[Bf-committers] Some Maths / ideas / solutions for Softbodies.

[frédéric van der essen]

hello, thanks for the precisions. :)

But i have some objections ;)
In every paper that i could found on softtbodies/cloth it was written 
clearly
"Do not use explicit Euler"
"Use adaptive method or Implicit method"
Instinctive softbodies use explicit euler, and if you take a too thin 
mesh, it explodes.
Mabye cloth is not the guideline but people will certainly use it, and 
it will crash with forward euler.
Users will think it's a bug, and not a mathematical limitation. It is 
not controlable by the imput values since it depends on the mesh 
resolution.
Because users will want bendable but not elastic cloth, they will set a 
very low elasticity that will lead to stiff problem. Stiff problem leads 
to a really small step, and then lot of calculation.
Even if it seems fast, it will quicly become a slow method with stiff 
problems.
As in everywhere in pixar's paper i see 'do not use forword euler' i 
really doubt that maya does it.
Cloth and softbodies are the same algorithms, so i think we should adapt 
them to fit the most difficult case .
that was my point view :)

-Fred


bjornmose at gmx.net wrote:

>Hmm,  
>
>I'd  like to point to that fact, that i never talked about cloth simulation
>to be the guiding idea behind this approach. 
>
>When you look at the Maya docs there are two features we first want to have.
>1. Goals : The soft object is hanging elastic in its original static cage
>2. General vertex to vertex (object to object) springs to modify elastic
>properties of the object(s).
>
>If you look closer at the Maya docs and the parameters used for setting such
>systems up, may lead you to the suspect that Maya also is integrating the
>system of ODEs, most likely using brute force forward euler, since for 
>reasonalbe well behaving systems the calculation of the jakobian and setting
>up a sparse matrix solver would be numerical overkill for a little yelo
>wobble.
>
>so much for that
>/ole
>
>  
>
>>
>>
>>
>>
>>_______________________________________________
>>Bf-committers mailing list
>>Bf-committers at projects.blender.org
>>http://projects.blender.org/mailman/listinfo/bf-committers
>>
>>    
>>
>
>  
>