Spatial Scene Preview
Use an image to 3D world workflow on the Astorie canvas to turn a scene reference into a navigable 3D environment preview. Explore camera viewpoints, test spatial relationships around a location and capture reference angles for downstream image and video shots.
A 3D world model builds an explorable environment concept rather than an isolated prop mesh. Comparing a flat single-frame reference with a navigable world preview shows how spatial exploration helps you plan camera coverage across a scene.
Start with a location still or concept frame that clearly shows the layout of the space you want to explore.
Choose or generate an interior room, street, courtyard or landscape image where the floor plane, main structures and distance cues are easy to read. Avoid extreme close-ups or heavy motion blur.
Add the world-generation stage on the Astorie canvas, such as the World Labs workflow, and supply your scene reference image or short spatial brief according to the active node controls.
Once the interactive 3D world preview loads, pan and orbit the camera to inspect how the foreground, midrange and background hold together from nearby angles.
Save or screenshot the viewpoints that hold up visually and pass those captured frames into Image or Video nodes to keep location layout consistent across your sequence.
A generated 3D world is a pre-visualization and spatial reference environment. Check three details as you move the camera away from the starting view.
Notice how far you can pan or orbit before geometry stretches or unseen corners break down. Stay within the angle range where architectural lines and ground planes remain convincing.
Areas hidden behind walls, furniture or foreground trees in the original 2D image are inferred by the model. Inspect those zones before framing a shot that relies on them as a primary background.
Compare doorway heights, horizon lines and object spacing across your captured viewpoints. When a new angle distorts scale, adjust your camera position or refine the captured still in an Image node before generating video.
Use this overview to keep spatial exploration grounded in what world models actually produce on the canvas.
| Input | A clear 2D scene or location reference image (or supported spatial prompt) with readable foreground-to-background depth. | |
|---|---|---|
| Output | An interactive, navigable 3D world preview on the canvas from which you can inspect space and capture reference angles. | |
| Review | Check camera movement bounds, inferred areas behind foreground objects, perspective consistency and horizon alignment. | |
| Plans and credits | World generation and downstream image or video runs consume workspace credits by model; check the pricing page for current plans. | |
| Rights and boundaries | Treat the output as a pre-visualization scene rather than a surveyed architectural scan or a game-engine level export. |
Combine navigable world previews with single-object 3D models and reference-driven video tools on Astorie.
See how World Labs turns a single scene image into a navigable 3D environment for pre-visualization on the canvas.
Compare environment-level world exploration with single-subject 3D object previews.
Turn captured viewpoints from your world preview into guided video shots.
An AI 3D model generator focuses on a single isolated subject, such as a prop, product or character that you rotate from the outside. An image to 3D world workflow builds a navigable scene environment so you can move the camera inside a location and capture multiple shooting angles.
Choose a wide or medium-wide shot with a visible floor or ground plane, clear lighting and distinct foreground, midground and background elements. Flat wall textures, extreme macro shots or heavily obscured scenes give the model very little spatial structure to work from.
No. A generated 3D world is bounded by what can be inferred around your starting scene reference and is designed for spatial exploration and pre-visualization on the canvas. Check the active model node for any supported capture or export options before planning external 3D pipeline handoffs.
Lock down your location once in the world preview, orbit to the camera angles your storyboard needs and capture those views as reference stills. Feed each captured still into an Image node or Video node so every shot in the scene shares the same spatial layout.
Bring a scene reference onto the Astorie canvas, test camera viewpoints inside a 3D world preview and capture the angles that fit your story.
