Getting Creative with Perlin Noise Fields

It turns out that adding noise to gradients is a really useful thing to do - so many new effects can be created with just a few additional parameters. Sadly, CSS and Canvas API gradients (linear, radial, conic) are very basic implementations (and SVG is not much more advanced).

Recently I did some work to add software gradient enhancements to my canvas library. Because these run on the CPU rather than GPU they're computationally intense, but still worth the effort just to see what can be done with different spreads (pad, repeat, reflect, transparent) and noise engine operations inserted while calculating stuff like gradient color selection, and pixel positioning.

Linear gradient demo test - https://scrawl-v8.rikweb.org.uk/demo/canvas-003.html

Reminds me of this writeup that I read recently about using a different noise function (not Perlin) to emulate terrain erosion: https://blog.runevision.com/2026/03/fast-and-gorgeous-erosio...

Really nice exploration, and their other posts are interesting too!

We used Perlin noise for demos of our Golang/BubbleTea terminal Glyph heatmap widget and then later with our Picture widget.

Live WASM demos of the Golang terminal programs:

https://nimblemarkets.github.io/ntcharts/demos/heatmap-perli...

https://nimblemarkets.github.io/ntcharts/demos/heatpicture-p... Press 't' to switch between glyph/image modes

Author here. I wrote this many years ago (2017?) while exploring techniques to create art that I could put up on my walls :-) If you enjoyed this article there are more similar ones linked on the main page: https://sighack.com/

Great write up! I also dove in to this topic a while ago over at https://damoonrashidi.me/articles/flow-field-methods, but putting the live processing sketches in was a very nice touch! Good job, and nice outputs!

Very beautiful.

I am a complete newbie, so I might be asking about something obvious -- does anyone know how some of these would relate to plotting random Polya vector fields ?

For example, one can take a sum of complex rational functions of the form

    f_i(z) = r_i / (z - p_i) 

where (r_i,p_i) are complex numbers drawn from some random point process, say a generalized Poisson one.

One needs to plot the conjugate of sum_i f_i(z).

EDIT: so many lovely pages pointed to in the comments ! Let me convey collective thanks to all, rather than clutter this page with individual thanks.

@ttctciyf you are marked dead. Not sure why.

My old work using perlin noise https://a.tulv.in/noise-planets.html

This is super nice article. Will revisit this one more time and try that out as well. In the age of AI slop, it is such a pleasure to read actual people blogs about what they do and are passionate about. Thank you this one!

some beautiful effects and good ideas. My favs are iterations 8, 20, 25

This is awesome