I'm cutting a short documentary about loneliness in crowded cities for a festival submission. The opening had to establish exactly that paradox: a packed urban intersection where every single person is alone with their phone, shoulder to shoulder and a thousand miles apart. I wanted rain-slicked Tokyo at night, neon reflections pooling on the asphalt, a sea of people who never once look up. I live in Ohio. My travel-and-shoot budget for the Tokyo sequence was, if I'm being generous to myself, one coffee.
For weeks I'd planned to fake it with stock footage, but every clip I licensed felt like a tourism board made it. Bright, lively, 'visit Japan.' None of them carried the loneliness, because stock footage is sold to be cheerful, and my film was not cheerful.
Building a city I've never filmed
I ran the Cinematic Establishing Shot — Story Opener for Film and YouTube prompt on Wan v2.6 Text-to-Video: location 'a rain-soaked Tokyo crossing at night,' atmosphere 'lonely and vast despite the activity,' camera move 'a slow drone rise from street level,' teal-orange cinematic grade. What came back genuinely held up on a big screen — wet asphalt mirroring the signage, a crowd flowing through the frame as soft, unfocused background life exactly like the prompt asks for, and the slow rise gradually revealing the sheer indifferent scale of the place above all those tiny heads.
How it cut into the film
The clip opens the documentary cold — no title card, no music sting, just the city and the rain. Then I hard-cut to a quiet talking-head interview with a subject sitting in a small, lamp-lit room in Columbus. The collision between that vast neon scale and one still person in one small room is the entire thesis of the film, stated visually before a single word of narration. The establishing shot earns that cut in eight seconds. The festival screener singled out the opening in their written notes, which is a sentence I intend to remain insufferable about for at least a year.
- Got the shot in three runs — I dialed in the grain in run two and the grade in run three
- Used it as the opening 8 seconds of an 18-minute documentary
- The film was accepted to a regional festival
- I generated a matching dusk version as a backup in case the night read too dark on projection
- Tokyo sequence travel cost: still, gloriously, just the coffee
The single most important variable is the atmosphere line, not the location, and this took me an embarrassing number of runs to internalize. 'Lonely and vast despite the activity' is the exact phrase that produced a crowded street that still feels empty. Without it, the model hands you a generic busy-city postcard with the emotional content of a screensaver. I tested it: the identical Tokyo location with atmosphere set to 'bustling and alive' is a completely different, much happier, much less useful shot. You are not asking the model for a place. You are asking it for a feeling that happens to have an address and a weather forecast.
The camera move sets the emotion as much as the grade
After the atmosphere line, the camera move is the variable I now agonize over, because it carries meaning the grade can't. A slow drone rise from street level says 'small person, enormous world' — perfect for my loneliness theme. A gentle push forward says 'come closer, pay attention.' A locked-off wide says 'observe this from a distance, stay neutral.' I tested the same Tokyo crossing with a push instead of a rise, and the push made the city feel inviting, like it was welcoming me in — completely wrong for a film about isolation. The rise made me feel dwarfed, which was the point. Pick the move for the feeling, not for what looks coolest in isolation, because in an establishing shot the camera is the narrator and the move is its tone of voice.
Honest caveat, and it's a real one: this is for establishing shots and environments, full stop — not for anything with a face you need an audience to recognize and track. The second you put a human in sharp foreground focus, the cracks show and the spell breaks. So keep people soft and far, treat them as texture, and let the place itself be the star of the frame. Grab the Cinematic Establishing Shot — Story Opener for Film and YouTube prompt on Prompt Dock and build the location you can't afford to fly to. When I cut from that wide city to my host on camera, the host is sitting in front of a generated Talking-Head Background Loop — Seamless B-Roll Behind a Presenter — so even the cozy interior is built, not borrowed, and the whole film is shot entirely from a desk in Ohio.