Wow! Although this exhibits slightly the unfortunate 'rubbery' look of most CG cloths
I've
seen, I actually forgot for a moment that I was looking at CG cloth, and was just
enjoying the
way it was undulating and slipping along the floor.

Do you know what's causing the rubberiness, and do you think you'll be able to
overcome it? I
think you have almost achieved realistic cloth.

Is this in Pov SDL? I hope you will be sharing your code/software at some stage.

All the best,

Andy Cocker


news:3dd2f8e3@news.povray.org...
> In this time, I used a semi-fine mesh (33x45 cells)
>
> - timestep size : 1e-4 seconds per step
> - 600 frames, about 300,000 steps, that's about 500 steps per frame
> - simulation time : about 1 hour (about 6 seconds per frame)
> - render time:  7.5 hours
> - machine used Pentium4 1.8ghz, 256Mb
> - no cloth-cloth collision detection
>
> I used an area light and a spotlight to enhance the appearance of the
> creases in the cloth. A very slight constant wind, I can't perceive it
> though.
>
> Not an awfully complex animation, but I wanted to know how the cloth behaves
> when compressed and moved around.
>
> I hope you like it and I'll be happy to receive any comments/suggestions.
>
> Fernando.
>
>
>

Post a reply to this message

An excellent animation- very impressive.
    The cloth does look a bit dense or massive; perhaps reducing your inertia
somewhat would help with that, if it is possible.  The appearance is more
reminiscent of heavy draperies than of a layer of thinner fabric.  My only
"flaw" observed is that the bottom fold of the fabric should fold over itself
and move backwards perhaps, given its density.  However, in a lighter fabric, I
have seen this exact behavior.

Cheers!

Chip Shults
My robotics, space and CGI web page - http://home.cfl.rr.com/aichip

Post a reply to this message

> 
> [...]
> 
> I hope you like it and I'll be happy to receive any comments/suggestions.
> 

Very nice.  Some more information about the used parameters would be
great:

- how strong bending stiffness did you use compared to the linear
stiffness
- how is the cloth attached at the top (it looks like at various points
which move in both horizontal directions during the animation).

I'm not totally sure but i think even on a polished ground the cloth would
'turn around' where it touches the ground when the movement reverses due
to friction. Right now it always slides.

Christoph

-- 
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 02 Nov. 2002 _____./\/^>_*_<^\/\.______

Post a reply to this message

My mom doesn't trust your animation: she could see you hiding behind that
curtain.  :)


INVALID_ADDRESS wrote:
>In this time, I used a semi-fine mesh (33x45 cells)
>
>- timestep size : 1e-4 seconds per step
>- 600 frames, about 300,000 steps, that's about 500 steps per frame
>- simulation time : about 1 hour (about 6 seconds per frame)
>- render time:  7.5 hours
>- machine used Pentium4 1.8ghz, 256Mb
>- no cloth-cloth collision detection
>
>I used an area light and a spotlight to enhance the appearance of the
>creases in the cloth. A very slight constant wind, I can't perceive it
>though.
>
>Not an awfully complex animation, but I wanted to know how the cloth behaves
>when compressed and moved around.
>
>I hope you like it and I'll be happy to receive any comments/suggestions.
>
>Fernando.
>

Post a reply to this message

Hi Andrew,

Yes, I think I was able to reduce the rubbery look by setting the "spring
tension" quite high and using a 4th order ODE solver. It is supposed to get
better when one uses stiffer springs, but it also causes numerical
instability.

I'm glad you think it looks realistic. Thanks!

The simulation is coded in C, compiled with Intel C++ Compiler (whoa, it
makes FAST code) and rendered with POV 3.5

I made the code quite flexible to accept many shapes of cloth, but
obviously, the rectangle is the easiest to implement.

Thanks again,

Fernando.



news:3dd2ff25@news.povray.org...
> Wow! Although this exhibits slightly the unfortunate 'rubbery' look of
most CG cloths I've
> seen, I actually forgot for a moment that I was looking at CG cloth, and
was just enjoying the
> way it was undulating and slipping along the floor.
>
> Do you know what's causing the rubberiness, and do you think you'll be
able to overcome it? I
> think you have almost achieved realistic cloth.
>
> Is this in Pov SDL? I hope you will be sharing your code/software at some
stage.
>
> All the best,
>
> Andy Cocker
>

> news:3dd2f8e3@news.povray.org...
> > In this time, I used a semi-fine mesh (33x45 cells)
> >
> > - timestep size : 1e-4 seconds per step
> > - 600 frames, about 300,000 steps, that's about 500 steps per frame
> > - simulation time : about 1 hour (about 6 seconds per frame)
> > - render time:  7.5 hours
> > - machine used Pentium4 1.8ghz, 256Mb
> > - no cloth-cloth collision detection
> >
> > I used an area light and a spotlight to enhance the appearance of the
> > creases in the cloth. A very slight constant wind, I can't perceive it
> > though.
> >
> > Not an awfully complex animation, but I wanted to know how the cloth
behaves
> > when compressed and moved around.
> >
> > I hope you like it and I'll be happy to receive any
comments/suggestions.
> >
> > Fernando.
> >
> >
> >
>
>

Post a reply to this message

Ha ha ha... Mmmm... I thought I was very well concealed :)

Take care,

Fernando.



news:web.3dd3ac588d75d2c6863e7dd0@news.povray.org...
> My mom doesn't trust your animation: she could see you hiding behind that
> curtain.  :)
>
>
> INVALID_ADDRESS wrote:
> >In this time, I used a semi-fine mesh (33x45 cells)
> >
> >- timestep size : 1e-4 seconds per step
> >- 600 frames, about 300,000 steps, that's about 500 steps per frame
> >- simulation time : about 1 hour (about 6 seconds per frame)
> >- render time:  7.5 hours
> >- machine used Pentium4 1.8ghz, 256Mb
> >- no cloth-cloth collision detection
> >
> >I used an area light and a spotlight to enhance the appearance of the
> >creases in the cloth. A very slight constant wind, I can't perceive it
> >though.
> >
> >Not an awfully complex animation, but I wanted to know how the cloth
behaves
> >when compressed and moved around.
> >
> >I hope you like it and I'll be happy to receive any comments/suggestions.
> >
> >Fernando.
> >
>
>

Post a reply to this message

Thanks!

Yes, I think it should fold too... Perhaps I put too little friction to the
floor, that's why the cloth "prefers" to glide instead of fold over itself.

Really the cloth is model as an infinitely thin mesh, so I'm not surprised
it gives that impression... I remember that Apache was already thinking in
implementing an algorithm that really modelled cloths as thick objects. I
hope he'll find the time to do that and impress us all.

Best wishes,

Fernando.


news:3dd31811$1@news.povray.org...
>     An excellent animation- very impressive.
>     The cloth does look a bit dense or massive; perhaps reducing your
inertia
> somewhat would help with that, if it is possible.  The appearance is more
> reminiscent of heavy draperies than of a layer of thinner fabric.  My only
> "flaw" observed is that the bottom fold of the fabric should fold over
itself
> and move backwards perhaps, given its density.  However, in a lighter
fabric, I
> have seen this exact behavior.
>
> Cheers!
>
> Chip Shults
> My robotics, space and CGI web page - http://home.cfl.rr.com/aichip
>
>
>

Post a reply to this message

One of the "innovations" in this model was that I allowed the nodes in the
cloth to be attached to other nodes with variable strength.

In this particular example each node is attached to its "horizontal" and
"vertical" neighbor with a very strong force 1e5, that is, for example, if
we have a mesh with points x(i,j) then x(i,j) will be attached to x(i-1,j),
x(i+1,j), x(i,j-1), x(i,j+2). This is because cloths resist very much to
stretching.

They also resist (not so much) to shearing and bending, so that's why I also
connected them diagonally, but in this way: x(i,j) is connected to
x(i-2,j-2), x(i-2,j+2), x(i+2,j-2), x(i+2,j+2). The force associated to this
connections was weaker.

I think this helped to improve the appearance of the cloth.


affected to any force, I fix their positions arbitrarily. In this case I
force them to be where I want, that's why we see the formation of the
creases.

I'm glad you liked it and I hope I was able to answer your questions more or
less satisfactorily (it's 1:29am!)

Take care,

Fernando.



news:3DD### [at] gmxde...
>
>

> >
> > [...]
> >
> > I hope you like it and I'll be happy to receive any
comments/suggestions.
> >
>
> Very nice.  Some more information about the used parameters would be
> great:
>
> - how strong bending stiffness did you use compared to the linear
> stiffness
> - how is the cloth attached at the top (it looks like at various points
> which move in both horizontal directions during the animation).
>
> I'm not totally sure but i think even on a polished ground the cloth would
> 'turn around' where it touches the ground when the movement reverses due
> to friction. Right now it always slides.
>
> Christoph
>
> --
> POV-Ray tutorials, include files, Sim-POV,
> HCR-Edit and more: http://www.tu-bs.de/~y0013390/
> Last updated 02 Nov. 2002 _____./\/^>_*_<^\/\.______

Post a reply to this message

> 
> One of the "innovations" in this model was that I allowed the nodes in the
> cloth to be attached to other nodes with variable strength.
> 
> In this particular example each node is attached to its "horizontal" and
> "vertical" neighbor with a very strong force 1e5, that is, for example, if
> we have a mesh with points x(i,j) then x(i,j) will be attached to x(i-1,j),
> x(i+1,j), x(i,j-1), x(i,j+2). This is because cloths resist very much to
> stretching.
> 
> They also resist (not so much) to shearing and bending, so that's why I also
> connected them diagonally, but in this way: x(i,j) is connected to
> x(i-2,j-2), x(i-2,j+2), x(i+2,j-2), x(i+2,j+2). The force associated to this
> connections was weaker.

Thanks, that makes it quite clear.  Sounds logical to use the 'second
order' diagonals to introduce both bending and shearing stiffness with
very few springs.  Sim-POV BTW allows to specifically weight the
individual types of connection with in the standard patch macros:

http://www-public.tu-bs.de:8080/~y0013390/simpov/docu02.html#MechSim_Generate_Patch


> affected to any force, I fix their positions arbitrarily. In this case I
> force them to be where I want, that's why we see the formation of the
> creases.

It looks very good.  Often such 'fixed movements' look somewhat artificial
but in this case not.  

Christoph

-- 
POV-Ray tutorials, include files, Sim-POV,
HCR-Edit and more: http://www.tu-bs.de/~y0013390/
Last updated 15 Nov. 2002 _____./\/^>_*_<^\/\.______

Post a reply to this message

WARNING! WARNING! EXTREME COOLNESS DETECTED!!!

Nice work! A bit graney at this resolution, but still very real-looking.
Good it up!

Andrew.

PS. And what about that floor texture? That's just plain _lush_

Post a reply to this message