We introduce a method to convert stereo 360{\\deg} (omnidirectional stereo)\nimagery into a layered, multi-sphere image representation for six\ndegree-of-freedom (6DoF) rendering. Stereo 360{\\deg} imagery can be captured\nfrom multi-camera systems for virtual reality (VR), but lacks motion parallax\nand correct-in-all-directions disparity cues. Together, these can quickly lead\nto VR sickness when viewing content. One solution is to try and generate a\nformat suitable for 6DoF rendering, such as by estimating depth. However, this\nraises questions as to how to handle disoccluded regions in dynamic scenes. Our\napproach is to simultaneously learn depth and disocclusions via a multi-sphere\nimage representation, which can be rendered with correct 6DoF disparity and\nmotion parallax in VR. This significantly improves comfort for the viewer, and\ncan be inferred and rendered in real time on modern GPU hardware. Together,\nthese move towards making VR video a more comfortable immersive medium.\n