In recent years, generative AI has become incredibly advanced, capable of producing near-photorealistic videos with just a prompt. Furthermore, by preparing a reference image, you can generate consistent visuals and animations that match your vision.
That said, it’s best not to leave everything entirely to AI. Instead, use it selectively—for backgrounds, partial effects, or breaking down movements.
In this tutorial, we’ll prepare an illustration created in CorelDRAW and use Google’s generative AI Veo 2 to animate the character.
1. Steps for Using Generative AI
Step 1: Prepare the source image
- Create a PNG or JPEG using a graphic tool (CorelDRAW, Illustrator, CLIP STUDIO, Blender, etc.)
- Be mindful of copyright when preparing characters or logos.
- In this case, we used an illustration from our own content (Sekka-chan turning around) as a sample.
- We wanted to test whether even a rough monochrome design could be animated.
Step 2: Open Google AI Studio
- Access Google AI Studio from your browser (Chrome recommended).
- Log in with your Google account and select “Generate Media” from the left menu.
- To create a video, select “Generative Video.”
- Note: You can test this within the free plan.
2. Add a prompt and image
Enter a prompt that describes the video you want to generate.
- You can write it in Japanese, but English often conveys nuances better.
- Then, add the image you want to animate.
As of the time of writing, the free version of Google AI Studio only allowed the use of a single image.
To connect multiple cuts, you would need the paid plan or another tool.
3. This experiment: Animating Sekka-chan
We originally planned to create a short animation using four cuts:
- Sekka-chan standing with an umbrella
- Throwing the umbrella upward
- Performing a quick-draw sword slash
- Displaying the logo
However, we ended up testing with just one image.
- The prompt was short: “A samurai-style girl turning around.”
- Default settings were left unchanged except for:
- Number of results: 1
- Aspect ratio: 16:9 (landscape)
- Video length: 8 seconds
- Frame rate: 24 FPS
4. Results and observations
The animation was generated in about 20–30 seconds.
- Even with a rough illustration, natural motions like wind and blinking were added, showing a high level of shape recognition.
- However, some awkwardness remained in the turning motion and the movement of clothing.
The main reason was that we only provided a single reference image, which made it difficult for the AI to fully capture the character’s unique traits.
Using more detailed and information-rich images could lead to more natural animations.
5. AI quality and key points
After using Veo 2, we found it impressive that the AI could recognize shapes and animate them from just one illustration.
AI excels at:
- Adding simple movements (wind, blinking, swaying, etc.)
- Automatically enhancing visuals with a cinematic feel
- Adding background effects and subtle details
But for:
- Complex choreography
- Character-specific motions
…multiple cuts and human adjustments are still necessary.
6. Strengths of CorelDRAW and graphic tools
Instead of relying entirely on AI, it’s best to combine it with 2D/3D graphic tools:
- Design characters and logos with CorelDRAW, Illustrator, or CLIP STUDIO
- Create 3D models in Blender to control angles and movements
- Add animation and effects at the final stage with generative AI
By separating what you control yourself from what AI handles, you can efficiently create highly original content.
Summary
Generative AI can animate even rough sketches or monochrome images, but relying solely on one image still has limitations.
The best workflow for now is to:
- Prepare solid materials with CorelDRAW or other graphic tools
- Use AI as a support tool for animation and effects
This approach will give you unique, high-quality results while maximizing efficiency.
The article was not found.
Comments