In this experiment I'm trying to use a real photograph as a background for
my cloth animation. I'm using a photogram from the Ministry of
Communications building in the historical center of Mexico City. Now it is
serves as an important museum. I took this photograph almost a year ago, and
yesterday I thought it would be a good idea if I added a simulated cloth as
a flag.

I think the wind should have been weaker because the cloth vibrates a bit
too much for my taste.

Technical specs:
* 49x49 mesh
* no cloth-cloth collision detection
* 10 hours of simulation (601 frames)
* 1:15 hours of rendering

I hope you liked it and I'll be glad to receive any comments. Unfortunately,
work has been building up lately, so I might not be able to play with cloths
as much as I would like to...

Fernando.

Post a reply to this message

Attachments:
Download 'mexflag09.mpg' (569 KB)

Looks nice.



> I think the wind should have been weaker because the cloth vibrates a bit
> too much for my taste.

Cloth in wind should have these small scale riples IMO. The ripples here look a
little odd in some way, though I can't seem to give a reason. The flag seems to
hang in one place for too long near the start, so maybe that is throwing me off.
Does your wind gust or stay constant?

> * 49x49 mesh

But the flag is not square. I wonder what the simulation would look like if four
nodes in the cloth defined a square section, instead of a rectangular one. This
would mean more nodes (say 49*98 for example) and a longer calculation time, but
I would be interested in seeing and comparing the results.

MJL
--
light_source{9*(y#macro A(H,B,R,T)prism{0,1H+4,0u*9,9v*9,0#local I=1;#while(I<H)
#local V=asc(substr(B,I,1))-33;<div(V,10)mod(V,10)>#local I=I+1;#end pigment{rgb
y}rotate-<90,R>translate-T}#end-z)rgb 9}A(14",6;MWmhryXN3,"60<15,0,8>)box{-99(x+
z)*99pigment{rgb 1}}fog{distance 11}camera{location(y-z)*25look_at 0}A(8"6hiAG=6
"-60x*-10)A(16"/.@VZno=<PLA89/"0x*5) // MJL

Post a reply to this message

I think the wind is too weak. You also need to make the air "thicker". More
oily. I don't know if I'm explaining myself...

Post a reply to this message

"Tony[B]" wrote:
> 
> I think the wind is too weak. You also need to make the air "thicker". More
> oily. I don't know if I'm explaining myself...

Close. Actually I think the flags movement is a bit to fluid. Outdoor
flags are usually constructed of rather thick material to keep them
from tearing apart in the wind. The movement of this flag makes it
look like it is paper thin and made of silk. It should act like it
is made of thicker material.

-- 
Ken Tyler

Post a reply to this message

That's exactly what I meant. Ken-AI is working well.

Post a reply to this message

I know what you folks mean. To make the cloth stiffer/thicker the amount of
spring connections between the cloth particles should be increased.
Unfortunately more springs connections will take more time to simulate.
I was lucky today and managed to get my totally renewed clothing code
working. My previous versions were able to simulate stiff clothing as you
can see in my previously posted animations.

The new version will be able to accept more detailed cloth-descriptions, so
the user will be able to describe stiff clothes or clothes that are
partially stiff or clothes that tear or shrink.
I'm still figuring out a way to handle friction forces in a both realistic
and fast way. Especially the cloth-cloth friction forces are driving me
nuts.

Post a reply to this message

This is really exciting! It seems your physical model is able to realize
some kind of turbulent behaviour!

Air is viscous, which imply non-linear equations for the flow. Thus, with
many degrees of freedom (here, we're talking thousands!) turbulence and
chaotic behaviour could show.

Though without knowing the details of your physical model, I'd say that the
air flow seems a bit _too_ laminar, and that wind speed is too slow. This
shows in the slow waving of the flag. (Then again, speeding up would
probably make the wind look really strong --- more like a storm!)

But I'm no expert... Keep up your very clever experiments. We really enjoy
it here. :P

Oh, and just a suggestion: To analyze the behaviour of the model, why not
increase the resolution/tone down the background.

Regards,
Simen Kvaal.


news:3c745116@news.povray.org...

>
> Technical specs:
> * 49x49 mesh
> * no cloth-cloth collision detection
> * 10 hours of simulation (601 frames)
> * 1:15 hours of rendering
>
> I hope you liked it and I'll be glad to receive any comments.
Unfortunately,
> work has been building up lately, so I might not be able to play with
cloths
> as much as I would like to...
>

Post a reply to this message

Sorry for answering a little bit too late, but I've been VERY busy lately. I
don't think I'll have too much free time in the following days (weeks?), but
I thank everyone who gave me very valuable comments and suggestions...

Thanks again,

Fernando.

Post a reply to this message