I've experimented a little with my water-system, and tried
leaving the dampening of the waves out. I would have
thought that the distribution of the wave would take care
of enough dampening, obviously, I was wrong.

But, as a side effect, it is clearly visible that
a) surrounding walls need to be high enough, and
b) that certain sections are saved from the parsing-process.

In all those places where the water pops over the rim and
then drops down to the walls, thats where a node is set
to "off" and isn't parsed at all. It just stays put, since
"normally" it wouldn't be needed (unless serious
problem rise up, like this animation shows).

Well, I found this rather an amusing look (like a cake
in an oven) and thus thought, some might want to see
it... :-)

--
Tim Nikias v2.0
Homepage: http://www.digitaltwilight.de/no_lights
Email: Tim### [at] gmxde

Post a reply to this message

Attachments:
Download 'water_b4.mpg' (353 KB)

Ah! It cuts off just as it gets interesting. Would it stabilise or get chaotic?
Looks good though, very SF'ish. 

Regards
        Stephen

Post a reply to this message

I bet it will get chaotic more and more until there are overflows
(numerically).

Post a reply to this message

I'm not quiet sure what would've happened. I'm
rendering another simulation right now, but I want
to use a similiar setup again and see what
actually happened.
For one, I realised that the very first position of the
wave never stopped jiggling about, and sort of
bumped itself upwards. Without dampening, the
reflection of the waves in such a tight corner might
add to itself in such a way, that the smoothing
part of the algorithm can't smooth it away. Dampening
definitely took care of that issue.
What I don't like about that, is that I can't control that.
There's no way, other than telling you to not use no
dampening, to avoid this effect.

Perhaps I'll add the "interesting" part tomorrow, we'll
see.

--
Tim Nikias v2.0
Homepage: http://www.digitaltwilight.de/no_lights
Email: Tim### [at] gmxde

Post a reply to this message

--------------( oo )--------------

Your simulation are very interresting.
You said that it tooks a lot of calculation to render this. How much time
does it tooks? Is the Parsing is longer than the render time?

--------------( oo )--------------



"Tim Nikias v2.0" <tim### [at] gmxde> wrote in message
news:3e837f7e@news.povray.org...
> I've experimented a little with my water-system, and tried
> leaving the dampening of the waves out. I would have
> thought that the distribution of the wave would take care
> of enough dampening, obviously, I was wrong.
>
> But, as a side effect, it is clearly visible that
> a) surrounding walls need to be high enough, and
> b) that certain sections are saved from the parsing-process.
>
> In all those places where the water pops over the rim and
> then drops down to the walls, thats where a node is set
> to "off" and isn't parsed at all. It just stays put, since
> "normally" it wouldn't be needed (unless serious
> problem rise up, like this animation shows).
>
> Well, I found this rather an amusing look (like a cake
> in an oven) and thus thought, some might want to see
> it... :-)
>
> --
> Tim Nikias v2.0
> Homepage: http://www.digitaltwilight.de/no_lights
> Email: Tim### [at] gmxde
>
>
>

Post a reply to this message

It depends on several factors. For the parsing time,
the resolution of the array which holds the heights
is the main factor, along with the amount of steps that
have to be calculated per frame.
After the simulation, my Mesh-Modifying-Macros kick
in to generate the mesh and its surface-normals. These
also take a little time, but very little compared to the
simulation.
The tracing time depends on the features I'm using. The
mesh itself can be traced fairly quick, but if I'm using
a transparent surface, variable reflection and photons,
render time rises of course. Along with that, I use antialiasing
for the the final renders, so that even the large high-quality
MPGs I keep for myself look smooth.
On a render I'm doing right now, I'm using 60*60 Nodes, so
thats 3600 nodes for the simulation. Mind that the simulation
sort of "clips" dead nodes and doesn't touch them. On these
spherical pools, this saves a lot, since the corners are
completely left out. The mesh uses a reflecting surface to get
that liquid metal look. Additionally, I use antialiasing on the
current render.
Thus, on my Athlon 1.4GHz, it parses at about 13 seconds
per frame, and traces with about 23 seconds per frame.
So thats 36 seconds per frame. For 250 frames, this takes
roughly 2h.
Still, parsing time wouldn't go up with more or less waves,
their height doesn't matter. So, thats an advantage, and
can guesstimate fairly good how long an animation
might take. OTOH, it'll take that long for parsing even
if there's no wave present.



--
Tim Nikias v2.0
Homepage: http://www.digitaltwilight.de/no_lights
Email: Tim### [at] gmxde

>
> --------------( oo )--------------
>
> Your simulation are very interresting.
> You said that it tooks a lot of calculation to render this. How much time
> does it tooks? Is the Parsing is longer than the render time?
>
> --------------( oo )--------------

>
>

Post a reply to this message

Here's a longer version. As mentioned before,
it seems that the water juggles itself upwards
in the tight back corner, where smoothening
doesn't "eliminate" height (which is normally
a wanted side-effect). I've stopped the simulation
when the liquid metal reached the clipping-object,
since that's just about where the interesting part
ends. After that, it'll just fill up indefinitely.

Nontheless, the waves look convincing, how
they swirl around the cylinder... I like it. :-)

--
Tim Nikias v2.0
Homepage: http://www.digitaltwilight.de/no_lights
Email: Tim### [at] gmxde

Post a reply to this message

Attachments:
Download 'water_b4b.mpg' (397 KB)

hey tim

that's really cool. i like espescially the moment just when it reaches the
clipping container, with only the "side walls" of the water visible. looks
really nice.

keep up the good work
florian

Post a reply to this message

So it stabilizes by growing upwards in the shape of the vessel. The last six
frames are intriguing. 
Regards
        Stephen

Post a reply to this message

Very cool, but as for a liquid metal, per se,  the surface tension is wrong.

Post a reply to this message