Why have you changed the range to be negative?

In D14096#376482, @Sebastian Parborg (zeddb) wrote:

Why have you changed the range to be negative?

Yeah, I guess that doesn't make much sense now that I think about it. If you set the values in the negatives It would produce falling fireballs. But maybe someone would want to do something more abstract/unrealistic like that? Should the lowest possible value just be set to 0? When it comes to simulations I'm not sure what kind of limits should be enforced. Only physically accurate ones?

Usually we don't try too hard to stick to physically accurate limits.
It is mainly there to enforce that numbers that would make the simulation unstable or simply crash can't be set.

Have you made any tests with negative values?

Yes the gas falls when the temperatures are in the negatives and it doesn't break the simulation (at least it didn't in my tests)

But the thing that I found out now is that the workbench/solid/wireframe view doesn't shade the lighting properly which is baked into the simulation. If I go below zero it loses shadows and if I go above the original limits it starts to darken the smoke.

Min(5) Max(10):

Min(-1000) Max(-100)

Min(100) Max(1000)

I'm starting to think that this would require a lot more work for proper implementation and I'm not experienced enough to fix the viewport shading problem that gets introduced.

Yepp, then I guess that is why the limits where there to begin with (to make sure it renders properly).

I'm guessing that I can close this then?
We can talk more on blender.chat if you like.

Yeah you can close it then since the solution would introduce a bug.

I just hope that if/when the workbench smoke rendering gets revisited this will be thought of.

If you know about Mantaflow stuff we can talk more about it there :)