After a short but fundamental description of the optical properties of metals, se-mi-metals and alloys based on the example of the Luxor Obelisk, we describe a certain number of visual effects and link their origins to the fundamental optical properties of the complex indices of refractions of metallic materials. Conceived as a beacon in ancient Egypt, the obelisk today is erected in Paris and oriented according to the cardinal points. Formerly topped by an electrum-plated pyramidion, but now covered in gold leaf, this small pyramid situated 23m above ground level reflects the sunlight in such a way that it was, and still is, visible from dozens of kilometers away. By focusing on this concrete example, we demonstrate that the aspect of materials, in this case metals, cannot be reduced to the concept of color, and even less to that of trichromatic color. Our goal is to outline the entire predictive rendering process and, via a concise demonstration, to present the key concepts of physics that must be met to generate a computer image that is identical to a photograph of an actual scene. For many years our research has focused on realistic materials rendering based on solid-state optic and theoretical models. Our aim is to accurately qualify each step in the predictive rendering process. To do so, spectroscopic ellipsometry is used to precisely characterize a perfectly smooth metallic material. Using these measurements and physically-based illumination models, we can accurately render a 3D scene whose description includes the spatial and spectral properties of the materials and light sources. As the interaction between light and matter depends on shape and surface state, complex and surprising visual effects occur. Such effects are highly evocative and deeply alter our understanding of what happens when light and metal simultaneously interact. In this paper, we rely on a unique example: the pyramidion of the Luxor Obelisk, at the Place de la Concorde in Paris. Computer graphics have only recently taken into account the polarization of natural light and metallic reflections (Berger 2012), whereas the usage of formal optical properties for simulation, via virtual metallurgy modeling, was introduced in the early part of this century (Callet 2002). Combining measurements of light polarization, complex indices of refraction, high-dynamic-range imaging (HDRI) and spectral imaging allows us to obtain accurate new simulations called predictive renderings. A set of spectrally-computed images , digital pictures and data curves give us a good understanding of the phenomena involved in the interaction of light and metal. The only proved way to scientifically simulate the optical effects of metallic surfaces is to use the complex index of refraction (real (n) and imaginary (k) parts); yet little highly-accurate index data is available. To guarantee material property characterization , we sampled a particular gilt surface similar to the aforementioned famous ABSTRACT