Vidu Q4: cinematic video with synced sound from a single photo
Vidu Q4 is a new video generation model in NeuralSpace: it turns a single photo or a set of reference images into a cinematic scene where characters speak with perfectly synced sound. Dialogue, footsteps, gravel crunch — everything is voiced within the same generation, with no separate dubbing or stitching in an editor. Clips run from 3 to 16 seconds, at up to 4K resolution.
What Vidu Q4 can do
- Image → video. Upload one photo — the model brings the scene to life: coordinated camera work, natural motion, realistic light and atmosphere.
- Multiple references → one scene. You can provide up to 15 photos: characters, locations, style. The model combines them into a single coherent scene and keeps faces, hairstyles and clothing consistent across shots.
- Synced sound and dialogue. Speech, effects and background music are generated together with the video and land on the lips and movements. A single checkbox turns the sound off.
- Cinematic camera. Push-ins, tracking shots and dollies — the camera «works» with the scene instead of panning over a static frame.
- Up to 4K. 540p, 720p, 1080p, 2K and 4K are available — good for social feeds and for big screens alike.
How to generate video from a photo
Open the Vidu Q4 page, upload a photo and describe the scene: what happens, how the camera moves, what you hear. For a scene with several characters, upload reference photos — they become character1, character2 and so on, and you can address these roles directly in the prompt. Then pick the resolution and duration (3–16 seconds) and hit Generate.
How much it costs
Pricing is per second and depends on the resolution. Examples from the site catalog at the time of writing Цены в токенах приведены по каталогу сайта на момент публикации; стоимость зависит от выбранного разрешения и длительности.
- 540p — from 19 tokens for 3 seconds (≈ $0.23), about 32 tokens for 5 seconds (≈ $0.38);
- 720p — 40 tokens for 3 seconds (≈ $0.48), 67 tokens for 5 seconds (≈ $0.80), about 215 tokens for 16 seconds (≈ $2.56);
- 1080p — 51 token for 3 seconds (≈ $0.61), 84.5 tokens for 5 seconds (≈ $1.01);
- 2K — about 112 tokens for 5 seconds (≈ $1.34);
- 4K — 229 tokens for 5 seconds (≈ $2.73), about 733 tokens for a full 16-second scene (≈ $8.74).
Dollar equivalents are approximate, converted at the current exchange rate (1 token ≈ 1 ₽). Tokens are charged only for successful generations: if something fails, they return to your balance automatically.
What it is best for
Vidu Q4 is built for short narrative video: ad inserts, music videos, character animation, «living» photos for social media, and short scenes with dialogue. It is especially strong when characters speak — voice, facial expressions and camera movement stay in sync within one generation.
FAQ
Do I need several photos?
No. The basic mode needs one photo and a scene description. Multiple references matter when your scene has several characters or a specific location.
What is the maximum duration?
Up to 16 seconds in a single generation, with a 3-second minimum.
Can I turn the sound off?
Yes — the «Synced sound» checkbox disables the voice track and leaves a silent cinematic clip.
What happens if a generation fails?
Tokens return to your balance automatically — a generation either succeeds or costs nothing.