← All prompts

Fill In the Story

How to make the viewer fill in the story. Four stills, three prompts, Midjourney.

commented HIDDEN

The prompts

Copy them, run them, change whatever you want. No signup.

Drowned city
Drowned city · Midjourney
cinematic film still, the upper floors of a drowned city skyline standing out of a flat calm sea at dawn, birds nesting on the ruined towers, one small boat crossing between them leaving a wake, warm gold first light on the eastern faces against the cool blue of the water and the shadowed west faces, immense scale and stillness, photorealistic, shot on 70mm, film grain, cinematography in the style of Erik Messerschmidt --ar 9:16 --style raw --stylize 250 --v 8.2 --relax --no painting, illustration, video game
Empty room
Empty room · Midjourney
Hyperreal architectural stillness, photoreal cinematic still. An empty room with a bare lime-plaster floor and unpainted plaster walls, filling the vertical frame edge to edge with no borders, mattes, or bars. A single hard shaft of late afternoon sun enters from a tall window out of frame at camera-right and lays a bright trapezoid of light across the floor and up the base of the far wall. Dust drifts through the beam as distinct suspended motes. Nothing moves. One plain oak chair stands just outside the light, its shadow stretching long toward the lower edge of the frame.

Light is 3200K low-angle direct sun, hard-edged, 6:1 ratio, the only source, with soft warm bounce off the plaster carrying the shadow side. The floor patch of sunlight is the brightest object in the frame, luminance anchor. Over the chair back hangs a folded rust-orange wool blanket, coarse loose weave, the only saturated color in the frame and the chromatic anchor against the bone and gray plaster.

Materials are specific: chalky troweled lime plaster with visible float marks and small imperfections, pale oak with open grain and a worn satin sheen, coarse slubbed wool. Dark corners of the room hold faint tonal gradient and dust haze rather than flat black.

40mm spherical at f/5.6, deep focus, camera at chest height, straight-on to the far wall. Palette restricted to bone, warm gray, honey and rust. Fine 35mm grain, gently lifted blacks, high contrast between the sunlit patch and the surrounding room. The light patch sits low in the vertical frame with tall empty wall breathing above it.

--ar 9:16 --v 8.2 --raw --stylize 100 --no letterbox, black bars, borders, matte, split screen, diptych, triptych, collage, grid
Snow plaza
Snow plaza · Midjourney
cinematic film still, two tiny figures in long coats crossing an enormous empty snow-covered plaza between brutalist government blocks at night, their footprints the only marks in the snow, warm sodium lamplight pooling on the snow against the cold blue-gray of the concrete and the sky, immense scale and exposure, high wide angle, photorealistic, shot on 35mm, film grain, cinematography in the style of Bruno Delbonnel --ar 9:16 --style raw --stylize 250 --v 8.2 --relax --no painting, illustration

Why it works

A viewer fills in what a frame leaves out. Give them one clue, and they work out the rest on their own. Each of these prompts puts one clue in the frame and leaves the rest unseen.

Show only the top. In the drowned city, the upper floors of a drowned city skyline stand out of the sea. You see the tops of the towers, so you picture the streets under the water.

Leave something behind. In the empty room, a folded rust-orange wool blanket hangs on the chair. An object a person used tells you someone was here, even though nobody is in the room.

Leave footprints. In the snow plaza, the prompt asks for their footprints. Tracks in the snow tell you people walked here before you looked.

Show one sign of life. In a second render of the drowned city, one small boat crossing between them carries one person. One person on the water tells you people still live in this city.

Before you run it

Check the settings on each prompt. The drowned city and the snow plaza were made with --style raw --stylize 250 --v 8.2. The empty room was made with --raw --stylize 100 --v 8.2.

Run the drowned city more than once. One run gave four renders. Three show a real skyline you might recognize, and in one the boat is tiny. Pick the render where the clue is clearest.

Expect more footprints than you asked for. The snow prompt asks for their footprints the only marks in the snow, and the render filled the whole plaza with tracks. That still works for this lesson. If you want one clean line of tracks, run it again.

To make it yours, keep the clue and change the rest. Swap the place and the people, and keep one thing in the frame that points at something the frame never shows.

midjourneyvisual storytellingcompositionpromptstill