2.35:1, exactly 10 seconds. Same Kensington corridor (real street), overcast winter daylight. Cold, high-contrast cinematic documentary look, long-lens compression (85–135mm). Third-person handheld tension; controlled framing; no UI/text/subtitles.
Wardrobe lock unchanged; @konglingchao still in the SAME black down jacket.
Action & cuts (6–8 cuts, continuous timeline; HOW they get “gang” attention without weapons/violence):
(0.0–1.6s) 135mm tele: as they pass a corner storefront edge, the handheld operator accidentally pans across a group area (faces remain unidentifiable blur), then quickly whips away—an obvious “we filmed something we shouldn’t.”
(1.6–3.2s) 120mm tele: one figure in the group turns and locks eyes on them; the figure’s face stays non-identifiable (hood/cap, shallow DOF), but posture is assertive.
(3.2–4.8s) 135mm tele: the figure steps out and mirrors their direction; the camera catches them in compressed background following at a distance.
(4.8–6.6s) 100mm tele: @konglingchao subtly lowers the camera angle and keeps walking; his black down jacket and glasses visible; he avoids eye contact and does not speak.
(6.6–8.4s) 135mm tele: @konglingc.noirwander moves slightly closer to @konglingchao , guiding the path toward a more trafficked corner.
(8.4–10.0s) 85–100mm tele: the follower remains behind as a blurred shape, still tracking them—clear “they’re being watched.”
Dialogue: none (only ambient).
Continuity lock: this directly sets up the confrontation in PART 5; no jump.
Negative: no UI/text; no music/narration/subtitles; no weapons; no violence; no doxxing; no identifiable bystander faces.