Lifecast Volurama

Lifecast is dedicated to pushing the limits of immersive volumetric media. They created a new format for volumetric photos and videos called 'ldi3' which enables real-time 6DOF photorealistic rendering on a wide variety of platforms.

Lifecast's main product is Volurama - a Windows/Mac GUI for the NeRF engine for reconstructing static scenes from iPhone video input.

They're most widely known for being able to train with GPU acceleration on Mac. This enables the rapid training on NeRFs on Mac devices. They are currently the only platform to offer this ability at the time of writing. Volurama can be downloaded from the Lifecast website and also explored on Github.

Lifecast was founded in 2021 by Forrest Briggs, Ph.d. (formerly of JPL, Facebook, Lyft, and Google X), and is venture-funded by Y Combinator. Additionally, Volurama is compatible with Looking Glasses's Quilt format.

Most recently, they have implemented Tiny Cuda NN to help accelerate training on Windows in V1.6.

Their Values are:

1. Open: Much of their software is open source

2. Practical: They don't just write papers for academia and build prototypes. They are making volumetric video for VR happen right now, with realistic hardware and processing requirements.

They are interested in the following topics:

1. Immersive: They are interested in volumetric media for VR, which means capturing and delivering whole environments that a user can be in.

2. Video: A lot of volumetric capture (NeRF/photogrammetry) focuses on static capture. We do video.

3. Portable: They strive to make our software work just about anywhere. For example, the NeRF video engine (C++) compiles on Mac, Windows, and Linux, and our volumetric video player works on the web in JavaScript, and in Unity and Unreal (all open source).

4. Inside-out: Most volumetric video capture systems are outside-in (lightstages) focussing on a single subject. They capture inside-out, both the subject, and the environment.