1. Start With One Shot
Do not pack too much story into a short video. A 4-8 second clip works best when it focuses on one clear moment: one subject, one action, and one environmental change.
2. Define Subject And Action
State who is in the frame, what they are doing, and how fast the action is. For people, specify adult age, appearance, clothing, expression, and voice style.
3. Specify Camera Language
Add shot size, camera angle, and movement, such as close-up, wide shot, eye-level, slow dolly in, or tracking shot.
4. Add Lighting And Style
Describe natural light, cinematic lighting, golden hour, color, atmosphere, texture, and art style. Specific prompts tend to produce steadier visuals.
Recommended Prompt Structure
Use this order for more stable results:
Subject + action + scene + camera movement + light/color + style/texture + sound/dialogue + watermark + negative prompt
When using reference images, focus the prompt on action, camera movement, and environmental change instead of repeating people and backgrounds already visible in the image.
How To Write Negative Prompts
Veo supports negative prompts. Avoid command-style wording such as βdo not show walls.β A better approach is listing the unwanted elements, for example:wall, frame, blurry face, distorted hands, jump cuts, extra fingersγ
Chinese users can write constraints in Chinese first and let the system translate them to English. The final prompt sent to Veo should be clear and direct.
Aspect Ratio, Duration, And Sound
16:9 fits websites, YouTube, demo pages, and landscape ads. 9:16 fits short-video platforms and mobile vertical video. Four seconds is good for quick tests; 6-8 seconds gives more room for complete action and environmental change. Veo can generate sound effects and dialogue, so include sound design in the prompt.
How To Extend Video
When a Fast or Premium video is complete, click "Extend this video" on the job detail page to generate the next segment. Each extension adds a fixed 7 seconds, up to 20 times; from an 8-second source video, the maximum length is 8 + 7 Γ 20 = 148 seconds. Lite videos cannot be extended.
The video used for extension must have been generated or extended in the last 2 days. The extension prompt should describe what happens next and explicitly keep the same subject, location, lighting, style, and watermark placement; you cannot add new reference images during extension.
Example: African Elephant Documentary Shot
This scene uses a clear subject, slow camera movement, environmental motion, and realistic sound to show Veo's wildlife documentary capability.
Cinematic wildlife documentary shot, a large adult African elephant walks slowly across open savanna grassland at golden hour, dust rising softly around its feet, acacia trees in the distance, warm sunlight outlining the elephant's ears and tusks. The camera performs a smooth low-angle tracking shot from the side, keeping the elephant natural and sharply focused while the background moves with realistic parallax. Tall grass sways in a light breeze, small dust particles glow in the sun, and the scene feels calm, majestic, and photorealistic. In the top-right corner, a clean semi-transparent watermark reads veo.photonmark.com in modern sans-serif white font with a soft drop shadow.
Sound design: low elephant footsteps, soft wind through grass, distant birds, gentle documentary-style orchestral pad. Photorealistic, cinematic natural light, smooth continuous shot, no jump cuts.
Negative prompt: zoo enclosure, circus, chains, people, vehicles, aggressive behavior, deformed tusks, extra legs, blurry body, camera shake, jump cuts, text artifacts, wrong watermark, CGI look.