As a continuation to: Point one object to face a distant object
I'm trying to get a mesh's surface to point to another object. I want a far away object to point to a close object, whenever the far object moves.
Here is a sample of what is happening: http://jsfiddle.net/UsVUv/
1) I create 2 mesh objects.
2) Since the far mesh is facing up, I rotate it so that it can be seen from the camera.
-> farMesh.rotation = new THREE.Vector3(Math.PI/2, 0, 0);
3) When the far mesh is moved, I call the following to get it to keep looking at the near mesh. The face of the far mesh does not look at the near mesh, but the edge of its plane does.
-> farMesh.lookAt(nearMesh.position);
4) I then try to rotate the mesh back so it faces the camera again, but that removes the rotation that the lookAt did.
-> farMesh.rotation = new THREE.Vector3(Math.PI/2, 0, 0);
You can comment out the lines 1), 2), and 3) in the sample to see what is happening.
Yes, the farMesh.lookAt() and farMesh.rotation are competing with each other.
What you need to do, instead, is rotate the farMesh plane geometry as soon as it is created with applyMatrix() like so:
farMesh.geometry.applyMatrix( new THREE.Matrix4().makeRotationX( Math.PI / 2 ) );
So to summarize, applyMatrix() transforms the object's geometry, and lookAt() sets the object's rotation vector.
Here is an updated Fiddle: http://jsfiddle.net/UsVUv/1/
Related
In three.js, when we define the camera we use something like camera = new THREE.PerspectiveCamera(fov, window.innerWidth / window.innerHeight, near, far); If the object is moved outside the bounds set by the two planes, far and near then it is clipped.
Suppose now I rotate the object and I zoom in/out.
How can I find the current plane my camera is on please? I am learning three.js, hence I dont know If I am explaining this clear enough.
I thought that it was camera.position.z that would give me this info. In fact, I think it gives the correct value when my camera looks down the z-axis. However when I rotate by 90 degrees, (effectively moving my camera on the x-axis) the value of camera.position.z changes by a lot.
I have added a graph in case it helps. The plane defined by the blue outline cuts through the data and is parallel to the far and near planes. As I zoom in and out, I am moving the plane forward and backward, right? Do I understand this correct or I have totally missed it? I would like to know the value (a float between far and near) indicating how far that blue plane is from the camera. If you rotate that distance shouldn't change
My end goal is to be able to find-out how close the scene is to the viewer. If it gets too close then I will be adding some more elements to the scene. If it is too far these elements will be removed.
I need to get the camera up direction and i've tried many ways with no luck, i'm not an expert of quaternions so i'm doubting i did it right.
I've tried:
camera.up
camera.up.applyMatrix4(camera.matrixWorld);
new THREE.Vertex3(0,1,0).applyMatrix4(camera.matrixWorld);
camera.up.normalize().applyMatrix4(camera.matrixWorld);
after this i create two planes passing by two points of my interest, and add the plane helper to the scene and i can see they are very far from where i was expecting them. (i'm expecting two planes that looks like the top and bottom of the camera frustum).
P.s. the camera is a shadow camera of a directional light so an orthographic camera, and i manipulate the directional light position and target before doing this operation, but i've called updateMatrixWorld on the light, on it's target and the camera, on the camera i've called also updateProjectionMatrix... still no results
I've made a sandbox to see what i've tried till now, and better visualize what i want to achieve:
https://codesandbox.io/embed/throbbing-cache-j5yse
once i manage to get the green arrow to point to the top of the blue triangle of the camera helper i'm good to go
In the normal render flow, shadow camera matrices are updated as part of rendering the shadow map (WebGLShadowMap.render).
However, if you want the updated matrix values before the render, then you'll need to update them manually (you already understand this part).
The shadow camera is a property of (not a child of) the DirectionalLight. As such, it doesn't follow the same rules as other scene objects when it comes to updating its matrices (because it's not really a child of the scene). Instead, you need to call the shadow property's updateMatrices method (inherited from LightShadow.updateMatrices).
const dl = new THREE.DirectionalLight(0xffffff, 1)
dl.shadow.updateMatrices(dl) // <<------------------------ Updates the shadow camera
This updates the shadow camera with information from the DirectionalLight's own matrix, and its target's matrix, to properly orient the shadow camera.
Finally, it looks like you're trying to get the "world up" of the camera. Personally, I'd use the convenience function localToWorld:
let up = new THREE.Vector3(0, 1, 0)
dl.shadow.camera.localToWorld(up) // destructively converts "up" from local-to-camera into world coordinates
via trial and errors i've figured out that what gave me the correct result was:
calling
directionalLight.shadow.updateMatrices(...)
and then
new THREE.Vector3(0,1,0).applyQuaternion(directionalLight.shadow.camera.quaternion)
Let's say I have a vertical list of meshes created from PlaneBufferGeometry with ShaderMaterial. The meshes are distributed vertically and evenly spaced.
The list will have two states:
Displaying the meshes as they are
Displaying meshes with each object's vertices transformed by the vertex shader to the same arbitrary value, let's say z = -50. This gives a zoomed out effect and the user can scroll through this list (in the code we do this by moving the camera y position)
In my app I'm trying to make my mouseover events work for the second state but it's tricky since the GPU transforms the vertices so the updated vertices are not reflected in the attributes on the JS side.
*Note I've looked into GPU picking and do not want to use it because I believe there should be a simpler way to do this without render targets
Attempted Solution
My current approach is to manually change the boundingBox of each plane when we are in the second state like so:
var box = new THREE.Box3().setFromObject(plane);
box.min.z = -50;
box.max.z = -50;
plane.geometry.boundingBox = box;
And then to change the boundingSphere's center to have the same z position of -50 after computing it.
I did this approach because I looked into the Raycaster and Mesh code for THREE.js and it seems like they check both boundingSphere and boundingBox for object intersections. So I thought if I modified both of them to reflect the transforms done by the GPU, the raycaster would work fine but it doesn't seem to be working for me.
The relevant raycaster code is here:
// mouse being vec2 of normalized coordinates and camera being a perspective camera
raycaster.setFromCamera( mouse, camera );
const intersects = raycaster.intersectObjects( planes );
Possible Theories
The only thing I can think of that's wrong about this approach is maybe I'm not projecting the mouse coords right? Since all the objects now lie on the plane z = -50 would I need to project those mouse coordinates to that plane?
Inspired by the link posted by #prisoner849 I found a working solution to just create additional transparent planes equal to the number of planes in the scene. In these planes, I set the z position to -50 and just intersect with these when in state #2.
A bit hacky, but works for now.
I have initialised a perspective camera at a position looking at the origin (0,0,0). Reading around the most common solution to this I've found is the one described here https://stackoverflow.com/a/27412386/1330719.
From my understanding of the project method is, I should get a vector where the x,y coordinates are between -1 and 1. This doesn't seem to be the case at all and I end up getting coordinates that are completely out of bounds.
Furthermore, if the original vector point is at (0,0,0) I seem to get (NaN, NaN) back. If my camera is looking at position (0,0,0) I expect the Vector3 (0,0,0) to return (width/2, height/2).
In case it is needed, this is how I'm initialising my camera:
this.camera = new THREE.PerspectiveCamera(90, window.innerWidth / window.innerHeight, 1, 10E5);
this.camera.position.set(0, 500, -500);
this.camera.lookAt(new THREE.Vector3(0,0,0));
this.camera.updateProjectionMatrix();
Does anyone have a reason why this might not be working? Or alternatively a recommended way of mapping a Vector3 point to the screen space given a camera?
A jsfiddle of what I mean:
https://jsfiddle.net/m78wjLyc/
You should also use camera.updateMatrixWorld(true) before projecting.
Usually this is done automatically by renderer, but you don't use any, so the camera.matrixWorld stays untouched after you change the position, and makes the camera project things as if it was at the world origin.
I have a geometry object, and I'm trying to add a Torus mesh that goes around that geometry. What I'm trying to do is have the original geometry, and then when the geometry is clicked, it adds a Torus shape on the line around the location that was clicked. However, I'm having trouble getting it to rotate correctly.
I get the torus to show up at the correct place, but I can't orient it around the line. I'm using a raycaster to get the point clicked, so I have the face and the faceindex of the point clicked. On every implementation I try using rotation (using setEulerFromRotationMatrix), it simply moves the location of the torus mesh, not actually rotate it to allow the line to go through the torus.
This seems like it would be trivial, but it's giving me a lot of trouble. What am I doing wrong? Two methods I tried, both unsucessful and exhibiting the behavior above:
var rotationMatrix = new THREE.Matrix4();
rotationMatrix.makeRotationAxis(geometry.faces[fIndex].centroid.normalize(), Math.PI/2);
torusLoop.matrix.multiply(rotationMatrix);
torusLoop.rotation.setEulerFromRotationMatrix(torusLoop.matrix);
//attempt two, similar results to above attempt
tangent = geometry.tangents[segments/radiusSegments].normalize();
axis.crossVectors( up, tangent ).normalize();
var radians = Math.acos( up.dot( tangent ) );
matrix.makeRotationAxis( axis, radians );
torusLoop.rotation.setEulerFromRotationMatrix( matrix );
I need the torus knot to follow the curve of the spline, but it will only stay flat, and rotations simply cause it to move around, not change angles.
Never mind, I figured it out. For those wondering, I translated before rotating, which caused my figure to be rotating around a different axis. My solution was to rotate first, and then translate, and then after creating the mesh, moving that to the position I needed it to be.