There is a company that does plug-ins for high end pricey renderers,
there is an animation of a box being filled with a liquid that is beyond
belief, for me anyway.  The website is www.areteis.com, click on the
picture of the half filled blue liquid box and download the animation.
Then let's figure out a way to do it......  Notice that just before the
"water" hits the floor of the box there is a induced split in the liquid
it seems, in order to make it flow around the box.  Anyone know of
literature on how this would be attempted?

Shaw

Post a reply to this message

Shaw <sha### [at] activech> wrote:
> There is a company that does plug-ins for high end pricey renderers,
> there is an animation of a box being filled with a liquid that is
beyond
> belief, for me anyway.  The website is www.areteis.com, click on the
> picture of the half filled blue liquid box and download the animation.
> Then let's figure out a way to do it......  Notice that just before
the
> "water" hits the floor of the box there is a induced split in the
liquid
> it seems, in order to make it flow around the box.  Anyone know of
> literature on how this would be attempted?

Certainly a very impressive animation (as is the company's list of
projects) - what particularly impressed me was the turbulence and the
'clinginess' of the liquid to the box's sides.  This leads me to guess
that the algorithm they used was a particle system that takes into
account the surface tension of the liquid, perhaps interpolating the
surface to join the particle points in space (they describe it as an
implicit surface rendering).  This might explain the split that seems to
occur before the initial splash: the water on the outside of the column
would be held back by the surface tension, causing a sort of peeling
effect (at least that's what it looks like to me).

To recreate such an effect in POV-Ray... well, it's definitely an
ambitious idea (but never say never!).  The idea of smoothly
interpolating the liquid surface was what lead me to create the Liquid
Spray include file using blobs, and to my mind such a technique still
holds plenty of possibilities (particularly given that blobs are
perfectly suited to raytracing, even with thousands of components).
Collision detection is obviously a must: Ron's trace {} function is
again particularly suited to the raytracing environment, and I trust
(hope?!) that inclusion in the official POV-Ray is not far off.

This leaves the accurate (or at least convincing) simulation of surface
tension: perhaps adding friction to those particles in contact with
air/other objects (rather than other particles) would be feasible.  As
far as I know the actual physics of liquid turbulence are (as yet) far
from fully understood, but obviously realistic simulations (as evidenced

Post a reply to this message

Hey, Chris, I remember you mentioned the possibility of using the
SuperPatch's abilities to improve your #includes. Have you made any attempts
at doing this yet? It would be really great to see a trace()-enabled version
of your object exploder and particle system includes. =)

Post a reply to this message

(reference image posted in images.binaries under title Blue Liquid II)

That got me thinking.
Water in my experience seems to be made of more concave surfaces than
positive curved ones.  The blob surfaces that I have built, and they are
many, tend toward convex. Flowing blobs look like mercury. I have put a lot
of work in to build flat and concave ones, but there is usually an easier
way.

Looking at the flowing water image it struck me that there was an easier
logic in what was happening above the liquid than in the liquid itself.  For
example to model the liquid clinging to the wall with blobs would be nasty,
stacked up in a pyramid configuration, sliding off the wall, surface tension
pulling the border back.  But if the negative space was modeled then only
one element would give the same curve against the wall, a sphere, or a
collection of sphere's blobbed together.

So giving it a shot, the image is a first attempt to recreate the Arete
animation.  The base element is a box, differenced with a blob, consisting
randomly placed, randomly sized and randomly strengthed spherical elements,
the "pour" is a cylindrical blob element that is negative to the other
positive blob elements.   The whole thing is given a media content and
rendered.  No animation steps, rendering time is five hours on a pentium II
400.

Now, since the direction is to model the space above the blob, I have in my
head a mental image of a box filled with ping pong balls, into which is
poured a stream of large ball bearings.  The "ping pong balls" are the blob
elements they form essentially a solid of the top of the "liquid".  For the
sake of this first concept the "ball bearings", are not rendered, they serve
to displace the larger and convex "ping pong balls", or air molecules, only
the lower surface of the blob "cloud" is rendered, as it boolean's the box.

I had seen this approach somewhere and it took racking my brain to remember,
there is a small shareware program that does beautiful animations of
particles and I highly recommend downloading it and playing with it, it will
open your eyes.   The program is called "Really" and it is written by
Giovanni Tummarelli in Italy, the program is a little bit rough in places,
but it lets one design arrangements of "atoms" and their character and let
them go at each other the objects collide and shiver and break apart in real
time, nothing short of amazing.    Load the "World" called Diver2 and watch,
think of the application of this in a Povray animation with blobs. It is all
of 450K big, and here is the URL;
http://www.ascu.unian.it/~tummarel/really.shtml  (considering the trouble I
had to go through to find it again, I would humbly nominate it for inclusion
in Ken's wondrous page of links, if only to make it easier to find the next
time.)

So, in conclusion, to render the water, it may be simple to render the
invisible air.

Shaw

Post a reply to this message

Shaw wrote:
> 
> There is a company that does plug-ins for high end pricey renderers,
> there is an animation of a box being filled with a liquid that is beyond
> belief, for me anyway.  The website is www.areteis.com, click on the
> picture of the half filled blue liquid box and download the animation.
> Then let's figure out a way to do it......  Notice that just before the
> "water" hits the floor of the box there is a induced split in the liquid
> it seems, in order to make it flow around the box.  Anyone know of
> literature on how this would be attempted?
> 
> Shaw

I downloaded the MOV but it wouldn't play for some reason. I tried both
MediaPlayer and Quicktime, but nothing showed. I'm assuming it was terrific 8)
I did see the fire-animation they had. It seemed small and simple but _very_
realistic. Not like just a special effect, but a thorough simulation.
With the liquid I guess it is much more complicated. 
You should be able to do that in POV, as I'm convinced you can do anything in
POV (I mean it), as long as you can get the scene-file and have the occasional
patch.
The problem with this is that I think the calculations for such a liquid
animation would be so complicated, with lots of parameters, and you'd need so
many that it would probably take POV over a week to parse a frame (so to speak).
It would be rather inefficient to do this in POV; a program (in C or Pascal)
would do this much faster and wouldn't be more complicated to make (au contraire
- I think).

Remco

Post a reply to this message

Shaw <sha### [at] activech> wrote:
> That got me thinking.
> Water in my experience seems to be made of more concave surfaces than
> positive curved ones.  The blob surfaces that I have built, and they
are
> many, tend toward convex. Flowing blobs look like mercury. I have put
a lot
> of work in to build flat and concave ones, but there is usually an
easier
> way.
>
> Looking at the flowing water image it struck me that there was an
easier
> logic in what was happening above the liquid than in the liquid
itself.  For
> example to model the liquid clinging to the wall with blobs would be
nasty,
> stacked up in a pyramid configuration, sliding off the wall, surface
tension
> pulling the border back.  But if the negative space was modeled then
only
> one element would give the same curve against the wall, a sphere, or a
> collection of sphere's blobbed together.
>
> So giving it a shot, the image is a first attempt to recreate the
Arete
> animation.  The base element is a box, differenced with a blob,
consisting
> randomly placed, randomly sized and randomly strengthed spherical
elements,
> the "pour" is a cylindrical blob element that is negative to the other
> positive blob elements.   The whole thing is given a media content and
> rendered.  No animation steps, rendering time is five hours on a
pentium II
> 400.
>
> Now, since the direction is to model the space above the blob, I have
in my
> head a mental image of a box filled with ping pong balls, into which
is
> poured a stream of large ball bearings.  The "ping pong balls" are the
blob
> elements they form essentially a solid of the top of the "liquid".
For the
> sake of this first concept the "ball bearings", are not rendered, they
serve
> to displace the larger and convex "ping pong balls", or air molecules,
only
> the lower surface of the blob "cloud" is rendered, as it boolean's the
box.
>
> I had seen this approach somewhere and it took racking my brain to
remember,
> there is a small shareware program that does beautiful animations of
> particles and I highly recommend downloading it and playing with it,
it will
> open your eyes.   The program is called "Really" and it is written by
> Giovanni Tummarelli in Italy, the program is a little bit rough in
places,
> but it lets one design arrangements of "atoms" and their character and
let
> them go at each other the objects collide and shiver and break apart
in real
> time, nothing short of amazing.    Load the "World" called Diver2 and
watch,
> think of the application of this in a Povray animation with blobs. It
is all
> of 450K big, and here is the URL;
> http://www.ascu.unian.it/~tummarel/really.shtml  (considering the
trouble I
> had to go through to find it again, I would humbly nominate it for
inclusion
> in Ken's wondrous page of links, if only to make it easier to find the
next
> time.)
>
> So, in conclusion, to render the water, it may be simple to render the
> invisible air.

An interesting exercise in lateral thinking, that's for sure!  Talking
of concave blob surfaces, perhaps negative components could be used to
get more of the shape you're after?  Beyond that, maybe it's time to
look at isosurfaces...?  (And no, Tony, I've yet to get 'round to giving
the SuperPatch a proper workout... but if all the currently custom
features become official in POV-Ray 3.5 I won't have any excuses not
to!).

Post a reply to this message

I acheived something remotely similar.  I used isosurfaces. First I had a
general shape (this could be animated). Then I added some 3dnoise to it.  I
posted to p.b.a.  an animation which looks more like frothing magma, but I
think that with a little bit of tweaking you could get something that looks
appropriate.

I couldn't download the movie either.....

Shaw wrote:

> There is a company that does plug-ins for high end pricey renderers,
> there is an animation of a box being filled with a liquid that is beyond
> belief, for me anyway.  The website is www.areteis.com, click on the
> picture of the half filled blue liquid box and download the animation.
> Then let's figure out a way to do it......  Notice that just before the
> "water" hits the floor of the box there is a induced split in the liquid
> it seems, in order to make it flow around the box.  Anyone know of
> literature on how this would be attempted?
>
> Shaw

Post a reply to this message