You are interested in how Nvidia’s Medusa tech demo, which was published for the GT200 launch, was created? PCGH takes a look behind the scenes.
You probably remember: For the launch of the Geforce GTX 280 and GTX 260 Nvidia had a new tech demo. The Medusa demo used Direct X 10.0 effects. Nvidia emphasized the use of DX10’s Geometry Shading and Stream Out features. When compared to the predecessor (Adrianne) the face animations were said to be much better. Medusa’s Blendshapes (the individual animations) were said to be four to five times as complex.
But how was this tech demo created? A presentation on Nvidia’s Nvision gives the answer. First of all the graphics artists put all their ideas into concept drawings. Those drawings, by the way, are real art themselves already. All characters were individually drawn at first.
Medusa tech demo: Artwork vs. real-timeMedusa tech demo: Artwork vs. real-time
For the Skin Shading some of the shaders of the Human Head Demo were used, but the face of Medusa had to be modeled from scratch. Therefore a cast of the actress’s face was taken and she was photographed from different points of view with five high resolution cameras. Those pictures were the base for the three-dimensional head that could be animated.
Medusa tech demo: The actress was photographed from different points of view.Medusa tech demo: Making a cast of the head
The visualization of the crystal was another important point on Nvidia’s list. Several levels of reflections and transparency were necessary to receive the final appearance visible in the tech demo.
In the picture gallery you can find additional Medusa making-of pictures.