Perlin noise was developed in 1983 by Ken Perlin, Computer Science Professor at NYU. The algorithm is ultra-fast and generates a random field that has controllable spectral dyanmics. It can be used to simulate realistic smoke (among other uses). Way before AI came along. Adjust the number of spectral OCTAVES to see the effect of smaller scales in the simulated smoke. A SEED number is used to start the simulation. The spectrum calculation can be checked by specifying a wavenumber, then CHECK FFT.