Hello,

I post a new water animation. As in the previous post, it is an object
falling in a tank of water. I tried to use a more complex mesh for the
object. The passage from a sphere to a coffe cup gives me a lot of problem
!

For the rendering, I tried the blur effects. I also changed the peusdo hdri
background.

The video is mpeg1 encoded.

The simulation time was around 12 hours and the rendering around 24 hours. I
use a P4 3.2 GHz with 1 Go of RAM.

Any comment is wellcome.

Regards,
Fidos.

Post a reply to this message

Attachments:
Download 'coffeecup.mpg' (629 KB)

Very nice !
I really like the focusing of the light on the table, near the end of the
animation.

Since the blur sampling is random in each frame, it looks a bit strange, but
i think there is no way to avoid this. (apart from using very high blur
quality settings)

maybe it would be nice to just make multiple copies if the first and last
frame, and add them at the start end end of the animation, to make it
easier to see the first part of the animation.
maybe it's possible possible to render some extra frames that show a falling
cup, that ends up right where your simulation starts. the rendering of
those frames should be fast, since you can just model the water surface as
a flat plane.
Or maybe do it in your simulator: a simple (just falling object) simulation
until some object-water contact is detected ?

Jaap.

Post a reply to this message

fidos wrote:

> Any comment is wellcome.

Is there a particular reason the volume of the liquid *decreases* after 
the mug falls in?

Regards,
John

Post a reply to this message

The water seems a little over viscous (too much damping?)

Otherwise... looks really quite impressive, if a little brief. (But with 
a render time like that, I can well understand why it's this short!)

Post a reply to this message

John VanSickle <evi### [at] hotmailcom> wrote:
> Is there a particular reason the volume of the liquid *decreases* after
> the mug falls in?

anti-matter mug ? :-)

Post a reply to this message

"Jaap" <jws### [at] yahoocom> wrote:
> Very nice !
> I really like the focusing of the light on the table, near the end of the
> animation.
>
> Since the blur sampling is random in each frame, it looks a bit strange, but
> i think there is no way to avoid this. (apart from using very high blur
> quality settings)
I use radiosity with very low count. I think it also contribute to introduce
more change from frame to frame. I think it was a mistake to use radiosity
for that type of scene.
>
> maybe it would be nice to just make multiple copies if the first and last
> frame, and add them at the start end end of the animation, to make it
> easier to see the first part of the animation.
> maybe it's possible possible to render some extra frames that show a falling
> cup, that ends up right where your simulation starts. the rendering of
> those frames should be fast, since you can just model the water surface as
> a flat plane.
> Or maybe do it in your simulator: a simple (just falling object) simulation
> until some object-water contact is detected ?
In the previous animation, I animated the object alone in MegaPov up to the
time the rendering reach the start time of the simulation. During the fly
of the object, I use the initial water object for the redering.
>
> Jaap.

Thanks for your comments.

Fidos

Post a reply to this message

John VanSickle <evi### [at] hotmailcom> wrote:
> fidos wrote:
>
> > Any comment is wellcome.
>
> Is there a particular reason the volume of the liquid *decreases* after
> the mug falls in?
>
> Regards,
> John

Very good remarks. It's a problem that I have from the start but with the
introduce of the object <-> water interaction, it seems to increase. I
suppose that some water enter a little bit in the object at each iteration
(this animation has 341 simulation iterations for 35 rendered frames).
Because I cut the water volume with the object one at each iteration (csg
of the isosurface), I loose some little mass of water at each iteration.
With the accumulation, the water lost finish to be important. The solution
should be a rewrite of the object <-> water interaction.

Thank you for your comment.

Regards,
Fidos

Post a reply to this message

Love these simulations.
An observation. Refractive index of water is ~ 1.335 it looks to be set to
air/vacuum ~1.0 as the side panels of the tank don't change as the water
level changes. The tank wall seem to be plausible though.

DLM

"fidos" <fid### [at] wanadoofr> wrote in message
news:web.4373b49739469c3d7ad89d270@news.povray.org...
> Hello,
>
> I post a new water animation. As in the previous post, it is an object
> falling in a tank of water. I tried to use a more complex mesh for the
> object. The passage from a sphere to a coffe cup gives me a lot of problem
> !
>
> For the rendering, I tried the blur effects. I also changed the peusdo
hdri
> background.
>
> The video is mpeg1 encoded.
>
> The simulation time was around 12 hours and the rendering around 24 hours.
I
> use a P4 3.2 GHz with 1 Go of RAM.
>
> Any comment is wellcome.
>
> Regards,
> Fidos.
>

Post a reply to this message

"fidos" <fid### [at] wanadoofr> wrote:
> suppose that some water enter a little bit in the object at each iteration

Can you calculate the volume of the water during the simulation? Then you
can just add the "lost" waster to the bottom of the tank after each
iteration.
There may be a mistake like intersecting the object with the water of the
next frame, since you seem to loose quite a lot of water.

Speaking of water simulations, have you seen:
http://www.richardrosenman.com/plumber.htm  (I liked it a lot :-)

Post a reply to this message

"Jaap" <jws### [at] yahoocom> wrote:
> "fidos" <fid### [at] wanadoofr> wrote:
> > suppose that some water enter a little bit in the object at each iteration
>
> Can you calculate the volume of the water during the simulation? Then you
> can just add the "lost" waster to the bottom of the tank after each
> iteration.
I had a similar idea : to add a source term in the bulk of the water body in
order to create water volume in order to compensate the lost. It should be
very easy to do.
> There may be a mistake like intersecting the object with the water of the
> next frame, since you seem to loose quite a lot of water.
I just found a problem of offset between the mesh of the object and the
isosurface representing it ...
>
> Speaking of water simulations, have you seen:
> http://www.richardrosenman.com/plumber.htm  (I liked it a lot :-)
Amazing. Beautifull and funny. They used the realflow tool of the NextLimit
company. It's an impressive tool
(http://www.nextlimit.com/realflow/index.html and see the Galery). It uses
particules and gives more freedom than in my "small" simulation (I use an
isosurface representing the water body).

Regards,
Fidos

Post a reply to this message