Veo 3.1's advantage is the face and the voice: native audio generation, believable lip sync, and enough textural fidelity to hold a close-up. These prompts are written for performance — spoken lines, micro-expression, and lighting that rewards skin detail rather than hiding it. Dialogue is written inline so the model has something specific to synchronize to.
Extreme close-up on a woman in her sixties in a dimly lit confession booth, shot on an 85mm lens at shallow depth of field, only the lattice screen light falling across her face in narrow bars. Every pore, fine line, and the wet rim of her eyes is rendered in high fidelity. She speaks quietly, with a slight tremor, lips fully synchronized to the line: "I told him it was the last time. That was nineteen years ago." Her jaw tightens after the final word. Warm single-source practical from above, heavy chiaroscuro, deep shadow filling the frame. Faint sound of a church organ two rooms away and her own unsteady breath.
A medium two-shot across a diner booth at 3am, 40mm lens, shallow focus favoring the man on the left. He is tired, unshaven, the fluorescent overhead catching the sweat at his hairline with realistic skin sheen. He says, evenly, lips precisely synced: "You knew before I did. Just say it." The woman opposite does not answer immediately; her eyes move first, then her mouth opens slightly and closes. Cold fluorescent key mixed with warm neon spill from the window. Faint hum of a refrigeration unit, the clink of a spoon somewhere off-frame, no music. Naturalistic performance, micro-expression detail on both faces.
A tight close-up of a young radio broadcaster in a soundproofed booth, lit by the amber glow of the console below her, 100mm lens, extremely shallow depth of field with only her eyes in critical focus. Skin texture, the fine hairs at her temple, and the condensation of breath on the microphone windscreen are all rendered at high fidelity. She leans in and speaks with practiced warmth, fully lip-synced: "If you're still awake out there — this one is for you." Her expression shifts from performance to something genuine on the last three words. Low-key amber and deep blue palette, soft top light, natural film grain, the ambient hush of an isolated booth.
A medium close-up of a man in his thirties standing in a doorway at dusk, half in warm interior lamplight and half in cool blue exterior shadow, 50mm lens. Rain falls behind him, out of focus. His face carries visible fatigue — reddened eyes, stubble, a small healing cut on the eyebrow, all rendered with high textural fidelity. He speaks with difficulty, lips fully synchronized: "I'm not asking you to wait. I'm asking you to remember." He swallows after the line and looks down. Naturalistic performance, split lighting, shallow focus, ambient rain and distant traffic, no score.
A locked-off medium shot of an elderly man seated at a bare table in a fluorescent-lit deposition room, 35mm lens, flat frontal lighting that deliberately reveals every crease, liver spot, and tremor in his hands. He waits, breathing audibly, then answers with quiet defiance, lips precisely synced: "I signed it. I read every word before I did." A slight tightening at the corner of his mouth follows. The unforgiving overhead light and pale institutional walls remove all flattery from the image. High-fidelity facial texture, subtle involuntary micro-movements, ambient air-conditioning drone and the tap of a stenographer's keys.
A tight close-up of a woman in her forties seated in a rain-streaked car at night, face lit only by the shifting magenta and cyan of a holographic billboard outside, 85mm lens, extremely shallow depth of field. Rain shadows crawl across her skin; every pore, the smudged liner beneath one eye, and the faint scar along her jaw render at high fidelity. She speaks low and fast, lips fully synchronized: "You get the file, I get a new name, and neither of us was ever here." Her eyes flick once to the mirror on the final word. Neon-noir palette, hard side key, deep black fill, ambient rain on the roof and the low idle of the engine.
A slow push-in from medium to close-up on a grey-bearded warlord standing over a candlelit war table in a stone hall, 50mm lens. Firelight flickers across weathered skin, rendering deep creases, wind-burn, and the wet shine of tired eyes at high fidelity. He looks up and speaks with weary authority, lips precisely synced: "We hold the bridge until sundown. After that, it will not matter who holds it." A muscle works in his jaw after the line. Warm single-source candlelight against cold stone, heavy chiaroscuro, shallow focus, ambient crackle of flame, distant rain, and the shift of armor off-frame.
A locked-off medium close-up of a man in his fifties seated against a blank grey wall, flat frontal lighting, 40mm lens. The image is rendered with unforgiving textural fidelity — sweat at the temple, a nervous swallow, the slight asymmetry of a forced smile. He begins confidently, lips fully synced: "There was never any building on that site. I would remember." Mid-sentence his own reflection in the dark glass behind him fails to move with him. His expression does not change; he keeps talking. Clinical documentary lighting, no score, the faint hiss of a recording device and room tone. Naturalistic performance, subtle involuntary micro-movements.
An extreme close-up of a woman's face floating in a featureless white void, 100mm lens, lit by an impossible soft light with no visible source and no cast shadow. Skin texture, fine hairs, and the moisture line of the lower lid are rendered at extreme fidelity against the total emptiness behind her. She speaks slowly, as if remembering the words rather than choosing them, lips fully synchronized: "I drew the map before there was any country. That was the mistake." Her pupils dilate slightly on the final word. Flat shadowless illumination, monochrome bone-white palette, no ambient sound at all except her voice and a single sustained low tone.
A handheld close-up of a young man holding a camcorder at arm's length in a dark stairwell, his face lit hard and unevenly from below by the camera's onboard light, 24mm lens with visible barrel distortion. The harsh underlighting reveals every detail — dilated pupils, sweat, a split lip, the tremor in his breathing. He whispers directly into the lens, lips precisely synced: "If anyone finds this, don't come looking. Just don't come looking." He glances off-frame, then kills the light. Overexposed hotspot falling to crushed black, heavy digital noise, timecode overlay, only breath and distant structural creaking in the audio.
Build your own prompt free — no account needed.