noisepad

The bits I kept copying between sketch scripts, cleaned up into a small package: 2D gradient noise with fractal octaves, and a supersampled Pillow canvas that gives you antialiased lines without thinking about it. All the images in the gallery are made with (an older copy of) this. How and why: Sunday plotting.

from noisepad import Noise, Canvas

n = Noise(seed=4)
c = Canvas(1500, 1000, bg=(20, 24, 28))
for y in range(0, 1000, 20):
    c.line([(x, y + n.fbm(x / 300, y / 300) * 80) for x in range(0, 1500, 4)],
           fill=(230, 220, 200), width=1.5)
c.save("lines.jpg")

The tarball contains the package and a flow-field example.

Changelog

• 0.1.2 (2026-04-14) — Canvas.save returns the downscaled image.
• 0.1.1 (2026-04-13) — fix JPEG save for very large canvases.
• 0.1.0 (2026-04-12) — first version.

Download

Full checksums: SHA256SUMS