Matrices And You, Part 1 : What are they?

What Are They?

That’s a good question. Here’s the usual 4-dimensional matrix used in 3D, like TV_3DMATRIX in TV6.5 :

| 1   2   3   0 |
| 5   6   7   0 |
| 9   10  11  0 |
| 13  14  15  1 |

It’s scary. It makes no sense. So let’s take it another way.

First, let’s think of a matrix as a combination of vectors. You know vectors, right? ...right? Alright, let’s take it from there.

What Really Are Vectors

The intuitive idea about vectors is that they’re a point in space. They’re not.

Well, what are they?

Vectors are lines in space... well almost. They’re lines without an origin; just a direction and a length.

For example, take this 3D vector here : V = [ 1 2 3 ] Don’t like him like that? It can also be expressed like this :

V = (1, 2, 3)
V = { 1, 2, 3 }
V = | 1  2  3 |
TV_3DVECTOR V = new TV_3DVECTOR(1, 2, 3)

The important part is that it’s called V, and has three components.

Three components... That’s fairly easy, it’s a 3D vector. One component for each dimension. Since the convention for Direct3D is X-Y-Z worlds where X is left-to-right, Y is bottom-to-top and Z is backward-to-forward, that means that V has a component of 1 for X, 2 for Y and 3 for Z.

That does not mean that its position in 3D space is (1, 2, 3). It has no position at all. Remember, a line without origin... a force in other words. A difference between two locations, wherever might those two locations be.

A good example vector would be a normal vector. A normal is the vector that pops out of a 3D triangle or a face and that is perpendicular to the surface of this triangle/face. It expresses the orientation of that face, and is used mainly for lighting and physics calculation. Since it only represents the orientation, its position is negligible.

We did mention vectors being a direction and a length. How would we fetch them from V?

Calculating the length and direction of a vector

Length of V : ||V||
||V|| = Sqrt(V.x² + V.y² + V.z²)

Direction of V : Normalized V
Normalized V = V / ||V|| 
             = V / Sqrt(V.x² + V.y² + V.z²)

Hope this didn’t scare you too much.

What were those symbols?

  • Sqrt is the square root operation
  • / is the division operator
  • ||x|| is just a notation for the length of a vector
  • ² is the square... x * x

Why this square root of the sum of the squares of the three components?

  • This is the pythagorean theorem, a pretty simple theorem to express the length of a line (V) if we know its separate components (x, y and z)... which we do. The length is then a single number.

And why would we divide the vector by the length, and what's a vector divided by a single number anyway?

  • If we take the vector and divide it by its length, and calculate the length of the resulting vector, it will always be 1. This is the definition of a normalized vector; a vector whose length is 1. It can also be called a unit vector.
  • A vector multiplied or divided by a single value, like the length here, will multiply or divide each of components by this value.

What does this have to do with anything?

Once we know all of that terminology, we can start expressing matrices as vector arrays. Which is covered in part two, actually.

tutorialsarticlesandexamples/matrices_and_you1.txt · Last modified: 2013/11/22 13:32