An 'initial proof-of-concept' for a roiling gasoline fire or explosion. Included
is a representative image from the 'net to compare my trial against.

Currently, I'm using individual media spheres that overlap somewhat, with  just
emission and absorption medias (and only 'half a donut's worth' of spheres, for
testing.) I toyed with the idea of using a large torus object instead (with a
toroidal warp of the media); but using individual spheres gives better
'erratic/random' control of the overall roiling fire look, and I can tweak the
micro-details more easily.

(For those interested, here's a video and an academic paper I came across,
describing a modern CGI/mathematical 'conceptual' approach to creating this type
of thing. )

https://www.youtube.com/watch?v=vBLZ4jja8No

http://www.cs.cornell.edu/~tedkim/wturb/

For my own MUCH simpler test, there still remains a lot of stuff to work out
(and all this just for the 'top' of the roiling fire!):

* adding more spheres!

* more intense fire colors

* making the individual 'fires' in the spheres random-looking, and maybe with
different rotation speeds. (Currently, it's the same single sphere, duplicated.)
I'll probably turn the code into a macro, to make it easier to work with for
multiple different spheres.

* making the yellow-to-red-to-black  color-gradations 'sharper' and less
'blurry' looking-- a matter of tweaking the color-maps

* addind scattering media, to get self-shadowing of the gray/black smoke from
the light source(s). That's currently missing (I may have to fade out both the
emission and absorption medias along the way, somewhat.) As a practical matter,
the initial fire doesn't need self-shadowing (real  fire is probably brighter
than any incoming light source.) But that needs to change as the fire 'does
down' and turns into smoke.

* thinning (and 'graying') of some of the smoke as it turns black, with a more
'wispy' nature

* a slow-down of the rotational roiling as the cloud progresses upwards

* expanding the size of the spheres as they turn black/gray.

Post a reply to this message

Attachments:
Download 'roiling fire cloud exp 2.mp4.mpg' (3223 KB)

"Kenneth" <kdw### [at] gmailcom> wrote:
>... (real  fire is probably brighter
> than any incoming light source.) But that needs to change as the fire 'does
> down' and turns into smoke.
>

Uh, "...as the fire DIES down". Sorry :-O

Post a reply to this message

"Kenneth" <kdw### [at] gmailcom> wrote:
>
> Currently, I'm using individual media spheres that overlap somewhat, with  just
> emission and absorption medias (and only 'half a donut's worth' of spheres, for
> testing.) I toyed with the idea of using a large torus object instead (with a
> toroidal warp of the media); but using individual spheres gives better
> 'erratic/random' control of the overall roiling fire look, and I can tweak the
> micro-details more easily.

It's a great start.

Maybe you can use the phase keyword for density color changes, with a clock
variable for that. Or maybe not... I was just trying it myself to see what I
could do making a fireball, not much success only a kludge. But that's me. :)

I like the big and little swirls idea.

Bob

Post a reply to this message

Hey -
that is REALLY nice!   :)

If that's what you're _starting_ with, it ought to be really impressive once you
apply all of your improvements, refinements, and tweaks!

That ought to be really nice to apply to volcanoes, dust, steam, water, nukes,
and all sorts of other fluidized thingamabobs.  :D

Excellent, excellent work.

Post a reply to this message

"omniverse" <omn### [at] charternet> wrote:

>
> Maybe you can use the phase keyword for density color changes, with a clock
> variable for that.

I actually forgot about the phase keyword :-/ I'll give that a try; sounds like
a good idea.

Currently, I'm animating warp{turbulence...} through the spheres, which gives a
nice roiling look-- but it's not ideal for this. Turbulence can be moved in x, y
or z (or all three), but that's still a single vector direction. What would be
more appropriate here (although probably impossible) would be to have turbulence
moving outward from the *center* of the sphere, radially, in all directions at
once (and maybe scaling up as it does so.) But that would require the turbulence
pattern to begin at what is essentially a 'singularity' at the center of the
sphere! Oh well.. it's still a 'fantasy feature' that I wish POV-Ray had.

Post a reply to this message

On 6/21/2017 11:46 PM, Kenneth wrote:
> An 'initial proof-of-concept' for a roiling gasoline fire or explosion. Included
> is a representative image from the 'net to compare my trial against.
>

I think it is pretty realistic and I've spent hours looking at them when 
I should have been working. :)


-- 

Regards
     Stephen

Post a reply to this message

"Bald Eagle" <cre### [at] netscapenet> wrote:
> Hey -
> that is REALLY nice!   :)

>
> That ought to be really nice to apply to volcanoes, dust, steam, water, nukes,
> and all sorts of other fluidized thingamabobs.  :D
>

Thanks!

Years ago, I actually worked out a much more complex version of this kind of
thing-- a full explosion-- but I didn't use media for it. Instead, to save
rendering time, the spheres used image_maps (simple ones, but heavily distorted
with animated turbulence), along with some crazy overlayed pigments that changed
colors. It actually looked OK-- from a distance-- but depended on the image_maps
and pigments constantly facing the camera. *If* I can find the old animation
video that I made, I'll post it... but don't hold your breath ;-)

Media just 'looks better', although it's a bear to code all the complexities--
and to render all the necessary but slow tests!

Post a reply to this message

Wondering if you've seen this:

http://www.cs.cornell.edu/~tedkim/WTURB/

Post a reply to this message

Am 30.06.2017 um 21:06 schrieb Bald Eagle:
> Wondering if you've seen this:
> 
> http://www.cs.cornell.edu/~tedkim/WTURB/

I guess he has - he mentioned that page earler ;)

Post a reply to this message

clipka <ano### [at] anonymousorg> wrote:

> I guess he has - he mentioned that page earler ;)

Ooooohhhh - OKAY.  It was the Friday before Independence Day, and it was hot,
and I posted it real quick before we had 20,000 lbs of stuff to move.   :P

Post a reply to this message