Re: Rigids bodies and intercollision.

[Sylvain Moreau <sylvain.moreau(at)videotron.ca>]

In 10 years I guess most 3D apps will have a "solid geometry" mode 
simulating the real world where objects don't go through each other too 
easily. At that point robot joints rotations will be limited according to 
RBD sims and robot arms will stop intersecting with their head if you don't 
watch them.
Hierarchical propagation of RBD properties will eventually be quite 
important, but for now, I agree with Kim, parenting is the most basic 
organisation tool. If I put my RBs under a null I expect properties 
propagation from the null.

Imagine you spent hours tweaking a sim than you need to rotate the whole 
result to match your scene. What can you do now?

syl


----- Original Message ----- 
From: "kim aldis" <kim(at)cg-soup.com>
To: <XSI(at)Softimage.COM>
Sent: Tuesday, November 01, 2005 4:29 PM
Subject: RE: Rigids bodies and intercollision.


> I'm not so bothered about parenting RBDs to each other but I am bothered
> about the loss of parenting as both an organisation tool and a positioning
> tool. The thought of arbitrary objects hanging off the root isn't a 
> pleasant
> one. It's something I continually give people a hard time about and the
> thought of dozens, or even hundreds at this level is horrifying. Parenting
> things up for ease of selection, collapsing to keep the explorer tidy,
> inheritance of properties, etc, is so fundamental to everday workflow that 
> I
> was a bit dumbstruck when I first heard about this.
>
> Ideally, I guess, I can't see a need to have RBDs parented to RBDs but I
> would like to be able to parent them to a null and have them move when I
> move the null, just like any other animated object.
>
> > -----Original Message-----
> > From: owner-xsi(at)Softimage.COM
> > [mailto:owner-xsi(at)Softimage.COM] On Behalf Of Jeffrey Smith
> > Sent: 01-November-2005 20:53
> > To: XSI(at)Softimage.COM
> > Subject: RE: Rigids bodies and intercollision.
> >
> > So, in the future, what sort of behavior would you expect if
> > you parented rigid bodies under one another?  Specifically,
> > what would you expect if you parented a passive rigid body
> > under an active one? Or vice versa?  What about a rigid body
> > parented under a rigid body parented under another, and so on?
> >
> > Do you use parenting for organization? If parenting rigid
> > bodies did something to their behavior (for example, created
> > compound bodies) would this cheese you off because you could
> > no longer use heirarchies as organizing groups?
> >
> > Just curious for no specific reason...
> >
> > -jeff
> >
> > -----Original Message-----
> > From: owner-xsi(at)Softimage.COM
> > [mailto:owner-xsi(at)Softimage.COM]On Behalf Of Harry BARDAK
> > Posted At: October 29, 2005 8:38 AM Posted To: xsi
> > Conversation: Rigids bodies and intercollision.
> > Subject: Re: Rigids bodies and intercollision.
> >
> >
> > yeah this make me laugh too !
> >
> > kim aldis a écrit :
> >
> > >Also - and this one made me laugh. Or cry ;) - don't parent
> > your rigid
> > >bodies. It really messes with them.
> > >
> > >
> > >
> > >>-----Original Message-----
> > >>From: owner-xsi(at)Softimage.COM
> > >>[mailto:owner-xsi(at)Softimage.COM] On Behalf Of Jeffrey Smith
> > >>Sent: 28-October-2005 17:46
> > >>To: XSI(at)Softimage.COM
> > >>Subject: RE: Rigids bodies and intercollision.
> > >>
> > >>Sylvain has the right idea here.  Actual shape collisions
> > like rich,
> > >>dense tesselations, so upping the subdivision is always a
> > good start.
> > >>Also, the detail level rarely needs to be set above "medium."
> > >> Higher settings can capture more intricate details of an
> > object, but
> > >>won't increase stability noicably.
> > >>
> > >>I'd try upping the number of substeps to something like 20,
> > and see if
> > >>that makes a difference.  You could also place some of the
> > bodies in
> > >>different collision layers, so they will not collide with
> > each other
> > >>at all. Finally, you could use proxies for visualization;
> > i.e. scale
> > >>down your rigid bodies a bit and then parent slightly larger
> > >>geometries (not rigid bodies!) underneath them.  This way
> > it will look
> > >>like your geometries interlock tightly, but there will actually be
> > >>sufficient space between them to prevent explosions.
> > >>
> > >>-jeff
> > >>
> > >>
> > >>-----Original Message-----
> > >>From: owner-xsi(at)Softimage.COM
> > >>[mailto:owner-xsi(at)Softimage.COM]On Behalf Of Sylvain Moreau
> > Posted At:
> > >>October 28, 2005 11:59 AM Posted To: xsi
> > >>Conversation: Rigids bodies and intercollision.
> > >>Subject: Re: Rigids bodies and intercollision.
> > >>
> > >>
> > >>Two parameters you should try:
> > >>
> > >>Rigid body properties -> Subdivision level Increase this level is
> > >>similar to having a high res mesh but is more performant
> > than a real
> > >>high res mesh. Very useful if you start with low res geometry.
> > >>
> > >>Modify -> rigid body -> edit simulation properties -> simulation
> > >>accuracy This is the control for temporal sampling, how
> > often do you
> > >>calculate the sim state, once every frame, 16 time per
> > frame, 100 time
> > >>per frame, etc. If you have an explosion and a lot of stuff happens
> > >>between frames you may need a high value here.
> > >>
> > >>Finally don't use the "coffee brake" option. With rigid bodies
> > >>simulation "longer" is not always "better". If it takes an
> > hour every
> > >>time you change a parameters you will most likely get bad
> > results by
> > >>the end of the day. Find a reasonable performance level, run many
> > >>sims, pick the best one.
> > >>
> > >>
> > >>syl
> > >>~~~~~~~~~~
> > >>wotomoro.com
> > >>
> > >>
> > >>----- Original Message -----
> > >>From: "Harry BARDAK" <harry(at)def2shoot.com>
> > >>To: <XSI(at)Softimage.COM>
> > >>Sent: Friday, October 28, 2005 9:54 AM
> > >>Subject: Re: Rigids bodies and intercollision.
> > >>
> > >>
> > >>
> > >>
> > >>>i was aware about using cube instead of plane. But my
> > >>>
> > >>>
> > >>problem isn't on this
> > >>
> > >>
> > >>>box/plane.
> > >>>In fact i would like to know the distance where i should
> > >>>
> > >>>
> > >>place all my
> > >>
> > >>
> > >>>object without to get my object exploding
> > >>>
> > >>>Rainer Schmidt a écrit :
> > >>>
> > >>>
> > >>>
> > >>>>Maybe you can make a denser mesh?  And I did not use a
> > plane in my
> > >>>>previous example. Plane's are 'evil'. Always use a box,
> > >>>>
> > >>>>
> > >>regardless how
> > >>
> > >>
> > >>>>thin it migth be. My stuff always falls through planes
> > >>>>
> > >>>>
> > >>except they are
> > >>
> > >>
> > >>>>ridiculously dense.
> > >>>>
> > >>>>Harry BARDAK wrote:
> > >>>>
> > >>>>
> > >>>>
> > >>>>>If you re read my first mail collision type was set to
> > >>>>>
> > >>>>>
> > >>Actual shape with
> > >>
> > >>
> > >>>>>high details.
> > >>>>>I guess i will get a big headaches tonight !
> > >>>>>
> > >>>>>Rainer Schmidt a écrit :
> > >>>>>
> > >>>>>
> > >>>>>
> > >>>>>>Also set the collision shape to the real geometry and
> > not to the
> > >>>>>>bounding box or the other aproximate shapes. Otherwise
> > >>>>>>
> > >>>>>>
> > >>you think you do
> > >>
> > >>
> > >>>>>>not collide with your geometry but the bounding boxes of
> > >>>>>>
> > >>>>>>
> > >>your irregular
> > >>
> > >>
> > >>>>>>shapes do resulting in a blow out. I had the same
> > >>>>>>
> > >>>>>>
> > >>problem. I had a
> > >>
> > >>
> > >>>>>>plane as colision object, and had the collision shape
> > >>>>>>
> > >>>>>>
> > >>set to bounding
> > >>
> > >>
> > >>>>>>box to do the sim a favor. Then I scattered it and
> > >>>>>>
> > >>>>>>
> > >>forgot about that
> > >>
> > >>
> > >>>>>>setting and fiddled for an hour till I found that
> > >>>>>>
> > >>>>>>
> > >>setting. And voila...
> > >>
> > >>
> > >>>>>>instant success.
> > >>>>>>Good luck!
> > >>>>>>
> > >>>>>>Harry BARDAK wrote:
> > >>>>>>
> > >>>>>>
> > >>>>>>
> > >>>>>>>Thanks for your advice.
> > >>>>>>>i ve taken care to not have doubles objects.
> > >>>>>>>In fact i don't have really interpenatrated object.
> > >>>>>>>My object bound stop where the other object start.
> > >>>>>>>Anyways i moved object to get an offset and to be sure
> > >>>>>>>
> > >>>>>>>
> > >>to remove any
> > >>
> > >>
> > >>>>>>>interpenatration and still have the problem.
> > >>>>>>>Even with only 2 objects. They are shattered. I used
> > >>>>>>>
> > >>>>>>>
> > >>the shatter
> > >>
> > >>
> > >>>>>>>script that you can find in Netview.
> > >>>>>>>
> > >>>>>>>It's a clever idea to use boxes as reference models but
> > >>>>>>>
> > >>>>>>>
> > >>i don't think
> > >>
> > >>
> > >>>>>>>it will work because i have ramdom shapes who doesn't
> > >>>>>>>
> > >>>>>>>
> > >>really fit with
> > >>
> > >>
> > >>>>>>>others boxes. In others word i will still get interpenetration
> > >>>>>>>
> > >>>>>>>
> > >>>>>>>Michael Klein a écrit :
> > >>>>>>>
> > >>>>>>>
> > >>>>>>>
> > >>>>>>>>I'm sure that interpenetration is the problem.
> > >>>>>>>>
> > >>>>>>>>Try to modify all objects that they have no
> > >>>>>>>>
> > >>>>>>>>
> > >>interpenetration and use
> > >>
> > >>
> > >>>>>>>>those
> > >>>>>>>>for the simulation.
> > >>>>>>>>But I would use those objects as a hidden root for
> > the perfect
> > >>>>>>>>matching but interpenetrating ones.
> > >>>>>>>>
> > >>>>>>>>Sounds probably a bit complicated because of double
> > >>>>>>>>
> > >>>>>>>>
> > >>objects in your
> > >>
> > >>
> > >>>>>>>>scene
> > >>>>>>>>but I had the same problem with the sim engine in 4.2
> > >>>>>>>>
> > >>>>>>>>
> > >>- which is not
> > >>
> > >>
> > >>>>>>>>comparable to the new cool one. At the end I used
> > >>>>>>>>
> > >>>>>>>>
> > >>simple bounding
> > >>
> > >>
> > >>>>>>>>boxes as
> > >>>>>>>>reference models for the simulation of an explosion of
> > >>>>>>>>
> > >>>>>>>>
> > >>those pieces
> > >>
> > >>
> > >>>>>>>>and used
> > >>>>>>>>the boxes as hiddem roots for my complex ones.
> > >>>>>>>>
> > >>>>>>>>That worked well because there was no complex obstacle stuff.
> > >>>>>>>>
> > >>>>>>>>president
> > >>>>>>>>
> > >>>>>>>>-----Original Message-----
> > >>>>>>>>From: owner-xsi(at)Softimage.COM
> > >>>>>>>>
> > >>>>>>>>
> > >>[mailto:owner-xsi(at)Softimage.COM] On
> > >>
> > >>
> > >>>>>>>>Behalf Of
> > >>>>>>>>Harry BARDAK
> > >>>>>>>>Sent: Friday, October 28, 2005 12:07 PM
> > >>>>>>>>To: XSI(at)Softimage.COM
> > >>>>>>>>Subject: Rigids bodies and intercollision.
> > >>>>>>>>
> > >>>>>>>>Hi everyone.
> > >>>>>>>>
> > >>>>>>>>I'm a bit stuck with Rigid bodies and i'm new with.
> > >>>>>>>>I have shattered an object who should explode.
> > >>>>>>>>I guess there are some interpenetration  because when
> > >>>>>>>>
> > >>>>>>>>
> > >>i run the
> > >>
> > >>
> > >>>>>>>>simulation
> > >>>>>>>>everything explode at the 1st frame.
> > >>>>>>>>I set the collision to actual shape and even with the
> > accuracy
> > >>>>>>>>"coffee break". I still get this problem.
> > >>>>>>>>
> > >>>>>>>>In new features video on the site of softimage,
> > >>>>>>>>
> > >>>>>>>>
> > >>interpenetration
> > >>
> > >>
> > >>>>>>>>doesn't
> > >>>>>>>>seem to be a problem.
> > >>>>>>>>
> > >>>>>>>>May be i should change my approach but in this case
> > >>>>>>>>
> > >>>>>>>>
> > >>what is the best
> > >>
> > >>
> > >>>>>>>>strategies to explode a ground ?
> > >>>>>>>>
> > >>>>>>>>Thanks by advance.
> > >>>>>>>>
> > >>>>>>>>Harry.
> > >>>>>>>>
> > >>>>>>>>
> > >>>>>>>>
> > >>>>>>>>---
> > >>>>>>>>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
> > >>>>>>>>
> > >>>>>>>>
> > >>>>>>>>
> > >>>>>>>>
> > >>>>>>>
> > >>>>>>>---
> > >>>>>>>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
> > >>>>>>
> > >>>>>>
> > >>>>>>
> > >>>>>
> > >>>>>---
> > >>>>>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
> > >>>>
> > >>>>
> > >>>>
> > >>>
> > >>>---
> > >>>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
> > >>
> > >>---
> > >>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
> > >
> > >
> > >
> >
> >
> >
> > ---
> > 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
>
>
> ---
> 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