
Vidu Reference-to-Video Q1 API by Vidu
Open and Advanced Large-Scale Video Generative Models.
Vidu Reference-to-Video Q1 został opracowany przez ShengShu Technology. Atlas Cloud (prowadzona przez Atlas Cloud AI LLC) zapewnia dostęp do modelu, ale nie jest jego właścicielem. Wszystkie znaki towarowe należą do ich odpowiednich właścicieli.
Reference to Video Q1
Features
Multi-Entity Consistency
-
Maintains visual identity of characters and objects
-
Preserves style and appearance across frames
-
Supports multiple reference subjects
Flexible Input
-
Accepts 1-3 reference images
-
Supports various aspect ratios (16:9, 9:16, 1:1)
-
Adjustable movement amplitude
Requirements
Images
-
Number: 1-3 reference images
-
Format: PNG, JPEG, JPG, WebP
-
Size: Minimum 128x128 pixels
-
Aspect ratio: Less than 1:4 or 4:1
-
File size: Maximum 50MB per image
Prompt
-
Maximum length: 1500 characters
-
Should describe desired motion and scene
Best Practices
-
Use clear, high-quality reference images
-
Provide detailed prompts for better results
-
Consider scene composition in reference images
-
Test different movement amplitudes for optimal output
Effect Boundaries
-
Best results with clear, well-lit reference images
-
Movement amplitude affects transition smoothness
-
Complex scenes may require multiple reference images


















