Learning Timeline
Key Insights
Cinematic Camera Movement Tips
It is highly recommended to include camera movement commands in your prompt. Kling 3.0 Omni is very proficient at understanding instructions like 'fast zoom' and 'orbit' to make videos look more professional.
How to Reference Assets (Tagging)
Use the '@' symbol in the prompt field to specifically 'tag' your reference images. This ensures the AI knows exactly which object or character to use in the scene.
Model Limitations
Facial detail quality may decrease or become less clear when the camera performs a far 'zoom out'. Monitor your 'Generate' results to ensure facial consistency.
Prompts
Video Generation Prompt from Reference Images
Target:
Kling 3.0 Omni
The woman wearing the dress in image [NUMBER] is walking up to the car in image [NUMBER] and touching it. Put them in a desert background. Dynamic camera movements, fast zoom, and orbit. Ultra detailed.
Video Element Editing Prompt
Target:
Kling 3.0 Omni
Make the woman wear a kimono instead. Make the car red.
Step by Step
Steps to Generate Video Using Kling 3.0 Omni
- Open the Kling dashboard and click on the 'Omni' tab at the top (make sure you are not in the standard 'Generate' tab).
- In the model selection menu at the bottom, select 'Video 3 Omni', which is the latest omnimodal model.
- Click the upload button to upload your reference image (e.g., an image of a car or a character) into the system.
- Type your prompt in the text field. To reference an uploaded image, type the '@' symbol and select the relevant asset (e.g., '@image1').
- Include specific instructions regarding actions and the background in your prompt (e.g., a character walking toward a car in a desert setting).
- Add camera movement commands like 'dynamic camera movements', 'fast zoom', or 'orbit' for a more cinematic result.
- Select your desired video duration (e.g., 5 seconds) in the duration settings.
- Click the 'Generate' button and wait for the process to complete until the video appears in your results gallery.
Steps for Editing Video Elements (Video-to-Video Editing)
- Take a previously generated video and upload (input) it back into the 'Omni' model.
- In the prompt field, use natural language to describe the changes you want to make to the video.
- Write specific instructions to change elements such as clothing (e.g., change the outfit to a kimono) or object colors (e.g., change the car color to red).
- Click the 'Generate' button to produce a version of the video edited according to the new prompt.
