The current trend for employing three-dimensional computer graphics to represent archaeological sites is limited because their level of realism cannot be guaranteed. The images that are generated may look realistic, but often no attempt has been made to validate their accuracy. In order for the archaeologist to benefit from computer-generated representations and use them in a meaningful way, virtual past environments must be more than pretty pictures - they must accurately simulate all the physical evidence for the site being modelled.
An often neglected facet of reconstructions is the lighting. Standard three-dimensional modelling packages do not allow precise control over lighting values, and as a result, many simulations show the archaeological site under inappropriate or false lighting conditions such as the bright, steady light of the present. Experimental archaeology and realistic lighting simulation allow us to recreate the original lighting of an archaeological site and show it how it might have looked to those who built and used it. Predictive lighting also opens up new avenues of exploring how past environments may have been perceived, allowing us to investigate on a computer new hypotheses about architecture, art, and artefacts in a safe, non-invasive manner.
This paper outlines the need for perceptual realism in virtual heritage, and examines case studies where the application of predictive lighting techniques have enhanced archaeological interpretation.