Aframe/threejs: calculate a point "behind" another element - vector

Lets say we have two elements, A and B, both with known positions; I need get the position of a point that is facing the direction between A and B, and X meters distant from B.
I guess I need to create a "line" from A to B, extend it of another X meters, and get the point at the end of the line.
How to translate this in vector?

Maybe like so:
const x = 2;
const a = new THREE.Vector3();
const b = new THREE.Vector3( 0, 0, 1 );
const v = new THREE.Vector3(); // the result vector
v.subVectors( a, b );
v.normalize();
v.multiplyScalar( x );
v.add( b ); // v is now (0, 0, - 1)
Full working example: https://jsfiddle.net/f2Lommf5/15820/

Related

Finding a specific point on an OBB (Oriented Bounding Box)

Hello, I've got a question which I cannot solve so I need a bit help.
In the picture above you can see an Oriented Bounding Box specified by 4 points (A, B, C, D). There is also a point in space called P. If I cast a ray from P against the OBB the ray is going to intersect the OBB at some point. This point of intersection is called Q in the picture. By the way the ray is always going to be x-axis aligned which means its directional vector is either (1, 0) or (-1,0) if normalized. My goal is to find the point of intersection - Q. Is there a way (if possible computationaly inexpensive) to do so?
Thanks in advance.
One way to do this is to consider each side of the bounding box to be a linear equation of the form y = ax + b, where a is the slope and b is the y-intercept. Then consider the ray from P to be an equation of the form y = c, where c is a constant. Then compare this equation to each of the four other equations to see where it intersects each one. One of these intersections will be our Q, if a Q exists; it's possible that the ray will miss the bounding box entirely. We will need to do a few checks:
Firstly, eliminate all potential Q's that are on the wrong side of P.
Secondly, check each of the four intersections to make sure they are within the bounds of the lines that they represent, and eliminate the ones that are not.
Finally, if any potential Q's remain, the one closest to P will be our Q. If no potential Q's remain, this means that the ray from P misses the bounding box entirely.
For example...
The line drawn from D to B would have a slope equal to (B.y - D.y) / (B.x - D.x) and a y-intercept equal to B.y - B.x * slope. Then the entire equation is y = (B.y - D.y) / (B.x - D.x) * x + B.y - B.x * (B.y - D.y) / (B.x - D.x). Set this equation equal to y = P.y and solve for x:
x = (P.y - B.y + B.x*(B.y - D.y)/(B.x - D.x))*(B.x - D.x)/(B.y - D.y)
The result of this equation will give you the x-value of the intersection. The y-value is P.y. Do this for each of the other 3 lines as well: A-B, A-C, C-D. I will refer to these intersections as Q(D-B), Q(A-B), Q(A-C), and Q(C-D) respectively.
Next, eliminate candidate-Q's that are on the wrong side of P. In our example, this eliminates Q(A-B), since it is way off to the right side of the screen. Mathematically, Q(A-B).x > P.x.
Then eliminate all candidate-Q's that are not on the line they represent. We can do this check by finding the lowest and highest x-values and y-values given by the two points that represent the line. For example, to check that Q(A-C) is on the line A-C, check that C.x <= Q(A-C).x <= A.x and C.y <= Q(A-C).y <= A.y. Q(A-C) passes the test, as well as Q(D-B). However, Q(C-D) does not pass, as it is way off to the left side of the screen, far from the box. Therefore, Q(C-D) is eliminated from candidacy.
Finally, of the two points that remain, Q(A-C) and Q(D-B), we choose Q(D-B) to be our winner, because it is closest to P.
We now can say that the ray from P hits the bounding box at Q(D-B).
Of course, when you implement this in code, you will need to account for divisions by zero. If a line is perfectly vertical, there does not exist a point-slope equation of the line, so you will need to create a separate formula for this case. If a line is perfectly horizontal, it's respective candidate-Q should be automatically eliminated from candidacy, as the ray from P will never touch it.
Edit:
It would be more efficient to only do this process with lines whose two points are on vertically opposite sides of point P. If both points of a line are above P, or they are both below P, they would be eliminated from candidacy from the beginning.
Find the two sides that straddle p on Y. (Test of the form (Ya < Yp) != (Yb < Yp)).
Then compute the intersection points of the horizontal by p with these two sides, and keep the first to the left of p.
If the ray points to the left(right) then it must intersect an edge that connects to the point in the OOB with max(min) x-value. We can determine which edge by simply comparing the y-value of the ray with the y value of the max(min) point and its neighbors. We also need to consider OBBs that are actually axis-aligned, and thus have two points with equal max(min) x-value. Once we have the edge it's simple to confirm that the ray does in fact intersect the OBB and calculate its x-value.
Here's some Java code to illustrate (ideone):
static double nearestX(Point[] obb, int y, int dir)
{
// Find min(max) point
int n = 0;
for(int i=1; i<4; i++)
if((obb[n].x < obb[i].x) == (dir == -1)) n = i;
// Determine next or prev edge
int next = (n+1) % 4;
int prev = (n+3) % 4;
int nn;
if((obb[n].x == obb[next].x) || (obb[n].y < y) == (obb[n].y < obb[next].y))
nn = next;
else
nn = prev;
// Check that the ray intersects the OBB
if(Math.abs(y) > Math.abs(obb[nn].y)) return Double.NaN;
// Standard calculation of x from y for line segment
return obb[n].x + (y-obb[n].y)*(obb[nn].x-obb[n].x)/(obb[nn].y-obb[n].y);
}
Test:
public static void main(String[] args)
{
test("Diamond", new Point[]{p(0, -2), p(2, 0), p(0, 2), p(-2,0)});
test("Square", new Point[]{p(-2, -2), p(2, -2), p(2, 2), p(-2,2)});
}
static void test(String label, Point[] obb)
{
System.out.println(label + ": " + Arrays.toString(obb));
for(int dir : new int[] {-1, 1})
{
for(int y : new int[] {-3, -2, -1, 0, 1, 2, 3})
System.out.printf("(% d, % d) = %.0f\n", y , dir, nearestX(obb, y, dir));
System.out.println();
}
}
Output:
Diamond: [(0,-2), (2,0), (0,2), (-2,0)]
(-3, -1) = NaN
(-2, -1) = 0
(-1, -1) = 1
( 0, -1) = 2
( 1, -1) = 1
( 2, -1) = 0
( 3, -1) = NaN
(-3, 1) = NaN
(-2, 1) = 0
(-1, 1) = -1
( 0, 1) = -2
( 1, 1) = -1
( 2, 1) = 0
( 3, 1) = NaN
Square: [(-2,-2), (2,-2), (2,2), (-2,2)]
(-3, -1) = NaN
(-2, -1) = 2
(-1, -1) = 2
( 0, -1) = 2
( 1, -1) = 2
( 2, -1) = 2
( 3, -1) = NaN
(-3, 1) = NaN
(-2, 1) = -2
(-1, 1) = -2
( 0, 1) = -2
( 1, 1) = -2
( 2, 1) = -2
( 3, 1) = NaN

Linear Interpolation of 2 Vector3 by unit of distance rather than factor in THREE.js

Apologies for any wording and incorrect references, 3 dimensional maths is new to me!
Problem:
I have two points in 3D space A: (1,2,3) and B: (6,5,4)
I have a distance L: 10
What is the 3D point C as defined as A TOWARDS B at a UNIT of L?
Notes:
Distance of L is not a factor of the distance between A or B, there .lerp() is not the right tool here
L can be both greater and smaller than the distance between A and B.
I'm using THREE.js, but any language examples are welcome!
Workings so far:
Calculate distance between A & B -> D
Get factor of D in relation to our target value L -> R
Apply .lerp() with A towards B with factor R
function lerpUnit (x1,y1,z1,x2,y2,z3,distance) {
... ???
return {x3,y3,z3}
}
You could achieve this with just the methods provided by THREE.Vector3()
First you could subtract vecB - vecA to get a parallel vector that originates at (0, 0, 0).
Second, use normalize() to turn this vector into a "unit vector", which means no matter which way it's pointing, its length will always be 1.
Third, multiply the vector by L, so length = 1 * L
Lastly, move this new parallel vector back to its initial position with + vecA to get the final position. It will land somewhere inside or beyond the line segment created by A and B
var vecA = new THREE.Vector3(1, 2, 3);
var vecB = new THREE.Vector3(6, 5, 4);
var L = 10;
// Create new vector
var direction = new THREE.Vector3();
// Assign positionB
direction.copy(vecB);
// Subtract positionA to get a vector starting at (0, 0, 0)
direction.sub(vecA);
// Normalize to get a vector of lengh = 1
direction.normalize();
// Multiply by new length
direction.multiplyScalar(L);
// Add positionA to return vector to initial start point
direction.add(vecA);
// Output to log to see result
console.log(direction.toArray());
<script src="https://cdn.jsdelivr.net/npm/three#0.117.1/build/three.min.js"></script>
Thanks, I just expanded on by workings far by using .lerp() but passing in a ratio's value, eg:
function lerpUnit (vectorA, vectorB, intendedDistance ) {
const aToBDistance = vectorA.distanceTo(vectorB)
const adjustedFactor = intendedDistance / aToBDistance
return new THREE.Vector3().lerpVectors(vectorA, vectorB, adjustedFactor)
}
or 1 liner: const vectorC = new THREE.Vector3().lerpVectors(vectorA, vectorB, intendedDistance / vectorA.distanceTo(vectorB))

2D Game, calculate the value for transform.Rotate out of 2 vectors

first I wanna say I am really new to unity3d. I have done some tutorials and now I playing around a bit. So here is my problem. I have a 2D scene with a gravity source in the middle (lets say its a planet). I have a spaceship in his orbit. The gravity is simulated with:
var myVector = GameObject.Find("middle").transform.position - transform.position;
rigidbody2D.velocity += 0.2 * Time.deltaTime * myVector;
I can rotate the spaceship with:
if(Input.GetKey(moveLeft)) {
transform.Rotate(Vector3.back * -turnSpeed * Time.deltaTime);
//this value is something like (0, 0, -8)
}
if(Input.GetKey(moveRight)) {
transform.Rotate(Vector3.back * turnSpeed * Time.deltaTime);
//this value is something like (0, 0, 8)
}
What I want is that the spaceship automaticly rotates when it changes the angle to the planet. So that if it is in his orbit the same side allways looks forward. I have done a small sketch for that:
http://snag.gy/AGJMR.jpg
(The Arrow is the spaceship with his direction, it should rotate while the angle between the spaceship and the planet changes)
Basicly: when the spaceship flies around the planet 1 time it also makes a 360° rotation.
I have the old Vector (from spaceship to planet) saved and also have the actual vector:
var myVector = GameObject.Find("middle").transform.position - transform.position;
lastVector = myVector;
//for example:
myVector is (-1, 1, 0)
lastVector is (-1, -1, 0)
Out of this 2 value I should be able to get the value for transform.Rotate (something like (0, 0, 2). But I have no idea how I get there.
The solution is pretty easy:
var angle = Vector3.Angle(lastVector, myVector);
var cross = Vector3.Cross(lastVector, myVector);
if(cross.z >0) {
angle = -angle;
}
transform.Rotate(Vector3.back * angle);

How to compute the cross-product?

I have the following piece of pseudo-C/Java/C# code:
int a[]= { 30, 20 };
int b[] = { 40, 50 };
int c[] = {12, 12};
How do I compute the cross-product ABxAC?
The solution that was given to you in your last question basically adds a Z=0 for all your points. Over the so extended vectors you calculate your cross product. Geometrically the cross product produces a vector that is orthogonal to the two vectors used for the calculation, as both of your vectors lie in the XY plane the result will only have a Z component. The Sign of that z component denotes wether that vector is looking up or down on the XY plane. That sign is dependend on AB being in clockwise or counter clockwise order from each other. That in turn means that the sign of z component shows you if the point you are looking at lies to the left or the right of the line that is on AB.
So with the crossproduct of two vectors A and B being the vector
AxB = (AyBz − AzBy, AzBx − AxBz, AxBy − AyBx)
with Az and Bz being zero you are left with the third component of that vector
AxBy - AyBx
With A being the vector from point a to b, and B being the vector from point a to c means
Ax = (b[x]-a[x])
Ay = (b[y]-a[y])
Bx = (c[x]-a[x])
By = (c[y]-a[y])
giving
AxBy - AyBx = (b[x]-a[x])*(c[y]-a[y])-(b[y]-a[y])*(c[x]-a[x])
which is a scalar, the sign of that scalar will tell you wether point c lies to the left or right of vector ab
Aternatively you can look at stack overflow or gamedev
Assuming whether you're asking whether the angle between AB and AC is acute or obtuse, you want this:
int a[]= { 30, 20 };
int b[] = { 40, 50 };
int c[] = {12, 12};
int ab_x = b[0] - a[0];
int ab_y = b[1] - a[1];
int ac_x = c[0] - a[0];
int ac_x = c[1] - a[1];
int dot = ab_x*ac_x + ab_y*ac_y;
boolean signABxAC = dot > 0; // pick your preferred comparison here
The cross product is a vector, it doesn't have "a sign".
Do you mean the scalar (dot) product? If you do, then that is computed as for a pair of vectors [a,b,c]•[d,e,f] as ad + be + cf, so the sign of that expression is the sign of the dot product.
Figuring out the sign without doing the multiplications and adds is probably not faster than just doing them.
Since all three points have just two components, I'll assume that the z-component for all three is zero. That means that the vectors AB and BC are in the xy-plane, so the cross-product is a vector that points in the z-direction, with its x and y components equal to zero.
If by "sign" you mean whether it points in the positive or negative z-direction, the computation will tell you that.
In your case, the two vectors are AB = (10, 30, 0) and AC = (-18, -8, 0). If I take the cross-product of those two, I get vector AB X AC = (0, 0, 460). Do you mean to say that this has a positive sign because the z-component is positive? If yes, that's your answer.
UPDATE: If it's the scalar product you want, it's negative in this case:
AB dot AC = -180 -240 + 0 = -420.
From reading the question you linked, it seems you want the sign of the z-component of the cross-product (assuming 0 z-value for AB and AC); to indicate which side of the line AB the point C lies.
Assuming that's the case, all you need is the sign of the determinant of the matrix with AB and AC as its rows.
xAB = b[0] - a[0]
yAB = b[1] - a[1]
xAC = c[0] - a[0]
yAC = c[1] - a[1]
detABxAC = (xAB * yAC) - (yAB * xAC)
if (detABxAC < 0)
// sign is negative
elif (detABxAC > 0)
// sign is positive
else
// sign is 0, i.e. C is collinear with A, B

Finding quaternion representing the rotation from one vector to another

I have two vectors u and v. Is there a way of finding a quaternion representing the rotation from u to v?
Quaternion q;
vector a = crossproduct(v1, v2);
q.xyz = a;
q.w = sqrt((v1.Length ^ 2) * (v2.Length ^ 2)) + dotproduct(v1, v2);
Don't forget to normalize q.
Richard is right about there not being a unique rotation, but the above should give the "shortest arc," which is probably what you need.
Half-Way Vector Solution
I came up with the solution that I believe Imbrondir was trying to present (albeit with a minor mistake, which was probably why sinisterchipmunk had trouble verifying it).
Given that we can construct a quaternion representing a rotation around an axis like so:
q.w == cos(angle / 2)
q.x == sin(angle / 2) * axis.x
q.y == sin(angle / 2) * axis.y
q.z == sin(angle / 2) * axis.z
And that the dot and cross product of two normalized vectors are:
dot == cos(theta)
cross.x == sin(theta) * perpendicular.x
cross.y == sin(theta) * perpendicular.y
cross.z == sin(theta) * perpendicular.z
Seeing as a rotation from u to v can be achieved by rotating by theta (the angle between the vectors) around the perpendicular vector, it looks as though we can directly construct a quaternion representing such a rotation from the results of the dot and cross products; however, as it stands, theta = angle / 2, which means that doing so would result in twice the desired rotation.
One solution is to compute a vector half-way between u and v, and use the dot and cross product of u and the half-way vector to construct a quaternion representing a rotation of twice the angle between u and the half-way vector, which takes us all the way to v!
There is a special case, where u == -v and a unique half-way vector becomes impossible to calculate. This is expected, given the infinitely many "shortest arc" rotations which can take us from u to v, and we must simply rotate by 180 degrees around any vector orthogonal to u (or v) as our special-case solution. This is done by taking the normalized cross product of u with any other vector not parallel to u.
Pseudo code follows (obviously, in reality the special case would have to account for floating point inaccuracies -- probably by checking the dot products against some threshold rather than an absolute value).
Also note that there is no special case when u == v (the identity quaternion is produced -- check and see for yourself).
// N.B. the arguments are _not_ axis and angle, but rather the
// raw scalar-vector components.
Quaternion(float w, Vector3 xyz);
Quaternion get_rotation_between(Vector3 u, Vector3 v)
{
// It is important that the inputs are of equal length when
// calculating the half-way vector.
u = normalized(u);
v = normalized(v);
// Unfortunately, we have to check for when u == -v, as u + v
// in this case will be (0, 0, 0), which cannot be normalized.
if (u == -v)
{
// 180 degree rotation around any orthogonal vector
return Quaternion(0, normalized(orthogonal(u)));
}
Vector3 half = normalized(u + v);
return Quaternion(dot(u, half), cross(u, half));
}
The orthogonal function returns any vector orthogonal to the given vector. This implementation uses the cross product with the most orthogonal basis vector.
Vector3 orthogonal(Vector3 v)
{
float x = abs(v.x);
float y = abs(v.y);
float z = abs(v.z);
Vector3 other = x < y ? (x < z ? X_AXIS : Z_AXIS) : (y < z ? Y_AXIS : Z_AXIS);
return cross(v, other);
}
Half-Way Quaternion Solution
This is actually the solution presented in the accepted answer, and it seems to be marginally faster than the half-way vector solution (~20% faster by my measurements, though don't take my word for it). I'm adding it here in case others like myself are interested in an explanation.
Essentially, instead of calculating a quaternion using a half-way vector, you can calculate the quaternion which results in twice the required rotation (as detailed in the other solution), and find the quaternion half-way between that and zero degrees.
As I explained before, the quaternion for double the required rotation is:
q.w == dot(u, v)
q.xyz == cross(u, v)
And the quaternion for zero rotation is:
q.w == 1
q.xyz == (0, 0, 0)
Calculating the half-way quaternion is simply a matter of summing the quaternions and normalizing the result, just like with vectors. However, as is also the case with vectors, the quaternions must have the same magnitude, otherwise the result will be skewed towards the quaternion with the larger magnitude.
A quaternion constructed from the dot and cross product of two vectors will have the same magnitude as those products: length(u) * length(v). Rather than dividing all four components by this factor, we can instead scale up the identity quaternion. And if you were wondering why the accepted answer seemingly complicates matters by using sqrt(length(u) ^ 2 * length(v) ^ 2), it's because the squared length of a vector is quicker to calculate than the length, so we can save one sqrt calculation. The result is:
q.w = dot(u, v) + sqrt(length_2(u) * length_2(v))
q.xyz = cross(u, v)
And then normalize the result. Pseudo code follows:
Quaternion get_rotation_between(Vector3 u, Vector3 v)
{
float k_cos_theta = dot(u, v);
float k = sqrt(length_2(u) * length_2(v));
if (k_cos_theta / k == -1)
{
// 180 degree rotation around any orthogonal vector
return Quaternion(0, normalized(orthogonal(u)));
}
return normalized(Quaternion(k_cos_theta + k, cross(u, v)));
}
The problem as stated is not well-defined: there is not a unique rotation for a given pair of vectors. Consider the case, for example, where u = <1, 0, 0> and v = <0, 1, 0>. One rotation from u to v would be a pi / 2 rotation around the z-axis. Another rotation from u to v would be a pi rotation around the vector <1, 1, 0>.
I'm not much good on Quaternion. However I struggled for hours on this, and could not make Polaris878 solution work. I've tried pre-normalizing v1 and v2. Normalizing q. Normalizing q.xyz. Yet still I don't get it. The result still didn't give me the right result.
In the end though I found a solution that did. If it helps anyone else, here's my working (python) code:
def diffVectors(v1, v2):
""" Get rotation Quaternion between 2 vectors """
v1.normalize(), v2.normalize()
v = v1+v2
v.normalize()
angle = v.dot(v2)
axis = v.cross(v2)
return Quaternion( angle, *axis )
A special case must be made if v1 and v2 are paralell like v1 == v2 or v1 == -v2 (with some tolerance), where I believe the solutions should be Quaternion(1, 0,0,0) (no rotation) or Quaternion(0, *v1) (180 degree rotation)
Why not represent the vector using pure quaternions? It's better if you normalize them first perhaps.
q1 = (0 ux uy uz)'
q2 = (0 vx vy vz)'
q1 qrot = q2
Pre-multiply with q1-1
qrot = q1-1 q2
where q1-1 = q1conj / qnorm
This is can be thought of as "left division".
Right division, which is not what you want is:
qrot,right = q2-1 q1
From algorithm point of view , the fastest solution looks in pseudocode
Quaternion shortest_arc(const vector3& v1, const vector3& v2 )
{
// input vectors NOT unit
Quaternion q( cross(v1, v2), dot(v1, v2) );
// reducing to half angle
q.w += q.magnitude(); // 4 multiplication instead of 6 and more numerical stable
// handling close to 180 degree case
//... code skipped
return q.normalized(); // normalize if you need UNIT quaternion
}
Be sure that you need unit quaternions (usualy, it is required for interpolation).
NOTE:
Nonunit quaternions can be used with some operations faster than unit.
Some of the answers don't seem to consider possibility that cross product could be 0. Below snippet uses angle-axis representation:
//v1, v2 are assumed to be normalized
Vector3 axis = v1.cross(v2);
if (axis == Vector3::Zero())
axis = up();
else
axis = axis.normalized();
return toQuaternion(axis, ang);
The toQuaternion can be implemented as follows:
static Quaternion toQuaternion(const Vector3& axis, float angle)
{
auto s = std::sin(angle / 2);
auto u = axis.normalized();
return Quaternion(std::cos(angle / 2), u.x() * s, u.y() * s, u.z() * s);
}
If you are using Eigen library, you can also just do:
Quaternion::FromTwoVectors(from, to)
Working just with normalized quaternions, we can express Joseph Thompson's answer in the follwing terms.
Let q_v = (0, u_x, v_y, v_z) and q_w = (0, v_x, v_y, v_z) and consider
q = q_v * q_w = (-u dot v, u x v).
So representing q as q(q_0, q_1, q_2, q_3) we have
q_r = (1 - q_0, q_1, q_2, q_3).normalize()
According to the derivation of the quaternion rotation between two angles, one can rotate a vector u to vector v with
function fromVectors(u, v) {
d = dot(u, v)
w = cross(u, v)
return Quaternion(d + sqrt(d * d + dot(w, w)), w).normalize()
}
If it is known that the vectors u to vector v are unit vectors, the function reduces to
function fromUnitVectors(u, v) {
return Quaternion(1 + dot(u, v), cross(u, v)).normalize()
}
Depending on your use-case, handling the cases when the dot product is 1 (parallel vectors) and -1 (vectors pointing in opposite directions) may be needed.
The Generalized Solution
function align(Q, u, v)
U = quat(0, ux, uy, uz)
V = quat(0, vx, vy, vz)
return normalize(length(U*V)*Q - V*Q*U)
To find the quaternion of smallest rotation which rotate u to v, use
align(quat(1, 0, 0, 0), u, v)
Why This Generalization?
R is the quaternion closest to Q which will rotate u to v. More importantly, R is the quaternion closest to Q whose local u direction points in same direction as v.
This can be used to give you all possible rotations which rotate from u to v, depending on the choice of Q. If you want the minimal rotation from u to v, as the other solutions give, use Q = quat(1, 0, 0, 0).
Most commonly, I find that the real operation you want to do is a general alignment of one axis with another.
// If you find yourself often doing something like
quatFromTo(toWorldSpace(Q, localFrom), worldTo)*Q
// you should instead consider doing
align(Q, localFrom, worldTo)
Example
Say you want the quaternion Y which only represents Q's yaw, the pure rotation about the y axis. We can compute Y with the following.
Y = align(quat(Qw, Qx, Qy, Qz), vec(0, 1, 0), vec(0, 1, 0))
// simplifies to
Y = normalize(quat(Qw, 0, Qy, 0))
Alignment as a 4x4 Projection Matrix
If you want to perform the same alignment operation repeatedly, because this operation is the same as the projection of a quaternion onto a 2D plane embedded in 4D space, we can represent this operation as the multiplication with 4x4 projection matrix, A*Q.
I = mat4(
1, 0, 0, 0,
0, 1, 0, 0,
0, 0, 1, 0,
0, 0, 0, 1)
A = I - leftQ(V)*rightQ(U)/length(U*V)
// which expands to
A = mat4(
1 + ux*vx + uy*vy + uz*vz, uy*vz - uz*vy, uz*vx - ux*vz, ux*vy - uy*vx,
uy*vz - uz*vy, 1 + ux*vx - uy*vy - uz*vz, uy*vx + ux*vy, uz*vx + ux*vz,
uz*vx - ux*vz, uy*vx + ux*vy, 1 - ux*vx + uy*vy - uz*vz, uz*vy + uy*vz,
ux*vy - uy*vx, uz*vx + ux*vz, uz*vy + uy*vz, 1 - ux*vx - uy*vy + uz*vz)
// A can be applied to Q with the usual matrix-vector multiplication
R = normalize(A*Q)
//LeftQ is a 4x4 matrix which represents the multiplication on the left
//RightQ is a 4x4 matrix which represents the multiplication on the Right
LeftQ(w, x, y, z) = mat4(
w, -x, -y, -z,
x, w, -z, y,
y, z, w, -x,
z, -y, x, w)
RightQ(w, x, y, z) = mat4(
w, -x, -y, -z,
x, w, z, -y,
y, -z, w, x,
z, y, -x, w)

Resources