*@dean* He uploaded an explosion example in Chapter 4-4

*@Huân Lê-Vương* That example was to show baked simulations, it's not just a procedural shader!

Technically you could do something similar with nodes, but it would be much more difficult than this simple fire and would probably never look as good as actual simulation.

But if you want something more stylized then it might be a valid option.

*@Simon Thommes* oh sorry!

*@Aron* scale is "linear". power isn't linear. when scaling, the vector remain the same "shape" of it, just "bigger"/"smaller". power is we multiply the value to itself multiple time, imagine each coordinate will be scaled in a different way than the others may result in stretching the vector.

*@Huân Lê-Vương* oh thanks for answer my question. 0~0.1 to power of 2 is 0~0.01 so 0~0.1 value range get shorter but 0.9~1 to power of 2 is 0.81~1 so 0.9~1 value range get longer , generally 0~1 value range is stretched lengthen. so top of fire animation is fatster then bottom is it correct?

*@Aron* for example, we create a plane. we take X value of the Generated. on the graph, the function of X is y=x with the limit from 0 to 1.

We power the value of X with the exponent of 2. now the function is y=x^2. limit from 0^2 to 1^2. the value still goes from 0 to 1 but look at the graph, it's stretched.

to see how it's stretched, just separate the y=x^2 into 10 steps of 0.1 using a snap operation (add a math node, change to snap, increment 0.1)

in the graph, as you mention above, at the first step of 0.1, y=0.1^2=0.01. so when x goes from 0 to 0.1, y only goes from 0 to 0.01.

At the last step 0.9 to 1, you see from 0.9^2 to 1^2 is 0.19. so the y goes 19 time faster than it does at the first step.

*@Huân Lê-Vương* great! thank you to solve my question!