Neural Radiance Fields
A technique that learns a scene from photographs as a neural function, enabling smooth new camera views. Ben Mildenhall and colleagues introduced NeRFs in 2020, and they became a landmark of 3D capture.
Spot it
- 01Foggy floating artifacts
- 02Smooth orbiting views
- 03Realistic shifting reflections
Why it happens
A network maps a 3D position and viewing direction to color and density, and volume rendering along camera rays composes those values into photoreal images.
Recreate it in Pixio
Image-to-video prompt“Animate this image as a neural radiance field capture: a smooth handheld orbit around the subject with convincing parallax, realistic view-dependent reflections, slightly foggy semi-transparent floaters near the edges, soft detail in distant areas and gently blurred, incomplete geometry at the borders. Keep the original subject and composition.”
Open PixioUpload a still and pick a video model.