One of my top games of 2005 was Shadow of the Colossus (SOTC) on the PS2. The developers at Sony Computer Entertainment (SCE) in Japan really took the PS2 to its limit. Features like HDR rendering – albeit somewhat faked due to hardware limitations – and extensive use of inverse kinematics in their animation controllers really gave the game an advanced look on the aging platform. Here is a great article on the development of some of the main features in the game along with developer interviews that was translated from a Japanese website:
Making of Shadow of the Colossus
Of particular interest to me was how they handled collision detection between the player and the colossi while the player was mantling around all over their body. Instead of using highly simplified collision models like spheres, cylinders, or boxes as in most games, the colossi used a hidden simplified secondary mesh that mostly followed the outline of their body. The bones that were driving the animation of the rendered mesh were also weighted to the vertices of the collision mesh allowing it to transform perfectly along with the colossus. Here is a debug rendering of how the collision mesh is layered along with the rendered mesh:
I thought this was a very clever way of using the bone transforms to also control the collision. The whole article is very interesting if you are interested in game development or just SOTC itself. Check it out!
give me some excellent ideas. thanks man