How Does Lytro Capture Light Fields for Virtual Reality?
Using an array of cameras set up in a predesigned capture matrix, each can be programmed to “see” different perspectives—exposure, shutter timing, focal length and position all carefully measured and quantized sequentially.
This array-based system of light-field technologies allows Lytro to capture a light field, calculate ray angles and then manifest a virtual representation for interactive immersion.
Light rays pass through the lens and aperture of the each Lytro camera, along with their color characteristics, angular directions and intensity. An array of microlenses gathers captured rays into image discs. Each image disc covers its own part of the sensor, and bundles of light rays are gathered up. The bundles are then processed by a computational engine within the hardware, using a geometric model to create accurate light behavior in virtual reality.
See the full story here: http://www.engineering.com/ARVR/ArticleID/15165/How-Does-Lytro-Capture-Light-Fields-for-Virtual-Reality.aspx