly 10 seconds. Same real Kensington area, overcast winter daylight, cold high-contrast documentary look. Third-person handheld retreat sequence, long-lens compression (100–135mm) with shallow depth; operator walking backward carefully, natural shake and breath. No UI/text/subtitles.
Wardrobe lock unchanged.
Action & cuts (4–6 cuts, continuous timeline immediately after PART 4): (0.0–2.6s) 120mm tele: @konglingchao in foreground, breathing faster, glasses visible; he looks over shoulder once, then forward; camera keeps him sharp, background compressed blur. (2.6–5.2s) 135mm tele: the threatening figures remain behind as indistinct shapes (no clear faces), not closing in dramatically—just pressure implied. (5.2–7.8s) 100mm tele: @konglingc.noirwander speaks into her phone while moving; keep phone screen unreadable; stay near a more trafficked corner (cars passing). (7.8–10.0s) 85mm tele: they reach a safer edge with more witnesses; siren sound grows nearer.
Dialogue (minimal, sync sound):
• @konglingc.noirwander (English, into phone): “We’re at Kensington—two men are threatening us. We’re backing away. Please send police.”
Audio: fast footsteps, heavy breathing, traffic, distant siren approaching.
Continuity lock: police are already nearby (siren nearing); next segment begins with arrival without a time/location jump.
Negative: no UI/text; no music/narration/subtitles; no readable address; no violence; no weapons; no identifiable bystander faces.