Japanese Meal by Mount Fuji — Continuous FPV Video

Format

Text Parameterized

Inputs

character_reference (image) · Required

Outputs

generated_video (video)

Replace the ${…} placeholders with your values before using this prompt. See parameters and defaults below.

Create a 10-second video in 16:9. One continuous shot, no cuts. No music.

CHARACTER REFERENCE
Use ${character-reference} as the strict character reference. Keep the same recognizable adult woman: same face, facial proportions, hairstyle, red ribbon, body proportions, and natural appearance. She is 30 years old, 165 cm tall, with a slim build and full adult proportions. Do not change her apparent age. Keep her freckles visible across the nose and upper cheeks. Match the same warm 90s color grade so this feels like another memory from the same film.

SETTING
A quiet traditional Japanese room with warm wood, tatami, and paper screens pushed aside. One large window stands fully open. The woman sits alone at a low table beside it. Late-afternoon light falls sideways across the table.

The table holds a simple, partly eaten Japanese meal: rice, grilled fish, small side dishes in ceramic bowls, faintly steaming miso soup, and tea. Keep the meal honest and unstyled, like home or inn food. Chopsticks rest unevenly, one dish is half-finished, and a few grains of rice are on the table edge. Do not make it look like a glossy restaurant presentation.

ACTION
She eats and genuinely enjoys the food. She lifts a piece with her chopsticks, brings it to her mouth, eats, and briefly closes her eyes with a small, involuntary moment of pleasure. She chews naturally, swallows, breathes out, and settles into a quiet closed-mouth smile. She glances out the window once, unhurried, then looks back to the bowl and reaches for the next bite.

Keep her performance small and unperformed. She never reacts to or looks into the camera. No posing, theatrical delight, wide grin, or hand-to-mouth gesture. Her enjoyment is conveyed through her eyes and the pause after the bite.

CAMERA — ONE CONTINUOUS FPV FLIGHT
00:00–00:03.5: Start close to her, slightly above the table and angled down across it. Keep her face and hands in frame, close enough to see soup steam and the rice texture. Drift slowly with the light instability of a small aircraft holding position. Hold on her through the bite and closed-eye pause.

00:03.5–00:06: Rise and draw back across the table, revealing the whole meal, the low table, her seated position, and the room. Accelerate smoothly from almost nothing; the move should begin before the viewer notices it.

00:06–00:07.5: Turn toward the open window and accelerate. Let the window frame rush toward the lens and pass on all four sides. The camera must pass cleanly through the opening in the same shot: no cut, dissolve, wipe, black frame, or other transition. Let exposure shift briefly from interior shade to daylight, as it would with a real lens.

00:07.5–00:10: Outside, continue forward and lift. Reveal roofs and trees below, then the land falling away. Mount Fuji appears ahead, snow-capped and hazy with distance above the horizon. The mountain must not be visible through the window from inside; it is revealed only after the camera is outside. Slow and settle facing Fuji, but keep drifting during the final second rather than locking off.

Maintain the character of a small flying camera throughout: continuous micro-instability, slight roll corrections, tiny overshoots, a touch of wobble during acceleration, a wide lens with mild barrel distortion at the edges, and deep focus from near to far. The flight is fast but never frantic. Avoid gimbal-perfect stabilization, a dead-smooth digital glide, or an obviously rendered camera path.

WORLD DETAILS
Outside, use late-afternoon haze to layer the distance. Fuji’s snow catches warm light, and cloud shadow moves slowly across the lower slopes. Inside, soup steam rises and bends in the window draft; the tea surface trembles slightly; loose strands of hair and the ribbon tails move in the air from the open window. Nothing freezes: the last frame should feel as alive as the first.

COLOR GRADE
Lock the image to a warm late-afternoon 90s-film character: creamy natural skin, muted earthy greens, soft golden-amber sunlight, warm neutral shadows, slightly faded saturation, gently lifted blacks, restrained contrast, smooth highlight roll-off, fine organic grain, subtle analog noise, mild vintage-lens softness, delicate halation, and naturally imperfect exposure. Fuji’s snow should read as warm ivory, not sterile digital white. No modern HDR, teal or cyan grading, neon greens, or ultra-sharp digital appearance.

SOUND
Absolutely no music, score, or instrumental sound. Use only natural diegetic audio: chopsticks against ceramic, quiet sounds of eating and breathing, tea being poured or set down, and the room’s soft interior tone. At the exact frame the camera passes through the window, swing the sound field completely: the interior falls away and open air arrives, with wind over the lens, distant birds, and the thin ambience of height. No crossfade and no dialogue.

PRIORITIES
1. The camera never cuts; it passes through the window rather than cutting around it.
2. Her enjoyment is small and real, expressed in the pause after the bite rather than an exaggerated expression.

FAIL CONDITIONS
The result is wrong if there is a cut, dissolve, wipe, or black frame at the window; if the camera stops at the window and the exterior becomes a new shot; if Fuji is visible through the window at the beginning; if she looks into the lens; or if the food is arranged like a styled restaurant presentation.

NEGATIVE CONSTRAINTS
No age drift or childlike proportions. No second person, staff, diners, or crowd. No extra limbs, warped hands, six fingers, morphing face, or chopsticks fused to the hand. No window cut or transition effect, speed ramp, slow motion, or freeze frame. No gimbal-perfect stabilization or rendered flight path. No HDR, cyan or teal grading, oversaturation, or modern ultra-sharp look. No text, signage, logos, subtitles, or music.