Aspects of the present disclosure relate to a method for producing an output image representing a scene. The method comprises rendering a plurality of component images. Each component image corresponds to an associated depth within the scene. The method comprises determining one or more elements of a view pose to which an output image is to correspond, and deriving an output image part from each of the plurality of component images based on the determined one or more elements. The method then comprises overlaying each of the output image parts, to produce the output image.