The use of videos as an input for a rendering process (video-based rendering, VBR) has recently been started to be looked upon with greater interest, and has added many other challenges and also solutions to classical image-based rendering (IBR). Although the general goal of VBR is shared by different applications, approaches widely differ regarding methodology, setup, and data representation. Previous attempts on classifying VBR techniques used external aspects as classification parameters, providing little insight on the inner similarities between works, and not defining clear lines of research. We found that the chosen navigation paradigm for a VBR application is ultimately the deciding factor on several details of a VBR technique. Based on this statement, this article presents the state of art on video-based rendering and its relations and dependencies to the used data representation and image processing techniques. We present a novel taxonomy for VBR applications with the navigation paradigm being the topmost classification attribute, and methodological aspects further down in the hierarchy. Different view generation methodologies, capture baselines and data representations found in the body of work are described, and their relation to the chosen classification scheme is discussed.