An SoMaterial^{( C++ | Java | .NET )}node includes the following fields:

	Tip: The transparency type is specified in the render action (see Chapter 8, Applying Actions).

An example of setting values in an SoMaterial^{( C++ | Java | .NET )} node is the following:

SoMaterial *gold	  = new SoMaterial;
//Set material values
gold->ambientColor.setValue(.3, .1, .1);
gold->diffuseColor.setValue(.8, .7, .2);
gold->specularColor.setValue(.4, .3, .1);
gold->shininess = .4;
    

SoMaterial gold	  = new SoMaterial();
//Set material values
gold.ambientColor.SetValue(.3f, .1f, .1f);
gold.diffuseColor.SetValue(.8f, .7f, .2f);
gold.specularColor.SetValue(.4f, .3f, .1f);
gold.shininess.SetValue(.4f);
    

SoMaterial gold	  = new SoMaterial();
//Set material values
gold.ambientColor.setValue(.3f, .1f, .1f);
gold.diffuseColor.setValue(.8f, .7f, .2f);
gold.specularColor.setValue(.4f, .3f, .1f);
gold.shininess.setValue(.4f);
    

Since gold is opaque, you can use the default value of 0.0 for the transparency field.

SoBaseColor^{( C++ | Java | .NET )}, another class derived from SoNode^{( C++ | Java | .NET )}, replaces only the diffuse color field of the current material and has no effect on other material fields.

Tip: If you are changing only the diffuse color of an object, use an SoBaseColor( C++ | Java | .NET ) node in place of an SoMaterial( C++ | Java | .NET ) node. For example, to represent a complex terrain that uses many different diffuse colors, use one SoMaterial( C++ | Java | .NET ) node for the ambient, specular, and emissive color values, and then use one SoBaseColor( C++ | Java | .NET ) node with multiple values for the changing diffuse colors. The SoBaseColor( C++ | Java | .NET ) class is also useful when the light model is BASE_COLOR (see the section called “Light-Model Node”).

An SoDrawStyle^{( C++ | Java | .NET )} node includes the following fields:

	Tip: Draw-style LINES and POINTS look best with a BASE_COLOR lighting model.

Figure 5.17, “ Drawing Styles (FILLED, LINES, POINTS)” shows the same object rendered in different drawing styles.

Figure 5.17.  Drawing Styles (FILLED, LINES, POINTS)

An SoLightModel^{( C++ | Java | .NET )} node includes the following field:

	Note: In Inventor, shading (such as Gouraud or flat) is dictated by the combination of the material specification of the object, the lighting model, and the normal bindings. A shading model is not explicitly specified.

Figure B.4, “ Plate 4 ” and Figure B.5, “ Plate 5 ” show the same scene with the different lighting models. (Figure B.4, “ Plate 4 ” uses BASE_COLOR, and Figure B.5, “ Plate 5 ” uses PHONG.)

SoMaterial^{( C++ | Java | .NET )} and SoBaseColor^{( C++ | Java | .NET )} can be used along with any drawing style and any lighting model. In some cases, however, some of the material attributes might be ignored. For example, if you specify BASE_COLOR for the SoLightModel model field, only the diffuse color and transparency of the current material are used. But what happens if you specify only a base color (with SoBaseColor^{( C++ | Java | .NET )}) and subsequently select the Phong lighting model for SoLightModel^{( C++ | Java | .NET )}? In this case, Inventor uses the base color for the diffuse color and the default or current material element values for the other SoMaterial^{( C++ | Java | .NET )} fields.

	Note: By default, the light model is PHONG. For images to render correctly, you need to specify normals and light sources. If you want to see only colored objects, change the light model to BASE_COLOR and use SoBaseColor( C++ | Java | .NET ) to specify only the base (diffuse) color.

You can use the SoEnvironment^{( C++ | Java | .NET )} node to simulate various atmospheric effects such as fog, haze, pollution, and smoke. For general purposes, these atmospheric effects are grouped under the term fog. The difference between fog and haze, for example, is simply the color and density.

Specifically, the SoEnvironment^{( C++ | Java | .NET )} node allows you to specify the color and intensity of the ambient lighting, the light attenuation for point lights and spotlights, and the type, color, and visibility factor for fog. Figure B.6, “ Plate 6 ” shows the effects of an SoEnvironment^{( C++ | Java | .NET )} node. This image uses a value of FOG for the fog type. The fogColor is (0.2, 0.2, 0.46).

An SoEnvironment^{( C++ | Java | .NET )}node includes the following fields:

	Tip: For realistic scenes, clear the window to the fog color before drawing the fogged objects (see the SoXtRenderArea( C++ )::setBackgroundColor() method.)

By default, Inventor does not assume anything about how the vertices in a vertex shape are ordered, whether its surface is closed or open, or whether the faces of the shape are convex or concave. If you know that the vertices are in a consistent order, that the shape is closed, or that the shape faces are convex, you can use the SoShapeHints^{( C++ | Java | .NET )} node to notify Inventor so that it can optimize certain rendering features.

The SoShapeHints^{( C++ | Java | .NET )} node has four fields:

If the shapeType is SOLID and the vertexOrdering is either CLOCKWISE or COUNTERCLOCKWISE, Inventor turns on backface culling and turns off two-sided lighting. If the shapeType is not SOLID and the vertexOrdering is either CLOCKWISE or COUNTERCLOCKWISE, Inventor turns off backface culling and turns on two-sided lighting. In all other cases, backface culling and two-sided lighting are both off. If you use the SoShapeHints^{( C++ | Java | .NET )} node, be sure to describe the object accurately; otherwise, objects may be rendered incorrectly.

	Tip: In general, the more information you specify with the shape-hints node, the faster the rendering speed. The exception to this rule is when shapeType is not SOLID and the vertexOrdering is either CLOCKWISE or COUNTERCLOCKWISE. In this case, rendering may be slower because two-sided lighting is automatically turned on and backface culling is turned off.

Use the SoComplexity^{( C++ | Java | .NET )} node to indicate the amount of subdivision into polygons for subsequent shape nodes in the scene graph. This node has three fields:

Figure 5.18, “ Specifying Different Levels of Complexity (left: OBJECT_SPACE; right: SCREEN_SPACE)” shows the same object with different levels of complexity. The spheres at the left use object-space complexity and a complexity value of .5. The spheres at the right use screen-space complexity and a complexity value of .06. The NURBS examples in Chapter 20, Curves and Surfaces use the SoComplexity^{( C++ | Java | .NET )} node.

Figure 5.18.  Specifying Different Levels of Complexity (left: OBJECT_SPACE; right: SCREEN_SPACE)

	Tip: Simpler scenes render more quickly than complex scenes. For example, to increase rendering speed, use fewer lights, turn off textures or specify a lower texture-quality value, and choose a simpler drawing style, such as wireframe, and a lower complexity value. The viewer pop-up menu allows you to disable certain of these features for faster rendering.

Inventor lets you define your data in a variety of different units. It uses meters as its default units, but you can use the SoUnits^{( C++ | Java | .NET )} node to specify a different unit of measurement. The units node acts like a scale node by scaling subsequent shapes into the specified units. SoUnits^{( C++ | Java | .NET )} can adjust the amount it scales an object by checking to see if any other units have been defined. The units node adjusts the scale so that the previously defined units are no longer in effect.

The SoUnits^{( C++ | Java | .NET )} node has one field:

To render your data in units other than these, use an SoUnits^{( C++ | Java | .NET )} node to set the current units back to meters, followed by a scale node that scales from meters into the desired units.

It is often useful to draw multiple primitives on the same plane in 3D space. For example we might want to draw lines on top of a polygonal surface to show the edges of the faces that form the surface. We might also want to draw text or additional polygons on top of a polygon. In a perfect world co-planar geometry would be resolved by the default “less than or equal” depth test so that whatever is drawn last appears on top. Unfortunately this is not reliable in practice. Usually the result is the depth buffer artifacts commonly called “stitching”. Open Inventor provides two nodes that allow you to introduce a small depth offset in order to reliably separate co-planar geometry. They are SoPolygonOffset^{( C++ | Java | .NET )} and SoDepthOffset^{( C++ | Java | .NET )}.

The SoPolygonOffset^{( C++ | Java | .NET )} property node provides a convenient implementation of the OpenGL polygon offset feature (glPolygonOffset), which is typically implemented in hardware.

The styles field is a bit-wise combination of FILLED, LINES, and POINTS, and determines which polygon draw style mode should be offset. For instance, if styles = LINES, only polygons whose drawstyle mode is wireframe (SoDrawStyle::LINES) are offset.

The two fields, factor and units, are used to calculate an offset added to the depth value of each fragment of polygons. The offset value is computed as follows:

offset = m * factor + r * units

where r is an OpenGL implementation-specific constant representing the smallest resolvable unit of distance in the depth buffer and m is the maximum depth slope of the polygon. For polygons that are parallel to the near and far camera clipping planes (perpendicular to the view vector), the depth slope is zero. So for polygons in your scene graph with a depth slope near zero, a small value for units is usually sufficient to separate co-planar geometry. But for polygons that are angled away from the view vector, a larger offset is usually needed (see the diagrams in the SoDepthOffset^{( C++ | Java | .NET )} section). Because the depth slope increases with the angle of the polygon, you can use the factor field to add an additional offset that increases with the angle of the polygon. Because the angle of the polygon can change constantly as the camera and/or scene rotate, you should start by setting both units and factor to a small value, e.g. 1. Larger values may be required in some cases.

If the offset is positive, polygons are pushed away from the camera (drawn closer to the far plane). If the offset is negative, polygons are pulled toward the camera (drawn closer to the near plane). Therefore if we wanted to draw lines on top of a polygon, we could use an SoLineSet^{( C++ | Java | .NET )} for the lines (which is not affected by SoPolygonOffset^{( C++ | Java | .NET )}) and put an SoPolygonOffset^{( C++ | Java | .NET )} node with units and factor set to 1 before the polygon, as shown in the following example program.

Note that SoPolygonOffset^{( C++ | Java | .NET )} only affects polygonal geometry. It does not affect lines, text or other types of geometry. When an SoPolygonOffset^{( C++ | Java | .NET )} node is traversed, the values of its fields replace the current values in the traversal state. The polygon offset values are pushed and popped by SoSeparator^{( C++ | Java | .NET )} like other properties.

Example 5.7. How to use SoPolygonOffset

Without polygon offset

With polygon offset

This example shows a cube with a diagonal line on each face of the cube. The program source code can be found in:$OIVHOME/src/Inventor/examples/Features/PolygonOffset/polygonOffset.cxx.

On the left, you can see the artifacts (“stitching”) that appear when the SoPolygonOffset^{( C++ | Java | .NET )}node is not used.

---

C++

int
main(int argc, char **argv)
{
static SbVec3f vertices[12] =
  {
  SbVec3f(1, 1, 1),    SbVec3f(-1, -1, 1),  SbVec3f(1, 1, 1), 
  SbVec3f(-1, 1, -1),  SbVec3f(1, 1, 1),    SbVec3f(1, -1, -1), 
  SbVec3f(-1, -1, -1), SbVec3f(1, 1, -1),   SbVec3f(-1, -1, -1), 
  SbVec3f(1, -1, 1),   SbVec3f(-1, -1, -1), SbVec3f(-1, 1, 1)
  };

// Initialize Inventor and Xt
Widget myWindow = SoXt::init(argv[0]);

  SoSeparator *root = new SoSeparator;
  SoSeparator *sep = new SoSeparator;
  SoCube *cube = new SoCube;
  SoCoordinate3 *coords = new SoCoordinate3;
  coords->point.setValues(0, 12, (const class SbVec3f *)vertices);

  SoLineSet *lineSet = new SoLineSet;
  lineSet->numVertices.set1Value(0, 12);

  SoPolygonOffset *polygonOffset = new SoPolygonOffset;
  polygonOffset->factor = 1.;
  polygonOffset->units = 1.;

  SoMaterial *mat = new SoMaterial;
  mat->diffuseColor.setValue(1, 0, 0);

  SoSeparator *cubeSep = new SoSeparator;
  cubeSep->addChild(cube);
  cubeSep->addChild(mat);
  cubeSep->addChild(coords);
  cubeSep->addChild(lineSet);

  root->ref();
  root->addChild(sep);

  sep->addChild(polygonOffset);
  sep->addChild(cubeSep);

// Create a viewer
  SoXtExaminerViewer *myViewer = new SoXtExaminerViewer(myWindow);

// attach and show viewer
  myViewer->setSceneGraph(root);
  myViewer->setTitle("Polygon Offset");
  myViewer->show();

// Loop forever
  SoXt::show(myWindow);
  SoXt::mainLoop();
  return 0;
}
               

---

.NET

SbVec3f[] vertices = new SbVec3f[12] 
{
  new SbVec3f(1, 1, 1),    new SbVec3f(-1, -1, 1),
  new SbVec3f(1, 1, 1),    new SbVec3f(-1, 1, -1), 
  new SbVec3f(1, 1, 1),    new SbVec3f(1, -1, -1),
  new SbVec3f(-1, -1, -1), new SbVec3f(1, 1, -1), 
  new SbVec3f(-1, -1, -1), new SbVec3f(1, -1, 1), 
  new SbVec3f(-1, -1, -1), new SbVec3f(-1, 1, 1)
};

// Initialize Inventor

SoSeparator root = new SoSeparator();
SoSeparator sep = new SoSeparator();
SoCube cube = new SoCube();

SoCoordinate3 coords = new SoCoordinate3();
coords.point.SetValues(0, vertices);

SoLineSet lineSet = new SoLineSet();
lineSet.numVertices[0] = 12;

SoPolygonOffset polygonOffset =  new SoPolygonOffset();
polygonOffset.factor.Value = 1f;
polygonOffset.units.Value = 1f;

SoMaterial mat = new SoMaterial();
mat.diffuseColor.SetValue(1, 0, 0);

SoSeparator cubeSep = new SoSeparator();
cubeSep.AddChild(cube);
cubeSep.AddChild(mat);
cubeSep.AddChild(coords);
cubeSep.AddChild(lineSet);

root.AddChild(sep);

sep.AddChild(polygonOffset);
sep.AddChild(cubeSep);   

// Create a viewer
SoWinExaminerViewer myViewer = new SoWinExaminerViewer(this, "", true,
	    SoWinFullViewer.BuildFlags.BUILD_ALL, SoWinViewer.Types.BROWSER);

// attach and show viewer
myViewer.SetSceneGraph(root);
myViewer.SetTitle("Polygon Offset");
myViewer.Show();
            

---

Java

SbVec3f[] vertices = new SbVec3f[]
{
  new SbVec3f(1, 1, 1),    new SbVec3f(-1, -1, 1),  new SbVec3f(1, 1, 1),
  new SbVec3f(-1, 1, -1),  new SbVec3f(1, 1, 1),    new SbVec3f(1, -1, -1),
  new SbVec3f(-1, -1, -1), new SbVec3f(1, 1, -1),   new SbVec3f(-1, -1, -1),
  new SbVec3f(1, -1, 1),   new SbVec3f(-1, -1, -1), new SbVec3f(-1, 1, 1)
};

// Initialize Inventor and Xt

SoSeparator root = new SoSeparator();
SoSeparator sep = new SoSeparator();
SoCube cube = new SoCube();

SoCoordinate3 coords = new SoCoordinate3();
coords.point.setValues(0, vertices);

SoLineSet lineSet = new SoLineSet();
lineSet.numVertices.set1Value(0, 12);

SoPolygonOffset polygonOffset = new SoPolygonOffset();
polygonOffset.factor.setValue(1f);
polygonOffset.units.setValue(1f);

SoMaterial mat = new SoMaterial();
mat.diffuseColor.setValue(1, 0, 0);

SoSeparator cubeSep = new SoSeparator();
cubeSep.addChild(cube);
cubeSep.addChild(mat);
cubeSep.addChild(coords);
cubeSep.addChild(lineSet);

root.addChild(sep);

sep.addChild(polygonOffset);
sep.addChild(cubeSep);

// Create a viewer
myViewer = new SwSimpleViewer(SwSimpleViewer.EXAMINER);
// attach and show viewer
myViewer.setSceneGraph(root);

myViewer = new SwSimpleViewer();
myViewer.setSceneGraph(root);
myViewer.setName("Polygon offset");
myViewer.viewAll();

panel.add(myViewer);
add(panel);

The SoDepthOffset^{( C++ | Java | .NET )} node can also be used to resolve depth buffer artifacts (“stitching”) for co-planar geometry. This node uses a different algorithm (based on Lengyel’s method from Game Programming Gems) in which the projection matrix is modified to offset subsequent geometry. As a result it has advantages over SoPolygonOffset^{( C++ | Java | .NET )} in some cases. Specifically, SoDepthOffset^{( C++ | Java | .NET )} applies its offset value to all types of geometry, not just polygons, and also does not require any additional per-vertex calculations on the GPU.

Note! The most important difference between SoPolygonOffset^{( C++ | Java | .NET )} and SoDepthOffset^{( C++ | Java | .NET )} is the direction of the offset. For SoPolygonOffset^{( C++ | Java | .NET )} a positive value moves the geometry away from the camera (farther in depth). For SoDepthOffset^{( C++ | Java | .NET )} a positive value moves the geometry toward the camera (closer in depth).

SoDepthOffset^{( C++ | Java | .NET )} has two other differences. First, since SoDepthOffset^{( C++ | Java | .NET )} requires the projection matrix to be modified during traversal, it may prevent building a render cache for part of the scene graph. Second, the offset does not take into account the depth slope of the geometry, so a larger offset may be required for geometry that is not perpendicular to the view vector.

The render caching issue is easily handled, by adjusting your scene graph structure, just as you would for any non-cacheable node, to ensure that the actual shape nodes are cached even if the parent group node cannot cache:

shapes are not cached Figure 5.21. Not recommended structure:

shapes are cached Figure 5.22. Recommended structure:

Understanding Depth offset:

Consider two shapes A and B that have a co-planar (front) face.

Front view

Top view

Defining a positive offset for the blue shape will cause it to be moved toward the camera. The visual artifact for coplanar faces is removed.

An important point to keep in mind when using the SoDepthOffset^{( C++ | Java | .NET )} node is that only a simple depth modification is done. The depth slope of the geometry is not considered. As a consequence, the full offset is realized when a polygon face is perpendicular to the view vector, but while rotating the scene, the offset effectively becomes smaller as the face becomes closer to parallel to the view vector.

Figure 5.23. depthOffset_offset