GIG reference manual

the parametric raytracer submenu

____________________________________________________________________________________

Section contents

Introduction
The parametric raytracer allows for changes in the attributes of an object within an already rendered image. In addition to color and intensity, maps can be added or redefined with a minimum of rendering time.

Raytracing is a rendering technique in which the color of a pixel is calculated by a simulated backtracking of light rays from the viewpoint. These rays consist of three components: diffuse light, reflected, light and refracted light. If the geometrical and optical relations are known, the contributions of each of these components can be calculated. This calculation can be represented by a tree that is generated for each ray and deleted when the calculation for a single ray is completed. In parametric raytracing, these trees are saved during raytracing and are used to update the rendered picture, thus saving time, for the compositions of the trees are known beforehand.

For best results, you must avoid updating that involves new raytracing, because then the involved trees must be recalculated. So, for instance, if a solid has the attribute transparency and you want to see how the solid looks without this attribute, do not delete the transparency (as this will trigger a new raytracingof the solid), but instead give the transparency intensity a value of 0.

There are cases where recalculations of trees are unavoidable: changing the refraction index of an object, adding the attribute glass, etc. Worst cases involve the adding, deleting, or moving of a light source, in which case all trees are deleted and recalculated, so here extra care is needed. Of course you could add all possible attributes to a solid, then raytrace the scene, and then adjust the attributes. But doing this has the disadvantage of claiming a lot of memory, and that will reduce performance. So the best way to use the parametric raytracer is to decide first which attributes a solid really needs, raytrace it, and then make the desired changes, if any.

Comments
The parametric raytracer is meant as a tool to speed up the adjustment of attributes. Currently, it is not meant for changes in geometrics. If get is used, the attributes of the objects of the last update are compared with the selected set. As a consequence the involved trees will be deleted and recalculated if, for instance, the refraction index of an object differs.

The Parametric Raytracer will need at least 64 mb of internal memory to run properly.