Assimp  v3.0 (July 2012)
Public Member Functions | Public Attributes
aiCamera Struct Reference

Helper structure to describe a virtual camera. More...

List of all members.

Public Member Functions

 aiCamera ()
void GetCameraMatrix (aiMatrix4x4 &out) const
 Get a right-handed camera matrix from me.

Public Attributes

float mAspect
 Screen aspect ratio.
float mClipPlaneFar
 Distance of the far clipping plane from the camera.
float mClipPlaneNear
 Distance of the near clipping plane from the camera.
float mHorizontalFOV
 Half horizontal field of view angle, in radians.
aiVector3D mLookAt
 'LookAt' - vector of the camera coordinate system relative to the coordinate space defined by the corresponding node.
aiString mName
 The name of the camera.
aiVector3D mPosition
 Position of the camera relative to the coordinate space defined by the corresponding node.
aiVector3D mUp
 'Up' - vector of the camera coordinate system relative to the coordinate space defined by the corresponding node.

Detailed Description

Helper structure to describe a virtual camera.

Cameras have a representation in the node graph and can be animated. An important aspect is that the camera itself is also part of the scenegraph. This means, any values such as the look-at vector are not absolute, they're relative to the coordinate system defined by the node which corresponds to the camera. This allows for camera animations. For static cameras parameters like the 'look-at' or 'up' vectors are usually specified directly in aiCamera, but beware, they could also be encoded in the node transformation. The following (pseudo)code sample shows how to do it:

// Get the camera matrix for a camera at a specific time
// if the node hierarchy for the camera does not contain
// at least one animated node this is a static computation
get-camera-matrix (node sceneRoot, camera cam) : matrix
{
node cnd = find-node-for-camera(cam)
matrix cmt = identity()
// as usual - get the absolute camera transformation for this frame
for each node nd in hierarchy from sceneRoot to cnd
matrix cur
if (is-animated(nd))
cur = eval-animation(nd)
else cur = nd->mTransformation;
cmt = mult-matrices( cmt, cur )
end for
// now multiply with the camera's own local transform
cam = mult-matrices (cam, get-camera-matrix(cmt) )
}
Note:
some file formats (such as 3DS, ASE) export a "target point" - the point the camera is looking at (it can even be animated). Assimp writes the target point as a subnode of the camera's main node, called "<camName>.Target". However this is just additional information then the transformation tracks of the camera main node make the camera already look in the right direction.

Constructor & Destructor Documentation

aiCamera::aiCamera ( )
inline

Member Function Documentation

void aiCamera::GetCameraMatrix ( aiMatrix4x4 out) const
inline

Get a right-handed camera matrix from me.

Parameters:
outCamera matrix to be filled

todo: test ... should work, but i'm not absolutely sure

We don't know whether these vectors are already normalized ...


Member Data Documentation

float aiCamera::mAspect

Screen aspect ratio.

This is the ration between the width and the height of the screen. Typical values are 4/3, 1/2 or 1/1. This value is 0 if the aspect ratio is not defined in the source file. 0 is also the default value.

float aiCamera::mClipPlaneFar

Distance of the far clipping plane from the camera.

The far clipping plane must, of course, be further away than the near clipping plane. The default value is 1000.f. The ratio between the near and the far plane should not be too large (between 1000-10000 should be ok) to avoid floating-point inaccuracies which could lead to z-fighting.

float aiCamera::mClipPlaneNear

Distance of the near clipping plane from the camera.

The value may not be 0.f (for arithmetic reasons to prevent a division through zero). The default value is 0.1f.

float aiCamera::mHorizontalFOV

Half horizontal field of view angle, in radians.

The field of view angle is the angle between the center line of the screen and the left or right border. The default value is 1/4PI.

aiVector3D aiCamera::mLookAt

'LookAt' - vector of the camera coordinate system relative to the coordinate space defined by the corresponding node.

This is the viewing direction of the user. The default value is 0|0|1. The vector may be normalized, but it needn't.

aiString aiCamera::mName

The name of the camera.

There must be a node in the scenegraph with the same name. This node specifies the position of the camera in the scene hierarchy and can be animated.

aiVector3D aiCamera::mPosition

Position of the camera relative to the coordinate space defined by the corresponding node.

The default value is 0|0|0.

aiVector3D aiCamera::mUp

'Up' - vector of the camera coordinate system relative to the coordinate space defined by the corresponding node.

The 'right' vector of the camera coordinate system is the cross product of the up and lookAt vectors. The default value is 0|1|0. The vector may be normalized, but it needn't.


The documentation for this struct was generated from the following file: