Instead of rotating the camera with camera.rotation or with the lookAt() function I\'d like to pass a look vector directly to the camera... Is it possible to set a camera lo
The camera does not have a "look vector", so you cannot set it.
You can, however, construct a point
to look at by adding your look vector to the camera's position, and then calling
camera.lookAt( point );
Here is how to determine the direction in which the camera is looking, assuming the camera either has no parent (other than the scene).
The camera is looking down its internal negative z-axis, so create a vector pointing down the negative z-axis:
var vector = new THREE.Vector3( 0, 0, - 1 );
Now, apply the same rotation to the vector that is applied to the camera:
vector.applyQuaternion( camera.quaternion );
The resulting vector will be pointing in the direction that the camera is looking.
Alternatively, you can use the following method, which works even if the camera is a child of another object:
camera.getWorldDirection( dirVector );
three.js r.73
What I did was to use method lookAt(Vector) just before render the scene like in below code , just try it using it on a new html file :)
<!DOCTYPE html>
<html>
<head>
<meta charset="utf-8">
<title></title>
<style >
*{
margin: 0 ;
}
#WebGlElement {
height: 500px;
width: 500px;
background-color: red
}
</style>
</head>
<body>
<script src="js/three.min.js"></script>
<div id="WebGlElement"></div>
<script>
var contianer = document.getElementById("WebGlElement");
var origin = new THREE.Vector3(0,0,0);
// CREATE THREE BASIC ELEMENTS
var scene = new THREE.Scene();
var camera = new THREE.PerspectiveCamera( 40, 300/300, 1, 1000 );
var renderer = new THREE.WebGLRenderer();
// SET RENDERER AND ATTACH IT TO BODY
renderer.setSize( 500, 500 );
//document.body.appendChild( renderer.domElement );
contianer.appendChild( renderer.domElement);
// CREATE A GEOMETRY AND ADD IT TO SCENE
/*var geometry = new THREE.BoxGeometry( 1, 1, 5 );
var material = new THREE.MeshBasicMaterial( { color: 0x1abc9c } );
material.wireframe = true ;
material.wireframeLinewidth = 0.1 ;
var cube = new THREE.Mesh( geometry, material );
scene.add( cube );*/
var geometry = new THREE.Geometry();
geometry.elementsNeedUpdate=true;
geometry.vertices.push(
new THREE.Vector3( -0, 1, 0 ),
new THREE.Vector3( -1, -1, 0 ),
new THREE.Vector3( 1, -1, 0 )
);
geometry.faces.push(
new THREE.Face3( 0, 1, 2 ),
new THREE.Face3( 2, 1, 0 )
);
var material = new THREE.MeshBasicMaterial( { color: 0x1abc9c } );
//material.wireframe = true ;
//material.wireframeLinewidth = 0.1 ;
var cube = new THREE.Mesh( geometry, material );
scene.add( cube );
var axisHelper = new THREE.AxisHelper( 1 );
scene.add( axisHelper );
// POSITION OF CAMER ON Z
camera.position.z = 5;
camera.position.y = 5;
camera.up = new THREE.Vector3(0,0,1);
var dir = 1;
var number = 0.115
// CREATE FUNCTION FOR RENDER
var render = function () {
requestAnimationFrame( render );
//circle.rotation.x += 0.01;
if ( camera.position.x> 5) {
dir = -1;
}
if ( camera.position.x< -5) {
dir = 1;
}
camera.lookAt(cube.position);
//camera.rotation.y += 0.015 * dir;
camera.position.x += number * dir;
renderer.render(scene, camera);
};
// EXECUTE FIRST RENDER
render();
</script>
</body>
</html>
These other answers are very insightful, but not completely correct. The code returns a vector that points in the same direction that the camera is pointing at. That's a great start!
But unless the camera is at the origin (0, 0, 0) (or lies exactly on the line segment that connects the origin to the rotated vector point without passing beyond that point so that the point is behind the camera) the camera won't be looking at that point. Clearly -- just common sense -- the position of the camera also influences what points are being looked at. Just think about it!!
The camera method lookAt()
looks at a point in 3D space regardless of where the camera is. So to get a point that the camera is looking at you need to adjust for the camera position by calling:
vec.add( camera.position );
It is also worth mentioning that the camera is looking not at a single point but is looking down a line of an infinite number of points, each at a different distance from the camera. The code from the other answers returns a vector that is exactly one unit in length because the application of a quaternion to the normalized z-axis vector (0, 0, -1) returns another normalized vector (in a different direction). To calculate the look at point at an arbitrary distance x
from the camera use:
THREE.Vector3( 0, 0, -x ).applyQuaternion( camera.quaternion ).add( camera.position );
This takes a point along the z-axis at a distance x
from the origin, rotates it to the direction of the camera and then creates a "world point" by adding the camera's position. We want a "world point" (and not just a "relative to the camera point") since camera.lookAt()
also takes a "world point" as a parameter.
The above answer wrapped as a util, this is what I do with my Three Utils:
THREE.Utils = {
cameraLookDir: function(camera) {
var vector = new THREE.Vector3(0, 0, -1);
vector.applyEuler(camera.rotation, camera.eulerOrder);
return vector;
}
};
Call it with THREE.Utils.cameraLookDir(camera);