This is release v2.0.0 of Intel® OSPRay. For changes and new features see the changelog. Visit http://www.ospray.org for more information.

Intel OSPRay is an open source, scalable, and portable ray tracing engine for high-performance, high-fidelity visualization on Intel Architecture CPUs. OSPRay is part of the Intel oneAPI Rendering Toolkit and is released under the permissive Apache 2.0 license.

The purpose of OSPRay is to provide an open, powerful, and easy-to-use rendering library that allows one to easily build applications that use ray tracing based rendering for interactive applications (including both surface- and volume-based visualizations). OSPRay is completely CPU-based, and runs on anything from laptops, to workstations, to compute nodes in HPC systems.

OSPRay internally builds on top of Intel Embree and ISPC (Intel SPMD Program Compiler), and fully exploits modern instruction sets like Intel SSE4, AVX, AVX2, and AVX-512 to achieve high rendering performance, thus a CPU with support for at least SSE4.1 is required to run OSPRay.

OSPRay is under active development, and though we do our best to guarantee stable release versions a certain number of bugs, as-yet-missing features, inconsistencies, or any other issues are still possible. Should you find any such issues please report them immediately via OSPRay’s GitHub Issue Tracker (or, if you should happen to have a fix for it,you can also send us a pull request); for missing features please contact us via email at [email protected].

For recent news, updates, and announcements, please see our complete news/updates page.

Join our mailing list to receive release announcements and major news regarding OSPRay.

The latest OSPRay sources are always available at the OSPRay GitHub repository. The default master branch should always point to the latest bugfix release.

OSPRay currently supports Linux, Mac OS X, and Windows. In addition, before you can build OSPRay you need the following prerequisites:

• You can clone the latest OSPRay sources via:

  git clone https://github.com/ospray/ospray.git
  

• To build OSPRay you need CMake, any form of C++11 compiler (we recommend using GCC, but also support Clang, MSVC, and Intel® C++ Compiler (icc)), and standard Linux development tools. To build the interactive tutorials, you should also have some version of OpenGL and GLFW.

• Additionally you require a copy of the Intel® SPMD Program Compiler (ISPC), version 1.9.1 or later. Please obtain a release of ISPC from the ISPC downloads page. The build system looks for ISPC in the PATH and in the directory right “next to” the checked-out OSPRay sources.¹ Alternatively set the CMake variable ISPC_EXECUTABLE to the location of the ISPC compiler.

  NOTE: OSPRay is incompatible with ISPC v1.11.0.

• OSPRay builds on top of a small C++ utility library called ospcommon. The library provides abstractions for tasking, aligned memory allocation, vector math types, among others. For users who also need to build ospcommon, we recommend the default the Intel® Threading Building Blocks (TBB) as tasking system for performance and flexibility reasons. Alternatively you can set CMake variable OSPCOMMON_TASKING_SYSTEM to OpenMP or Internal.

• OSPRay also heavily uses Intel Embree, installing version 3.7.0 or newer is required. If Embree is not found by CMake its location can be hinted with the variable embree_DIR.

• OSPRay also heavily uses Intel Open VKL, installing version 0.7.0 or newer is required. If Open VKL is not found by CMake its location can be hinted with the variable openvkl_DIR.

• OSPRay also provides an optional module that adds support for Intel Open Image Denoise, which is enabled by OSPRAY_MODULE_DENOISER. When loaded, this module enables the denosier image operation. You may need to hint the location of the library with the CMake variable OpenImageDenoise_DIR.

Depending on your Linux distribution you can install these dependencies using yum or apt-get. Some of these packages might already be installed or might have slightly different names.

Type the following to install the dependencies using yum:

sudo yum install cmake.x86_64
sudo yum install tbb.x86_64 tbb-devel.x86_64

Type the following to install the dependencies using apt-get:

sudo apt-get install cmake-curses-gui
sudo apt-get install libtbb-dev

Under Mac OS X these dependencies can be installed using MacPorts:

sudo port install cmake tbb

Under Windows please directly use the appropriate installers for CMake, TBB, ISPC (for your Visual Studio version) and Embree.

For convenience, OSPRay provides a CMake Superbuild script which will pull down OSPRay’s dependencies and build OSPRay itself. By default, the result is an install directory, with each dependency in its own directory.

Run with:

mkdir build
cd build
cmake [<OSPRAY_SOURCE_LOC>/scripts/superbuild]
cmake --build .

The resulting install directory (or the one set with CMAKE_INSTALL_PREFIX) will have everything in it, with one subdirectory per dependency.

CMake options to note (all have sensible defaults):

• CMAKE_INSTALL_PREFIX will be the root directory where everything gets installed.
• BUILD_JOBS sets the number given to make -j for parallel builds.
• INSTALL_IN_SEPARATE_DIRECTORIES toggles installation of all libraries in separate or the same directory.
• BUILD_EMBREE_FROM_SOURCE set to OFF will download a pre-built version of Embree.
• BUILD_OIDN_FROM_SOURCE set to OFF will download a pre-built version of OpenImageDenoise.
• BUILD_OIDN_VERSION determines which verison of OpenImageDenoise to pull down.

For the full set of options, run:

ccmake [<OSPRAY_SOURCE_LOC>/scripts/superbuild]

or

cmake-gui [<OSPRAY_SOURCE_LOC>/scripts/superbuild]

Assuming the above requisites are all fulfilled, building OSPRay through CMake is easy:

• Create a build directory, and go into it

  mkdir ospray/build
  cd ospray/build
  

  (We do recommend having separate build directories for different configurations such as release, debug, etc.).

• The compiler CMake will use will default to whatever the CC and CXX environment variables point to. Should you want to specify a different compiler, run cmake manually while specifying the desired compiler. The default compiler on most linux machines is gcc, but it can be pointed to clang instead by executing the following:

  cmake -DCMAKE_CXX_COMPILER=clang++ -DCMAKE_C_COMPILER=clang ..
  

  CMake will now use Clang instead of GCC. If you are ok with using the default compiler on your system, then simply skip this step. Note that the compiler variables cannot be changed after the first cmake or ccmake run.

• Open the CMake configuration dialog

  ccmake ..
  

• Make sure to properly set build mode and enable the components you need, etc.; then type ’c’onfigure and ’g’enerate. When back on the command prompt, build it using

  make
  

• You should now have libospray.[so,dylib] as well as a set of example application. You can test your version of OSPRay using any of the examples on the OSPRay Demos and Examples page.

On Windows using the CMake GUI (cmake-gui.exe) is the most convenient way to configure OSPRay and to create the Visual Studio solution files:

• Browse to the OSPRay sources and specify a build directory (if it does not exist yet CMake will create it).

• Click “Configure” and select as generator the Visual Studio version you have (OSPRay needs Visual Studio 14 2015 or newer), for Win64 (32 bit builds are not supported by OSPRay), e.g., “Visual Studio 15 2017 Win64”.

• If the configuration fails because some dependencies could not be found then follow the instructions given in the error message, e.g., set the variable embree_DIR to the folder where Embree was installed and openvkl_DIR to where Open VKL was installed.

• Optionally change the default build options, and then click “Generate” to create the solution and project files in the build directory.

• Open the generated OSPRay.sln in Visual Studio, select the build configuration and compile the project.

Alternatively, OSPRay can also be built without any GUI, entirely on the console. In the Visual Studio command prompt type:

cd path\to\ospray
mkdir build
cd build
cmake -G "Visual Studio 15 2017 Win64" [-D VARIABLE=value] ..
cmake --build . --config Release

Use -D to set variables for CMake, e.g., the path to Embree with “-D embree_DIR=\path\to\embree”.

You can also build only some projects with the --target switch. Additional parameters after “--” will be passed to msbuild. For example, to build in parallel only the OSPRay library without the example applications use

cmake --build . --config Release --target ospray -- /m

Client applications using OSPRay can find it with CMake’s find_package() command. For example,

find_package(ospray 2.0.0 REQUIRED)

finds OSPRay via OSPRay’s configuration file osprayConfig.cmake². Once found, the following is all that is required to use OSPRay:

target_link_libraries(${client_target} ospray::ospray)

This will automatically propagate all required include paths, linked libraries, and compiler definitions to the client CMake target (either an executable or library).

Advanced users may want to link to additional targets which are exported in OSPRay’s CMake config, which includes all installed modules. All targets built with OSPRay are exported in the ospray:: namespace, therefore all targets locally used in the OSPRay source tree can be accessed from an install. For example, ospray_module_ispc can be consumed directly via the ospray::ospray_module_ispc target. All targets have their libraries, includes, and definitions attached to them for public consumption (please report bugs if this is broken!).

The following API documentation of OSPRay can also be found as a pdf document.

For a deeper explanation of the concepts, design, features and performance of OSPRay also have a look at the IEEE Vis 2016 paper “OSPRay – A CPU Ray Tracing Framework for Scientific Visualization” (49MB, or get the smaller version 1.8MB). The slides of the talk (5.2MB) are also available.

To access the OSPRay API you first need to include the OSPRay header

#include "ospray/ospray.h"

where the API is compatible with C99 and C++.

To use the API, OSPRay must be initialized with a “device”. A device is the object which implements the API. Creating and initializing a device can be done in either of two ways: command line arguments using ospInit or manually instantiating a device and setting parameters on it.

The first is to do so by giving OSPRay the command line from main() by calling

OSPError ospInit(int *argc, const char **argv);

OSPRay parses (and removes) its known command line parameters from your application’s main function. For an example see the tutorial. For possible error codes see section Error Handling and Status Messages. It is important to note that the arguments passed to ospInit() are processed in order they are listed. The following parameters (which are prefixed by convention with “--osp:”) are understood:

: Command line parameters accepted by OSPRay’s ospInit.

The second method of initialization is to explicitly create the device and possibly set parameters. This method looks almost identical to how other objects are created and used by OSPRay (described in later sections). The first step is to create the device with

OSPDevice ospNewDevice(const char *type);

where the type string maps to a specific device implementation. OSPRay always provides the “cpu” device, which maps to a fast, local CPU implementation. Other devices can also be added through additional modules, such as distributed MPI device implementations.

Once a device is created, you can call

void ospDeviceSet1i(OSPDevice, const char *id, int val);
void ospDeviceSetString(OSPDevice, const char *id, const char *val);
void ospDeviceSetVoidPtr(OSPDevice, const char *id, void *val);

to set parameters on the device. The following parameters can be set on all devices:

: Parameters shared by all devices.

Once parameters are set on the created device, the device must be committed with

void ospDeviceCommit(OSPDevice);

To use the newly committed device, you must call

void ospSetCurrentDevice(OSPDevice);

This then sets the given device as the object which will respond to all other OSPRay API calls.

Users can change parameters on the device after initialization (from either method above), by calling

OSPDevice ospGetCurrentDevice();

This function returns the handle to the device currently used to respond to OSPRay API calls, where users can set/change parameters and recommit the device. If changes are made to the device that is already set as the current device, it does not need to be set as current again.

OSPRay allows applications to query runtime properties of a device in order to do enhanced validation of what device was loaded at runtime. The following function can be used to get these device-specific properties (attiributes about the device, not paramter values)

int64_t ospDeviceGetProperty(OSPDevice, OSPDeviceProperty);

It returns an integer value of the queried property and the following properties can be provided as parameter:

OSP_DEVICE_VERSION
OSP_DEVICE_VERSION_MAJOR
OSP_DEVICE_VERSION_MINOR
OSP_DEVICE_VERSION_PATCH
OSP_DEVICE_SO_VERSION

OSPRay’s generic device parameters can be overridden via environment variables for easy changes to OSPRay’s behavior without needing to change the application (variables are prefixed by convention with “OSPRAY_”):

: Environment variables interpreted by OSPRay.

Note that these environment variables take precedence over values specified through ospInit or manually set device parameters.

The following errors are currently used by OSPRay:

: Possible error codes, i.e., valid named constants of type OSPError.

These error codes are either directly return by some API functions, or are recorded to be later queried by the application via

OSPError ospDeviceGetLastErrorCode(OSPDevice);

A more descriptive error message can be queried by calling

const char* ospDeviceGetLastErrorMsg(OSPDevice);

Alternatively, the application can also register a callback function of type

typedef void (*OSPErrorFunc)(OSPError, const char* errorDetails);

via

void ospDeviceSetErrorFunc(OSPDevice, OSPErrorFunc);

to get notified when errors occur.

Applications may be interested in messages which OSPRay emits, whether for debugging or logging events. Applications can call

void ospDeviceSetStatusFunc(OSPDevice, OSPStatusFunc);

in order to register a callback function of type

typedef void (*OSPStatusFunc)(const char* messageText);

which OSPRay will use to emit status messages. By default, OSPRay uses a callback which does nothing, so any output desired by an application will require that a callback is provided. Note that callbacks for C++ std::cout and std::cerr can be alternatively set through ospInit() or the OSPRAY_LOG_OUTPUT environment variable.

Applications can clear either callback by passing nullptr instead of an actual function pointer.

OSPRay’s functionality can be extended via plugins (which we call “modules”), which are implemented in shared libraries. To load module name from libospray_module_<name>.so (on Linux and Mac OS X) or ospray_module_<name>.dll (on Windows) use

OSPError ospLoadModule(const char *name);

Modules are searched in OS-dependent paths. ospLoadModule returns OSP_NO_ERROR if the plugin could be successfully loaded.

When the application is finished using OSPRay (typically on application exit), the OSPRay API should be finalized with

void ospShutdown();

This API call ensures that the current device is cleaned up appropriately. Due to static object allocation having non-deterministic ordering, it is recommended that applications call ospShutdown() before the calling application process terminates.

All entities of OSPRay (the renderer, volumes, geometries, lights, cameras, …) are a logical specialization of OSPObject and share common mechanism to deal with parameters and lifetime.

An important aspect of object parameters is that parameters do not get passed to objects immediately. Instead, parameters are not visible at all to objects until they get explicitly committed to a given object via a call to

void ospCommit(OSPObject);

at which time all previously additions or changes to parameters are visible at the same time. If a user wants to change the state of an existing object (e.g., to change the origin of an already existing camera) it is perfectly valid to do so, as long as the changed parameters are recommitted.

The commit semantic allow for batching up multiple small changes, and specifies exactly when changes to objects will occur. This can impact performance and consistency for devices crossing a PCI bus or across a network.

Note that OSPRay uses reference counting to manage the lifetime of all objects, so one cannot explicitly “delete” any object. Instead, to indicate that the application does not need and does not access the given object anymore, call

void ospRelease(OSPObject);

This decreases its reference count and if the count reaches 0 the object will automatically get deleted. Passing NULL is not an error.

Sometimes applications may want to have more than one reference to an object, where it is desirable for the application to increment the reference count of an object. This is done with

void ospRetain(OSPObject);

It is important to note that this is only necessary if the application wants to call ospRelease on an object more than once: objects which contain other objects as parameters internally increment/decrement ref counts and should not be explicitly done by the application.

Parameters allow to configure the behavior of and to pass data to objects. However, objects do not have an explicit interface for reasons of high flexibility and a more stable compile-time API. Instead, parameters are passed separately to objects in an arbitrary order, and unknown parameters will simply be ignored (though a warning message will be posted). The following function allows adding various types of parameters with name id to a given object:

void ospSetParam(OSPObject, const char *id, OSPDataType type, const void *mem);

The valid parameter names for all OSPObjects and what types are valid are discussed in future sections.

Note that mem must always be a pointer to the object, otherwise accidental type casting can occur. This is especially true for pointer types (OSP_VOID_PTR and OSPObject handles), as they will implicitly cast to void *, but be incorrectly interpreted. To help with some of these issues, there also exist variants of ospSetParam for specific types, such as ospSetInt and ospSetVec3f in the OSPRay utility library (found in ospray_util.h).

Users can also remove parameters that have been explicitly set from ospSetParam. Any parameters which have been removed will go back to their default value during the next commit unless a new parameter was set after the parameter was removed. To remove a parameter, use

void ospRemoveParam(OSPObject, const char *id);

OSPRay consumes data arrays from the application using a specific object type, OSPData. There are several components to describing a data array: element type, 1/2/3 dimensional striding, and whether the array is shared with the application or copied into opaque, OSPRay-owned memory.

Shared data arrays require that the application’s array memory outlives the lifetime of the created OSPData, as OSPRay is referring to application memory. Where this is not preferable, applications use opaque arrays to allow the OSPData to own the lifetime of the array memory. However, opaque arrays dictate the cost of copying data into it, which should be kept in mind.

Thus the most efficient way to specify a data array from the application is to created a shared data array, which is done with

OSPData ospNewSharedData(const void *sharedData,
                   OSPDataType,
  uint64_t numItems1,
  int64_t byteStride1 = 0,
  uint64_t numItems2 = 1,
  int64_t byteStride2 = 0,
  uint64_t numItems3 = 1,
  int64_t byteStride3 = 0);

The call returns an OSPData handle to the created array. The calling program guarantees that the sharedData pointer will remain valid for the duration that this data array is being used. The number of elements numItems must be positive (there cannot be an empty data object). The data is arranged in three dimensions, with specializations to two or one dimension (if some numItems are 1). The distance between consecutive elements (per dimension) is given in bytes with byteStride and can also be negative. If byteStride is zero it will be determined automatically (e.g., as sizeof(type)). Strides do not need to be ordered, i.e., byteStride2 can be smaller than byteStride1, which is equivalent to a transpose. However, if the stride should be calculated, then an ordering like byteStride1 < byteStride2 is assumed to disambiguate.

The enum type OSPDataType describes the different element types that can be represented in OSPRay; valid constants are listed in the table below.

: Valid named constants for OSPDataType.

An opaque OSPData with memory allocated by OSPRay is created with

OSPData ospNewData(OSPDataType,
  uint32_t numItems1,
  uint32_t numItems2 = 1,
  uint32_t numItems3 = 1);

To allow for (partial) copies or updates of data arrays use

void ospCopyData(const OSPData source,
  OSPData destination,
  uint32_t destinationIndex1 = 0,
  uint32_t destinationIndex2 = 0,
  uint32_t destinationIndex3 = 0);

which will copy the whole³ content of the source array into destination at the given location destinationIndex. The OSPDataTypes of the data objects must match. The region to be copied must be valid inside the destination, i.e., in all dimensions, destinationIndex + sourceSize <= destinationSize. The affected region [destinationIndex, destinationIndex + sourceSize) is marked as dirty, which may be used by OSPRay to only process or update that sub-region (e.g., updating an acceleration structure). If the destination array is shared with OSPData by the application (created with ospNewSharedData), then

• the source array must be shared as well (thus ospCopyData cannot be used to read opaque data)
• if source and destination memory overlaps (aliasing), then behaviour is undefined
• except if source and destination regions are identical (including matching strides), which can be used by application to mark that region as dirty (instead of the whole OSPData)

To add a data array as parameter named id to another object call also use

void ospSetObject(OSPObject, const char *id, OSPData);

Volumes are volumetric data sets with discretely sampled values in 3D space, typically a 3D scalar field. To create a new volume object of given type type use

OSPVolume ospNewVolume(const char *type);

Note that OSPRay’s implementation forwards type directly to Open VKL, allowing new Open VKL volume types to be usable within OSPRay without the need to change (or even recompile) OSPRay.

Structured volumes only need to store the values of the samples, because their addresses in memory can be easily computed from a 3D position. A common type of structured volumes are regular grids.

Structured regular volumes are created by passing the structuredRegular type string to ospNewVolume. Structured volumes are represented through an OSPData 3D array data (which may or may not be shared with the application), where currently the voxel data needs to be laid out compact in memory in xyz-order⁴

The parameters understood by structured volumes are summarized in the table below.

: Additional configuration parameters for structured regular volumes.

The size of the volume is inferred from the size of the 3D array data, as is the type of the voxel values (currently supported are: OSP_UCHAR, OSP_SHORT, OSP_USHORT, OSP_FLOAT, and OSP_DOUBLE).

Structured spherical volumes are also supported, which are created by passing a type string of "structuredSpherical" to ospNewVolume. The grid dimensions and parameters are defined in terms of radial distance $r$, inclination angle $\theta$, and azimuthal angle $\phi$, conforming with the ISO convention for spherical coordinate systems. The coordinate system and parameters understood by structured spherical volumes are summarized below.

: Additional configuration parameters for structured spherical volumes.

The dimensions $(r, \theta, \phi)$ of the volume are inferred from the size of the 3D array data, as is the type of the voxel values (currently supported are: OSP_UCHAR, OSP_SHORT, OSP_USHORT, OSP_FLOAT, and OSP_DOUBLE).

These grid parameters support flexible specification of spheres, hemispheres, spherical shells, spherical wedges, and so forth. The grid extents (computed as [gridOrigin, gridOrigin + (dimensions - 1) * gridSpacing]) however must be constrained such that:

• $r \geq 0$
• $0 \leq \theta \leq 180$
• $0 \leq \phi \leq 360$

OSPRay currently supports block-structured (Berger-Colella) AMR volumes. Volumes are specified as a list of blocks, which exist at levels of refinement in potentially overlapping regions. Blocks exist in a tree structure, with coarser refinement level blocks containing finer blocks. The cell width is equal for all blocks at the same refinement level, though blocks at a coarser level have a larger cell width than finer levels.

There can be any number of refinement levels and any number of blocks at any level of refinement. An AMR volume type is created by passing the type string "amr" to ospNewVolume.

Blocks are defined by three parameters: their bounds, the refinement level in which they reside, and the scalar data contained within each block.

Note that cell widths are defined per refinement level, not per block.

: Additional configuration parameters for AMR volumes.

Lastly, note that the gridOrigin and gridSpacing parameters act just like the structured volume equivalent, but they only modify the root (coarsest level) of refinement.

In particular, OSPRay’s AMR implementation was designed to cover Berger-Colella [1] and Chombo [2] AMR data. The method parameter above determines the interpolation method used when sampling the volume.

• OSP_AMR_CURRENT finds the finest refinement level at that cell and interpolates through this “current” level
• OSP_AMR_FINEST will interpolate at the closest existing cell in the volume-wide finest refinement level regardless of the sample cell’s level
• OSP_AMR_OCTANT interpolates through all available refinement levels
• at that cell. This method avoids discontinuities at refinement level boundaries at the cost of performance

Details and more information can be found in the publication for the implementation [3].

1. M. J. Berger, and P. Colella. “Local adaptive mesh refinement for shock hydrodynamics.” Journal of Computational Physics 82.1 (1989): 64-84. DOI: 10.1016/0021-9991(89)90035-1
2. M. Adams, P. Colella, D. T. Graves, J.N. Johnson, N.D. Keen, T. J. Ligocki. D. F. Martin. P.W. McCorquodale, D. Modiano. P.O. Schwartz, T.D. Sternberg and B. Van Straalen, Chombo Software Package for AMR Applications - Design Document, Lawrence Berkeley National Laboratory Technical Report LBNL-6616E.
3. I. Wald, C. Brownlee, W. Usher, and A. Knoll. CPU volume rendering of adaptive mesh refinement data. SIGGRAPH Asia 2017 Symposium on Visualization on - SA ’17, 18(8), 1–8. DOI: 10.1145/3139295.3139305

Unstructured volumes can have their topology and geometry freely defined. Geometry can be composed of tetrahedral, hexahedral, wedge or pyramid cell types. The data format used is compatible with VTK and consists of multiple arrays: vertex positions and values, vertex indices, cell start indices, cell types, and cell values. An unstructured volume type is created by passing the type string “unstructured” to ospNewVolume.

Sampled cell values can be specified either per-vertex (vertex.data) or per-cell (cell.data). If both arrays are set, cell.data takes precedence.

Similar to a mesh, each cell is formed by a group of indices into the vertices. For each vertex, the corresponding (by array index) data value will be used for sampling when rendering, if specified. The index order for a tetrahedron is the same as VTK_TETRA: bottom triangle counterclockwise, then the top vertex.

For hexahedral cells, each hexahedron is formed by a group of eight indices into the vertices and data values. Vertex ordering is the same as VTK_HEXAHEDRON: four bottom vertices counterclockwise, then top four counterclockwise.

For wedge cells, each wedge is formed by a group of six indices into the vertices and data values. Vertex ordering is the same as VTK_WEDGE: three bottom vertices counterclockwise, then top three counterclockwise.

For pyramid cells, each cell is formed by a group of five indices into the vertices and data values. Vertex ordering is the same as VTK_PYRAMID: four bottom vertices counterclockwise, then the top vertex.

To maintain VTK data compatibility an index array may be specified via the indexPrefixed array that allow vertex indices to be interleaved with cell sizes in the following format: $n, id_1, ..., id_n, m, id_1, ..., id_m$.

: Additional configuration parameters for unstructured volumes.

Transfer functions map the scalar values of volumes to color and opacity and thus they can be used to visually emphasize certain features of the volume. To create a new transfer function of given type type use

OSPTransferFunction ospNewTransferFunction(const char *type);

The returned handle can be assigned to a volumetric model (described below) as parameter “transferFunction” using ospSetObject.

One type of transfer function that is supported by OSPRay is the linear transfer function, which interpolates between given equidistant colors and opacities. It is create by passing the string “piecewiseLinear” to ospNewTransferFunction and it is controlled by these parameters:

: Parameters accepted by the linear transfer function.

Volumes in OSPRay are given volume rendering appearance information through VolumetricModels. This decouples the physical representation of the volume (and possible acceleration structures it contains) to rendering-specific parameters (where more than one set may exist concurrently). To create a volume instance, call

OSPVolumetricModel ospNewVolumetricModel(OSPVolume volume);

: Parameters understood by VolumetricModel.

Geometries in OSPRay are objects that describe intersectable surfaces. To create a new geometry object of given type type use

OSPGeometry ospNewGeometry(const char *type);

Note that in the current implementation geometries are limited to a maximum of 2^32^ primitives.

A mesh consiting of either triangles or quads is created by calling ospNewGeometry with type string “mesh”. Once created, a mesh recognizes the following parameters:

: Parameters defining a mesh geometry.

The data type of index arrays differentiates between the underlying geometry, triangles are used for a index with vec3ui type and quads for vec4ui type. Quads are internally handled as a pair of two triangles, thus mixing triangles and quads is supported by encoding some triangle as a quad with the last two vertex indices being identical (w=z).

The vertex.position and index arrays are mandatory to create a valid mesh.

A mesh consisting of subdivision surfaces, created by specifying a geometry of type “subdivision”. Once created, a subdivision recognizes the following parameters:

: Parameters defining a Subdivision geometry.

The vertex and index arrays are mandatory to create a valid subdivision surface. If no face array is present then a pure quad mesh is assumed (the number of indices must be a multiple of 4). Optionally supported are edge and vertex creases.

A geometry consisting of individual spheres, each of which can have an own radius, is created by calling ospNewGeometry with type string “sphere”. The spheres will not be tessellated but rendered procedurally and are thus perfectly round. To allow a variety of sphere representations in the application this geometry allows a flexible way of specifying the data of center position and radius within a data array:

: Parameters defining a spheres geometry.

A geometry consisting of multiple curves is created by calling ospNewGeometry with type string “curve”. The parameters defining this geometry are listed in the table below.

: Parameters defining a curves geometry.

Depending upon the specified data type of vertex positions, the curves will be implemented Embree curves or assembled from rounded and linearly-connected segments.

Positions in vertex.position_radius format supports per-vertex varying radii with data type vec4f[] and instantiate Embree curves internally for the relevant type/basis mapping (See Embree documentation for discussion of curve types and data formatting).

If a constant radius is used and positions are specified in a vec3f[] type of vertex.position format, then type/basis defaults to OSP_ROUND and OSP_LINEAR (this is the fastest and most memory efficient mode). Implementation is with round linear segements where each segment corresponds to a link between two vertices.

The following section describes the properties of different curve basis’ and how they use the data provided in data buffers:

OSP_LINEAR : The indices point to the first of 2 consecutive control points in the vertex buffer. The first control point is the start and the second control point the end of the line segment. The curve goes through all control points listed in the vertex buffer.

OSP_BEZIER : The indices point to the first of 4 consecutive control points in the vertex buffer. The first control point represents the start point of the curve, and the 4th control point the end point of the curve. The Bézier basis is interpolating, thus the curve does go exactly through the first and fourth control vertex.

OSP_BSPLINE : The indices point to the first of 4 consecutive control points in the vertex buffer. This basis is not interpolating, thus the curve does in general not go through any of the control points directly. Using this basis, 3 control points can be shared for two continuous neighboring curve segments, e.g. the curves $(p0, p1, p2, p3)$ and $(p1, p2, p3, p4)$ are C1 continuous. This feature make this basis a good choice to construct continuous multi-segment curves, as memory consumption can be kept minimal.

OSP_HERMITE : It is necessary to have both vertex buffer and tangent buffer for using this basis. The indices point to the first of 2 consecutive points in the vertex buffer, and the first of 2 consecutive tangents in the tangent buffer. This basis is interpolating, thus does exactly go through the first and second control point, and the first order derivative at the begin and end matches exactly the value specified in the tangent buffer. When connecting two segments continuously, the end point and tangent of the previous segment can be shared.

OSP_CATMULL_ROM : The indices point to the first of 4 consecutive control points in the vertex buffer. If $(p0, p1, p2, p3)$ represent the points then this basis goes through $p1$ and $p2$, with tangents as $(p2-p0)/2$ and $(p3-p1)/2$.

The following section describes the properties of different curve types’ and how they define the geometry of a curve:

OSP_FLAT : This type enables faster rendering as the curve is rendered as a connected sequence of ray facing quads.

OSP_ROUND : This type enables rendering a real geometric surface for the curve which allows closeup views. This mode renders a sweep surface by sweeping a varying radius circle tangential along the curve.

OSP_RIBBON : The type enables normal orientation of the curve and requires a normal buffer be specified along with vertex buffer. The curve is rendered as a flat band whose center approximately follows the provided vertex buffer and whose normal orientation approximately follows the provided normal buffer.

OSPRay can directly render axis-aligned bounding boxes without the need to convert them to quads or triangles. To do so create a boxes geometry by calling ospNewGeometry with type string “box”.

: Parameters defining a boxes geometry.

OSPRay can directly render multiple isosurfaces of a volume without first tessellating them. To do so create an isosurfaces geometry by calling ospNewGeometry with type string “isosurface”. Each isosurface will be colored according to the transfer function assigned to the volume.

: Parameters defining an isosurfaces geometry.

Geometries are matched with surface appearance information through GeometricModels. These take a geometry, which defines the surface representation, and applies either full-object or per-primitive color and material information. To create a geometric model, call

OSPGeometricModel ospNewGeometricModel(OSPGeometry geometry);

Color and material are fetched with the primitive ID of the hit (clamped to the valid range, thus a single color or material is fine), or mapped first via the index array (if present). All paramters are optional, however, some renderers (notably the path tracer) require a material to be set. Materials are either handles of OSPMaterial, or indices into the material array on the renderer, which allows to build a world which can be used by different types of renderers.

: Parameters understood by GeometricModel.

To create a new light source of given type type use

OSPLight ospNewLight(const char *type);

All light sources accept the following parameters:

: Parameters accepted by all lights.

The following light types are supported by most OSPRay renderers.

The distant light (or traditionally the directional light) is thought to be far away (outside of the scene), thus its light arrives (almost) as parallel rays. It is created by passing the type string “distant” to ospNewLight. In addition to the general parameters understood by all lights the distant light supports the following special parameters:

: Special parameters accepted by the distant light.

Setting the angular diameter to a value greater than zero will result in soft shadows when the renderer uses stochastic sampling (like the path tracer). For instance, the apparent size of the sun is about 0.53°.

The sphere light (or the special case point light) is a light emitting uniformly in all directions from the surface towards the outside. It does not emit any light towards the inside of the sphere. It is created by passing the type string “sphere” to ospNewLight. In addition to the general parameters understood by all lights the sphere light supports the following special parameters:

: Special parameters accepted by the sphere light.

Setting the radius to a value greater than zero will result in soft shadows when the renderer uses stochastic sampling (like the path tracer).

The spotlight is a light emitting into a cone of directions. It is created by passing the type string “spot” to ospNewLight. In addition to the general parameters understood by all lights the spotlight supports the special parameters listed in the table.

: Special parameters accepted by the spotlight.

Setting the radius to a value greater than zero will result in soft shadows when the renderer uses stochastic sampling (like the path tracer).

The quad⁵ light is a planar, procedural area light source emitting uniformly on one side into the half-space. It is created by passing the type string “quad” to ospNewLight. In addition to the general parameters understood by all lights the quad light supports the following special parameters:

: Special parameters accepted by the quad light.

The emission side is determined by the cross product of edge1×edge2. Note that only renderers that use stochastic sampling (like the path tracer) will compute soft shadows from the quad light. Other renderers will just sample the center of the quad light, which results in hard shadows.

The HDRI light is a textured light source surrounding the scene and illuminating it from infinity. It is created by passing the type string “hdri” to ospNewLight. In addition to the general parameters the HDRI light supports the following special parameters:

: Special parameters accepted by the HDRI light.

Note that the currently only the path tracer supports the HDRI light.

The ambient light surrounds the scene and illuminates it from infinity with constant radiance (determined by combining the parameters color and intensity). It is created by passing the type string “ambient” to ospNewLight.

Note that the SciVis renderer uses ambient lights to control the color and intensity of the computed ambient occlusion (AO).

The path tracer will consider illumination by geometries which have a light emitting material assigned (for example the Luminous material).

Groups in OSPRay represent collections of GeometricModels and VolumetricModels which share a common local-space coordinate system. To create a group call

OSPGroup ospNewGroup();

Groups take arrays of geometric models and volumetric models, but they are optional. In other words, there is no need to create empty arrays if there are no geometries or volumes in the group.

: Parameters understood by groups.

Note that groups only need to re re-committed if a geometry or volume changes (surface/scalar field representation). Appearance information on OSPGeometricModel and OSPVolumetricModel can be changed freely, as internal acceleration structures do not need to be reconstructed.

Instances in OSPRay represent a single group’s placement into the world via a transform. To create and instance call

OSPInstance ospNewInstance(OSPGroup);

: Parameters understood by instances.

Worlds are a container of scene data represented by instances. To create an (empty) world call

OSPWorld ospNewWorld();

Objects are placed in the world through an array of instances. Similar to [group], the array of instances is optional: there is no need to create empty arrays if there are no instances (though there will be nothing to render).

Applications can query the world (axis-aligned) bounding box after the world has been committed. To get this information, call

OSPBounds ospGetBounds(OSPObject);

This call can also take OSPGroup and OSPInstance as well: all other object types will return an empty bounding box.

Finally, Worlds can be configured with parameters for making various feature/performance trade-offs (similar to groups).

: Parameters understood by worlds.

A renderer is the central object for rendering in OSPRay. Different renderers implement different features and support different materials. To create a new renderer of given type type use

OSPRenderer ospNewRenderer(const char *type);

General parameters of all renderers are

: Parameters understood by all renderers.

OSPRay’s renderers support a feature called adaptive accumulation, which accelerates progressive rendering by stopping the rendering and refinement of image regions that have an estimated variance below the varianceThreshold. This feature requires a framebuffer with an OSP_FB_VARIANCE channel.

Per default the background of the rendered image will be transparent black, i.e., the alpha channel holds the opacity of the rendered objects. This eases transparency-aware blending of the image with an arbitrary background image by the application. The parameter backgroundColor or map_backplate can be used to already blend with a constant background color or backplate texture, respectively, (and alpha) during rendering.

OSPRay renderers support depth composition with images of other renderers, for example to incorporate help geometries of a 3D UI that were rendered with OpenGL. The screen-sized texture map_maxDepth must have format OSP_TEXTURE_R32F and flag OSP_TEXTURE_FILTER_NEAREST. The fetched values are used to limit the distance of primary rays, thus objects of other renderers can hide objects rendered by OSPRay.

The SciVis renderer is a fast ray tracer for scientific visualization which supports volume rendering and ambient occlusion (AO). It is created by passing the type string “scivis” to ospNewRenderer. In addition to the general parameters understood by all renderers, the SciVis renderer supports the following parameters:

: Special parameters understood by the SciVis renderer.

The path tracer supports soft shadows, indirect illumination and realistic materials. This renderer is created by passing the type string “pathtracer” to ospNewRenderer. In addition to the general parameters understood by all renderers the path tracer supports the following special parameters:

: Special parameters understood by the path tracer.

The path tracer requires that materials are assigned to geometries, otherwise surfaces are treated as completely black.

The path tracer supports volumes with multiple scattering. The scattering albedo can be specified using the transfer function. Extinction is assumed to be spectrally constant.

Materials describe how light interacts with surfaces, they give objects their distinctive look. To let the given renderer create a new material of given type type call

OSPMaterial ospNewMaterial(const char *renderer_type, const char *material_type);

The returned handle can then be used to assign the material to a given geometry with

void ospSetObject(OSPGeometricModel, "material", OSPMaterial);

The OBJ material is the workhorse material supported by both the SciVis renderer and the path tracer. It offers widely used common properties like diffuse and specular reflection and is based on the MTL material format of Lightwave’s OBJ scene files. To create an OBJ material pass the type string “obj” to ospNewMaterial. Its main parameters are

: Main parameters of the OBJ material.

In particular when using the path tracer it is important to adhere to the principle of energy conservation, i.e., that the amount of light reflected by a surface is not larger than the light arriving. Therefore the path tracer issues a warning and renormalizes the color parameters if the sum of Kd, Ks, and Tf is larger than one in any color channel. Similarly important to mention is that almost all materials of the real world reflect at most only about 80% of the incoming light. So even for a white sheet of paper or white wall paint do better not set Kd larger than 0.8; otherwise rendering times are unnecessary long and the contrast in the final images is low (for example, the corners of a white room would hardly be discernible, as can be seen in the figure below).

If present, the color component of geometries is also used for the diffuse color Kd and the alpha component is also used for the opacity d.

Note that currently only the path tracer implements colored transparency with Tf.

Normal mapping can simulate small geometric features via the texture map_Bump. The normals $n$ in the normal map are with respect to the local tangential shading coordinate system and are encoded as $½(n+1)$, thus a texel $(0.5, 0.5, 1)$⁶ represents the unperturbed shading normal $(0, 0, 1)$. Because of this encoding an sRGB gamma texture format is ignored and normals are always fetched as linear from a normal map. Note that the orientation of normal maps is important for a visually consistent look: by convention OSPRay uses a coordinate system with the origin in the lower left corner; thus a convexity will look green toward the top of the texture image (see also the example image of a normal map). If this is not the case flip the normal map vertically or invert its green channel.

All parameters (except Tf) can be textured by passing a texture handle, prefixed with “map_”. The fetched texels are multiplied by the respective parameter value. Texturing requires geometries with texture coordinates, e.g., a [triangle mesh] with vertex.texcoord provided. The color textures map_Kd and map_Ks are typically in one of the sRGB gamma encoded formats, whereas textures map_Ns and map_d are usually in a linear format (and only the first component is used). Additionally, all textures support texture transformations.

The Principled material is the most complex material offered by the path tracer, which is capable of producing a wide variety of materials (e.g., plastic, metal, wood, glass) by combining multiple different layers and lobes. It uses the GGX microfacet distribution with approximate multiple scattering for dielectrics and metals, uses the Oren-Nayar model for diffuse reflection, and is energy conserving. To create a Principled material, pass the type string “principled” to ospNewMaterial. Its parameters are listed in the table below.

: Parameters of the Principled material.

All parameters can be textured by passing a texture handle, prefixed with “map_” (e.g., “map_baseColor”). texture transformations are supported as well.

The CarPaint material is a specialized version of the Principled material for rendering different types of car paints. To create a CarPaint material, pass the type string “carPaint” to ospNewMaterial. Its parameters are listed in the table below.

: Parameters of the CarPaint material.

All parameters can be textured by passing a texture handle, prefixed with “map_” (e.g., “map_baseColor”). texture transformations are supported as well.

The path tracer offers a physical metal, supporting changing roughness and realistic color shifts at edges. To create a Metal material pass the type string “metal” to ospNewMaterial. Its parameters are

: Parameters of the Metal material.

The main appearance (mostly the color) of the Metal material is controlled by the physical parameters eta and k, the wavelength-dependent, complex index of refraction. These coefficients are quite counter-intuitive but can be found in published measurements. For accuracy the index of refraction can be given as an array of spectral samples in ior, each sample a triplet of wavelength (in nm), eta, and k, ordered monotonically increasing by wavelength; OSPRay will then calculate the Fresnel in the spectral domain. Alternatively, eta and k can also be specified as approximated RGB coefficients; some examples are given in below table.

: Index of refraction of selected metals as approximated RGB coefficients, based on data from https://refractiveindex.info/.

The roughness parameter controls the variation of microfacets and thus how polished the metal will look. The roughness can be modified by a texture map_roughness (texture transformations are supported as well) to create notable edging effects.

The path tracer offers an alloy material, which behaves similar to Metal, but allows for more intuitive and flexible control of the color. To create an Alloy material pass the type string “alloy” to ospNewMaterial. Its parameters are

: Parameters of the Alloy material.

The main appearance of the Alloy material is controlled by the parameter color, while edgeColor influences the tint of reflections when seen at grazing angles (for real metals this is always 100% white). If present, the color component of geometries is also used for reflectivity at normal incidence color. As in Metal the roughness parameter controls the variation of microfacets and thus how polished the alloy will look. All parameters can be textured by passing a texture handle, prefixed with “map_”; texture transformations are supported as well.

The path tracer offers a realistic a glass material, supporting refraction and volumetric attenuation (i.e., the transparency color varies with the geometric thickness). To create a Glass material pass the type string “glass” to ospNewMaterial. Its parameters are

: Parameters of the Glass material.

For convenience, the rather counter-intuitive physical attenuation coefficients will be calculated from the user inputs in such a way, that the attenuationColor will be the result when white light traveled trough a glass of thickness attenuationDistance.

The path tracer offers a thin glass material useful for objects with just a single surface, most prominently windows. It models a thin, transparent slab, i.e., it behaves as if a second, virtual surface is parallel to the real geometric surface. The implementation accounts for multiple internal reflections between the interfaces (including attenuation), but neglects parallax effects due to its (virtual) thickness. To create a such a thin glass material pass the type string “thinGlass” to ospNewMaterial. Its parameters are

: Parameters of the ThinGlass material.

For convenience the attenuation is controlled the same way as with the Glass material. Additionally, the color due to attenuation can be modulated with a texture map_attenuationColor (texture transformations are supported as well). If present, the color component of geometries is also used for the attenuation color. The thickness parameter sets the (virtual) thickness and allows for easy exchange of parameters with the (real) Glass material; internally just the ratio between attenuationDistance and thickness is used to calculate the resulting attenuation and thus the material appearance.

The path tracer offers a metallic paint material, consisting of a base coat with optional flakes and a clear coat. To create a MetallicPaint material pass the type string “metallicPaint” to ospNewMaterial. Its parameters are listed in the table below.

: Parameters of the MetallicPaint material.

The color of the base coat baseColor can be textured by a texture map_baseColor, which also supports texture transformations. If present, the color component of geometries is also used for the color of the base coat. Parameter flakeAmount controls the proportion of flakes in the base coat, so when setting it to 1 the baseColor will not be visible. The shininess of the metallic component is governed by flakeSpread, which controls the variation of the orientation of the flakes, similar to the roughness parameter of Metal. Note that the effect of the metallic flakes is currently only computed on average, thus individual flakes are not visible.

The path tracer supports the Luminous material which emits light uniformly in all directions and which can thus be used to turn any geometric object into a light source. It is created by passing the type string “Luminous” to ospNewMaterial. The amount of constant radiance that is emitted is determined by combining the general parameters of lights: color and intensity.

: Parameters accepted by the Luminous material.

OSPRay currently implements two texture types (texture2d and volume) and is open for extension to other types by applications. More types may be added in future releases.

To create a new texture use

OSPTexture ospNewTexture(const char *type);

The texture2d texture type implements an image-based texture, where its parameters are as follows

: Parameters of texture2d texture type.

The supported texture formats for texture2d are:

: Supported texture formats by texture2d, i.e., valid constants of type OSPTextureFormat.

The size of the texture is inferred from the size of the 2D array data, which also needs have a compatible type to format. The texel data in data starts with the texels in the lower left corner of the texture image, like in OpenGL. Per default a texture fetch is filtered by performing bi-linear interpolation of the nearest 2×2 texels; if instead fetching only the nearest texel is desired (i.e., no filtering) then pass the OSP_TEXTURE_FILTER_NEAREST flag.

The volume texture type implements texture lookups based on 3D world coordinates of the surface hit point on the associated geometry. If the given hit point is within the attached volume, the volume is sampled and classified with the transfer function attached to the volume. This implements the ability to visualize volume values (as colored by its transfer function) on arbitrary surfaces inside the volume (as opposed to an isosurface showing a particular value in the volume). Its parameters are as follows

: Parameters of volume texture type.

TextureVolume can be used for implementing slicing of volumes with any geometry type. It enables coloring of the slicing geometry with a different transfer function than that of the sliced volume.

All materials with textures also offer to manipulate the placement of these textures with the help of texture transformations. If so, this convention shall be used. The following parameters (prefixed with “texture_name.”) are combined into one transformation matrix:

: Parameters to define texture coordinate transformations.

The transformations are applied in the given order. Rotation, scale and translation are interpreted “texture centric”, i.e., their effect seen by an user are relative to the texture (although the transformations are applied to the texture coordinates).

To create a new camera of given type type use

OSPCamera ospNewCamera(const char *type);

All cameras accept these parameters:

: Parameters accepted by all cameras.

The camera is placed and oriented in the world with position, direction and up. OSPRay uses a right-handed coordinate system. The region of the camera sensor that is rendered to the image can be specified in normalized screen-space coordinates with imageStart (lower left corner) and imageEnd (upper right corner). This can be used, for example, to crop the image, to achieve asymmetrical view frusta, or to horizontally flip the image to view scenes which are specified in a left-handed coordinate system. Note that values outside the default range of [0–1] are valid, which is useful to easily realize overscan or film gate, or to emulate a shifted sensor.

The perspective camera implements a simple thin lens camera for perspective rendering, supporting optionally depth of field and stereo rendering, but no motion blur. It is created by passing the type string “perspective” to ospNewCamera. In addition to the general parameters understood by all cameras the perspective camera supports the special parameters listed in the table below.

: Parameters accepted by the perspective camera.

Note that when computing the aspect ratio a potentially set image region (using imageStart & imageEnd) needs to be regarded as well.

In architectural photography it is often desired for aesthetic reasons to display the vertical edges of buildings or walls vertically in the image as well, regardless of how the camera is tilted. Enabling the architectural mode achieves this by internally leveling the camera parallel to the ground (based on the up direction) and then shifting the lens such that the objects in direction dir are centered in the image. If finer control of the lens shift is needed use imageStart & imageEnd. Because the camera is now effectively leveled its image plane and thus the plane of focus is oriented parallel to the front of buildings, the whole façade appears sharp, as can be seen in the example images below.

The orthographic camera implements a simple camera with orthographic projection, without support for depth of field or motion blur. It is created by passing the type string “orthographic” to ospNewCamera. In addition to the general parameters understood by all cameras the orthographic camera supports the following special parameters:

: Parameters accepted by the orthographic camera.

For convenience the size of the camera sensor, and thus the extent of the scene that is captured in the image, can be controlled with the height parameter. The same effect can be achieved with imageStart and imageEnd, and both methods can be combined. In any case, the aspect ratio needs to be set accordingly to get an undistorted image.

The panoramic camera implements a simple camera without support for motion blur. It captures the complete surrounding with a latitude / longitude mapping and thus the rendered images should best have a ratio of 2:1. A panoramic camera is created by passing the type string “panoramic” to ospNewCamera. It is placed and oriented in the scene by using the general parameters understood by all cameras.

To get the world-space position of the geometry (if any) seen at [0–1] normalized screen-space pixel coordinates screenPos use

void ospPick(OSPPickResult *,
             OSPFrameBuffer,
             OSPRenderer,
             OSPCamera,
             OSPWorld,
             osp_vec2f screenPos);

The result is returned in the provided OSPPickResult struct:

typedef struct {
    int hasHit;
    osp_vec3f worldPosition;
    OSPGeometricModel GeometricModel;
    uint32_t primID;
} OSPPickResult;

Note that ospPick considers exactly the same camera of the given renderer that is used to render an image, thus matching results can be expected. If the camera supports depth of field then the center of the lens and thus the center of the circle of confusion is used for picking.

The framebuffer holds the rendered 2D image (and optionally auxiliary information associated with pixels). To create a new framebuffer object of given size size (in pixels), color format, and channels use

OSPFrameBuffer ospNewFrameBuffer(osp_vec2i size,
                                 OSPFrameBufferFormat format = OSP_FB_SRGBA,
                                 uint32_t frameBufferChannels = OSP_FB_COLOR);

The parameter format describes the format the color buffer has on the host, and the format that ospMapFrameBuffer will eventually return. Valid values are:

: Supported color formats of the framebuffer that can be passed to ospNewFrameBuffer, i.e., valid constants of type OSPFrameBufferFormat.

The parameter frameBufferChannels specifies which channels the framebuffer holds, and can be combined together by bitwise OR from the values of OSPFrameBufferChannel listed in the table below.

: Framebuffer channels constants (of type OSPFrameBufferChannel), naming optional information the framebuffer can store. These values can be combined by bitwise OR when passed to ospNewFrameBuffer.

If a certain channel value is not specified, the given buffer channel will not be present. Note that OSPRay makes a clear distinction between the external format of the framebuffer and the internal one: The external format is the format the user specifies in the format parameter; it specifies what color format OSPRay will eventually return the framebuffer to the application (when calling ospMapFrameBuffer): no matter what OSPRay uses internally, it will simply return a 2D array of pixels of that format, with possibly all kinds of reformatting, compression/decompression, etc., going on in-between the generation of the internal framebuffer and the mapping of the externally visible one.

In particular, OSP_FB_NONE is a perfectly valid pixel format for a framebuffer that an application will never map. For example, an application driving a display wall may well generate an intermediate framebuffer and eventually transfer its pixel to the individual displays using an OSPImageOperation image operation.

The application can map the given channel of a framebuffer – and thus access the stored pixel information – via

const void *ospMapFrameBuffer(OSPFrameBuffer,
                              OSPFrameBufferChannel = OSP_FB_COLOR);

Note that OSP_FB_ACCUM or OSP_FB_VARIANCE cannot be mapped. The origin of the screen coordinate system in OSPRay is the lower left corner (as in OpenGL), thus the first pixel addressed by the returned pointer is the lower left pixel of the image.

A previously mapped channel of a framebuffer can be unmapped by passing the received pointer mapped to

void ospUnmapFrameBuffer(const void *mapped, OSPFrameBuffer);

The individual channels of a framebuffer can be cleared with

void ospResetAccumulation(OSPFrameBuffer);

This function will clear all accumulating buffers (OSP_FB_VARIANCE, OSP_FB_NORMAL, and OSP_FB_ALBEDO, if present) and resets the accumulation counter accumID. It is unspecified if the existing color and depth buffers are physically cleared when ospResetAccumulation is called.

If OSP_FB_VARIANCE is specified, an estimate of the variance of the last accumulated frame can be queried with

float ospGetVariance(OSPFrameBuffer);

Note this value is only updated after synchronizing with OSP_FRAME_FINISHED, as further described in asynchronous rendering.

The framebuffer takes a list of pixel operations to be applied to the image in sequence as an OSPData. The pixel operations will be run in the order they are in the array.

: Parameters accepted by the framebuffer.

Image operations are functions that are applied to every pixel of a frame. Examples include post-processing, filtering, blending, tone mapping, or sending tiles to a display wall. To create a new pixel operation of given type type use

OSPImageOperation ospNewImageOperation(const char *type);

The tone mapper is a pixel operation which implements a generic filmic tone mapping operator. Using the default parameters it approximates the Academy Color Encoding System (ACES). The tone mapper is created by passing the type string “tonemapper” to ospNewImageOperation. The tone mapping curve can be customized using the parameters listed in the table below.

: Parameters accepted by the tone mapper.

To use the popular “Uncharted 2” filmic tone mapping curve instead, set the parameters to the values listed in the table below.

: Filmic tone mapping curve parameters. Note that the curve includes an exposure bias to match 18% middle gray.

OSPRay comes with a module that adds support for Intel® Open Image Denoise. This is provided as an optional module as it creates an additional project dependency at compile time. The module implements a “denoiser” frame operation, which denoises the entire frame before the frame is completed.

Rendering is by default asynchronous (non-blocking), and is done by combining a frame buffer, renderer, camera, and world.

What to render and how to render it depends on the renderer’s parameters. If the framebuffer supports accumulation (i.e., it was created with OSP_FB_ACCUM) then successive calls to ospRenderFrame will progressively refine the rendered image. If additionally the framebuffer has an OSP_FB_VARIANCE channel then ospRenderFrame returns an estimate of the current variance of the rendered image, otherwise inf is returned. The estimated variance can be used by the application as a quality indicator and thus to decide whether to stop or to continue progressive rendering.

To start an render task, use

OSPFuture ospRenderFrame(OSPFrameBuffer,
                         OSPRenderer,
                         OSPCamera,
                         OSPWorld);

This returns an OSPFuture handle, which can be used to synchronize with the application, cancel, or query for progress of the running task. When ospRenderFrame is called, there is no guarantee when the associated task will begin execution.

Progress of a running frame can be queried with the following API function

float ospGetProgress(OSPFuture);

This returns the progress of the task in [0-1].

Applications can wait on the result of an asynchronous operation, or choose to only synchronize with a specific event. To synchronize with an OSPFuture use

void ospWait(OSPFuture, OSPSyncEvent = OSP_TASK_FINISHED);

The following are values which can be synchronized with the application

: Supported events that can be passed to ospWait.

Currently only rendering can be invoked asynchronously. However, future releases of OSPRay may add more asynchronous versions of API calls (and thus return OSPFuture).

Applications can query whether particular events are complete with

int ospIsReady(OSPFuture, OSPSyncEvent = OSP_TASK_FINISHED);

As the given running task runs (as tracked by the OSPFuture), applications can query a boolean [0,1] result if the passed event has been completed.

The use of either ospRenderFrame or ospRenderFrame requires that all objects in the scene being rendered have been committed before rendering occurs. If a call to ospCommit() happens while a frame is rendered, the result is undefined behavior and should be avoided.

For convenience in certain use cases, ospray_util.h provides a synchronous version of ospRenderFrame:

float ospRenderFrameBlocking(OSPFrameBuffer,
                             OSPRenderer,
                             OSPCamera,
                             OSPWorld);

This version is the equivalent of:

- `ospRenderFrame`
- `ospWait(f, OSP_TASK_FINISHED)`
- return `ospGetVariance(fb)`

This version is closest to ospRenderFrame from OSPRay v1.x.

The OSPRay MPI module is now a stand alone repository. It can be found on GitHub here, where all code and documentation can be found.

A minimal working example demonstrating how to use OSPRay can be found at apps/tutorials/ospTutorial.c⁷.

An example of building ospTutorial.c with CMake can be found in apps/tutorials/ospTutorialFindospray/.

To build the tutorial on Linux, build it in a build directory with

gcc -std=c99 ../apps/ospTutorial/ospTutorial.c \
-I ../ospray/include -L . -lospray -Wl,-rpath,. -o ospTutorial

On Windows build it can be build manually in a “build_directory\$Configuration” directory with

cl ..\..\apps\ospTutorial\ospTutorial.c -I ..\..\ospray\include -I ..\.. ospray.lib

Running ospTutorial will create two images of two triangles, rendered with the Scientific Visualization renderer with full Ambient Occlusion. The first image firstFrame.ppm shows the result after one call to ospRenderFrame – jagged edges and noise in the shadow can be seen. Calling ospRenderFrame multiple times enables progressive refinement, resulting in antialiased edges and converged shadows, shown after ten frames in the second image accumulatedFrames.ppm.

Apart from tutorials, OSPRay comes with a C++ app called ospExamples which is an elaborate easy-to-use tutorial, with a single interface to try various OSPRay features. It is aimed at providing users with multiple simple scenes composed of basic geometry types, lights, volumes etc. to get started with OSPRay quickly.

ospExamples app runs a GLFWOSPRayWindow instance that manages instances of the camera, framebuffer, renderer and other OSPRay objects necessary to render an interactive scene. The scene is rendered on a GLFW window with an imgui GUI controls panel for the user to manipulate the scene at runtime.

The application is located in apps/ospExamples/ directory and can be built with CMake. It can be run from the build directory via:

./ospExamples <command-line-parameter>

The command line parameter is optional however.

Different scenes can be selected from the scenes dropdown and each scene corresponds to an instance of a special detail::Builder struct. Example builders are located in apps/common/ospray_testing/builders/. These builders provide a usage guide for the OSPRay scene hierarchy and OSPRay API in the form of cpp wrappers. They instantiate and manage objects for the specific scene like cpp::Geometry, cpp::Volume, cpp::Light etc.

The detail::Builder base struct is mostly responsible for setting up OSPRay world and objects common in all scenes (for eg: lighting and ground plane), which can be conveniently overridden in the derived builders.

Given below are different scenes listed with their string identifiers:

boxes : A simple scene with box geometry type.

cornell_box : A scene depicting a classic cornell box with quad mesh geometry type for rendering two cubes and a quad light type.

curves : A simple scene with curve geometry type and options to change curveBasis. For details on different basis’ please check documentation of curves.

gravity_spheres_volume : A scene with structuredRegular type of volume.

gravity_spheres_isosurface : A scene depicting iso-surface rendering of gravity_spheres_volume using geometry type isosurface.

perlin_noise_volumes : An example scene with structuredRegular volume type depicting perlin noise.

random_spheres : A simple scene depicting sphere geometry type.

streamlines : A scene showcasing streamlines geometry derived from curve geometry type.

subdivision_cube : A scene with a cube of subdivision geometry type to showcase subdivision surfaces.

unstructured_volume : A simple scene with a volume of unstructured volume type.

This app comes with three renderer options: scivis, pathtracer and debug. The app provides some common rendering controls like pixel samples and other more specific to the renderer type like aoIntensity for scivis renderer.