[Sidefx-houdini-list] Smorganic

Ozgur Yılmaz eoyilmaz at gmail.com
Sun Apr 18 18:54:57 EDT 2010


again I was wrong about the usage of divergence and curl in fluid dynamics,
it seems that, divergence is a scalar field which is generated from a vector
field like velocity but it is not limited with velocity, it measures how
much the vectors are converging or diverging at any point, and curl is the
rotation of a vector field around a point and it is another vector field.

As the Navier-Strokes equations states (especially the incompressibility
equation) the divergence of the velocity field should be zero. So I've read
some articles where they are updating particle positions and velocities so
that the divergence be zero and the total volume of the fluid is preserved
while having a nice fluid. On the other hand, the Fusion CI guys are filling
new particles in high divergent points which leads a fluid with more volume.
But I think they don't care about it, because it is visually more
satisfactory.

an some good news, I was going to try to create nodes to let me calculate
the divergence, but hopefully there is a node among micro solvers called
"Gas Analysis" which does the divergence calculation among the other things.

So I'm on my way...

E.Ozgur Yilmaz
Lead Technical Director
www.ozgurfx.com


On Sun, Apr 18, 2010 at 8:23 PM, Ozgur Yılmaz <eoyilmaz at gmail.com> wrote:

> sorry I was wrong about vorticity. It is the divergence and curl of the
> velocity field that you use as parameters while modelling a fluid flow.
> Vorticity is another concept that happens in boundaries of viscous fluids
> when the fluid is in touch with a solid material.
>
>
> E.Ozgur Yilmaz
> Lead Technical Director
> www.ozgurfx.com
>
>
> On Sun, Apr 18, 2010 at 11:02 AM, Ozgur Yılmaz <eoyilmaz at gmail.com> wrote:
>
>> Pablo, if you can download the videos in their site, you can actually see
>> the thin layer of particles that they are talking about. So it doesn't seem
>> to me like that they are doing some trick in surfacing phase, but instead
>> they are filling new particles in specific places, specifically the places
>> where the holes are going to form.
>>
>> As they have explained in the interview, this places where a hole is going
>> to form actually can be found by calculating the divergences of the
>> particles. I haven't tested this idea though, so I'm sure there are some
>> other things that they keep as a secret.
>>
>> I've generated particles where the density gets lower a threshold, but as
>> the Fusion CI guys said there are sudden explosions when you do something
>> like that. So you also need to keep track of the inner pressure to prevent
>> explosions in the simulation.
>>
>> So the whole idea should be something like that, calculate each particles
>> divergence, and if the divergence goes below a threshold (may be negative
>> divergence is enough, I need to check it), then check the pressure at that
>> position and if the pressure goes also below a threshold emit a new particle
>> at this position, and may inherit the attributes from the neighbour
>> particles ( again I haven't tested anything). Though, these are bare ideas,
>> this Monday I'm going to look at these ideas, when I get back to work.
>>
>>
>> Cheers.
>>
>> E.Ozgur Yilmaz
>> Lead Technical Director
>> www.ozgurfx.com
>>
>>
>> 2010/4/18 Pablo Giménez <pablogipi at gmail.com>
>>
>> I remembered this article about Smorganic.
>>> After reading it and having done some similar work before in Houdini, the
>>> main thing the people from Fusion CI is doing is just slving the lack of
>>> deontrol in the RF mesher, basically doing stuff you can do in SOPs.
>>> Fluid solvers can get nice results at mid scales but not a so small
>>> scales,
>>> that´s truth, but you can get the overall motion sing SPH or grid
>>> solvers,
>>> the trick is the look. Their main achieve is to get a smooth thin layer
>>> of
>>> geometry from the particles created in RF. This is perfectly doable in
>>> SOP
>>> using particles created from DOPS or even RF. I remember doing simialr
>>> stuff
>>> some years ago in the H8 times using RF particles in Houdini to get a
>>> decent
>>> mesh.
>>> Then you have a plethora of tools like Particle surfacer, Smooth, Preak,
>>> magnet, Convert, etc ... to get a nice mesh in Houdini.
>>> Finally to get the finest and high frequency details you have
>>> displacement
>>> shaders on your arsenal.
>>> Voila Smorganic!
>>> I have seen lots of amazing sims from these guys but the stuff  around
>>> smorganinc is just marketing around some techniques to fix and massage
>>> meshes created from particles simulated using SPH. Stuf that Houdini TDs
>>> have been done for years.
>>> We have been living with Smorganics and we didn´t know, Oh my godness!
>>> My 2 cents.
>>>
>>> 2010/4/17 Ozgur Yılmaz <eoyilmaz at gmail.com>
>>>
>>> > first of all, David Johnson; you were write about being not specific,
>>> sorry
>>> > about that, as you correctly guessed the thing I want to achieve was to
>>> > have
>>> > a continuous thin sheet of water.
>>> >
>>> > Alex and Szymon, thank you for your links, those were very helpful...
>>> >
>>> > After reading the article about smorganic, I remembered that the fluid
>>> > motion is estimated as vorticities and divergences about a point, so
>>> the
>>> > motion of moving away or moving towards a point called the divergence,
>>> and
>>> > moving or spinning around a point is called vorticity. I had a book
>>> > (probably left at the studio) where it was explaining the whole idea
>>> behind
>>> > the fluid dynamics, I need to look at that book again (need to wait
>>> till
>>> > Monday :) ).
>>> >
>>> > As Szymon suggested, I need to add my own nodes to the particle fluid
>>> > solver, nodes that searches for low divergent areas and creates new
>>> > particles.
>>> >
>>> > So thank you guys again, you enlightened me.
>>> >
>>> > Cheers...
>>> >
>>> > E.Ozgur Yilmaz
>>> > Lead Technical Director
>>> > www.ozgurfx.com
>>> >
>>> >
>>> > On Sat, Apr 17, 2010 at 9:09 PM, Szymon Kapeniak
>>> > <szymon.kapeniak at gmail.com>wrote:
>>> >
>>> > > Go it. Thanks!:
>>> > >
>>> > >
>>> >
>>> http://www.cgw.com/Publications/CGW/2009/Volume-32-Issue-12-Dec-2009-/Viewpoint.aspx
>>> > >
>>> > > 2010/4/17 Alex Czetwertynski <alex at franktheplumber.com>:
>>> > > > If you can find it, get the CGW (Computer Graphics World) issue
>>> from
>>> > > > december 09 (the one with Avatar on the cover).  The Fusion CS
>>> people
>>> > > > wrote an article on smorganics and how it works within realflow
>>> > > >
>>> > > >
>>> > > > On Apr 17, 2010, at 7:12 AM, François Duchesneau wrote:
>>> > > >
>>> > > >> And with a Vop Pop you can have lots of control on the behavior of
>>> the
>>> > > >> global force and letting the sph make the particles stick
>>> together.
>>> > > >>
>>> > > >> François
>>> > > >>
>>> > > >> Szymon Kapeniak wrote:
>>> > > >>>>> I've looked at the orbit node but it seems like it orbits a
>>> bunch
>>> > > >>>>> of
>>> > > >>>>> particles around another particle, my problem is to fill new
>>> > > >>>>> particles
>>> > > >>>>> (probably to a position where the density is below a threshold)
>>> > > >>>>> into a
>>> > > >>>>> fluid
>>> > > >>>>> simulation...
>>> > > >>>>>
>>> > > >>>>> I'm quite new in DOPs/POPs in houdini (although I'm much more
>>> > > >>>>> experienced
>>> > > >>>>> with SOPs), so I don't know how to setup a system that adds new
>>> > > >>>>> particles
>>> > > >>>>> to
>>> > > >>>>> an ongoing simulation...
>>> > > >>>>>
>>> > > >>>>> Thank you...
>>> > > >>>>>
>>> > > >>>
>>> > > >>> How close you can come to Fusion demos this is another story of
>>> > > >>> course, but my suggestion would be to dive into sph solver in
>>> DOPs
>>> > > >>> and
>>> > > >>> look around. Basically it consists with a series of microsolvers
>>> > > >>> computing different qualities of fluids. You can turn on / turn
>>> off
>>> > > >>> any of them (open asset and start play with it) - I think only
>>> > > >>> density
>>> > > >>> is mandatory for other solvers to work properly. You can for
>>> example
>>> > > >>> turn off all external forces and pipe a pop solver instead to a
>>> > > >>> chain.
>>> > > >>> If you point it to a pop network (like the one created for you by
>>> a
>>> > > >>> shelf tool "Particle Fluid from Object", you can control fluids
>>> like
>>> > > >>> they were particles, but they look rather like computed by sph. I
>>> > > >>> suspect this is what is happening in Fusion demos inside RF.
>>> > > >>>
>>> > > >>>
>>> > > >>>> 2010/4/17 Alex Czetwertynski <alex at franktheplumber.com>:
>>> > > >>>> I was at a presentation at DMALA where the Fusion guy was
>>> showing
>>> > > >>>> this.  He made a big point about it being a Realflow solution
>>> only.
>>> > > >>>>
>>> > > >>>
>>> > > >>> Perhaps in terms of amount of particles which practically can be
>>> > > >>> computed, this can be done by RF only.
>>> > > >>>
>>> > > >>>
>>> > > >>>> I think you should have been more specific in what about the
>>> > > >>>> example you
>>> > > >>>> were looking to recreate. François' suggestion of the orbit pop
>>> > > >>>> would be for
>>> > > >>>> making the particles move in the vortex, but my guess is that
>>> > > >>>> you're trying
>>> > > >>>> to make the water stay in a nice thin sheet as it moves. That is
>>> a
>>> > > >>>> much
>>> > > >>>> harder task. Maybe you could contact the guy from Fusion CIS to
>>> > > >>>> see if you
>>> > > >>>> could license his solver and use it in DOPs. Just a thought.
>>> > > >>>>
>>> > > >>>>
>>> > > >>>
>>> > > >>> Aren't they using vanilla Realflow with custom scripting?
>>> > > >>> _______________________________________________
>>> > > >>> Sidefx-houdini-list mailing list
>>> > > >>> Sidefx-houdini-list at sidefx.com
>>> > > >>>
>>> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
>>> > > >>>
>>> > > >>>
>>> > > >> _______________________________________________
>>> > > >> Sidefx-houdini-list mailing list
>>> > > >> Sidefx-houdini-list at sidefx.com
>>> > > >> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
>>> > > >
>>> > > > _______________________________________________
>>> > > > Sidefx-houdini-list mailing list
>>> > > > Sidefx-houdini-list at sidefx.com
>>> > > > https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
>>> > > >
>>> > > _______________________________________________
>>> > > Sidefx-houdini-list mailing list
>>> > > Sidefx-houdini-list at sidefx.com
>>> > > https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
>>> > >
>>> > _______________________________________________
>>> > Sidefx-houdini-list mailing list
>>> > Sidefx-houdini-list at sidefx.com
>>> > https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
>>> >
>>>
>>>
>>>
>>> --
>>> Un saludo
>>> Best Regards
>>> Pablo Giménez
>>> _______________________________________________
>>> Sidefx-houdini-list mailing list
>>> Sidefx-houdini-list at sidefx.com
>>> https://lists.sidefx.com:443/mailman/listinfo/sidefx-houdini-list
>>>
>>
>>
>



More information about the Sidefx-houdini-list mailing list