[opensource-dev] Soft body physics
gcanaday at gmail.com
Wed Apr 7 17:25:50 PDT 2010
I'm only playing with the idea. Soft-body would have to be client-side
only. if you think your SL is laggy NOW, imagine if it were also sending
the vertex data back and forth to the sim. That would make SL unusable.
It's a fun idea, though... I can imagine fields of waving grass, rubber
couches and trampolines, and well as a little bit o' jiggle when ya
walk, ya know, jes for fun :) Prim clothes and flexi hair that actually
I'm only dreaming it up. The client would have to anticipate collisions
a few frames before they occur to keep it more realistic, and then the
server would have a few server frames to report that the collision
actually happened like it already does. No server changes in mind.
It would require that changes be made to the avatar meshes.
On 04/07/2010 07:39 PM, Jacek Antonelli wrote:
> Personally, I would keep soft body effects as purely client-side eye
> candy, without any impact on the world, like flexi prims and particles
> are today. Sending the deformation data from the server to the client,
> or vice versa, would be very bandwidth intensive, and a huge headache
> to keep in sync. And some users would need to disable the feature
> altogether, because it would also be costly to process the effect.
> - Jacek
> On Wed, Apr 7, 2010 at 2:48 PM, Glen Canaday<gcanaday at gmail.com> wrote:
>> I came to the same conclusion, with the exception that a lookahead for a
>> collision would be helpful in the client. If they use the same physics
>> engine and the client does no more than implement a collision check, I think
>> it could be a good thing. I'm looking at several free physics engines atm
>> and thought I would clear my thinking up some. Collisions with soft bodies
>> deform the collision mesh itself so yeah it gets a little tricky when you
>> split them.
>> On 04/07/2010 11:14 AM, Moriz Gupte wrote:
>> I feel you are right. Makes more sense to have it implemented client side
>> for many soft body dynamic behaviors... eg cloth, hair etc...
>> but I think in areas where rigid body behaviors impact local soft body
>> dynamics, there will be lots of timing and synch problems to deal with.
>> So there's where I think that perhaps all physics need to be done at the
>> same site.
>> On Wed, Apr 7, 2010 at 9:03 AM, Glen Canaday<gcanaday at gmail.com> wrote:
>>> Soft body physics are best implemented in a local viewer, leaving the
>>> rigid-body collision detection to the server, am I right in this?
>>> Policies and (un)subscribe information available here:
>>> Please read the policies before posting to keep unmoderated posting
>> 'Consider how the lilies grow. They do not labor or spin.'
>> Rameshsharma Ramloll PhD Research Assistant Professor Idaho State
>> University, PocatelloTel: 208-282-5333
>> More info at http://tr.im/RRamloll
>> Policies and (un)subscribe information available here:
>> Please read the policies before posting to keep unmoderated posting
More information about the opensource-dev