VRML Class Summary

History of VRML

Capabilities

• Selective Refinement
• Hyperlinks
• Texture Maps
• Level of Detail

System Requirements

The Future of VRML

• VRML+
• Embedded VRML
• Java

VRML Browsers

• SDSC repository

Introduction to the Webspace Browser

• Examiner Viewer
• Walk Viewer

Setting up your VRML environment

• VRML mime-type
• VRML mailcap for unix

Converting 3D data into VRML

• Repository site for VRML convertors
• DXF to VRML (from Intergraph)
• Inventor to VRML

Basic VRML File Editing

• Adding lights
• Adding Hyperlinks to Objects

Details of VRML File Format

• What is an Node.
• What is a Scene Graph
• Types of nodes
  • Objects (spheres, cones, cubes, text)
  • 3D primitives (lines, points, normals)
  • Materials
  • Lights
  • Groups/Separators
  • Transformations
  • Control
  • Cameras
• How does the renderer proceed through the scene.
  • Propagation of Transformations
  • Propagation of materials.
  • Control structures
  • Shape Types
• Exercises in building VRML files by hand.
  • Simple cone.
  • Change materials of cone
    • diffuse color
    • specular color
  • Change lighting of cone
    • types of lights
    • light properties
  • Making the cone a hyperlink
  • Applying a texture map

Other Examples

VRML Class

History of VRML

The VRML 1.0 standard was developed during discussions via an electronic mailing list involving representatives from industry and the web community and organized during the 1993 web conference. The first VRML browser, WebSpace, was unveiled during the 1995 web conference in Darmstadt Germany.

Capabilties

• Selective Refinement (WWW Inlines) (Palladium) and (stonehenge):
  It would take impractically long to download very complex scenes. Having all of the scene data at the same time would also severely impact rendering performance. VRML has the ability to download portions of the scene from the webserver independently so that sections of the scene can be selectively refined as you get close enough to see the details. This capability is works hand-in-hand with Level-of-Detail.
• Level of Detail: This has a two-fold purpose; to reduce network load (when used with WWWInlines) and also enable some browsers to select different levels of detail to improve their drawing speed. The 3D rendering capabilites can differ greatly from machine-to-machine. Slower machines can select a lower level-of-detail so that they can render the objects faster.
• HyperLinks (Palladium):
  Similar to text hyperlinks and active-maps, objects in the 3D VRML scenes can be hyperlinked to other VRML files or web pages. When you go inside of the palladium, you will notice that some of the objects highlight when you hold the mouse cursor over them. These are 3D hyperlinks that can be activated when you click on them.
• Texture Mapping (Paragraph):
  Applying images to surfaces of 3-D objects to give them texture and realistic detail.

System Requirements for VRML browsing

The Future of VRML

• VRML Worlds Ultimate 3D Chat Plus : A collaborative form of VRML being developed by Worlds Inc. Normally viewing a VRML geometry is a private experience. With VRML Worlds Ultimate 3D Chat Plus, each of the people entering a VRML scene can see each other.
• Embedded VRML: Most popular web browsers have plug-in's to embed the VRML viewer inside of the Web browser so you can embed geometries inside of your documents. You will see much more of this sort of embedding with Active-X and OpenDoc/Cyberdog containers for VRML viewers which will allow VRML scenes to be embedded in text documents and even your machine's desktop.
• Hot Java: Java is a machine-independent interpreted object-oriented language developed by Sun Microsystems. The web as it is now allows sharing of data (hypertext, images, video, geometries...etc.) Java allows small programs ("applets") to be shared on the Web. Not only would you be able to dynamically download your VRML viewer with the 3D scene it is intended to view, you could also make your 3D objects active programs. This also ties in with collaborative VRML since Java-based collaborative development frameworks like Habenero will allow you to create hybrid Java/VRML based collaborative environments.

VRML Browsers

The browser we used for example here is SGI/TGS Webspace which is integrated with SGI's Web Browser. Template Graphics Software (TGS), which has been discontinued, developed the SGI Webspace versions, which has been replace by Cosmo.

Introduction to the WebSpace Browser

The browser has two modes of operaton which are selected under the "options" menu (image)

• Examiner viewer
  

  Used for viewing objects.

  It has a globe which you can use to spin the object around and move it around the window.

  The pad with the arrows on it on the left moves you up,down, left, or right without rotating you as the handebars would.

  The crosshairs allow you to select the center of rotation for the scene.

  The thumbwheel to the right of the globe zooms you in and out of the scene.

• Walk viewer
  

  Used to walk through architectural models.

  The handlebars act like the controls of a motor-scooter that flys through the scene. It is not able to sense the location of walls or even the floor, so nothing will stop you from walking right through them.

  The pad with the arrows on it on the left moves you up,down, left, or right without rotating you as the handebars would.

  The crosshairs allow you to select a location to fly towards automatically.

  The Dial on the right side of the handlebars controls your vertical angle-of-view. (ie. looking up or down)

Setting up your System Environment for VRML

VRML mimetype

The following line in your .mimetypes file in Unix identifies .wrl/.vrml files as being this mimetype.

x-world/x-vrml exts=wrl,vrml

In addition, you need to add this mime type to your http server so that it will serve up the file as the proper mime-type. For the NCSA httpd, you can find the mime.types file in the conf subdirectory. You would need to add the following entry.

x-world/x-vrml wrl vrml The installation script for many VRML browsers will do this automatically.

VRML mailcap

x-world/x-vrml; /usr/sbin/webspace -remote %s ; \
description="VRML document";

With Macintosh and Windows Web browsers, you change your mailcap from within the browser using the preferences menu.

Converting Model Files to VRML

• DXF to VRML
  DXF is an old file format for interchange of 3D models. It is most commonly associated with AutoCad, however, many other CAD systems support it.
• Inventor to VRML
  Which converts from SGI's Inventor 3D model file format.

Basic VRML File Editing

Following are some common changes that you need to make to VRML files that don't require very detailed knowledge of the file format.

Adding lights

Sometimes a model you have imported from another file format appears black when you view it using the VRML browser (sample). This is because the VRML file lacks light sources. (note: the SGI Webspace browser will automatically insert lights if none are present in the file) A point light source provides the simplest way to light a scene (sample). To add a point light, add the following lines to the head of a VRML file.

PointLight {
	location 0 0 1
	color 1 1 1
}

This will add a white light to the scene at location x=0,y=0,z=0 in 3D space, thus illuminating any objects in the scene. There are many other lighting styles availible.

Adding Hyperlinks to Objects

This requires that we get into our files with more depth. We will use the following simplified demo file.

#VRML V1.0 ascii
Separator {
        PerspectiveCamera {
            position 0 0 10
        }
	#light the scene
	DirectionalLight { #default color and location
            direction 0 0 -1
        }
	#Draw 3 spheres side-by-side
	#The sphere in the middle is half the
	#size of the ones on the end.

	# Draw First Sphere is in the Center
	Sphere {
		radius 0.5
	}
	Translation { # move to the left by 1
		translation -4 0 0
	}
	# Draw Second Sphere usng above translation
	Sphere {
		radius 1.0
	}
	Translation { #now move right by 2
		# Which ends up being to the right of center by 1
		translation 8
	}
	#Draw the 3rd Sphere
	Sphere {
		radius 1.0
	}
}
	  

Now, we will make each of these spheres VRML hyperlinks to other URLs on the Web. To make an object a hyperlink, we encapsulate it with a WWWAnchor as follows

WWWAnchor {
	name "some URL"
	some object
}
	  

In this case, we will encapsulate each of the spheres with WWWAnchor's. For example, lets take the left sphere and make it a hyperlink to an html page using.

WWWAnchor {
	name "http://www.ncsa.uiuc.edu"
	#Draw the 2nd sphere
	Sphere {
		radius 1.0
	}
}
	  

And then make the second sphere a link to another VRML file using

 WWWAnchor {
	name "examples/LightedCone.wrl"
	#Draw the 2nd sphere
	Sphere {
		radius 1.0
	}
}
	  

When you reload the scene, you will notice that the leftmost sphere will highlight if you hold the mouse pointer over it. If you click on it, you will go to that URL.

More advanced Hyperlinks.

It is important to be able to manipulate VRML files without actually understanding their content. Often the content was computer- generated, so you cannot hope to completely understand it. In this example, we will take two computer-generated models, combine them into one scene, and then hyperlink them.

Starting with fred.iv and car.iv.

1. Convert fred.iv into a VRML file using ivtovrml
   ivToVRML -o fred.wrl fred.iv
   Convert car.iv using the same utility
2. Copy fred.wrl to a new file called both.wrl as follows
   cp fred.wrl both.wrl
   
3. Now we will combine the two files by cutting and pasting in an editor. On your SGI's commandline type;
   jot both.wrl
   then type
   jot car.wrl
   Copy the entire contents of the car.wrl editing window and paste it at the beginning of the both.wrl editing window. Then save the result, but don't exit the editors.
4. Open your browser to the both.wrl file. You will notice that the two objects are on top of one-another. We need to move the car to the left some to get it out of the way of the space-station. This is done by adding the following to the both.wrl file.
   

   	Translation {
   		translation 1 0 0
   	}
   	      

   
   Save both.wrl and reload it in Webspace. Now the objects are separated. 5) Now we will make each object a hyperlink to the individual object. Where the fred.wrl data begins in both.wrl, add the following

   WWWAnchor {
   	name "file://fred.wrl"
   	.....< the fred data>.....
   }
   	      

   
   Do the same for the car.wrl data in the both.wrl file using the url file://car.wrl.
   
5. Now when you reload the file both.wrl, each of the objects are hyperlinked to the .wrl file with the individual object. This would be a rudimentary method of creating a gallery of selectable objects.

Details of VRML File Format

What is an Node.

A VRML node is an individual object or operation in the 3D model. For example, a node can be a light, an object like a cube, or an operation like a rotation.

The VRML textual representation of a node is of the form
name
{

some data

......

}

The name identifies the node type. The contents (some data) can be descriptions of parameters for the node or other "nested" nodes. Paramters can be one of the following

Single-valued Property

propertynamevalue
such as the radius of a sphere
radius 1.0
or such as a URL
name "http://www.anywhere.com"

Mult-valued Property
propertyname value1 value2 .... valueN
such as a location in 3D
location 0 1 1

List of Properties propertyname [ value1a value2a value3a, value1b value2b value3b ]
such as a list of points in 3D space.
point [ 0 0 0, 0 0 1, 0 -3 0, 1 1 1 ]

What is a Scene Graph

A scene graph is a heirarchial description of the objects and operations in the VRML model. The term, graph, comes from the tree-like organization of objects in the model. The effect of operations is restricted only to objects in the same branch of the tree and all of their children.

Types of Inventor Nodes

• Objects (spheres, cones, cubes)
• 3D primitives
• Materials
• Lights
• Groups/Separators
• Transformations
• Control

How does the renderer proceed through the scene

• Propagation of Transformations
• Propagation of materials.
• Control structures
• Shape Types

Exercises in building VRML files by hand

Simple cone

1. Begin with the following scene graph

   #VRML V1.0 ascii
   Separator {
   	#PerspectiveCamera {}
   	DirectionalLight {
   	     direction 0 0 -1
   	}
   	Cone {
   		bottomRadius 1
   		height 2
   	}
   }
   	  

   
   
2. Change materials of cone Now we will change the material that the cone is made of using a "Material" node. Insert the material node right after the "PointLight". The following settings will made the node appear to be a bright red.
   

   Material {
   	diffuseColor 0.8 0 0 # r=.8 g=0 b=0 
   }
   	  

   
   

   diffuse color:

   This is the flat (matte) reflective color of the material.

   ambient color:

   This is the color of the material when unlighted (ie. the material will appear to glow) in the absence of light.

   specular color/shininess:

   The specular color of a material is the color of the highlights of a shiny/metallic surface. Hence, the shininess of the material and its specular color are linked. The shininess determines how pronounced the highlights are while the specular color determines the color of those highlights.

   Change lighting of cone

   Currently, the cone is lighted with a point-light. There are a wide variety of lights availible to duplicate real-word situations

   Types of lights

   PointLight:

   Similar to a bare lightbulb.

   DirectionalLight:

   Similar to a flashlight. Illuminates as a tube of light.

   SpotLight:

   Similar to a tracklight, floodlight, or lamp. Illuminates in a cone of light.

   Light properties

   This depends on the light. We will try lighting the cone with three different light sources. First, we will experiment with the PointLight.
   
   1. Make the object white again by changing the material properties to
      diffuseColor 1 1 1
   2. The point light has only 4 properties (on,intensity,color,location), really, only color and location are important. Lets change the color of the light to blue by adding the property
      color 0 0 1
   3. We can also move the light to a different location by adding the property
      location 10 0 0
      This places the light to the right of the cone.

   Using the SpotLight

   In addition to the parameters availible to the PointLight, the parameters of the SpotLight node basically define the dimensions and orientation of a cone using location, direction, and cutOffAngle.
   1. Replace the PointLight node with the following SpotLight node.
      

      SpotLight{
      	location 0 4 0     # Above the object by 4 units
      	direction 0 -1 0   # point downward
      	cutOffAngle 0.7
      }
      	  

   Applying a texture map

   Now we can investigate how to apply an image to these 3D surfaces. Texture maps can be used to improve the realism of a relatively simple scene.

   1. Begin with the following scene graph

      #VRML V1.0 ascii
      Separator {
      	PointLight { 
      		location 4 4 4
      	}
      	PointLight {
      		location -8 -8 -8
      		color .2 .2 .8   #blue light
      	}
      	Cube {
      		width 1
      		height 1
      		depth 2
      	}
      	Translation {
      		translation 3 0 0
      	}
      	Sphere{ radius 1 }
      }
            

   2. We need to have an image to apply to the surface. We will use "capture" to get ourselves an appropriate image. Type
      capture
      at the sgi commandline. Find some interesting pattern on the screen and grab it by pressing on the "sweep" button and selecting an area of the screen. Try to select a relatively square region.

      In practice, you will want to convert your texturemaps to jpeg or gif since not all machines support rgb images.

   3. Apply the texturemap to the the cube by using the Texture2 node before the Cube node.

      Texture2 {
      # for texture maps, you need to use the absolute path name to the
      # image
      	filename "file://texture.rgb"
      }
      	    

   4. Load the VRML scene.

   ---

   Other Examples

   ---

   maintained by John Shalf
   jshalf@ncsa.uiuc.edu