Animate on bézier with ScrollMagic: Initial state = first bézier point - animation

See a jsfiddle here
I am tweening along a bézier path with 3 points.
// bezier data
var bezierData = {
curviness: 1,
autoRotate: false,
values: [
{x: 0, y: 0, rotation:"40_cw"}, /* <-- The desired state of the object before any animation has happened */
{x: 20, y: 0, rotation:"0_ccw"},
{x: 40, y:0, rotation:"-20_ccw"}
]
};
// build tween
var tween = new TimelineMax()
.add(TweenMax.to("#testobj", 1, {css:{bezier: bezierData}, ease:Power1.easeInOut}));
// create scene
var scene = new ScrollMagic.Scene({
triggerElement: "#testobj",
duration: 100,
offset: 10
})
.setTween(tween)
.addTo(ctrl)
.addIndicators();
What I want: The initial state of my object (i.e. before any animation has happened) should be the first bézier point, {x: 0, y: 0, rotation:"40_cw"}.
What's happening: The initial state is the object's default style, i.e. the equivalent of {x: 0, y: 0, rotation:"0"}. Note how in the jsfiddle the green square starts out upright while I want it to start rotated 40° clock-wise.

Tahir Ahmed's answer works!
perhaps you can use .set() before doing the .to() tween? something like this.

Related

shadow and rippled-texture in threejs

I try to apply texture on ground (a THREE.BoxBufferGeometry) , with shadow of a sphere (sphere is animated by Ammo.js)
The presence of shadow seems to depends of two parameters:
the shadow.camera settings of the THREE.DirectionalLight :
thObject.shadow.camera.left = -d;
thObject.shadow.camera.right = d;
thObject.shadow.camera.top = d;
thObject.shadow.camera.bottom = -d;
the shadow.camera.far setting :
thObject.shadow.camera.far = 110;
(it seems to depend of light position regarding shadow projection on ground).
But "playing" with these two settings create the following situation :
ripples on texture
shadow or not shadow (see after a part of shadow).
How to set these parameters but not by trial-and-error ?
(This depends of light position too): for exemple if light is not pointing in vertical axle:
pos: { x: -20, y: 35, z: -20 },
lookAt: { x: 0, y: 0, z: 0 },

threejs applyMatrix4 appears to do nothing

The method applyMatrix4 seems like it does nothing...
Why can I not apply this transformation matrix to my vector?
const vec = new THREE.Vector3(1,1,1)
const geometry = new THREE.BoxGeometry(1,1,1)
const material = new THREE.MeshBasicMaterial({ color: 0xff0000 })
const mesh = new THREE.Mesh(geometry, material)
mesh.rotateX(Math.PI)
const rotatedVec = vec.applyMatrix4(mesh.matrix)
console.log(rotatedVec)
Expectation (taking the cross product):
{x: 1, y: -1, z: -1}
Reality (the vector is unchanged)
{x: 1, y: 1, z: 1}
My mesh's matrix has changed - it is not the identity matrix.
[
[1, 0, 0, 0],
[0, -1, 0, 0],
[0, 0, -1, 0],
[0, 0, 0, 1],
]
Object3D.rotateX() only affects the object's quaternion property. It does not update its local matrix. If you say your matrix has changed, I assume you have checked it at a later point when other engine logic triggers a recalculation.
You can solve this issue by adding mesh.updateMatrix(); after you have called Object3D.rotateX().
Or even better use Vector3.applyQuaternion(). In this way, you don't have to recompute the matrix because you don't need it anyway.
const rotatedVec = vec.applyQuaternion(mesh.quaternion)

Animate every point of a line to another line with three.js

I want to implement an animation.
The animation should be a line move to another line. There will be some deformation in the process of the line moving
There is a correspondence between the points of the two lines.
How to implements it with three.js?
I try to use the tween.js.It works.
const material = new THREE.LineBasicMaterial({ color: 0x0000ff })
const geometry = new THREE.BufferGeometry().setAttribute('position',
new THREE.Float32BufferAttribute([-2, 0, 0, -0.5, 0, -0.5, 0, 0, -2], 3))
const line = new THREE.Line(geometry, material)
var position2 = new Float32Array([5, 0, 0, 1, 1, 1, 0, 0, 5])
setupObjectPositionTween(line, geometry.attributes.position.array, position2,
10000, 0, TWEEN.Easing.Linear.None) // duration, delay, easing
scene.add(line)
function setupObjectPositionTween(object, source, target, duration, delay, easing) {
new TWEEN.Tween(source)
.to(target, duration)
.delay(delay)
.easing(easing)
.onUpdate(function () {
// console.log(object,source,target)
object.geometry.attributes.position.array = source
object.geometry.attributes.position.needsUpdate=true
})
.start()
}
and in the render function
TWEEN.update()
I suggest you implement the animation of the green line with morph targets. Meaning the green line represents your default geometry whereas the blue line represents a morph target (also known as blend shape). You can then animate the transformation from green to blue by modulating the morphTargetInfluences parameter from 0 to 1.
Morph targets are part of the geometry data and defined in the BufferGeometry.morphAttributes property. Since multiple meshes can share the same geometry, the morphTargetInfluences property belongs to 3D objects like a mesh.
I suggest you study how the official example webgl_morphtargets is implemented. You can apply the same principles on lines.

How to animate something at the same time in GASP?

I want to animate these two things at the same time, and not one by one.
t1.fromTo(searchForm, { scaleX: 0 }, { duration: 1, autoAlpha: 1, scaleX: 1 });
t1.fromTo(loupe, {x: '-=0'}, {duration: 1, x: '+=265'})
I want to move them together; how can I do that?
Just use the position parameter to position your tween(s) wherever you want in the timeline.
t1.fromTo(searchForm, { scaleX: 0 }, { duration: 1, autoAlpha: 1, scaleX: 1 });
t1.fromTo(loupe, {x: '-=0'}, {duration: 1, x: '+=265'}, 0);
There's more info about the position parameter at https://greensock.com/position-parameter.
Also, there's no reason to use a .fromTo() in the second case because x: "-=0" does absolutely nothing, so you can just to a normal .to() tween and omit that whole object.
Happy tweening!

find upper face of cube on demand

The general problem I'm trying to solve is to find out what face of a cube faces upwards. The cube can be rolled 90° at a time and in any direction. If a certain face faces up, the cube disappears. I'm working with tweens to rotate the cube and change the position of it.
I'm currently trying to solve this by creating a new ray, with its origin set just above the cube and its direction going downwards for a short distance, so it intersects with the upper face of the cube only.
violet thingy on top of die is ray cast downward into the cube
So far so good. I get my cube as the object of intersection when I check per console.log(), but as soon as I try to access the face of intersection by faceIntersect.face it seems to be undefined.
Function in question:
function checkUpperFace(posX, posZ) {
// get position from passed x- and z-values (y is always above cube)
// and set direction and length of ray
var position = new THREE.Vector3( posX, 3, posZ );
var direction = new THREE.Vector3(0, -1, 0);
var far = 2;
// create ray, that goes downwards from above the cube
var cubeRaycaster = new THREE.Raycaster( position, direction, 0, far );
// get intersection with upper face of rolled cube
var faceIntersect = cubeRaycaster.intersectObject( currentCube );
// add a helper to see the ray
var arrowHelper = new THREE.ArrowHelper( direction, position, far, 0x770077 );
scene.add( arrowHelper );
console.log(faceIntersect); // object is shown with everything I want to know
console.log(faceIntersect.face); // is shown to be undefined
}
In the end I did it in a way #unx recommended but I really wanted to avoid the huge if-else statement, so I did it with an array rotationLibrary that has all possible rotations with the corresponding top face of the die. But because of the tween I use to rotate and move the die its rotation values are not really on point and therefore hard to compare to fixed rotation values as I use them in the array.
So I "normalize" the rotation values of the die to values I can use to compare them to my values in rotationLibrary. The last step is to store/update the result on what face is on top in the cube object itself, so I can get it whenever I want.
// spawn condition:
// 1 on top, 2 facing camera, 3 facing right (seen from camera),
// 4 facing left (see 3), 5 facing away from camera, 6 facing down
var rotationLibrary = [
{x: 0, y: 0, z: 0, face: 1},
{x: 0, y: 90, z: 0, face: 1},
{x: 180, y: 0, z: 180, face: 1},
{x: 0, y: -90, z: 0, face: 1},
{x: -90, y: 0, z: 0, face: 2},
{x: -90, y: 0, z: 90, face: 2},
{x: -90, y: 0, z: 180, face: 2},
{x: -90, y: 0, z: -90, face: 2},
{x: 0, y: 0, z: 90, face: 3},
{x: 90, y: 90, z: 0, face: 3},
{x: -90, y: -90, z: 0, face: 3},
{x: -90, y: 90, z: 180, face: 3},
{x: 180, y: 0, z: -90, face: 3},
{x: 0, y: 0, z: -90, face: 4},
{x: 90, y: -90, z: 0, face: 4},
{x: -90, y: 90, z: 0, face: 4},
{x: 180, y: 0, z: 90, face: 4},
{x: 90, y: 0, z: 0, face: 5},
{x: 90, y: 0, z: -90, face: 5},
{x: 90, y: 0, z: 180, face: 5},
{x: 90, y: 0, z: 90, face: 5},
{x: 90, y: 90, z: 90, face: 5},
{x: 0, y: 0, z: 180, face: 6},
{x: 180, y: -90, z: 0, face: 6},
{x: 180, y: 90, z: 0, face: 6},
{x: 180, y: 0, z: 0, face: 6}
];
function checkRotationsToGetUpperFace(cube) {
// create object with "normalized" (brought to quarter-circle-degree-values) degrees
var normalizedRotation = {
x: 0,
y: 0,
z: 0
};
normalizedRotation.x = getNormalizedDegree(cube.rotation._x);
normalizedRotation.y = getNormalizedDegree(cube.rotation._y);
normalizedRotation.z = getNormalizedDegree(cube.rotation._z);
// go through the library that has all the degrees with the corresponding upper face
for (var i = 0; i < rotationLibrary.length; i++) {
// check if the objects match, then get the face
if (rotationLibrary[i].x == normalizedRotation.x &&
rotationLibrary[i].y == normalizedRotation.y &&
rotationLibrary[i].z == normalizedRotation.z) {
cube.face = rotationLibrary[i].face;
}
}
// reattach cube for correct movement later
THREE.SceneUtils.attach(cube, scene, pivot);
}
function getNormalizedDegree(rotationValue) {
// transform rotation value into degree value
var rotValue = rotationValue / (Math.PI / 180);
// default value is 0, so only check if it should be 90°, 180° or -90°
var normalizedDegree = 0;
// x between 45° and 135° ( ~ 90)
if (rotValue > 45 && rotValue < 135) {
normalizedDegree = 90;
}
// x between -45° and -135° ( ~ -90)
else if (rotValue < -45 && rotValue > -135) {
normalizedDegree = -90;
}
// x between 135° and 215° or x between -135° and -215° ( ~ 180)
else if ((rotValue > 135 && rotValue < 215) || (rotValue < -135 && rotValue > -215)) {
normalizedDegree = 180;
}
return normalizedDegree;
}
(http://jsfiddle.net/b2an3pq7/3/)
Might not be the solution to your raycast problem but another approach: Why don't you simply detect the upper face by comparing the rotation euler angles? E.g. (pseudo code):
if(cube.rotation.x % 360 == 0)
{
// upper face upwards
}
else if(cube.rotation.x % 360 == 90)
{
// left face upwards
}
You would have to deal with value tolerance (85° - 95°) negative rotation values and values out of the range of PI*2 but other than that, isn't that much easier?

Resources