About Geometry Shaders

 
 
 

Geometry shaders allow you to create geometry in the mental ray renderer at render time. All that's stored in your scene is a simple placeholder that defines the position of the geometry — represented by a randomized wireframe cube.

Below, the cube is a geometry shader object that appears fully rendered in the region (image on the left). But, in Shaded and OpenGL views, only a wireframe placeholder is visible (image on the right).

Geometry shaders are ideal for things like feathers that involve many instances of identical geometry. By using geometry shaders, Softimage keeps track of only the placeholder objects, reducing the complexity and memory requirements of the scene.

Geometry shaders are a good solution when you need many identical copies of simple geometry. However, if you use a geometry shader to create complex geometry, it can take a long time to render.

This is because the geometry must be flushed and regenerated for each frame. This extra processing time is more noticeable when the geometry is complex.

Creating Geometry Shaders

Typically, you write custom geometry shaders for the specific needs of your production: feathers, leaves, scales, and so on. To get information about creating your own geometry shaders, refer to Custom Shaders [SDK Guide] and the mental ray manual (available in the Doc\mental_ray subfolder of the directory where Softimage is installed).

There are also several basic geometry shaders provided as examples in the mental ray > Geometry category of the preset manager in the render tree view.