So, I've written spring cameras before in other APIs, but I can't seem to wrap my head around Unity's (I'm still fairly new to it). Here is my SpringFollow camera script in C#
using UnityEngine;
using System.Collections;
public class SpringFollow : MonoBehaviour
{
public Transform Target;
public Camera SpringCamera = Camera.mainCamera;
public float Stiffness = 1800.0f;
public float Damping = 600.0f;
public float Mass = 50.0f;
public Vector3 DesiredOffset = new Vector3(0.0f, 3.5f, -4.0f);
public Vector3 LookAtOffset = new Vector3(0.0f, 3.1f, 0.0f);
private Vector3 desiredPosition = Vector3.zero;
private Vector3 cameraVelocity = Vector3.zero;
// Use this for initialization
void Start ()
{
}
void FollowUpdate()
{
Vector3 stretch = SpringCamera.transform.position - desiredPosition;
Vector3 force = -Stiffness * stretch - Damping * cameraVelocity;
Vector3 acceleration = force / Mass;
cameraVelocity += acceleration * Time.deltaTime;
SpringCamera.transform.position += cameraVelocity * Time.deltaTime;
Matrix4x4 CamMat = new Matrix4x4();
CamMat.SetRow(0, new Vector4(-Target.forward.x, -Target.forward.y, -Target.forward.z));
CamMat.SetRow(1, new Vector4(Target.up.x, Target.up.y, Target.up.z));
Vector3 modRight = Vector3.Cross(CamMat.GetRow(1), CamMat.GetRow(0));
CamMat.SetRow(2, new Vector4(modRight.x, modRight.y, modRight.z));
desiredPosition = Target.position + TransformNormal(DesiredOffset, CamMat);
Vector3 lookat = Target.position + TransformNormal(LookAtOffset, CamMat);
SpringCamera.transform.LookAt(lookat, Target.up);
//SpringCamera.projectionMatrix = Matrix4x4.Perspective(SpringCamera.fieldOfView, SpringCamera.aspect, SpringCamera.near, SpringCamera.far);
print("cam = " + SpringCamera.transform.position.ToString());
}
// Update is called once per frame
void Update ()
{
FollowUpdate();
}
Vector3 TransformNormal(Vector3 normal, Matrix4x4 matrix)
{
Vector3 transformNormal = new Vector3();
Vector3 axisX = new Vector3(matrix.m00, matrix.m01, matrix.m02);
Vector3 axisY = new Vector3(matrix.m10, matrix.m11, matrix.m12);
Vector3 axisZ = new Vector3(matrix.m20, matrix.m21, matrix.m22);
transformNormal.x = Vector3.Dot(normal, axisX);
transformNormal.y = Vector3.Dot(normal, axisY);
transformNormal.z = Vector3.Dot(normal, axisZ);
return transformNormal;
}
}
I originally did this code in XNA a while back, which may be painfully evident with me manually adding in the TransformNormal(Vector3 normal, Matrix4x4 matrix) function. The result I'm getting from this is a camera that's attached to the side of my target and has twitchy behavior. So what am I doing horribly wrong?
