follow-one-encounter-and-departure · git:20260821.8fad868 · 2026-08-21 · sha256 41e169285f710c38
follow-one-encounter-and-departure git:20260821.8fad868A
Immutable. This exact content never changes and is served at /api/v1/blob/41e169285f710c38.
--- name: follow-one-encounter-and-departure description: | Generate a 1970s Greenwich Village street cinéma-vérité piece — a 16mm film camera tailing a twenty-something completely make-up-free young woman through streets and squares, brushing past a group of children, and dissolving into the moving street — as a 15s 4:3 24fps short with raw 16mm film grain, organic texture, pure natural daylight, heavy shoulder-held shake, no stabilization, occasional focus loss, slight light leaks, and an unsettled frame until the very end. Use when the user wants a street-follow / documentary-style long take, a 1970s analog film look, a cinéma-vérité chase after one person, or a "one encounter then departure" observational short. T2VA (text-to-video) only, no reference image. Hard rules: no subtitles and no logo, no music — only on-location sound; no modern digital quality; no on-screen text; keep 4:3 framing and 16mm grain. --- # Follow One Encounter and Departure — 1970s 16mm Street Chase Use this Skill when the user wants a **cinéma-vérité street follow**: a 16mm film camera tailing one person through a living 1970s city street, catching an encounter, then letting them dissolve back into the crowd. Core principle: **the camera stays unsettled and invisible-friendly until the last second — raw grain, shoulder-held shake and pure natural light carry the truth of the street, never modern digital sharpness.** The prompt follows the H3 timecoded structure (core concept → four-segment screen description with positive and negative per beat → hard negative list). T2VA only — pure text, no reference image. --- ## 1. Trigger Use this Skill when the user's request matches any of: - A street-follow / documentary-style long take after one person - "Make it look like 1970s analog 16mm film" - A cinéma-vérité (realism) observational short — one encounter, then departure - A period New York (or any 1970s city) street atmosphere piece Do **not** use this Skill when: - The user wants modern digital HD sharpness or a stabilized gimbal follow - The user wants a fully scripted, posed scene — this must feel improvised and observed - The user wants music or voice-over — only on-location sound is allowed - A reference image must drive the look (this is pure T2VA) ## 2. Inputs - A text description of the scene only: 1970s New York, Greenwich Village streets and a small square; a twenty-something completely make-up-free young woman in a long coat and printed dress walking through, brushing past children, and vanishing into the moving street. - Duration 15s, aspect 4:3, 24fps, raw 16mm film look. - No reference image, no first frame — T2VA. ## 3. Core Concept 15 seconds, 4:3, documentary short. A 1970s New York Greenwich Village street piece shot like a 16mm cinéma-vérité follow — a twenty-something completely make-up-free young woman walks through streets and a square, tailed closely by the lens, brushing past a group of children, before dissolving into the moving street. Raw 16mm film texture, heavy organic grain, pure natural daylight, shoulder-held camera with heavy shake, no stabilization, occasional focus loss, slight light leaks, real and elegant, rich with daily life; the frame stays unsettled until the very end. Ordinary cuts. ## 4. Screen Description | Time | Beat | What must happen | |---|---|---| | 0–3s | Close hand-held follow, medium-close | A twenty-something American woman, completely make-up-free, soft features, wind-blown hair, in a 1970s long coat and printed dress, walks past a bakery on a small Village street; the lens hugs her profile and upper body as it follows; she slows half a beat, glances at the bakery window, then looks to the lens with an unguarded, quiet expression. No posed eye contact, no stabilized shots, no crisp digital sharpness. | | 3–7s | Unstable tailing shot | She walks into a sun-bright small square; the lens trails slightly behind and to one side of her, bobbing with her steps; children run and play around her, pigeons scatter into flight, passers-by cross paths — the soft 1970s New York light and sense of motion fill the frame. No cleared empty square, no modern street elements. | | 7–11s | Hand-held medium shot | A group of children chasing a ball suddenly cuts across in front of her, one nearly bumping into her; she sidesteps gracefully, laughs softly, and exchanges a brief glance with that child; the square stays lively and chaotic around her. No falls or conflict, no posed third-person staging. | | 11–15s | Raw tracking shot | She continues across the square, the lens follows close and clumsily, catching the leftover energy of the children, the light moving across her face, the natural rhythm of a 1970s New York street — the frame closes unsettled. No settling the lens, no adding any modern elements. | ## 5. Hard Rules - **Unsettled frame until the end** — the camera never locks down; it stays shaky, improvised and observant throughout. - **Raw 16mm film texture** — heavy organic grain, pure natural daylight, shoulder-held heavy shake, no stabilization, occasional focus loss, slight light leaks. - **4:3 framing** — keep the 4:3 aspect and the 16mm grain at all times. - **On-location sound only** — no music; only street ambience (street voices, children calling, pigeon wings, footsteps, wind and distant traffic). - **No on-screen text** — clean frame; no subtitles, no logo. ## 6. Negative Rules (Do Not) Subtitles or Logo; music; modern digital image quality; any text on screen; stabilized shots; posed eye contact; crisp sharp digital detail; a cleared empty square; modern street elements; falls or conflict; posed third-person staging; a lens that settles down before the end; adding any modern elements. ## 7. Generation - Mode: T2VA (pure text-to-video) — no reference image, no first frame. - 15s, 4:3, 24fps, native H3 audio (on-location street ambience only, no music). ## 8. Verified Output A 15.08s / 960×696 / 24fps MP4 (4:3) was generated end-to-end with MiniMax H3 from the packaged prompt. Preview: `assets/preview.mp4`, poster: `assets/poster.webp`.