POV-Ray : Newsgroups : povray.binaries.animations : Continued cloth experiments (quicktime ~450 kb) : Re: Continued cloth experiments (quicktime ~450 kb) Server Time
20 Jul 2024 07:15:56 EDT (-0400)
  Re: Continued cloth experiments (quicktime ~450 kb)  
From: Chaps
Date: 22 Nov 2001 12:14:39
Message: <3bfd327f$1@news.povray.org>
"Reusser" <reu### [at] chorusnet> wrote in message
news:221120010908005894%reu### [at] chorusnet...
> In article <3bfcf6f4@news.povray.org>, Chaps <cha### [at] yahoocom> wrote:
>
> > I am very interested in your experiments. Sinc I saw your first
animation I
> > have tried to do the same. I've got some results, but the parsing time
is
> > huuuuuuuuuuuuuge. how do you handle that?
>
> I've wondered the same.  The cloth patches for pov-ray seem to be far,
> far faster than I can manage to make mine.  I'm can't help but wonder
> exactly how they are speeding things up.  Most likely it's just a lot
> faster not in pov-ray code. (?)  I have simplified things as much as I
> can, so I don't think I can speed it up too much more.  All I can
> really to do help the parsing time is just do low quality versions with
> the high-quality versions going overnight or when I'm gone.
>
My idea, for this point is to use the same algorithm for any piece of cloth.
To do so I create an object that is very similar to a mesh2 sructure. The
first atempt create the cloth object using 4 corners, a number of rows, a
number of columns, a "depth" (a point behaviour depend on its direct
neighbours, or neighbours of neighbours...), weight and viscosity. What I
would like to do with patches, is to start from a more sophisticated shape
(a shirt or a dress for example).
It is affected today by gravity, and a list of still objects (torus,
sphere,cylinder and box). I'd like to add wind interaction, objects with
movement, and CGS.

> > On my side I've started to write an external utility in C++ that can
perform
> > the cloth animation and provide mesh2 objects at each clock steps.
However,
> > I've still many "open points" like synchronizing the rendering and my
> > utility program, input the cloth description from a union of bicubic
> > patches...
>
> For mine, I have just been using megapov's persistent variables to
> store variables from one frame to the next.  It's been working pretty
> well for me.  Is the C++ version siginificantly faster?

It is faster. But today I do not read the scene from any file, it is all
"hard coded" in the main program. Nevertheless, I've prepared all objects in
order to be abble to build a dynamic scene from an external simplified pov
file. Let's see when everything is completed.
>
>  - Rico


Post a reply to this message

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.