rumore 0.1.6

A numpy-based noise generator

Table of Contents

1. Vectorized value & gradient noise with NumPy
   1. Install with PIP
   2. Install locally
   3. Examples

Vectorized value & gradient noise with NumPy

Rumore is a lightweight Python library for procedural noise. It provides value noise and gradient noise in 1D/2D/3D plus octave summation of these (fractal Brownian motion, fBm).

The library does not implement Ken Perlin’s original algorithm directly, but its gradient noise produces visually similar results.

Install with PIP

pip install rumore

Install locally

Clone the repo, navigate to the directory and from there

pip install -e .

Examples

Import necessary stuff

import matplotlib.pyplot as plt
import numpy as np
import rumore

Generate some 1d gradient noise with different number of octaves

x = np.linspace(-10, 10, 200)
plt.figure(figsize=(5,4))
for i in range(1, 8):
    plt.plot(x, rumore.grad_noise(x, octaves=i))
plt.show()

Perturb points along a circle with 2d noise

t = np.linspace(0, np.pi*2, 200)
x = np.cos(t)
y = np.sin(t)
r = rumore.grad_noise(x, y)*0.5+0.5
plt.figure(figsize=(4,4))
plt.plot(x*r, y*r)
plt.show()

Generate a grid of 2d noise value (a grayscale image)

x = np.linspace(0, 5, 300)
y = np.linspace(0, 3, 100)
img = rumore.noise_grid(x, y)
plt.imshow(img)
plt.show()

Set noise properties. Here we set the octave_map property, defining a function that is applied to the noise of each octave:

from importlib import reload
from rumore import rumore
reload(rumore)
rumore.cfg.lacunarity = 2.0
rumore.cfg.falloff = 0.5 # falloff for each octave (persistence)
rumore.cfg.set_degree(5) # Interpolation degree (3 or 5)
rumore.cfg.shift = 0 #120.321 # Shift between octaves
rumore.cfg.octave_map = lambda x: np.sin(x*np.pi*4)

x = np.linspace(0, 5, 300)
y = np.linspace(0, 3, 100)
img = rumore.noise_grid(x, y)
plt.imshow(img, cmap='gray')
plt.show()