Smokesolvers and Age fields

[Macha]

I'm thinking about good ways to make an age field for volume sims. Temperature doesn't work well and nor does heat. I'm hoping to find a way to track age through a smoke or fire sim. One of the ideas I have is by advecting a particle sim and transferring the life attributes back. It would be so great to have age fields. And so useful!

[anim]

not tested

but maybe you can create scalar field Age default to 0

and at each timestep do:

source value 0 from your sources to Age field

advect Age field by velocity

add small value (timestep) to whole Age field masked by Density

[bunker]

you can use a scalar field:

initial : age = 0

age = age + source

age = age * (1-aging_rate)

you get values from 1 to 0

Edited September 12, 2011 by bunker

[Macha]

Hm, good idea Bunker, but doesn't that do the reverse? If I try this, say with a pyroball, then the center of the sim becomes denser. But that is the youngest part, where the smoke emerges. The smoke gets older as it travels, so I would expect the age field to increase there.

[Macha]

Seems like if I make an age field that progressively adds temperature and density (if < lowvalue), and advect it, then I get a sort-ish age field.

Edited September 14, 2011 by Macha

[bunker]

yes it looks like it's the reverse (actually complement), but you can do a lot with that, and that's the most accurate and logical way I found - there is only age where there is density. since the age values are normalized, you can use (1-age) anywhere (eg: a gasfield vop). What are you actually doing with the age ? is it for shading purpose ?

[Benyee]

It would be hard to achieve in voxel based sim process as the properties carried by the voxels rather than carried by individual entities like that in classic particle sim.

[Klockworks]

Why not does something along the way a rest field is handled? Without the reinitialization of course. Where density is emitted you get a value of 0, and have the rest of the field be -1. At each timestep you could add a value to it and then advect it by the vel field. I would think using 100% of the vel to advect would give you a pretty accurate age track of the density as it moves. You will probably get some odd things happening where your density does not move entirely outside of emission.

Though I think particles would work as well.