问题
Maybe this is more of a math question, but I am hoping somebody can help me understand how I would compensate for changes to my camera's orientation.
What I would like to do is to be able to move my camera around the scene using game-like controls. I have W S A D keys set up to move the camera forward, backward, left and right respectively. I do this by changing the values of the camera node's position vector. I also have the right left up and down arrow keys mapped to turning and tilting the camera. I do this by adjusting the camera node's euler rules (pitch and yaw).
Things work well until I turn the camera with the right and left arrows. Pressing W after this moves the camera in the direction it was facing before I applied yaw to it. The behavior makes sense to me, but I cannot figure out what I need to do to my position-vector tweaking to compensate for the yaw I've applied.
Do I need to multiply my position vector somehow by new transform? Or better yet is there some other property like the pivot that I can change, so that I can keep my position code the same?
Thanks for any pointers.
Update here is the code I'm working with so far:
public func displayTimerDidFire(timer: MyCustomDisplayTimer!) {
if timer === cameraMoveTimer {
var cameraTransform = cameraNode.transform
var x = CGFloat(0.0)
var y = CGFloat(0.0)
var z = CGFloat(0.0)
var step : CGFloat = 10.0
if moveUpKeyDown {
y += step
}
if moveDownKeyDown {
y -= step
}
if moveLeftKeyDown {
x -= step
}
if moveRightKeyDown {
x += step
}
if moveForwardKeyDown {
z -= step
}
if moveBackwardKeyDown {
z += step
}
cameraTransform = SCNMatrix4Translate(cameraTransform, x, y, z)
cameraNode.transform = cameraTransform
}
else if timer === cameraTiltTimer {
var angles = cameraNode.eulerAngles
var stepAngle : CGFloat = CGFloat(M_PI_2 / 90.0)
if turnLeftKeyDown {
angles.y += stepAngle
}
if turnRightKeyDown {
angles.y -= stepAngle
}
if tiltForwardKeyDown {
angles.x -= stepAngle
}
if tiltBackwardKeyDown {
angles.x += stepAngle
}
cameraNode.eulerAngles = angles
}
}
I have two timers, one that updates the cameras position, and the other that updates its orientation.
The part I'm stuck on is that I know my new transform (old position plus the move increment from A S D W), and I know the camera's rotation, but I don't know how to combine those two into my true new transform. Transform is a SCNMatrix4, and rotation is a SCNVector4.
Update 2
Here is a reworked version of the code. I switched to working with the transform instead. I still cannot find the right manipulation to account for the rotation however.
public func displayTimerDidFire(timer: MyDisplayTimer!) {
var cameraTransform = cameraNode.transform
var x = CGFloat(0.0)
var y = CGFloat(0.0)
var z = CGFloat(0.0)
var step : CGFloat = 10.0
if moveUpKeyDown {
y += step
}
if moveDownKeyDown {
y -= step
}
if moveLeftKeyDown {
x -= step
}
if moveRightKeyDown {
x += step
}
if moveForwardKeyDown {
z -= step
}
if moveBackwardKeyDown {
z += step
}
cameraTransform = SCNMatrix4Translate(cameraTransform, x, y, z)
// Do something with the transform to compensate for rotation..
// ???
cameraNode.transform = cameraTransform
var angles = cameraNode.eulerAngles
var stepAngle : CGFloat = CGFloat(M_PI_2 / 90.0)
if turnLeftKeyDown {
angles.y += stepAngle
}
if turnRightKeyDown {
angles.y -= stepAngle
}
if tiltForwardKeyDown {
angles.x -= stepAngle
}
if tiltBackwardKeyDown {
angles.x += stepAngle
}
cameraNode.eulerAngles = angles
// printNode(cameraNode)
}
Last Update
Thanks for the suggestions. Here is the implementation that is working for me
I needed this objc helper, as some of these functions to not seem to be available in Swift:
@implementation MySceneKitUtils
+ (SCNVector3)position:(SCNVector3)position multipliedByRotation:(SCNVector4)rotation
{
if (rotation.w == 0) {
return position;
}
GLKVector3 gPosition = SCNVector3ToGLKVector3(position);
GLKMatrix4 gRotation = GLKMatrix4MakeRotation(rotation.w, rotation.x, rotation.y, rotation.z);
GLKVector3 r = GLKMatrix4MultiplyVector3(gRotation, gPosition);
return SCNVector3FromGLKVector3(r);
}
@end
And the implementation
public func displayTimerDidFire(timer: MyDisplayTimer!) {
var x = CGFloat(0.0)
var y = CGFloat(0.0)
var z = CGFloat(0.0)
var step : CGFloat = 10.0
if moveUpKeyDown {
y += step
}
if moveDownKeyDown {
y -= step
}
if moveLeftKeyDown {
x -= step
}
if moveRightKeyDown {
x += step
}
if moveForwardKeyDown {
z -= step
}
if moveBackwardKeyDown {
z += step
}
var cameraTransform = cameraNode.transform
var rotation = cameraNode.rotation
var rotatedPosition = MySceneKitUtils.position(SCNVector3Make(x, y, z), multipliedByRotation: cameraNode.rotation)
cameraTransform = SCNMatrix4Translate(cameraTransform, rotatedPosition.x, rotatedPosition.y, rotatedPosition.z)
cameraNode.transform = cameraTransform
// The rotation
cameraTransform = cameraNode.transform
var stepAngle = CGFloat(0.05)
if turnLeftKeyDown {
cameraTransform = SCNMatrix4Rotate(cameraTransform, stepAngle, 0.0, 1.0, 0.0);
}
if turnRightKeyDown {
cameraTransform = SCNMatrix4Rotate(cameraTransform, -stepAngle, 0.0, 1.0, 0.0);
}
if tiltForwardKeyDown {
cameraTransform = SCNMatrix4Rotate(cameraTransform, stepAngle, 1.0, 0.0, 0.0);
}
if tiltBackwardKeyDown {
cameraTransform = SCNMatrix4Rotate(cameraTransform, -stepAngle, 1.0, 0.0, 0.0);
}
cameraNode.transform = cameraTransform
}
回答1:
if t
is the translation you compute from the WSAD keys, then you should not simply add it (node.position += t
) but first apply a rotation and then add it : node.position += node.rotation * t
回答2:
My answer is add on to the above answer. I have used this for Pinch zoom.
//Creating unit vector
let unitVector = GLKVector4Make(0, 0, -1, 1)
//converting tranform matrix
let glktranform = SCNMatrix4ToGLKMatrix4(cameraNode.transform)
//multiply unit vector with transform matrix
let rotatedMatrix = GLKMatrix4MultiplyVector4(glktranform, unitVector)
//scale it with translation you compute from the WSAD key
let finalScaledVector = GLKVector4MultiplyScalar(rotatedMatrix, translation)
cameraNode.position.x = finalScaledVector.x
cameraNode.position.y = finalScaledVector.y
cameraNode.position.z = finalScaledVector.z
Hopefully this helps.
来源:https://stackoverflow.com/questions/27738794/scenekit-camera-how-to-compensate-for-euler-rules-orientation-changes