A handheld device having a display and a front-facing sensor and a back-facing sensor is able to render 3D content in a realistic and spatially correct manner using position-dependent rendering and view-dependent rendering. In one scenario, the 3D content is only computer-generated content and the display on the device is a typical, non-transparent (opaque) display. The position-dependent rendering is performed using either the back-facing sensor or a front-facing sensor having a wide-angle lens. In another scenario, the 3D content is composed of computer-generated 3D content and images of physical objects and the display is either a transparent or semi-transparent display where physical objects behind the device show through the display. In this case, position-dependent rendering is performed using a back-facing sensor that is actuated (capable of physical panning and tilting) or is wide-angle, thereby enabling virtual panning.