* The whole scene should be contained in a <scene> ... </scene> node.
* In a scene you can have (multiple of) the following nodes:
- <camera>
- <sceneGraphNode>
- <triangleMesh>
- <material>
- <light>
- <intersector>
- <texture>
- <renderer>
* Basic value types:
- vec3d "x y z" or "xyz". E.g., name="1.0 2.0 3.0" (vec3d(1,2,3)) or name="1.0" (vec3d(1,1,1)).
- vec2d "x y" or "xy"
- float "f"
- string "s"
- integer "i"
- bool "true" or "false"
- color "r g b" or "rgb"
* A <camera> node is of the form: <camera eye=vec3d view=vec3d up=vec3d fov=float width=integer height=integer auto=bool/>, with the following defaults:
- eye="0.0 0.0 0.0"
- view="0.0 0.0 1.0"
- up="0.0 1.0 0.0"
- fov="60"
- width="256"
- height="256"
- auto="false"
When 'auto' is set to true, the camera will be moved along the view direction (i.e., change 'eye') such that the scene's bounding box fully falls within the view.
* <sceneGraphNode> has the following form: <sceneGraphNode id=string ref=string> .. </sceneGraphNode>, where either id or ref occurs. 'id' allows you to name the node. 'ref' refers to a named sceneGraphNode, and the rest of the sceneGraphNode is ignored. A sceneGraphNode can contain (multiple of) the following nodes:
- <transformation>
- <sceneGraphNode>
- <triangleMesh>
- <material>
* A <transformation> node can have any of multiple of the following nodes:
- <rotationX>
- <rotationY>
- <rotationZ>
- <rotation>
- <translation>
- <scale>
The resulting transformation is the multiplication of the nodes.E.g., a <rotationX> followed by a <rotationY> and then a <scale>, will be equivalent to <rotationX>*<rotationY>*<scale>
* A <rotation?> has the following form: <rotation? angle=float />
* A <translation> has the form: <translation offset=vec3d />
* A <scale> has the form: <scale scale=vec3d />
* A <rotation> has the form: <rotation angle=float axis=vec3d />, with a default of axis="0.0 0.0 1.0".
* A <triangleMesh> has the following form: <triangleMesh id=string ref=string filename=string> .. </triangleMesh>, where 'ref' and 'id' are the same as with <sceneGraphNode>, and 'obj' can contain the name of an OBJ file. A triangleMesh can contain (multiple of) the following nodes:
- <transformation>
- <material>
- <triangle> or <polygon>
* A <triangle> (or <polygon> node can have multiple occurances of a <vertex> which can have any of the following forms:
- <vertex v=vec3d />
- <vertex v=vec3d t=vec2d />
- <vertex v=vec3d n=vec3d />
- <vertex v=vec3d n=vec3d t=vec2d />
where 'v' refers to the vertex position, 'n' to the vertex normal, and 't' to the vertex' texture coordinate. Vertices are expected to be provided in CCW order.
* A <material> node has the following form: <material id=string ref=string> which can have the following children:
- <shadingFrameTransform>
- <diffuse>
- <phong>
- <material> or <compound>
* A <shadingFrameTransformation> has the following form <shadingFrameTransform type=string> with the following general children:
- <shadingFrameTransform>
- <diffuse>
- <phong>
- <material> or <compound>
in addition to specific node dependent children depending on the 'type' which can be:
- normalMap
- bumpMap
* A <normalMap> node has a <texture> child node, and a <material> node to apply it too.
* A <bumpMap> node has a <texture> child node, and a <material> node to apply it too. In addition, the shadingFrameTransformation-node also has a 'scale' parameter defined that sets the height of the bump, and a 'channel' parameter that indicates which color channel in the texture defines the bump map.
* A <diffuse> node has the following form <diffuse id=string ref=string>...</difuse>, where it can have the following children:
- <albedo value=color> and which can have a <texture> as child. (preference if given of 'texture' over a fixed 'value'
* A <phong> node has the following form <phong id=string ref=string>...</phong>, where it can have the following children:
- <albedo value=color> and which can have a <texture> as child. (preference if given of 'texture' over a fixed 'value'
- <sharpness value=float channel=integer> and which can have a <texture> as child. (preference if given of 'texture' over a fixed 'value'
* A <light> node has the following form: <light type=string .../>. Depending on type the following additional parameters are defined:
- "directional": 'direction=vec3d power=color'
- "spot": 'direction=vec3d power=color position=vec3d cutoff=float sharpness=float attenutation=vec3d'
- "point": 'position=vec3d power=color attenuation=vec3d'
- "area": 'power=color attenuation=vec3d'. Furthermore, it can have a <material> and <sceneGraphNode>/<triangleMesh> children. The material is the surface reflectance of the light source, and the geometry represents the shape of the area light source.
* An <intersector> node has the following form: <intersector type=string />, where type can be any of the following:
- "linear" : brute force linear intersector O(N) time complexity.
- "bvh" : bounding volume hierarchy
* A <texture> node has the following form: <texture type=string filename=string repeat=bool/>, where type can be
- "nearest": nearest pixel look-up.
- "bilinear": bilinear interpolation.
Textures are cached (map based on filename & type) and only read once.
* A <renderer> node has the following forms:
- <renderer type="raycasting"/>
- <renderer type="recursiveRaytracing" maxDepth=int samplesPerPixel=int />
- <renderer type="pathtracing" samplesPerPixel=int directOnly=bool />
