William Tracy <wtr### [at] calpolyedu> wrote:
> Okay, I'm really liking that flame.
>
> I'm assuming emitting media? What sort of pattern?
>

My first try at animated flames! It's actually four different medias--two
emission and two absorption (which absorb only gray values, giving it a nice
opague density without altering the colors.) There are two 'different' flames
at work--the main one, plus a smaller stationary one at the base. Two things
give them 'life'--phase for the moving color_maps, and translate for the
turbulences.  (I discovered that phase doesn't affect turbulence; never knew
that before.) Basically, the repeated color_map moves upward, while turbulance
moves through it in a different direction. It looks OK as-is, but I continue to
experiment. The 'licking' of the flames is a hard nut to crack! I'm sure it can
all be improved.

Even though the color_maps are gradient y--which extend all the way 'across' the
isosurface and stop abruptly at the edges--the undulating, bumpy isosurface
shape and the flame turbulence make that less apparent. (It could probably
benefit from an additional all-encompassing faded-out pigment_map, using
a spherical pattern perhaps.)

Rather than describe the method in detail--difficult in any case!-- I'll just
post the code over in text.scene-files, including the isosurface container.
It's made for a 96-frame animation (at 24fps) with the clock going from 0 to 1.
Try commenting out sections to see what they do. (BTW, for my own torch flame, I
chopped off the lower/darker section of the isosurface.)

The code is actually somewhat forgiving, as far as exact 'movement' values are
concerned. Whatever looks visually appealing. If you're not using it in
animation, all of that code can be eliminated.

Ken W.

Post a reply to this message