Lifecast.ai
☰


We make software for capturing the world in 3D, using computer vision. The world doesn't stand still, and neither do we. We do volumetric video (not just static 3D models). Our mission is to make capturing immersive volumetric video with the cameras we carry and wear ubiquitous, using AI to fill in the missing details, and achieve cinematic results. Unlike traditional volumetric capture, we go wherever the action is, and capture the entire scene. We work with cameras ranging from iPhones, to AR glasses, and light-field cameras. We make tools for enhancing 2D videos with AI. We offer services for film, VFX, broadcast, and XR.

Products

UpscaleVideo.ai
Enhance video and photo quality with AI

Enhance videos and photos with AI. Upscaling (super resolution), de-noising, and sharpening. 10-bit color, 8K video support. Ideal for up-scaling 8K VR video to 16K, or AI-generated video to 4K.

Volurama
Capture 3D models with a phone and make fly-through videos

Record a short video with a phone. Volurama figures out the camera motion, then builds a 3D model of the scene. Define custom camera paths and render fly-through videos. Export 2D video, VR video, or holographic. Uses neural radiance fields (NeRF) for photorealistic 3D reconstruction.

Volumetric Video Editor
Capture volumetric video for VR with VR180 cameras

Film with VR180 cameras such as the Canon R5. Render and export to volumetric video using Lifecast's 'ldi3' format, with open source players for WebXR, Unreal Engine, and Unity. Create 6DOF VR video for Quest 3 and Vision Pro, or deploy on the web for 2D screens, and holographic displays.

4D neural radiance field (NeRF) video engine
Open source WebXR volumetric video for light field cameras

Our C++ pipeline for volumetric capture from light-field cameras reconstructs a 4D neural radiance field, then compresses into a web-streamable format, which can be embedded in a web page and viewed in VR. All with an MIT open source license. Read our paper.

Services

Volumetric capture with 4D Gaussian splatting video on-site

We offer volumetric capture services using our custom-built light field cameras, or your cameras. We have in-house software for 4D Gaussian splatting (state of the art volumetric video), which can be customize for specific applications. We do robotics-grade camera calibration, then our software estimates camera positions in 3D space automatically, enabling flexible deployment of multiple cameras. Our 4D Gaussian video engine renders 3D videos from arbitrary viewpoints with photorealistic quality ("bullet time"). Unlike traditional volumetric capture studios, our system is portable, and captures both the subject and environment. The results can be rendered to high quality 2D video for VFX, or embedded in a web page for interactive viewing.

Contact info@lifecastvr.com for more info.

About Us, Community, Contact


Lifecast Incorporated was founded in 2021, and is funded by YCombinator. Our team includes former members of Google X, Facebook, Meta, Lyft, Industrial Light & Magic, Lucas Films, and Paramount, and previously worked on self driving cars, robots, and Star Wars. We are experts in 3D computer vision, camera calibration, machine learning, high performance computing, video processing, XR, and VFX. We are located in the San Francisco Bay Area.