Open Inventor supplies a set of classes designed for real-time rendering of very large models. These tools can be divided into several categories:
Geometry simplification | Also known as mesh decimation, this technique allows parts of the geometry (triangles) to be removed from the shape before sending it to the graphics pipeline. The level of decimation (number of resulting triangles) can be controlled by the user. |
Automatic level of detail | Open Inventor already includes an LOD node (SoLOD). The geometry simplification feature allows Open Inventor to automatically build different levels of decimation as children of LOD nodes. |
Level of simplification node | This is a specialized LOD that also stores the decimation percentage for each child, which makes it very useful for maintaining frame rate. |
Adaptive viewer | The Open Inventor viewers have been extended to have a target frame rate. The viewer can force LOD and level of simplification nodes to use lower or higher levels of decimation. |
Get primitive count action | This action (SoGetPrimitiveCountAction) returns the number of triangles, lines, etc. in a scene graph or a path. |
Using multiple threads in an application can increase overall performance by making use of multiple processors, or better use of a single processor. This also enables use of multiple graphics pipes. Now, multiple graphics windows can each have their own rendering thread, multiplethreads can simultaneously traverse the same scene graph, and modification and traversal of the scene graph can safely occur in different threads. Open Inventor can also be used in immersive VR applications based on tools like CAVELib™ and Multipipe SDK™. Multithreading, gives you full details on multithreading.
The SoTrackerEvent and SoControllerButtonEvent classes have been designed to be useful in immersive VR applications that use tracking. For instance, a “wand” device can generate SoTrackerEvents when moving and SoControllerButtonEvents when a button is pressed. All classes that respond to SoMouseButtonEvent (i.e., SoSelection, SoExtSelection, etc.) also respond to SoControllerButtonEvents. All dragger classes also respond to SoTrackerEvents and SoControllerButtonEvents. So, for example, a wand input device can be used to interact with draggers and manipulators in an immersive environment. Immersive Virtual Reality discusses these events.
Open Inventor supports detection of collisions between the viewer (camera) and the scene, collisions between objects in the scene, as well as collsions between two scene graphs, one moving and one static. The following classes have been added to handle these kinds of collisions:
SoCollisionViewer | Prevents the camera from penetrating walls, floors, or other objects, making viewer control much more comfortable and realistic (“solid” camera). |
SoIntersectionDetectionAction | Analyzes a scene graph (or subgraph) to detect whether shapes are intersecting, and optionally return the intersections to the application. |
SoDualSceneCollider | Performs very fast collision detection for the specific case of two scene graphs, one static, and one moving. This class has been designed to detect collisions between very large scenes (i.e., containing millions of triangles) while still allowing interactive display of the two scenes. You can retrieve information about colliding triangles, coordinates of common points, and so forth. This is useful, for example, for walk-through of a complex factory environment. |
Collision Detection explains how to use these classes.
Open Inventor supports high quality stereo projection with different types of stereo modes to render your virtual model with high realism. Software stereo is supported as well as hardware stereo (OpenGL stereo). Furthermore, a new graphical user interface dialog box, accessible from the Stereo menu item of the viewer popup menu, allows the user to interactively select the stereo mode and make adjustments to various stereo parameters. Stereo Viewing , describes how to use and configure these different stereo types.
Open Inventor provides support for programmable vertex, geometry, and fragment (pixel) shaders, allowing you to take advantage of the latest generation of graphics cards to produce more advanced real-time rendering effects.
Programming in the following languages is supported (depending, of course, on support for these shader languages by your graphics card):
Remote rendering allows an Open Inventor application to both execute and render on a remote server machine. Only the optimized and compressed rendered UserGuide_Images are sent to the local machine. This allows very large data sets to be centralized and handled on a machine with more memory, processors, and graphics hardware.
Remote Rendering deals with this extension.
In addition to the actions for large model visualization and collision detection, the following new actions are available:
Open Inventor has followed the evolution of graphics boards and OpenGL in terms of texturing support. Furthermore, numerous other features have been added to provide more flexibility and improve performance. The following features have been added:
When creating your application with Open Inventor, you may need to use external dynamic libraries for specific purposes. However, the need for these libraries can depend on the use of the application. In order to save memory and loading time at start-up, you can load these dynamic libraries “on the fly” only when required.
The class SoDynamicLibManager manages the loading and the unloading of a dynamic library (cross-platform). You can query the list of all loaded libraries or whether a specific library is currently loaded. You can also inquire whether a specific library supports a particular function.
This class is used by Open Inventor to load the following libraries: Zlib, libJpeg, libTiff, libPng, libJasper, OpenAL (Windows only), FreeType.
The library name given to the load method can include an absolute or relative path. If no path is supplied, the library will be searched for in the following paths:
In addition to the extension nodes mentioned previously, the following new nodes have been added by FEI:
+ Annotation text, + Stroke (vector) 3D text, + Rendering mode for 3D text (outline and texture), + Text string orientation (left-to-right, top-to-bottom, etc.), + Horizontal and vertical justification, + Kerning.See @ref Inventor_Mentor_Core__Text__Chapter6Text).
+ Ability to write additional image formats. + Ability to render very large off-screen UserGuide_Images. + Pbuffer support to improve rendering speed and quality (the resulting image is more likely to match the on-screen image). + Ability to share OpenGL contexts to improve the rendering speed and reduce memory consumption.