Re: Re[2]: cloth rigging - how to?

["Tim Leydecker" <BauerOink(at)gmx.de>]

P.S: I don´t understand that there is so much buzz about
the various (indeed useful) pointcache implemenations
in all the latest 3D apps but nothing more than basic
ways to edit such data or animate ontop of it easily,
time-remap instead of just scale, re-orient, stuff to
add or remove worldspace "travel" to get a "local"
effect in relation to a basepose or series of blendshapes
that would allow to convert a pointcache into a series
of blendshapes or mixer clips assigned to a mesh that is
then "static" but still has the relative animation attached,
ready to be dragged and dropped on another mesh.

stuff like that.


Simple example, you solve some sort of splash effect
of a last drop into a cup of tea, then extend that to blend
nicely ontop of a huge bowl of soup using clips with
near realtime feedback, painting in the blends and
getting it to still look convincing...then animate over it again.
Without having to script anything...

Cheers

tim


----- Original Message ----- 
From: "Frank Lenhard" <franky(at)ixdream.com>
To: "Tim Leydecker" <XSI(at)Softimage.COM>
Sent: Thursday, September 27, 2007 11:31 AM
Subject: Re[2]: cloth rigging - how to?


> where is you "PS:" section tim...  :)
>
> ciao
> franky
>
>
> Thursday, September 27, 2007, 10:07:30 AM, you wrote:
>
> > > Hey Tim,
> > > wasnt wondering how to attach deformers to a mesh - thats easy enough.
> > > its rather how to have them attached to the deformed mesh, and have them
> > > deform the mesh in turn, without doubling the deformation.
>
> TL> The hands on approach I tried to outline does that,
> TL> it´s not a good idea to plot a point of the deformed
> TL> mesh and parent the deformer (like a lattice) to that
> TL> input (copied to a Null) if you plan to change the source
> TL> input anim again but it´s good enough if all you want is
> TL> to change something like the general rotation of a
> TL> mesh ontop of a sim in a hurry, like in "now".
>
> TL> I´ll also look into Steve´s article and scops today,
> TL> having the mixer, corrective shapes and working
> TL> out with pointcaches on my plate anyway in general.
>
> TL> Cheers
>
> TL> tim
>
>
> TL> ----- Original Message ----- 
> TL> From: "peter boeykens" <peter_b(at)skynet.be>
> TL> To: <XSI(at)Softimage.COM>
> TL> Sent: Tuesday, September 25, 2007 10:05 PM
> TL> Subject: Re: cloth rigging - how to?
>
>
> TL> I think if you're a smart cookie you can get straight to the
> TL> positions of points over time, using the object model.
>
> TL> But each time I needed this its fast enough making a constraint 
> manually or
> TL> even typing a few commands to set it up "en masse".
>
>
> TL> Hey Tim,
> TL> wasnt wondering how to attach deformers to a mesh - thats easy enough.
> TL> its rather how to have them attached to the deformed mesh, and have 
> them
> TL> deform the mesh in turn, without doubling the deformation.
>
> TL> looks like expressions on the static kinestate is going to be the way.
>
>
> TL>   ----- Original Message ----- 
> TL>   From: Steven Caron
> TL>   To: XSI(at)Softimage.COM
> TL>   Sent: Tuesday, September 25, 2007 9:14 PM
> TL>   Subject: Re: cloth rigging - how to?
>
>
> TL>   "No constraints required, no cycles, quite fast."
>
> TL>   without constraints how are you.. "copying translation values" ? and 
> how
> TL> are you "hanging" the deformer on to the null? expressions maybe?
>
> TL>   steven
>
>
> TL>   On 9/25/07, Tim Leydecker < BauerOink(at)gmx.de> wrote:
> TL>     A hands on approach to get a deformer to follow an
> TL>     allready animated - or worse - a vertice baked mesh
> TL>     is to copy the translation values of one vertice/point
> TL>     onto a Locator/Null object (after snapping it to that
> TL>     vertice) then "hang" the deformer onto that Null.
>
> TL>     It´ll travel with your simulated mesh (until you change
> TL>     the sim obviously) and by that allows "local" deformations
> TL>     regardless of stackorder or deformation order.
>
> TL>     No constraints required, no cycles, quite fast.
>
> TL>     Good enough to let´s say, rotate some sim´ed
> TL>     mesh with clean values, might even be good
> TL>     enough to substract "offset" from "origin" to
> TL>     get a "local" effect. Sorry for all those ".",
> TL>     can´t put it better, I´m still dizzy from brooding
> TL>     over the above the last few, long days. done.
>
> TL>     Cheers
>
> TL>     tim
>
> TL>     P.S: When you guys test Moondust, make sure
> TL>     animation plotted to points is _easily editable and
> TL>     layerable (time remap, SRT ontop, re-useable).
>
>
> TL>     ----- Original Message -----
> TL>     From: "Steven Caron" <carons(at)gmail.com>
> TL>     To: < XSI(at)Softimage.COM>
> TL>     Sent: Tuesday, September 25, 2007 8:24 PM
> TL>     Subject: Re: cloth rigging - how to?
>
>
> TL>     > using the static kine state is stable in my experience...here is 
> how
> TL> we
> TL>     > have
> TL>     > been using it here at blur.
> TL>     >
> TL>     > http://www.xsi-blog.com/archives/209
> TL>     >
> TL>     > steven
> TL>     >
> TL>     > On 9/25/07, peter boeykens <peter_b(at)skynet.be > wrote:
> TL>     >>
> TL>     >> Hi Ben,
> TL>     >> will see if I can get the driven static kinestate to work - the 
> way I
> TL>     >> understand you, this should be indeed what I'm after. Hope XSI 
> agrees
> TL> :-)
> TL>     >>
> TL>     >> Speaking of cloth sims - from what I understand syflex has such 
> a
> TL> thing
> TL>     >> as
> TL>     >> a
> TL>     >> target pose (not initial state!) - the shape a cloth would 
> assume if
> TL> not
> TL>     >> under influence of any forces. But not in the XSI 
> implementation :-(
> TL>     >> Anyone can confirm this?
> TL>     >>
> TL>     >> > You have to constrain two things to the cloth:
> TL>     >> > 1) Your deformer
> TL>     >> > 2) A "base" null that represents the "undeformed" position. 
> When
> TL> the
> TL>     >> > deformer hasn't been moved these two things follow the cloth 
> in the
> TL>     >> > exact same position.
> TL>     >> >
> TL>     >> > Then drive the static kinestate of your deformer with the 
> base
> TL> null.
> TL>     >> > This will make the deformations relative to the base null 
> rather
> TL> than
> TL>     >> > the predeformed position when the operator was applied, which 
> is
> TL> what
> TL>     >> > static kinestate is by default. This is analogous to bind 
> pose in
> TL>     >> > Maya.
> TL>     >> >
> TL>     >> > Also I've had luck with copying the cloth mesh and rigging 
> it, then
> TL>     >> > using a simple scripted op to do a "blend" between them. You 
> can
> TL> even
> TL>     >> > get really fancy and read a weight map if you like.
> TL>     >> >
> TL>     >> > for each point in position array:
> TL>     >> > alpha = weightmap * slider value
> TL>     >> > lerp(mesh1.point mesh2.point, alpha)
> TL>     >>
> TL>     >> what's "lerp"?
> TL>     >>
> TL>     >> and how can I access a point's weightmap value?
> TL>     >> (its probably RTFM - will dig into the scripting reference 
> tomorrow,
> TL>     >> sounds
> TL>     >> like it can come in handy.)
> TL>     >>
> TL>     >>
> TL>     >>
> TL>     >> >
> TL>     >> >
> TL>     >> > On 9/20/07, peter boeykens <peter_b(at)skynet.be > wrote:
> TL>     >> >> trying to find a good workflow for rigging cloth.
> TL>     >> >>
> TL>     >> >> at the moment I am cage deforming the highres cloth geometry 
> with
> TL> the
> TL>     >> >> syflex
> TL>     >> >> simulated cloth. Its all good.
> TL>     >> >> now Im trying to layer deformations on top of the 
> simulation.
> TL>     >> >>
> TL>     >> >> What I would like to get now, is a layer of (few) 
> manipulators,
> TL> driven
> TL>     >> by
> TL>     >> >> the simulated cloth.
> TL>     >> >> I guess I'll cluster constrain nulls to the cloth, and then 
> parent
> TL> my
> TL>     >> >> manipulators under those nulls - but Im open for other
> TL> suggestions.
> TL>     >> With
> TL>     >> >> those manipulators I'd like to "sculpt" on top of the 
> simulation.
> TL>     >> >>
> TL>     >> >> I do not want to simply envelope the highres to the 
> manipulators,
> TL>     >> >> since
> TL>     >> I
> TL>     >> >> wont get the precision I'm getting from the cage. But if I'm
> TL>     >> >> enveloping
> TL>     >> >> the
> TL>     >> >> highres to the manipulators on top of the cage I'l get a 
> "double"
> TL>     >> >> deformation.
> TL>     >> >>
> TL>     >> >> It works fine if the layer of my manipulators is not 
> deformed by
> TL> the
> TL>     >> >> cloth,
> TL>     >> >> sculpting locally around the default pose of the cloth, but 
> its
> TL> not
> TL>     >> >> intuitive at all working that way.
> TL>     >> >>
> TL>     >> >> Any suggestions how to approach this?
> TL>     >> >>
> TL>     >> >> ---
> TL>     >> >> Unsubscribe? Mail Majordomo(at)Softimage.COM with the following 
> text
> TL> in
> TL>     >> >> body:
> TL>     >> >> unsubscribe xsi
> TL>     >> >>
> TL>     >> >
> TL>     >> >
> TL>     >> > -- 
> TL>     >> > -Ben Barker
> TL>     >> > ---
> TL>     >> > Unsubscribe? Mail Majordomo(at)Softimage.COM with the following 
> text
> TL> in
> TL>     >> body:
> TL>     >> > unsubscribe xsi
> TL>     >>
> TL>     >> ---
> TL>     >> Unsubscribe? Mail Majordomo(at)Softimage.COM with the following 
> text in
> TL>     >> body:
> TL>     >> unsubscribe xsi
> TL>     >>
> TL>     >
>
>
> TL>     ---
> TL>     Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text 
> in
> TL> body:
> TL>     unsubscribe xsi
>
>
>
>
> TL> ---
> TL> Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in 
> body:
> TL> unsubscribe xsi
>
>
>
> ---
> Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in body:
> unsubscribe xsi 


---
Unsubscribe? Mail Majordomo(at)Softimage.COM with the following text in body:
unsubscribe xsi