Testing AI Video Control with After Effects Animation | ControlNet + WAN 2.1 Workflow
In this video, I continue exploring how to gain more control over AI-generated video workflows.
Starting with a simple After Effects animation — basic blocks, text, and motion — I tested different approaches to see how much structure and intention we can preserve when applying AI models like WAN 2.1 and ControlNet.
Key findings from this experiment:
Having a reference image aligned with your first frame is essential.
High contrast between elements helps the preprocessing phase work much better.
Depth preprocessing gives promising results for animated graphics, but still needs improvement to capture more detail.
More detailed input = better final output.
There’s still a lot to explore, especially with so many new models being released.
This is just the beginning!
👉 If you’re experimenting with AI-driven workflows for motion graphics, let’s connect and share insights!
📺 Full workflow includes:
After Effects animation setup
ControlNet preprocessing (Depth, Canny, Segmentation testing)
WAN 2.1 Fun Control
Video2Video AI generation based on simple animated assets