Experience gained in associating interpretative VRML modelling with real-time video has demonstrated that it is an effective technique for explaining the content of the video image. However searching the VRML model requires interactive exploration by the user. It is more difficult for the viewer to quickly identify the location of an object of interest in the video image, if a search is run in the database to return the location of an object or objects in the image. It is suggested that there are a number of instances where it would be useful to distort a photo realistic modelled scene to draw attention to specific-portions, ranging from the representation of uncertainty in archaeological presentations or planning proposals, to the identification of the location of an object of interest in a video clip or live video frame. The advantages and disadvantages of some possible approaches are discussed.