NodeBox

This version of NodeBox is no longer supported. (Why?) For the latest version, visit nodebox.net.

Controlling Randomness

Randomness in NodeBox is not really random. It produces pseudo-random values controlled by an initialization value called the seed.

Each seed produces a different range of random numbers. The same seed will always give the same collection of random numbers. That’s why we talk about controlled randomness: within a certain collection, the ordering appears random:

  • Seed 1: 14 85 77 26 50 45 66 79 10 3
  • Seed 2: 96 95 6 9 84 74 67 31 61 61
  • Seed 3: 24 55 37 61 63 7 2 84 26 24

This is actually very useful. As a designer, you might pick a certain random variation that you think looks good. It wouldn’t make much sense if NodeBox would change the output when you reopen the document. The seed gives us a way of creating variation without giving up control.

Everywhere a NodeBox node uses a random value, it provides a seed parameter for you. Nodes such as wiggle, scatter and sort will always produce the same result for the same seed.

Here’s an example:

  • Create a star node. Set the X to -120.00 to move it to the left.
  • Create a wiggle node. Set the Seed to 1.
  • Connect star1 to wiggle1.
  • Select the two nodes, copy them, and paste them again.
  • Double-click wiggle2 to make it rendered.
  • In star2, set the X to 120.00 to move it to the right.
  • Create a merge node.
  • Connect wiggle1 to merge1.
  • Connect wiggle2 to merge1.

If you look at the output of the merge node, you’ll see that they both have the same random variation. That’s because they share the same random seed.

Same wiggled stars

In wiggle2, change the seed to 2. Note that the two stars now look different.

Same wiggled stars

The actual number of the seed is unimportant. Seed 10 is not “more random” than seed 1. See it as the index number of a certain variation.

Using seed in your own nodes

When you’re programming your own nodes that use random values, add a “seed” parameter using the metadata menu and add the following line right below the cook method:

  from random import seed
  
    def cook(self):
        seed(self.seed)
        ...

NodeBox nodes are evaluated lazily, and producing “true” random numbers breaks this model. Using a seed parameter makes your nodes idempotent.