This article will guide you on creating high-quality, cinematic AI videos featuring consistent characters using AI tools. We will explore how to generate images of your characters and animate them into realistic videos.
Image Generation for Consistent Characters
Using Omni Reference in Midjourney
To ensure character consistency across multiple images, we'll use the Omni reference feature (orefs) in Midjourney. This allows you to upload a reference image of your character, and Midjourney will generate new images of that same character in different scenarios.
Creating the Initial Character Image
- In Midjourney, go to the prompt bar.
- Describe your character in detail. For example: "Full body shot photo of a samurai wearing armor, shot on Kodak Portrait Film, muted colors."
- Midjourney will generate four image options based on your prompt.
- Select the image that best represents your character.
Generating New Images with the Reference Image
- Drag the chosen character image into the prompt bar, ensuring it's within the Omni reference box (orefs).
- Enter a new prompt describing the desired scene. For example: "Photo of this samurai with skulls of dead demons laying on the ground around him in a post-apocalyptic landscape. Shot on Kodak Portrait Film, muted colors."
- Midjourney will generate new images of your character in the specified setting while maintaining visual consistency.
Facial Feature Consistency
Generating consistent faces can take a few tries. Focus on details like eyebrows, nose, brow line, lips, and jawline. Even if not perfectly identical, consistent clothing and details significantly contribute to the overall character presence.
Image Upscaling for Higher Quality
Upscaling in Midjourney
Before creating AI videos, upscale your character images to achieve higher resolution and detail.
-
Right-click the image in Midjourney and select "Upscale."
-
"Subtle" upscaling is recommended for maintaining details.
Using Magnific for Enhanced Upscaling
For even better results, consider using Magnific for upscaling.
- Select the desired upscaler in Magnific.
- Upload your character image.
- Choose a scale factor to determine the degree of upscaling.
- If upscaling a person, set "Optimize for" to "Portrait" and increase the resemblance setting.
While Magnific sharpens details, it might slightly alter facial features.
Creating Cinematic AI Videos with Vidu AI
Introducing Vidu AI
Vidu AI offers a powerful platform for generating cinematic AI videos. The new Vidu Q1 model is specifically designed for creating high-quality, cinematic shots.
Image to Video Generation
- Go to the "Image to Video" section in Vidu AI.
- Upload your upscaled character image.
- Describe the desired motion in the prompt. For example: "Pick up a skull to inspect it."
- Adjust settings like movement amplitude (usually "Auto" works best), video clip amount, and HD resolution.
Prompting Tips for AI Video
For best results, use prompts with a singular focus, describing a single motion or camera movement. Examples include bowing the head, camera rotating around an object, or zooming on a specific area.
Controlling Movement Amplitude
The movement amplitude setting (small, medium, large) controls the amount of motion in the video.
-
Use larger amplitudes for more dynamic movement, but this can introduce blurring.
-
Smaller amplitudes are generally better for preserving consistency and detail.
Reference to Video Feature
Vidu's "Reference to Video" feature allows direct video creation from reference images. 1. Upload an image of your character. 2. Add a prompt to describe the action, such as "She is in an Asian temple looking through." 3. Vidu will generate a four-second video, which you can increase to 720P.
You can also upload up to seven image references to use multiple characters. This allows the AI to put the characters into a video. You can also add a background by adding a reference image.
Start and End Frames Feature
The Start and End Frames features allow the AI generator to fill in what is in between the two frames, allowing for creative backgrounds.
AI Sound Effects
Enhance your AI videos with AI-generated sound effects in Vidu. 1. Go to the AI sound effects feature. 2. Enter a prompt describing the desired sound. For example: "Asian Buddhist chants." 3. Adjust the duration of the sound effect.