The Online POV-Ray Tutorial

Modifiers

The things in this file are used to modify the default behavior of certain objects. These modifiers include the following.

Bounded By

The only effect of bounding is to speed the rendering of a scene. Specifying a bounding object for another object tells POV-Ray the latter is entirely contained by the former. When rendering, POV-Ray first does any ray tests against the presumably simpler bounding object. If the ray does not intersect the bounding object, then POV-Ray assumes that it does not intersect the interior object either. POV-Ray then skips the potentially lengthy ray intersection tests with the interior object. If the ray does intersect the bounding object, POV-Ray then performs tests to see if the ray also intersects the interior object. Bounding objects do not affect the way the scene looks. Any object (except light sources) can be bounded, although bounding will show the most speed improvement when used with complex CSG or polynomial objects. Boxes and spheres make wonderful bounding objects as they are highly optimized, however, any object may be used to bound another.

Note that POV-Ray has a limited ability to do automatic bounding. Most finite objects can be automatically bounded with a reasonable amount of efficiency. Unions can also be bounded fairly efficently. Things that cannot be automatically bounded are infinite objects, differences, intersections, or merges.

Here's an example of the use of bounded_by. Really_Complex_Chair is presumed to be defined somewhere above this in the scene file. 

object {
   Really_Complex_Chair
   bounded_by {
      box {
         <-4, 0, -4>, >4, 8, 4>
      }
   }
}

Objects should never extend outside their bounds. Doing so will have undefined results. Sometimes the object will render ok, sometimes parts that stick out will be trimmed away, sometimes things will just get weird. If you want to slice away part of an object, use clipped_by or CSG.

If you have an object which you clipped, and want to use the same object to bound that you used to clip, you can do something like this . . . 

object {
   Some_Object_Or_Other
   clipped_by {
      sphere {
         <0, 2, 0>, 3
      }
   }
   bounded_by { clipped_by }
}

Clipped By

The clipped_by statement allows you slice away parts of objects to reveal a hollow interior. It is similar to intersection except it leaves holes where parts of the object were clipped away. Any object (except light sources) can be clipped, and any object (except for light sources again) can be used to clip. For example, here is a cube that is clipped with a sphere.

Only the portions of the cube that are on the inside of the sphere are retained. The notions of "inside" and "outside" in the clipping object can be switched with the inverse keyword. Note that you can specify multiple objects to clip with in one clipped_by statement. The result will be that the object is clipped first with first object, then the result is clipped with the second, and the result of that is clipped with the third, etc. Here's an example of a cube with two clipping spheres.

If you've bounded an object with a second object, and want to use that same second object to clip the first, you can do something like this (presumably This_Is_An_Object is declared elsewhere) 

object {
   This_Is_An_Object
   bounded_by {
      cylinder {
         <0, 0, 0>, <0, 1, 0>, 0.5
      }
   }
   clipped_by { bounded_by }
}

Map Type

The map_type keyword is only applicable to image maps, bump maps, and material maps. It is used to modify how the image is projected onto the surface in the mapping. The parameter for map_type is an integer which tells what kind of mapping to use. The default is "map_type 0". This is a planar mapping in which the image is projected along the z axis onto the x-y plane. The image is rescaled to fit into the unit square in the x-y plane. It also repeats infinitely in the x and y directions. Other mappings are as follows.
map_type 0
The planar mapping as described above.
map_type 1
A sperical mapping. The image is mapped Mercator-style around the origin. The right edge of the image ends up in the +x direction and is mapped clockwise around the y-axis. The top and bottom edges of the image are compressed down to points on the y-axis. The image shows a great deal of distortion around the poles of the mapping (the y-axis by default). Each pixel in the image is transformed into a three dimensional wedge shape which radiates from the origin.
map_type 2
A cylindrical mapping. The image is mapped around the y-axis like wrapping paper. As with the spherical mapping, the right edge of the image ends up in the +x direction, and the image is mapped around the y-axis in a clockwise sense. The image is still rescaled in the y-direction to be one unit tall and repeats infinitely up the y-axis. Each pixel in the image is transformed into a pie-slice shape which radiates from the y-axis.
map_type 5
A toroidal mapping. This type of mapping is the most difficult to describe. The image will fit around an untransformed torus (i.e. one that was created with the torus primitive and then left alone). Experiment with it.
Map types 3 and 4 are reserved for future use.

The following images show map types 0, 1, 2, and 5, respectively. Each map type is projected onto its corresponding shape. Note that for types 1, 2, and 5, the pigment was rotated to give a better view of what exactly was going on.

For example, if you wanted to make a globe, you might try something like this 

sphere {
   <0, 0, 0>, 4
   pigment {
      image_map {
         gif "earth.gif"
         map_type 1
      }
   }
}

No Shadow

The no_shadow keyword is applicable to any object (not light sources or cameras). When included in an object, that object will not cast a shadow on anything else. This is primarily useful for special effects, primarily those which simulate objects which "produce" light. Laser beams are a good example. The no_shadow keyword takes no parameter, it's either there, or it's not (in which case the object will cast a shadow like normal).

The following images show the effect of no_shadow. The first image is normal, the second has a sphere with the no_shadow keyword.

Once

The once keyword is only applicable to image maps, bump maps, and material maps. By default, the image which is mapped is repeated infinitely in the x and y directions (this varies with the map type). Adding the "once" keyword to a mapping specification will remove all the duplications of the image except the one in the (0, 0), (1, 1) unit square. With image maps, everywhere else becomes transparent; with bump maps, everywhere else becomes flat; with material maps, everywhere else becomes texture 0. The once keyword takes no parameter.

If, for example, you wanted to create a single image stamped into a mirror you might try 

box {
   <-1, -1, -1>, <2, 2, 2>
   pigment { color rgb <1, 1, 1> }
   finish {
      reflection 1
      ambient 0
      diffuse 0
   }
   normal {
      bump_map {
         gif "stamp.gif"
         once
      }
   }
}

Open

THe open keyword controls the existence of endcaps on cylinders and cones. It is not legal anywhere else. This keyword takes no parameter. If it is specified, then cones and cylinders will be rendered as hollow tubes. Otherwise they will have the circular ends in place (default). Here are two sample scenes, the first without open and the second with.

Sturm

Sturm is only applicable to blobs and polynomial objects (including tori). These objects require extremely accurate calculations to render. Normally they will look ok with POV-Ray's default root-solver. However, in certain cases they will render incorrectly. In these cases, you can specify the keyword "sturm" inside the object's declaration to tell POV-Ray to use its more accurate (but slower) Sturmian Root Solver. Another possible fix is to rotate, translate, or scale the object by a very small amount.

Reference Index Top of Document Beginning of the Tutorial

The Online POV-Ray Tutorial © 1996 The Online POV-Ray Tutorial ThinkQuest Team