*See also my notes on row-major and column-major conventions.*

## Types of matrices

There are four types of matrices we are concerned with in most OpenGL apps.

- Scaling – make bigger or smaller
- Translation – move
- Rotation – spin around.
- Projection – convert view-space (aka, camera space) coordinates to screen-space

By building and concatenating these types of matrices, we can create a single object (matrix) that performs all of these operations with a single (and fast on the GPU) matrix multiply.

Looking at 16 apparently-random floating point values in a 4×4 matrix can be confusing, we can simplify things a bit if we understand the meaning of each of the individual cells within the matrix.

## Meaning

The matrix is really a collection of 4, 4-component vectors. Each vector describes a** basis vector** of the transformed coordinate system. For instance, in a native, untransformed 3D Cartesian space, we know that the three axis are represented by unit vectors in each direction.

\dot{\vec{x}} = 1\hat{x} + 0\hat{y} + 0\hat{z} \\

\) \(

\dot{\vec{y}} = 0\hat{x} + 1\hat{y} + 0\hat{z} \\

\) \(

\dot{\vec{z}} = 0\hat{x} + 0\hat{y} + 1\hat{z} \\

\)

These can be conveniently represented by a 3×3 matrix and a vector:

\(\begin{bmatrix}

1 & 0 & 0 \\

0 & 1 & 0 \\

0 & 0 & 1 \\

\end{bmatrix} *

\begin{bmatrix}

v_x \\

v_y \\

v_z \\

\end{bmatrix}

\)

Multiplying a three component (x,y,z) vector(aka Vector3) * V* by this matrix would give us a three-component vector \(v_1\),

v^1_x = v_x * 1 + v_y * 0 + v_z * 0

\) \(

v^1_y = v_x * 0 + v_y * 1 + v_z * 0

\) \(

v^1_z = v_x * 0 + v_y * 0 + v_z * 1

\)

Note that this is the same as the dot product of \(v_1\) with each row of the matrix.

## How to multiply a 4×4 matrix with a 4 component vector

\(\begin{bmatrix}

a & b & c & d \\

e & f & g & h \\

i & j & k & l \\

m & n & o & p \\

\end{bmatrix}

\begin{bmatrix}

v_x \\

v_y \\

v_z \\

v_w \\

\end{bmatrix} =

\begin{bmatrix}

v_x * a + v_y * b + v_z * c + v_w * d \\

v_x * e + v_y * f + v_z * e + v_w * f \\

v_x * i + v_y * j + v_z * k + v_w * l \\

v_x * m + v_y * n + v_z * o + v_w * p \\

\end{bmatrix}

\)

## Why we use 4×4 Matrices

Translation – we need a 4th row or column to store translation values. See the Translation Matrix page for details.