POV-Ray : Documentation : 1.3.10.6 Initializing the raytracer
  POV-Ray 3.6 Documentation Online View  
1.3.10.5 Scene definition   1.3.10.7 Ray-sphere intersection

1.3.10.6 Initializing the raytracer

#declare MaxDist = 1e5;
#declare ObjAmnt = dimension_size(Coord, 1);
#declare LightAmnt = dimension_size(LVect, 1);

#declare Ind = 0;
#while(Ind < LightAmnt)
  #declare LVect[Ind][0] = vnormalize(LVect[Ind][0]);
  #declare Ind = Ind+1;
#end

Before being able to start the raytracing, we have to intialize a couple of things.

The MaxDist defines the maximum distance a surface can be from the starting point of a ray. This means that if a surface is farther away from the starting point of the ray than this value, it will not be seen. Strictly speaking this value is unnecessary and we can make the raytracer so that there is no such a limitation, but we save one extra step when we do it this way, and for scenes sized like ours it does not really matter. (If you really, really want to get rid of the limitation, I am sure you will figure out yourself how to do it after this tutorial.)

The ObjAmnt and LightAmnt identifiers are declared just to make it easier for us to see how many objects and lights are there (we need this info to loop through all the objects and lights). Calling the dimension_size() function is a really nice way of getting the number of items in an array.

All right, now we are getting to a bit more advanced stuff: What does the while-loop do there?

The #while-loop uses the Ind identifier as an index value going from 0 to LightAmnt-1 (yes, -1; when Ind gets the value LightAmnt the loop is ended right away). We also see that we are indexing the LVect array; thus, it is clear we are going through all the light sources (specifically through their direction vectors, as we only use the [0] part) and we assign something to them.

What we are doing is to assign a normalized version of each light source direction onto themselves, that is, just normalizing them.

Normalize is a synonym for "convert to unit vector", that is, convert to a vector with the same direction as the original but with length 1.

Why? We will later see that for illumination calculations we will be needing unit vectors. It is more efficient to convert the light source directions to unit vectors once at the beginning than every time for each pixel later.

1.3.10.5 Scene definition   1.3.10.7 Ray-sphere intersection


Copyright 2003-2021 Persistence of Vision Raytracer Pty. Ltd.