New isotropic scatterer material

[patrickwalz]

i was actually looking foward to user selectabilty -- figured that might get mie and raleigh scattering into the system quicker -- along with hg and schlick lol

[jeanphi]

Hi,

Whatever, it's your material, there'll always be use for it I guess, so make it user selectable. However the more options you have to deal with while rendering, the slower it gets (this should be marginal though but worth keeping in mind).

Jeanphi

[paco]

After a bit more investigation i've come up with a few things:

1) The last formula I posted (with the 1/(cosi+coso) ) term seems to be correct - and is the same one that tomb was using for the velvet material, except that it has a general phase function. Also if you expand/approximate the exponential you get a 1/( cosi coso) dependence

2) What I had originally wanted to do was fake an emitting gas - but instead of working with a new light I went for a material that might be able to fake it. But that really needs a 1/coso term in it - but ideally no 1/cosi term. The 1/(cosi+coso) doesn't look like an emitting gas.

3) The scattering bxdf only accounts for a single scatter event. You could get a much better fit by calculating what percentage of light gets scattered more than once and converting that to an isotropic component (not my original idea - read it somewhere)

So at this stage I'm a bit torn between implementing a realistic scattering material accounting for multiple scatters, or implementing a material to fake an emitting gas. In retrospect what I should have done originally was create a double-sided arealight and create an isotropic IES file to go with it.

[pablo]

If I may speak my mind, I really liked the image of your second post (http://www.luxrender.net/forum/viewtopic.php?f=10&t=5057#p52619), because until now there was NO way to get a similar effect in Luxrender, nor in Maxwell Render for what I know.
(BTW it is what is called "Fresnel effect" in Cinema4D).
I do not think that the current implementation achieves the effect that "you get most light where you are viewing the flame along an edge, rather than looking face on" (or absorption, or diffuse, i would say), as it clearly appears in that image.
just my two cents.

pablo