How to create a line between two points in 3d space with RealityKit?
There are examples of creating lines between two points in Scenekit, however, there are basically none using RealityKit.
To create the line, I've created a rectangle model entity and placed it between my first touched point and the current touched point. From here, all I would need to do is rotate the rectangle to face the current touched point. However, using the simd_quatf(from: to:) doesn't work as intended.
rectangleModelEntity.transform.rotation = simd_quatf(from: firstTouchedPoint,
to: currTouchedPoint)
If I were to touch a point and then drag directly downwards, the rectangle model should be to be a straight line between first touched and current touched point, but it stays horizontal with a slight tilt.
To solve this, I tried getting the angle between my initially horzontal line as a vector and a vector from the first touched to current touched point
let startVec = currTouchedPoint - firstTouchedPoint
let endVec = endOfModelEntityPoint - modelEntityCenterPoint
let lengthVec = simd_length(cross(startVec, endVec))
let theta = atan2(lengthVec, dot(startVec, endVec))
This gives me the angle between two vectors in 3d space, which seems correct, when I checked it gave me 90 degrees when touching and dragging directly between it.
The problem is I don't know what the axis to rotate it on should be. Since this is 3d space, the line doesn't need to be on a 2d plane, the current touched position can be downwards and in front of the starting touch position.
rectangleModelEntity = simd_quatf(angle: theta, axis: ???)
Personally, I'm not even too sure if the above is the correct solution to creating a line between two points. In theory it's rather basic, create a rectangle with low height/depth to mimic a line, position it in the center of the starting and current touch point then rotate it so it's oriented correctly.
What should be the axis for the above degrees between two vectors?
Is there a better method of creating two lines between points in 3d space with RealityKit/ARKit?
I have implemented using a box. Let me know if you have a better way.
let midPosition = SIMD3(x:(position1.x + position2.x) / 2,
y:(position1.y + position2.y) / 2,
z:(position1.z + position2.z) / 2)
let anchor = AnchorEntity()
anchor.position = midPosition
anchor.look(at: position1, from: midPosition, relativeTo: nil)
let meters = simd_distance(position1, position2)
let lineMaterial = SimpleMaterial.init(color: .red,
roughness: 1,
isMetallic: false)
let bottomLineMesh = MeshResource.generateBox(width:0.025,
height: 0.025/2.5,
depth: meters)
let bottomLineEntity = ModelEntity(mesh: bottomLineMesh,
materials: [lineMaterial])
bottomLineEntity.position = .init(0, 0.025, 0)
anchor.addChild(bottomLineEntity)
The axis is the cross product of the direction your object is facing at the beginning and the direction it should be facing now.
Like if it's at position p1=[x1,y1,z1], initially facing d1=[0, 0, -1], and you want it to face a point p2=[x, y, z] the axis would be the cross product: |d1|✕|p2 - p1|.
May have to swap the two around, or just negate the angle though, depending on how it works out.
Related
So,
I have the exact position I want to place the node at. If I test things with a sphere geometry I can place spheres in the world by telling the node:
node.simdPosition = position
(I provide the "position" as an input to the function).
That successfully places the object in the world exactly where I want it to go.
What I really want to do is placing a plane:
let plane = SCNPlane(width: 0.2, height: 0.3)
plane.cornerRadius = plane.width / 10
plane.firstMaterial?.diffuse.contents = UIColor.red
plane.firstMaterial?.specular.contents = UIColor.white
let node = SCNNode(geometry: plane)
Then telling it to be placed at the "position":
node.simdPosition = position
All this works with the plane as well. What I have problems with is the angle:
I want to tell the plane's node to be placed with a given "angle" (around Y) offset to the camera. I tried this but it's not working:
node.rotation = SCNVector4Make(0, 1, 0, currentFrame.camera.eulerAngles.z - angle)
So then, the question is, how can a node be placed at a certain position and at the moment it gets placed in the world, also have a certain Y angle offset from the perpendicular to the camera?
I was using the wrong Euler angle... (z)
This made it work:
node.eulerAngles = SCNVector3Make(0, cameraEulerAngles.y - Float(0.7), 0)
I've managed to get a cube (SCNNode) placed on a surface where the camera is pointed, however I am finding it very difficult to do the simple (?) task of also placing text in the same position.
I've created the SCNText and subsequent SCNNode, however when I add it to the rootNode the text always seems to be added above my head and off the camera to the right (which tells me thats the global origin point).
Even when I use the exact same values of position I used for the the cube, the SCNText node still gets placed above my head in the same spot.
Apologies if this is a basic question, I've never worked in SceneKit before.
The coordinate center for an SCNGeometry is its center point. But when you are creating a SCNText the center point is somewhere in the bottom left corner:
You need to center the text first. This can be done by checking the bounding box of the node containing your text and setting a pivot transform to change the texts center to its actual center:
func center(node: SCNNode) {
let (min, max) = node.boundingBox
let dx = min.x + 0.5 * (max.x - min.x)
let dy = min.y + 0.5 * (max.y - min.y)
let dz = min.z + 0.5 * (max.z - min.z)
node.pivot = SCNMatrix4MakeTranslation(dx, dy, dz)
}
Edit:
Also note this answer that explains some additional pitfalls:
A text with 16 pts font size is 16 SceneKit units tall. But in ARKit 1 SceneKit units = 1 meter!
The image below shows a rotated box that should be moved horizontally on the X and Z axes. Y should stay unaffected to simplify the scenario. The box could also be the SCNNode of the camera, so I guess a projection does not make sense at this point.
So lets say we want to move the box in the direction of the red arrow. How to achieve this using SceneKit?
The red arrow indicates -Z direction of the box. It also shows us it is not parallel to the camera's projection or to the global axes that are shown as dark grey lines of the grid.
My last approach is the product of a translation matrix and a rotation matrix that results in a new transformation matrix. Do I have to add the current transform to the new transform then?
If yes, where is the SceneKit function for the addition of matrices like SCNMatrix4Mult for multiplication or do I have to write it myself using Metal?
If no, what I'm missing out with the matrix calculations?
I don't want to make use of GLKit.
So my understanding is that you want to move the Box Node along its own X axis (not it's parents X axis). And because the Box Node is rotated, its X axis is not aligned with its parent's one, so you have the problem to convert the translation between the two coordinate systems.
The node hierarchy is
parentNode
|
|----boxNode // rotated around Y (vertical) axis
Using Transformation Matrices
To move boxNode along its own X axis
// First let's get the current boxNode transformation matrix
SCNMatrix4 boxTransform = boxNode.transform;
// Let's make a new matrix for translation +2 along X axis
SCNMatrix4 xTranslation = SCNMatrix4MakeTranslation(2, 0, 0);
// Combine the two matrices, THE ORDER MATTERS !
// if you swap the parameters you will move it in parent's coord system
SCNMatrix4 newTransform = SCNMatrix4Mult(xTranslation, boxTransform);
// Allply the newly generated transform
boxNode.transform = newTransform;
Please Note: The order matters when multiplying matrices
Another option:
Using SCNNode coordinate conversion functions, looks more straight forward to me
// Get the boxNode current position in parent's coord system
SCNVector3 positionInParent = boxNode.position;
// Convert that coordinate to boxNode's own coord system
SCNVector3 positionInSelf = [boxNode convertPosition:positionInParent fromNode:parentNode];
// Translate along own X axis by +2 points
positionInSelf = SCNVector3Make(positionInSelf.x + 2,
positionInSelf.y,
positionInSelf.z);
// Convert that back to parent's coord system
positionInParent = [parentNode convertPosition: positionInSelf fromNode:boxNode];
// Apply the new position
boxNode.position = positionInParent;
Building on #Sulevus's correct answer, here's an extension to SCNNode that simplifies things by using the convertVector rather than the convertPosition transformation, in Swift.
I've done it as a var returning a unit vector, and supplied an SCNVector3 overload of multiply so you can say things like
let action = SCNAction.move(by: 2 * cameraNode.leftUnitVectorInParent, duration: 1)
public extension SCNNode {
var leftUnitVectorInParent: SCNVector3 {
let vectorInSelf = SCNVector3(x: 1, y: 0, z: 0)
guard let parent = self.parent else { return vectorInSelf }
// Convert to parent's coord system
return parent.convertVector(vectorInSelf, from: self)
}
var forwardUnitVectorInParent: SCNVector3 {
let vectorInSelf = SCNVector3(x: 0, y: 0, z: 1)
guard let parent = self.parent else { return vectorInSelf }
// Convert to parent's coord system
return parent.convertVector(vectorInSelf, from: self)
}
func *(lhs: SCNVector3, rhs: CGFloat) -> SCNVector3 {
return SCNVector3(x: lhs.x * rhs, y: lhs.y * rhs, z: lhs.z * rhs)
}
func *(lhs: CGFloat, rhs: SCNVector3) -> SCNVector3 {
return SCNVector3(x: lhs * rhs.x, y: lhs * rhs.y, z: lhs * rhs.z)
}
}
The far easier way this is usually done:
The usual, normal, and extremely easy way to do this in any game engine or 3D engine is:
You simply have a wrapper node, which, holds the node in question.
This is indeed the entire point of transforms, they enable you to abstract out a certain motion.
That's the whole point of 3D engines - the GPU just multiplies out all the quaternions on the way down to the object; it's wholly pointless to (A) figure out in your head the math and (B) do it manually (indeed in the CPU).
In Unity it's "game objects", in scene kit it's "nodes" and so on.
In all 3D engines, including scene kit, almost everything has one or more "holders" around it.
To repeat, the reasons for this are (A) it's the entire raison d'etre of a game engine, to achieve performance in multiplying out the quaternions of every vertex and (B) sheer convenience and code solidity.
One of a million examples ...
Of course you can trivially do it in code,
cameraHolder.addChildNode(camera)
In the OP's example. It looks like you would use cameraHolder only to rotate the camera. And then for the motion the OP is asking about, simply move camera.
It's perfectly normal to have a chain of a number of nodes to get to an object.
This is often used for "effects". Say you have an object, which sometimes has to "vibrate up and down". You can have one node which only does that movement. Note that then, all the animations etc for that movement only have to be on that node. And critically, they can run independently of any other animations or movements. (And indeed you can just use the node elsewhere to jiggle something else.)
Generally what I'm trying to achieve: we have map data that historically was all 2D, and the coordinate system we use is the origin point (0,0) at the top left, positive x goes right, positive y goes down. We have now added 3D data by adding a z axis, positive z coming out of the screen towards you (think top-down map view). This is a left handed coordinate system, but SceneKit is a right handed coordinate system. I would like to apply some transform at the top level of my SceneKit Scene that will convert the Scene into a left handed coordinate system such that I can modify/position/add nodes to the scene in terms of our custom mapping coordinate system and things will just work.
So far I have this:
let scene = SCNScene()
let cameraNode = SCNNode()
cameraNode.camera = SCNCamera()
cameraNode.scale = SCNVector3(1,-1,1)
scene.rootNode.addChildNode(cameraNode)
This achieves exactly what I want, but has one big problem. It inverts all of the geometry faces, so my geometry's disappear unless I change their material's cullMode:
let mapLength = 1000 //max X axis
let mapWidth = 800 //max Y axis
let mapHeight = 100 //max Z axis
cameraNode.position = SCNVector3(mapLength / 2, mapWidth / 2, 2000)
let mapPlane = SCNNode()
mapPlane.position = SCNVector3(mapLength / 2, mapWidth / 2, 0)
mapPlane.geometry = SCNPlane(width: mapLength, height: mapWidth)
mapPlane.geometry?.firstMaterial?.diffuse.contents = UIColor.blackColor()
scene.rootNode.addChildNode(mapPlane)
mapPlane doesn't show at all! You have to rotate the camera to the underside of mapPlane in order to see it. You can easily fix this by adding a single line:
mapPlane.geometry?.firstMaterial?.cullMode = .Front
But I don't want to have to change the cullMode for every geometry/material. Is there a way to achieve this without requiring extra code at each geometry/material? Some transform that would invert the geometry face normals for all child nodes of rootNode? Ideally this would be achieved entirely by settings on the actual Scene, or by transforms on rootNode or the camera.
Suppose you have two points in 3-D space. Call the first o for origin and the other t for target. The rotation axes of each are alligned with the world/parent coordinate system (and each other). Place a third point r coincident with the origin, same position and rotation.
How, in Swift, can you rotate r such that its y-axis points at t? If pointing the z-axis is easier, I'll take that instead. The resulting orientation of the other two axes is immaterial for my needs.
I've been through many discussions related to this but none satisfy. I have learned, from reading and experience, that Euler angles is probably not the way to go. We didn't cover this in calculus and that was 50 years ago anyway.
Got it! Incredibly simple when you add a container node. The following seems to work for any positions in any quadrants.
// pointAt_c is a container node located at, and child of, the originNode
// pointAtNode is its child, position coincident with pointAt_c (and originNode)
// get deltas (positions of target relative to origin)
let dx = targetNode.position.x - originNode.position.x
let dy = targetNode.position.y - originNode.position.y
let dz = targetNode.position.z - originNode.position.z
// rotate container node about y-axis (pointAtNode rotated with it)
let y_angle = atan2(dx, dz)
pointAt_c.rotation = SCNVector4(0.0, 1.0, 0.0, y_angle)
// now rotate the pointAtNode about its z-axis
let dz_dx = sqrt((dz * dz) + (dx * dx))
// (due to rotation the adjacent side of this angle is now a hypotenuse)
let x_angle = atan2(dz_dx, dy)
pointAtNode.rotation = SCNVector4(1.0, 0.0, 0.0, x_angle)
I needed this to replace lookAt constraints which cannot, easily anyway, be archived with a node tree. I'm pointing the y-axis because that's how SCN cylinders and capsules are directed.
If anyone knows how to obviate the container node please do tell. Everytime I try to apply sequential rotations to a single node, the last overwrites the previous one. I haven't the knowledge to formulate a rotation expression to do it in one shot.