Question

Every implementation I’ve come across of perlin noise generation has been for the generation of 2D terrain, etc. I cannot find a decent example of point to point lightning generation anywhere.

Are there many other forms of generating ‘lightning’? I was told this is what I want. What algorithms exist for forked lightning, or 2D trees (I could turn this upside down for lightning maybe)

I work in multiple languages so examples in pseudo-code are OK also.

Answer 1

My understanding is that Perlin noise is designed so all of its patterns are close to one single “size”, but lightning is a fractal with patterns at every “size”.
There may be a way to use several Perlin noise systems at different scales to make lightning, but perhaps one of these other methods would work better:

The midpoint displacement method is a very fast algorithm
for generating shapes that look a lot like lightning.
Alas, it only generates a (very jagged) lines between two points, never forks.
There’s some source code and pictures at
http://www.krazydad.com/bestiary/bestiary_lightning.html

Probabilistic L-systems can be used to generate shapes that look like lightning.
I see that at least one person abandoned Perlin noise and switched to L-systems in order to simulate lightning.
http://www.grepart.com/showcase/content/lightning_paper.pdf

The stochastic Lichtenberg algorithm runs somewhat slower,
but it is more physically realistic model of lightning and generates lots of forks.
http://fisica.ciencias.uchile.cl/alejo/fractal_antenna/node7.html

Is this a dup of the “Best lightning generation\simulation algorithm?” question?