Non-linear depth rendering of stereoscopic animated images

Abstract

A method for rendering stereoscopic images with non-linear depth variation. The method includes storing content in memory that is ready for rendering, e.g., computer animated images including animated objects or models. A processor is operated to position stereo or horizontally offset cameras and to render the images based on a non-linear relationship between disparity assigned to one or more of the animated objects and a distance between the cameras and the objects. The non-linear relationship is defined by a function or algorithm callable by the processor such as a function that defines a curved depth variation for the computer animated scene. In other cases, the non-linear relationship is defined by stored table, and the rendering includes using the distance between the cameras and an object to retrieve the disparity value to assign to that object. More than one non-linear relationship may be used to render objects with differing depth variations.