📝 Post Body:
Hi all — I’m Dave Perry. I’m not a developer, just a user who enjoys playing with image generation tools like Midjourney and Stable Diffusion. One thing I really struggle with is getting perspective right — like setting the camera angle or creating a scene with a consistent viewpoint.
I had an idea and wanted to share to see what others think
🎯 The Idea: Perspective Cube
A simple tool built into the image generator interface — a 3D cube that lets the user set the camera angle visually. You’d rotate it like a floating dice to choose the direction of view: up, down, sideways, tilted, etc.
The cube wouldn’t appear in the image — it just sets the perspective for the scene.
🛠️ How It Works:
Rotate the cube → the generator understands where the “camera” is
It sets the horizon line, camera tilt, and vanishing points
The system adjusts the prompt or the generation engine to match the chosen viewpoint
📸 Example Use:
You want a knight seen from below. Instead of typing:
“low angle, looking up at a heroic knight, dramatic perspective”
…you just rotate the cube to tilt up. The generator now knows what you mean.
💡 Why It’s Useful:
Perspective becomes intuitive, not a guessing game
Great for non-technical users (like me)
Keeps images consistent across scenes
Helps with storytelling, layout, and concept art
🧠 Future Possibilities:
Save cube settings for continuity
Use it to block out object depth
Animate the camera through a scene using multiple cube views
I’d love to hear what others think — especially artists, prompt engineers, or anyone building tools like ComfyUI, Deforum, Leonardo, or Runway. Could this be doable? Would it be helpful?
Thanks for reading — and if someone builds it, I’d just appreciate the credit!
— Dave Perry
Top comments (0)