Click or drag to resize
SoVolumeRender Class

Renders a data volume using direct volume rendering.

Inheritance Hierarchy

Namespace: OIV.VolumeViz.Nodes
Assembly: OIV.VolumeViz (in OIV.VolumeViz.dll) Version: 9.9.19.0 (9.9.19.0)
Syntax
public class SoVolumeRender : SoVolumeShape

The SoVolumeRender type exposes the following members.

Constructors
  NameDescription
Public methodSoVolumeRender

Constructor.

Top
Methods
  NameDescription
Public methodAffectsState

Overrides default method on SoNode.

(Inherited from SoShape.)
Public methodBeginShape(SoAction, SoShapeTriangleShapes)
Calls BeginShape(action, shapeType, (OIV.Inventor.Details.SoFaceDetail ^)nullptr).
(Inherited from SoShape.)
Public methodBeginShape(SoAction, SoShapeTriangleShapes, SoFaceDetail)

These methods can be used by subclasses to generate triangles more easily when those triangles are part of a larger structure, such as a triangle strip, triangle fan, or triangulated polygon, according to the TriangleShape enumerated type.

(Inherited from SoShape.)
Public methodCallback

Implements primitive generation for all shapes.

(Inherited from SoShape.)
Public methodCallShouldGLRender Obsolete.
Calls CallShouldGLRender(action, false).
(Inherited from SoShape.)
Public methodComputeBBox(SoAction, SbXfBox3d, SbVec3d)

Compute object oriented bounding box (OOB) for subclass using information in the given action (which may not necessarily be an SoGetBoundingBoxAction).

(Inherited from SoShape.)
Public methodComputeBBox(SoAction, SbBox3f, SbVec3f)
Public methodCopy
Calls Copy(false).
(Inherited from SoNode.)
Public methodCopy(Boolean)

Creates and returns an exact copy of the node.

(Inherited from SoNode.)
Public methodCopyFieldValues(SoFieldContainer)
Calls CopyFieldValues(fc, false).
(Inherited from SoFieldContainer.)
Public methodCopyFieldValues(SoFieldContainer, Boolean)

Copies the contents of fc's fields into this object's fields.

(Inherited from SoFieldContainer.)
Public methodDispose
Releases all resources used by SoDisposable.
(Inherited from SoDisposable.)
Public methodDistribute
(Inherited from SoNode.)
Public methodDoAction
(Inherited from SoNode.)
Public methodEnableNotify

Notification at this Field Container is enabled (if flag == true) or disabled (if flag == false).

(Inherited from SoFieldContainer.)
Public methodEndShape

end shape previously started with beginShape().

(Inherited from SoShape.)
Public methodEquals
Determines whether the specified Object is equal to the current Object.
(Inherited from Object.)
Public methodFieldsAreEqual

Returns true if this object's fields are exactly equal to fc's fields.

(Inherited from SoFieldContainer.)
Public methodGet

Returns the values of the fields of this object in the Open Inventor ASCII file format in the given string.

(Inherited from SoFieldContainer.)
Public methodGetAllFields

Returns a list of fields, including the eventIn's and eventOut's.

(Inherited from SoFieldContainer.)
Public methodGetAlternateRep

This method is called by actions to allow the node to provide an "alternate representation" when appropriate (typically depending on the action type).

(Inherited from SoNode.)
Public methodGetBoundingBox

Implements bounding box method using virtual computeBBox() method.

(Inherited from SoVolumeShape.)
Public methodGetEventIn

Returns a the eventIn with the given name.

(Inherited from SoFieldContainer.)
Public methodGetEventOut

Returns the eventOut with the given name.

(Inherited from SoFieldContainer.)
Public methodGetField

Returns a the field of this object whose name is fieldName.

(Inherited from SoFieldContainer.)
Public methodGetFieldName

Returns the name of the given field in the fieldName argument.

(Inherited from SoFieldContainer.)
Public methodGetFields

Appends references to all of this object's fields to resultList, and returns the number of fields appended.

(Inherited from SoFieldContainer.)
Public methodGetHashCode
Overrides GetHashCode().
(Inherited from SoNetBase.)
Public methodGetMatrix
(Inherited from SoNode.)
Public methodGetName

Returns the name of an instance.

(Inherited from SoBase.)
Public methodGetPrimitiveCount

Counts number of primitives produced by this shape.

(Inherited from SoShape.)
Public methodGetRenderUnitID
(Inherited from SoNode.)
Public methodGetShapeType

Gets the current shape Full Scene Antialiasing type.

(Inherited from SoShape.)
Public methodGetStringName (Inherited from SoBase.)
Public methodGetType
Gets the Type of the current instance.
(Inherited from Object.)
Public methodGLRender
(Inherited from SoVolumeShape.)
Public methodGLRenderBelowPath
(Inherited from SoNode.)
Public methodGLRenderInPath
(Inherited from SoNode.)
Public methodGLRenderOffPath
(Inherited from SoNode.)
Public methodGrabEventsCleanup
(Inherited from SoNode.)
Public methodGrabEventsSetup
(Inherited from SoNode.)
Public methodHandleEvent
(Inherited from SoNode.)
Public methodHandleInvisibleOrBBox Obsolete.
Calls HandleInvisibleOrBBox(action, false).
(Inherited from SoShape.)
Public methodHasDefaultValues

Returns true if all of the object's fields have their default values.

(Inherited from SoFieldContainer.)
Public methodIntersect

Returns true if the shape intersects the given tile.

(Inherited from SoLdmShape.)
Public methodIsBoundingBoxIgnoring

This method is used by getBoundingBox action traversal to know if the current node must be traversed or not, ie the bounding should be ignored.

(Inherited from SoNode.)
Public methodIsNotifyEnabled

Notification is the process of telling interested objects that this object has changed.

(Inherited from SoFieldContainer.)
Public methodIsOverride

Returns the state of the override flag.

(Inherited from SoNode.)
Public methodIsSynchronizable

Gets the ScaleViz synchronizable state of this object.

(Inherited from SoBase.)
Public methodPick
(Inherited from SoNode.)
Public methodRayPick

Implements picking along a ray by intersecting the ray with each primitive generated by subclass.

(Overrides SoShapeRayPick(SoRayPickAction).)
Public methodSearch
(Inherited from SoNode.)
Public methodSet

Sets one or more fields in this object to the values specified in the given string, which should be a string in the Open Inventor file format.

(Inherited from SoFieldContainer.)
Public methodSetAbortCallback(SoVolumeRenderRenderAbortCB) Obsolete.
Use AbortCallback property instead.
Public methodSetAbortCallback(SoVolumeRenderRenderAbortUserDataCB, Object) Obsolete.
Use AbortCallback property instead.
Public methodSetName (Inherited from SoBase.)
Public methodSetOverride

Turns the override flag on or off.

(Inherited from SoNode.)
Public methodSetRenderProgress

Set an application defined SoProgressIndicator object which will raise an event before and after the volume rendering task, before and after each subtask (in this case: Texture creation and Geometry creation) and after each step in the subtasks which represents in most cases individual tiles of data.

Public methodSetShapeType

set the antialiasing type for this shape.

(Inherited from SoShape.)
Public methodSetSynchronizable

Sets this to be a ScaleViz synchronizable object.

(Inherited from SoBase.)
Public methodSetToDefaults

Sets all fields in this object to their default values.

(Inherited from SoFieldContainer.)
Public methodShapeVertex

add a primitive vertex to the shape prevously started with beginShape().

(Inherited from SoShape.)
Public methodShouldGLRender Obsolete.
Calls ShouldGLRender(action, false).
(Inherited from SoShape.)
Public methodShouldGLRenderCore(SoGLRenderAction) Obsolete.
Calls ShouldGLRenderCore(action, false, false).
(Inherited from SoShape.)
Public methodShouldGLRenderCore(SoGLRenderAction, Boolean) Obsolete.
Calls ShouldGLRenderCore(action, isPointsOrLines, false).
(Inherited from SoShape.)
Public methodToString
Converts this SoBase structure to a human readable string.
(Inherited from SoBase.)
Public methodTouch

Marks an instance as modified, simulating a change to it.

(Inherited from SoNode.)
Public methodWrite
(Inherited from SoNode.)
Top
Properties
  NameDescription
Public propertyAbortCallback

Sets delegateto call during texture map rendering to test for an abort condition.

Public propertyboundingBoxIgnoring

Whether to ignore this node during bounding box traversal.

(Inherited from SoShape.)
Public propertycomposition Obsolete.

Specifies color composition mode.

(Inherited from SoVolumeShape.)
Public propertyfixedNumSlicesInRoi

When this field is set to false (the default), the number of samples set by numSlices is the number of samples used for the region defined by the current ROI.

Public propertygpuVertexGen Obsolete.

If true, slice vertices used for volume rendering are generated by the GPU instead of the CPU.

Public propertyinterpolation

Interpolation mode.

(Inherited from SoVolumeShape.)
Public propertyIsDisposable
ISafeDisposable interface implementation.
(Inherited from SoDisposable.)
Public propertylightDirection Obsolete.

Light direction (relative to the volume).

Public propertylighting Obsolete.

Indicates if lighting is required.

Public propertylightIntensity Obsolete.

Light intensity in the range [0-1].

Public propertylowResMode

Sets the method to use when moving in low resolution.

Public propertylowScreenResolutionScale

If lowResMode is DECREASE_SCREEN_RESOLUTION, render the volume at a lower screen resolution.

Public propertynumEarlyZPasses Obsolete.

When useEarlyZ is true, pixels are tested for full opacity every N samples, where N is the value of this field.

Public propertynumSlices

Specifies the number of samples along each ray.

Public propertynumSlicesControl

Controls how the number of samples along each ray is determined.

Public propertyopacityCorrection

Controls whether opacity correction is done.

Public propertyopacityThreshold

Specifies a threshold opacity (alpha) value that defines voxels considered to be "solid" (non-transparent).

Public propertyprojectedTileSubdivision

When doing volume projection (see SoProjection), only the geometry (corner vertices) of the LDM tiles are projected, not the individual voxels.

Public propertyrenderMode

Specifies the composition mode.

Public propertysamplingAlignment

Specifies which technique to use to align rayCast samples.

Public propertysubdivideTile
Public propertyuseEarlyZ Obsolete.

If true, stop computing when the pixel's cumulative opacity reaches 1.

Public propertyUserData
Gets or sets the user data to be contained by the field container.
(Inherited from SoFieldContainer.)
Public propertyviewAlignedSlices Obsolete.
Top
Remarks

This node renders volume data using "direct volume rendering". The volume process involves sampling, interpolation, classification and composition. By default the rendering algorithm is a GPU-based technique called raycasting. Raycasting has similarities to the well known algorithm called raytracing, but is specialized for volume data (renders voxels not triangles) and does not currently implement reflection or refraction. One or more "rays" are cast through the volume from each relevant pixel on the screen. Samples are taken at intervals along each ray. The sample interval is controlled by the numSlices and numSlicesControl fields At each sample point a value is interpolated from the closest voxels. The interpolation technique is controlled by the interpolation field. Classification means that color and opacity are computed based on the current SoDataRange and SoTransferFunction (and possibly values from other volumes - see SoVolumeShader). Optional rendering effects may modify the base color and/or opacity. These effects are controlled by an SoVolumeRenderingQuality node and include lighting, shadows, edge coloring, boundary opacity and more. The sample is then composited with other samples along the ray. Composition is controlled by the renderMode field. By default colors are combined based on the opacity of the sample (alpha blending). The ray is terminated if it reaches full opacity or reaches the current depth in the depth buffer. If raycasting is not enabled (SoVolumeShader.raycasting field), the volume is rendered by drawing "slices" (texture mapped polygons) from the back to the front. This effectively samples the volume. Interpolation and classification are done the same way as raycasting, but composition is done by the graphics hardware, so only alpha blending is available.

For a volume containing scalar data values, each voxel's basic RGBA value is determined by the current SoDataRange and SoTransferFunction nodes. This value is combined with current diffuse color and transparency (set, for example, with an SoMaterial node). This means that, for example, the current transparency can be used as a global alpha scale factor to decrease the opacity of all voxels. By default, scalar values are loaded on the GPU and the GPU interpolates between the data values before applying the color map. See the interpolation field for options. To force RGBA values to be loaded (implies color map is applied on the CPU and GPU interpolates between color values), see the usePalettedTexture field in the SoVolumeData node.

For an RGBA volume, each voxel's base RGBA value comes directly from the volume data (SoDataRange and SoTransferFunction are ignored). However if lighting is enabled, the final voxel color is also affected by the emissiveColor, specularColor and shininess fields of SoMaterial.

The samplingAlignment field controls whether the samples are axis aligned (perpendicular to one axis of the volume), view aligned (perpendicular to the view direction) or boundary aligned (each ray starts at the first intersected voxel with alpha value > 0). Generally boundary aligned slices should be used (better image quality). Using SoVolumeGroup, SoVolumeIsoSurface or SoProjection nodes will automatically switch to view-aligned samples.

The property nodes SoVolumeIsoSurface and SoVolumeDataDrawStyle add additional rendering styles and can be used, for example, to force SoVolumeRender to draw a GPU computed isosurface instead of volume rendering.

Example rendering:

<TABLE border=1 cellspacing=0 cellpadding=5>
  <TR><TD valign=_top> lighting = false               <TD valign=_top> lighting = true
  <TR><TD valign=_top> @image html lightoff.jpg
            <TD valign=_top> @image html lighton.jpg
  <TR><TD valign=_top> composition = ALPHA_BLENDING   <TD valign=_top> composition = MAX_INTENSITY
  <TR><TD valign=_top> @image html compalpha.jpg
            <TD valign=_top> @image html compmax.jpg
</TABLE>

Multiple volumes:

VolumeViz provides several mechanisms for combining or rendering multiple volumes, depending on the application's specific requirements. There are several cases:

  • CPU data combining The data values from multiple volumes can be combined on the CPU during data loading, using the SoDataCompositor node. For example computing the "difference" between two volumes on the fly. SoLDMDataTransform and SoVolumeTransform can be used to modify data for a single volume, for example scaling data values or computing a derived data set. SoLDMDataTransform is applied when data is loaded and SoVolumeTransform is applied before data is transferred to the GPU.

  • Multiple independent volumes These are volumes in the same scene that are completely independent (but might overlap in space) or have different dimensions or different extents. For example, medical datasets from different modalities. No special handling is required for rendering slices, including volume skin, or when the color map is opaque. When transparent volumes overlap in space, use the SoVolumeGroup node above the volume data and rendering nodes to ensure that the rendering is properly interleaved.

  • Multiple data sets These are volumes that are really multiple data sets on the same "grid", in other words volumes having exactly the same dimensions and extent. For example, seismic attribute volumes. These volumes can be combined on the GPU using a simple fragment shader function. For example, replacing the VVizComputeFragmentColor (GLSL) function allows "co-blending" the colors from multiple volumes. See SoVolumeShader for more details. Use an SoMultiDataSeparator node to group the volume data nodes that should be combined.

Custom shaders:

The SoVolumeShader node allows a variety of custom shader functions to be defined for special computation or rendering effects on single volumes or multiple volumes. All of these features require programmable shader support on the GPU. Be sure to use an SoMultiDataSeparator (instead of SoSeparator) when combining multiple volumes.

Lighting:

The SoVolumeRenderingQuality property node allows you to to enable GPU computed lighting based on the first SoLight node in the scene graph. (Note that this is typically the viewer's "headlight".) VolumeViz supports two lighting algorithms. They are both computed on the GPU and are independent (but normally only one should be enabled).

  • Gradient lighting Computes the effect of lighting for every sample along the ray, similar to the computation for polygonal geometry, but using a gradient vector computed from the voxel data values instead of a normal vector. Gradient lighting is enabled by the SoVolumeRenderingQuality lighting field. Gradient lighting is the classic solution but can have problems in relatively homogeneous regions where the gradient magnitude is small and gradient direction somewhat random. Multiple fields affect this computation including interpolation, gradientThreshold, gradientQuality and surfaceScalarExponent.

  • Deferred lighting Computes the effect of lighting for every visible voxel, as a post-processing step, using a gradient vector computed from the (visible) voxel depth values. Deferred lighting is enabled by the SoVolumeRenderingQuality deferredLighting field. Deferred lighting generally provides better performance because fewer lighting computations are needed. Deferred lighting generally provides better image quality for volumes that inherently contain surfaces (sharp gradients) like medical and industrial scans. Deferred lighting is most effective when the opacity values in the transfer function are approximately binary (mostly 0 and 1 values). See also the opacityThreshold field.

Warning: CPU computed lighting can be enabled using the lighting field in this node. In this case the lightDirection and lightIntensity fields control the light. Note that when lighting is computed on the CPU, RGBA textures are loaded on the GPU, so color map changes generally require re-loading all the data textures. NOTE: This feature is obsolete. Use SoVolumeRenderingQuality.lighting instead.

Gradient lighting Deferred lighting

Shadows:

Open Inventor shadow rendering works for volume rendering similar to any other geometry. When shadow rendering is enabled (see SoShadowGroup), non-transparent voxels can cast and receive shadows (see SoShadowStyle). Shadow rendering is independent of whether lighting is enabled for the volume.

SoVolumeRender also supports "ambient occlusion" rendering (see ambientOcclusion field in SoVolumeRenderingQuality). This rendering mode is visually a kind of self-shadowing and represents an approximation of the effect of ambient global lighting in the scene. Ambient occlusion can be combined with gradient or deferred lighting and with shadow casting.

Shadow casting Ambient occlusion

Clipping:

VolumeViz provides multiple tools for clipping volume rendering. Any or all of these tools may be combined for more complex clipping situations. Note that these tools clip all volume shapes including slices.

  • The SoROI (Region of Interest) node limits volume rendering to a subvolume. The SoROI node's EXCLUSION_BOX mode can also be used to exclude a sub-region, forming what is sometimes called a "chair cut". Note that the Region of Interest also limits data loading , so it is valuable when the total size of the volume exceeds the available system memory.

  • The SoVolumeClippingGroup node clips volume rendering to any closed shape defined by a group of standard Open Inventor geometry nodes. Both "inside" and "outside" clipping are supported. The SoScreenDrawer, SbExtrusionGenerator and SoCSGShape nodes are useful for creating clipping geometry for interactive volume "sculpting".

  • The SoUniformGridClipping and SoUniformGridProjectionClipping nodes clip volume rendering against one or more surfaces defined by a height field. This is particularly useful in seismic applications for clipping against (or between) horizon surfaces.

  • The SoVolumeMask node can be used to clip volume rendering against a boolean mask volume on a per-voxel basis. But the mask mechanism is much more powerful than that. Each region can have its own transfer function (color map) using SoTransferFunction nodes. Each region can also have its own draw style (volume rendering, isosurface or boundary) using SoVolumeDataDrawStyle nodes. Each region, including the original unmasked volume, is only visible if there exists a transfer function (SoTransferFunction) with the same id value.

Picking:

SoRayPickAction handles picking of VolumeViz shapes similar to other geometry in the scene, but with additional features. Picking on an SoVolumeRender node can return the first non-transparent voxel "hit" or the entire set of intersected voxels along the pick ray. Similar to other geometry, SoPickedPoint can return a "detail" class specific to SoVolumeRender. SoVolumeRenderDetail returns the IJK (voxel coordinate) position of the pick and the data value at that point.

Since Open Inventor 8.6, the SoVolumeRender node (by default) uses the GPU to compute the picked voxel during an SoRayPickAction. For this to work, the SoRayPickAction must have its scene manager initialised using the method SoAction.setSceneManager(). SoHandleEventAction does this automatically, so it is not necessary for the application to take any action when using (for example) an SoEventCallback node and calling the getPickedPoint() method. However if the application creates its own SoRayPickAction then it should set the scene manager. If no scene manager is specified, a warning message is issued and software picking is done. If necessary, using the GPU for volume picking may be disabled by setting the environment variable IVVR_GPU_PICKING to 0 (see SoPreferences).

Projection:

The SoVolumeRender node supports projected volume rendering, for example rendering a volume defined on a grid of latitude / longitude coordinates. Projection is enabled by adding an SoProjection node before the SoVolumeRender node (see SoProjection for more information about supported coordinate systems, ellipsoids and map projections). The projection quality versus speed ratio can be controlled using the new projectedTileSubdivision field that defines how often each tile's geometry will be subdivided when projected. This is important because only the corner points of the tiles are projected, not the individual voxels. So subdividing the tiles provides a better approximation of the actual shape of the grid. Volume projection works with both regular (uniform voxel spacing) and rectilinear (non-uniform voxel spacing) grids. SoProjection automatically selects view-aligned sampling.

Warning:

  • Volume projection is incompatible with some rendering options enabled by this node. The useEarlyZ and gpuVertexGen fields are ignored.

  • Volume projection is incompatible with some options enabled by the VolumeRenderingQuality node. Do not enable the preIntegrated, jittering or edgeDetect2D fields.

  • Volume projection requires all culling to be disabled. The following options in class SoLDMGlobalResourceParameters should be disabled: setScreenResolutionCulling (default is false), setViewpointRefinement (default is true) and setViewCulling (default is true).

Performance:

Volume rendering performance is affected by many factors including the size of the volume and the rendering options selected. Image quality is also affected by many rendering options and, in general, higher quality implies lower performance. Some of the factors affecting volume rendering performance are:

  • Number of voxels: This mainly depends on the size of the volume, but can be reduced using an SoROI (region of interest) node.

  • Number of pixels: A larger number of pixels means a larger number of rays must be cast through the volume and therefore the shader execution time on the GPU will be longer. This effect is most noticeable when high quality rendering options are enabled. The number of pixels rendered can be temporarily reduced by setting the lowResMode field to DECREASE_SCREEN_RESOLUTION. This reduces the number of times the shader programs running on the GPU must be executed.

  • Number of samples (slices): This is controlled by the numSlices and numSlicesControl fields. Note that better image quality can obtained with the same number of samples by enabling options like preintegrated rendering (see SoVolumeRenderingQuality) and/or the BOUNDARY_ALIGNED setting for samplingAlignment. The number of samples can be automatically decreased when interacting using an SoInteractiveComplexity node. Since Open Inventor 9.2, we recommend to set the numSlicesControl field to AUTOMATIC and the numSlices field to -1. The number of samples will be computed based on the dimensions of the volume (number of voxels on each axis), the SoComplexity.value setting and the viewing direction. If the viewing direction changes, the number of samples will be automatically adjusted.

  • Opacity: Increasing the number of opaque, or nearly opaque, voxels in the volume (using SoTransferFunction) will generally improve performance because the sampling rays can terminate sooner. See also IVVR_ALPHA_THRESHOLD_INTERACTIVE in SoPreferences. If you are using a completely opaque transfer function, for example with a "volume probe", SoVolumeSkin will generate the same image much faster.

  • Rendering options: Many of the advanced rendering options and rendering effects enabled by SoVolumeRenderingQuality have an additional performance cost. These include lighting, edge coloring, boundary opacity, cubic interpolation and gradient quality. These settings can be automatically changed while interacting using an SoInteractiveComplexity node.

  • Tile size: For backward compatibility, the default tile size is still only 64. This is quite small for modern CPU/GPU hardware. The smaller the tile size, the larger the total number of tiles that must be managed by VolumeViz. This overhead can be significant, especially for operations that require reloading the data textures on the GPU, for example, changing the data range (SoDataRange). For smaller volumes, like 512^3, it can be efficient to set the tile size large enough to contain the entire volume. For very large volumes, larger tile sizes are efficient for SoVolumeRender but somewhat inefficient for slice rendering because complete tiles must be loaded even though the slice only uses part of the data (see also SoSlice.largeSliceSupport). Applications should experiment. For volumes stored in LDM file format, the tile size must be specified when the volume is converted to LDM (see SoConverter and the "-t" option). For other data data formats the tile size can be specified using the SoVolumeData node's ldmResourceParameters field, but only after setting the filename field or calling the SetReader()method.

  • LDM_USE_IN_MEM_COMPRESSION This environment variable (see SoPreferences) affects much more than its name implies. VolumeViz always manages data as "tiles", regardless of the data format. In many cases VolumeViz must create (or uncompress) the tiles at run time. These cases include in-memory volumes, any volume reader that does not implement the readTile() method (all built-in formats except LDM) and compressed LDM format files. If this variable is true ( the default value), then VolumeViz only keeps a small cache of tiles in memory. See the SoBufferObject method getBufferObjectCache() for the current default and note that this setting is separate from the max main memory parameter. If a tile's data is needed and that tile is not in the cache, the tile must be recreated. This overhead can be significant, especially for operations that require recreating data textures on the GPU, for example, changing the data range (SoDataRange). We recommend setting this variable to false unless the memory conserving feature for compressed tiles is critical.

    VolumeViz keeps tiles from a compressed LDM format file in CPU memory as compressed data. This option allows more tiles to be kept in the same amount of CPU memory, but incurs a performance penalty because a tile must be uncompressed when its data is needed, unless it is found in a small cache of uncompressed tiles (see SoBufferObject.getBufferObjectCache). This penalty is particularly noticeable for operations that require recreating data textures on the GPU, for example changing the data range (SoDataRange).

If rendering performance is too slow, it may be necessary to render with high quality settings when the user is not interacting with the scene, but temporarily switch to high performance (lower quality) settings when the user is interacting. Open Inventor automatically sets "interactive mode" when the user is moving the camera or moving a dragger. The application can explicitly set interactive mode, for example while the user is moving a slider in the user interface. Some important tools are:

  • LowResMode For example, set the lowResMode field to DECREASE_SCREEN_RESOLUTION and set the lowScreenResolutionScale field to 2 or 4. In interactive mode VolumeViz will render the volume at lower resolution (reducing the number of sample rays and shader executions).

  • SoInteractiveComplexity This node allows you to specify different values to use for certain fields in interactive mode. For example set a smaller value for SoComplexity.value (reduces the number of samples) or turn off an expensive rendering option. The values specified in SoInteractiveComplexity override the actual fields in the scene graph.

Limitations:

  • Multi-thread rendering: Unlike most Open Inventor nodes, VolumeViz nodes do not support simultaneous rendering in multiple threads (even when Open Inventor is initialized using one of the initThread() methods).

  • Geometric transforms: The volume size (extent in 3D space) and orientation can be modified by transformation nodes in the scene graph just like any geometry. For a volume this in turn modifies the appearance of volume rendering nodes like SoVolumeRender. However please note: The same transformation must be applied to the volume data node and all volume rendering nodes associated with that volume. So effectively any transformation nodes that affect the volume must be placed before the volume data node.

EXAMPLE

For simple data sets, a basic VolumeViz rendering could be achieved with only a few nodes: minimally an SoVolumeData node to identify the data set and one rendering node. However most data sets need at least some of the additional nodes shown here in order to get a correct and useful rendering. Most applications will need additional nodes to take advantage of region of interest, interaction, clipping and other VolumeViz features. Please consider the code shown here as simply a guideline and a starting point for exploring the many powerful features available in Open Inventor.

Note that some of the property nodes (data, material, color map, etc) will typically be shared by multiple rendering nodes. For example the volume data usually only needs to be loaded once, using a single SoVolumeData node. Multiple slices and/or regions can be rendered using that data node and they may use the same transfer function or each have their own.

Also note that this example is for a data volume, not a label volume. Please see the notes about label volumes following the code block.

SoPreferences.SetValue( "LDM_USE_IN_MEM_COMPRESSION", "0" );

SoSeparator volSep = new SoSeparator();
 root.AddChild( volSep );

SoInteractiveComplexity interact = new SoInteractiveComplexity();
 // Decrease “number of samples”
 interact.fieldSettings[0] = "SoComplexity value 0.2 0.5";
 // Decrease interpolation quality. 
 interact.fieldSettings[1] = "SoVolumeRender interpolation LINEAR CUBIC";
 // DonÂ’t wait before returning to full quality rendering. 
 interact.refinementDelay.Value = 0;
 volSep.AddChild( interact );

SoComplexity volComp = new SoComplexity();
 volSep.AddChild( volComp );

SoVolumeData volData = new SoVolumeData();
 volData.fileName.Value = "$OIVNETHOME/src/demos/data/VolumeViz/3DHead.ldm";
 volSep.AddChild( volData );

SoDataRange volRange = new SoDataRange();
if (volData.GetDatumSize() > 1)
{
   double minVal, maxVal;
   volData.GetMinMax( out minVal, out maxVal);
   volRange.min.Value = minVal;
   volRange.max.Value = maxVal;
}
volSep.AddChild( volRange );

SoTransferFunction volTF = new SoTransferFunction();
 volTF.predefColorMap.Value = SoTransferFunction.PredefColorMaps.GRAY;
 volTF.minValue.Value = 10; // Make "noise" voxels transparent
 volSep.AddChild( volTF );

SoMaterial volMat = new SoMaterial();
 volMat.diffuseColor.SetValue(1, 1, 1);
 volSep.AddChild( volMat );

SoVolumeRenderingQuality volQual = new SoVolumeRenderingQuality();
 // Remove border artifacts while moving. 
 volQual.interpolateOnMove.Value = true;
 // Higher quality rendering
 volQual.preIntegrated.Value = true;
 // Optional: Enable screen space lighting
 volQual.deferredLighting.Value = true;
 // Optional: If using gradient lighting, increase quality
 volQual.lightingModel.Value = SoVolumeRenderingQuality.LightingModels.OPENGL;
 volQual.surfaceScalarExponent.Value = 5;
 volSep.AddChild( volQual );

SoVolumeRender volRend = new SoVolumeRender();
 // Let Inventor compute best number of slices
 volRend.numSlicesControl.Value = SoVolumeRender.NumSlicesControls.AUTOMATIC;
 // Optional: Use lower screen resolution while moving.
 volRend.lowResMode.Value = SoVolumeRender.LowResModes.DECREASE_SCREEN_RESOLUTION;
 volRend.lowScreenResolutionScale.Value = 2;
 // Remove "slicing" artifacts
 volRend.samplingAlignment.Value = SoVolumeRender.SamplingAlignments.BOUNDARY_ALIGNED;
 // Ignore low visibility voxels (OIV 9.3)
 //volRend.opacityThreshold.Value = 0.1f;
 volSep.AddChild( volRend );

Label volumes

A label volume, also known as a label field, is usually the result of doing some sort of segmentation on a data volume. Each voxel value is an integer label (id) identifying which material, object, etc that the voxel belongs to. There could be 100’s or 1000’s of labels, but there might be as few as 8 label values. For example, a simple label volume might have 7 opaque materials plus plus an “exterior” material which is completely transparent. Conceptually, there is one big difference between a (typical) data volume and a label volume. A data volume is conceptually a set of discrete samples taken from a continuous scalar field. So we know the exact value at the center of each voxel and interpolate between those values to get the value at any position in between voxels. In a label volume we normally consider each voxel to belong completely to one material, so the value is constant until we cross the boundary into the next voxel. Therefore we do not want to interpolate the label values.

When rendering a label volume, make the following changes to the above example:

  • Set the SoVolumeRender interpolation field to NEAREST and

  • Leave the SoVolumeRenderingQuality preintegrated field set to false.

If rendering isosurfaces (SoVolumeIsosurface), set the SoVolumeRenderingQuality segmentedInterpolation field to true.

It is also important to set the data range, texture precision and color map size carefully. Please see the label volume discussion in SoTransferFunction.

FILE FORMAT/DEFAULT

VolumeRender {
fixedNumSlicesInRoi false
gpuVertexGen false
interpolation LINEAR
lighting false
lightDirection -1, -1, -1
lightIntensity 1
lowResMode DECREASE_NONE
lowScreenResolutionScale 1
numEarlyZPasses 30
numSlices -1
numSlicesControl ALL
opacityCorrection true
projectedTileSubdivision 1
renderMode VOLUME_RENDERING
samplingAlignment VIEW_ALIGNED
subdivideTile false
useEarlyZ false
}

ACTION BEHAVIOR

SoGLRenderAction Draws a volume-rendered image based on current SoVolumeData.

SoGetBoundingBoxAction Computes the bounding box that encloses the volume.

SoRayPickAction Since Open Inventor version 8.5, picking always returns the first non-transparent voxel intersected by the pick ray. The old behavior can be restored by using an SoPickStyle node set to BOUNDING_BOX.

See Also