HOME Mathematics Computer Science
Noise Functions
Show All
Hide All

  • Heightmaps are a way of representing a 3D surface as a 2D image.
  • Each pixel in the 2D image corresponds to a specific point on the 3D surface.
  • The brightness of the pixel determines the height of the point:
    • A dark pixel is a low point
    • A bright pixel is a high point
Image rendered in SketchUp
from PIL import Image

im = Image.new("L", (256, 256))
pix = im.load()

for x in range(256):
    for y in range(256):
        pix[x,y] = 256 - abs(x - 128) - abs(y - 128)

im.save("test.png", "PNG")

Noise Functions

Random Values
We can generate random values using a pseudorandom number generator:
X 1 2 3 4 ...
rand(X) 28 21 96 96 ...
import random

randX = [random.randint(0, 100) for x in range(0, 50)]
Many areas of procedural generation require random values, but the randomness needs to look more organic:
This is the job of a noise function. It produces random data, but the data has an underlying organic nature.
Higher Dimensions
  • The above examples were in one dimension
    • X is an array of numbers, for each number in X there is a corresponding value.
  • However it is often useful to generate noise for 2 or more dimensions:


    3D Terrain

    Image rendered in SketchUp
  • In the above example a random value is assigned to each (x, y) coordinate pair:
    • 2D noise functions are often used to generate terrains in which the height values are the result of the noise function applied to each point.
    • (x, y, z) = (x, y, noise(x, y))
  • Noise functions can be extended into 3 or more dimensions:
    • For example, a 3D noise function could be used for modelling gas. The random value at each 3D point could describe its density.
    • 3D noise functions can also be used to animate 2D noise functions:
      • A 3D array of data is produced and then cut into a series of 2D slices.
      • You can then convert each slice into an individual frame of an animation.
      • This can then be used for when you need 2D noise that changes over time, such as when animating the formation of clouds.