Powers of Ten style, FPS seemless zooming continuous shot into the earth, Japan, Tokyo, Shibuya, the scramble crossroad then catch the wink of a girl in the street then say "hi!" then zooming out to the out of the earth and galaxy.

https://twitter.com/AICUai/status/1951328268986228880

Veo3/FlowへのプロンプトはJSON形式で与えることでより高精度のコントロールができるといわれています。

{

"prompt_description": "Powers of Ten style, from a satellite view, a seamless continuous zoom shot into Earth, Japan, Tokyo, Shibuya. In the scramble crossing, the camera focuses on a girl looking up at the satellite. She winks, and says 'hi!' with a hand sign. The camera then seamlessly zooms back out from her to the galaxy.",

"global_settings": {

"style": "photorealistic, cinematic, 8K, high frame rate (FPS)",

"shot_style": "seamless continuous shot, no cuts, satellite point-of-view (POV)"

"sequence": [

{

"shot_id": "001",

"description": "Starting from a satellite's POV, a rapid, seamless zoom begins, descending through Earth's atmosphere to pinpoint Japan, Tokyo, and finally begins to descend directly over the Shibuya Scramble Crossing at night.",

"camera_instruction": {

"type": "continuous zoom-in with a descending path",

"speed": "very fast, smooth, exponential",

"path": [

"A satellite floating in orbit above Earth",

"Planet Earth (full view)",

"Japanese archipelago",

"Tokyo metropolitan area, glowing at night",

"High angle view, descending towards the center of Shibuya Scramble Crossing"

]

"audio": {

"sfx": "Starts with the silent hum of space, transitions into a powerful atmospheric entry whoosh, which then blends into the ambient noise of a bustling city."

}

{

"shot_id": "002",

"description": "The camera's descent, representing the satellite's focus, locks onto one of a young Japanese girl in the crowd of people in the crossroad and she is looking directly up at the camera. She smiles, winks, mouths 'hi!', and gives a friendly hand sign of a heart mark.",

"camera_instruction": {

"type": "Final approach of the satellite zoom, stopping at super close-up the face of the girl",

"angle": "Directly overhead, looking down (bird's-eye view)",

"focus": "Lock focus on the girl's upturned face. The surrounding crowd and traffic are in a slight motion blur."

"subject": {

"description": "A stylish young Japanese girl in her early 20s, standing in the middle of the Shibuya Scramble Crossing.",

"action": "Looks directly up towards the satellite (the camera), as if she knows she is being watched. She gives a bright smile, a playful wink, mouths the word 'hi!', and makes a peace sign (V sign) with her hand towards the camera."

"audio": {

"dialogue": {

"text": "Hi!",

"style": "silent, mouthed clearly"

"sfx": "The immersive, 360-degree sound of the Shibuya crossing: thousands of footsteps, distant traffic, overlapping chatter, and music from giant screens."

}

{

"shot_id": "003",

"description": "Immediately after her greeting, the camera reverses its path, pulling out in a continuous, rapid zoom from the girl holding her hand sign, ascending vertically and returning to the satellite's original orbit and beyond to the start position of this sequence to make a loop video.",

"camera_instruction": {

"type": "continuous zoom-out, ascending vertically",

"speed": "very fast, smooth, exponential",

"path": [

"Close-up high-angle of the girl's face and hand sign",

"Shibuya Scramble Crossing from directly above",

"Tokyo metropolitan area",

"Planet Earth",

"The satellite in orbit",

"View of the Milky Way galaxy"

]

"audio": {

"sfx": "The city sounds rapidly fade as the camera ascends, replaced by a powerful whooshing sound that diminishes back into the silent hum of space."

}

]

}

Gemini Ultraでは5回しか動画生成できないのですが、Flowは実はGoogle AI Proがお得です。毎月Flowと画像生成のWhiskが利用できる1000AIクレジットが付与されます。Fastなら20クレジット、Qualityなら100クレジットで動画生成できます。このサービスがGoogle AI Pro (2TB)プランなら月額2900円で利用できます。Veo3をGemini Ultraで使う制限に比べればかなり緩やかです！