A research team in Korea has taken a new approach to telepresence systems, that have traditionally worked in a way that does not reproduce the actual light field from a real object or person. Instead, they estimate shapes, surfaces, or depth maps and then render a 3D-looking image using computational models. This can create convincing visual effects, but it does not fully preserve the physical behavior of light, such as diffraction, phase, and natural depth-dependent focusing. A research team led by Professor YongKeun Park at Korea Advanced Institute of Science and Technology (KAIST), South Korea, has developed a different approach. Rather than first turning a real scene into a geometric model, the system directly captures the complex optical wavefront of the scene.