exploding ball, please help
22641 18 4- klasse
- Member
- 60 posts
- Joined: 6月 2008
- Offline
Hi.
I'm trying to do an exploding ball, so i would need some advice.
i've cut my ball in to peaces with the break tool.
my problem is to get the a explosion force on the separate peaces, whats the best way to do this?
at the moment i'm putting a fluid particle emitter inside my ball, so the ball breaks up from the inside.
is there a better way to do this?
thanks!
I'm trying to do an exploding ball, so i would need some advice.
i've cut my ball in to peaces with the break tool.
my problem is to get the a explosion force on the separate peaces, whats the best way to do this?
at the moment i'm putting a fluid particle emitter inside my ball, so the ball breaks up from the inside.
is there a better way to do this?
thanks!
- goldfarb
- スタッフ
- 3464 posts
- Joined: 7月 2005
- Offline
- klasse
- Member
- 60 posts
- Joined: 6月 2008
- Offline
- Allegro
- Member
- 696 posts
- Joined: 3月 2006
- Offline
- klasse
- Member
- 60 posts
- Joined: 6月 2008
- Offline
Allegro
I would create some particles, copy some spheres to them, cache them to disk, read them back in, and use these as an rbd fractured object.
This should prove slightly faster than trying to blow something apart with fluids.
do i need to cache out the geometry to be able to get it to work?
right now in the dop i only see one of my 100 spheres. it looks like it ignoring the copied spheres. any ideas?
- Allegro
- Member
- 696 posts
- Joined: 3月 2006
- Offline
- klasse
- Member
- 60 posts
- Joined: 6月 2008
- Offline
- Soothsayer
- Member
- 874 posts
- Joined: 10月 2008
- Offline
- Allegro
- Member
- 696 posts
- Joined: 3月 2006
- Offline
The fundamental difference is that I set it up using the shelf so the whole file took about a minute to set-up minus the commenting. There end up being some seemingly redundant nodes in there when it's such a small network, but the nodes don't really slow anything down and, I find, help for visual clarity down the road
Ideally, looking at your set-up it could also be set so that the torus is on the rightmost of the merge dop, and the collisions are set to left affect right so that dops can speed along faster without worrying about mutual collisions.
Ideally, looking at your set-up it could also be set so that the torus is on the rightmost of the merge dop, and the collisions are set to left affect right so that dops can speed along faster without worrying about mutual collisions.
Stephen Tucker
VFXTD
VFXTD
- dim1984kimo
- Member
- 84 posts
- Joined: 8月 2008
- Offline
- cklosters
- Member
- 224 posts
- Joined: 11月 2008
- Offline
The exploding ball in the Pyro demo's was made using two passes.
First I created a simple explosion to get the force needed to drive the rbd's. That was it's only purpose. After that I advected particles using the simulations velocity, copy stamping on some primitives at the end.
After that I created a shattered object. As I did not want the inside of the ball to contain any geometry, I used a for each loop to extrude every shard with a random offset. This also makes sure the geometry is still in separate groups.
At this stage I had some particles for pushing away the shards and the shards itself. Now you only have to set up your simulation in Dops. The shards get filled with fuel and I added an extra piece of geo for some extra fuel distribution inside the sphere. The particles get used to push away the shards and the temperature/fuel emission creates the explosion. It takes some time to find the right balance but things should work straight of the bat. Also, when shading, the scattering gave me the exploding kind of feel, as it lights up the smoke.
You also get the burning shards as they fly off (because the fuel is still being attached to your object as it flies off. )
There is also the option to use feedback for fluid/rbd interaction. But this takes a lot of time to tweak and simulates for a longer time. I got around the problem by using the particles driven by a similar simulation to push away the shards and use that data to create the explosion
Hope this helps!
Cheers Coen
First I created a simple explosion to get the force needed to drive the rbd's. That was it's only purpose. After that I advected particles using the simulations velocity, copy stamping on some primitives at the end.
After that I created a shattered object. As I did not want the inside of the ball to contain any geometry, I used a for each loop to extrude every shard with a random offset. This also makes sure the geometry is still in separate groups.
At this stage I had some particles for pushing away the shards and the shards itself. Now you only have to set up your simulation in Dops. The shards get filled with fuel and I added an extra piece of geo for some extra fuel distribution inside the sphere. The particles get used to push away the shards and the temperature/fuel emission creates the explosion. It takes some time to find the right balance but things should work straight of the bat. Also, when shading, the scattering gave me the exploding kind of feel, as it lights up the smoke.
You also get the burning shards as they fly off (because the fuel is still being attached to your object as it flies off. )
There is also the option to use feedback for fluid/rbd interaction. But this takes a lot of time to tweak and simulates for a longer time. I got around the problem by using the particles driven by a similar simulation to push away the shards and use that data to create the explosion
Hope this helps!
Cheers Coen
Senior Technical Artist Guerrilla Games
- sandzuro
- Member
- 24 posts
- Joined: 6月 2009
- Offline
Sorry for a silly question and bad english. When I make for each loop with extrusion, simulation have stoped working probably cause it recieving wrong group names. What have you done in this case? And another question: Have you drived shatters simulation by impacting it with geo attached to instances?
- thalie
- Member
- 14 posts
- Joined: 2月 2008
- Offline
Dear all,
I have read thoroughly all the posts! And I must admit that I'm really interesting in managing an explosion!
I have tried Allegro's file and it is exactly what I want!! But there's one thing I didn't manage to do and mainly understand, it's to “create some particles, copy some spheres to them, cache them to disk, read them back in, and use these as an rbd fractured object”, as Allegro put it.
I'm sorry but I'm new in that!!
Could anyone explain how do we do that? cache the particles on the disk?
Tx,
Thalie
I have read thoroughly all the posts! And I must admit that I'm really interesting in managing an explosion!
I have tried Allegro's file and it is exactly what I want!! But there's one thing I didn't manage to do and mainly understand, it's to “create some particles, copy some spheres to them, cache them to disk, read them back in, and use these as an rbd fractured object”, as Allegro put it.
I'm sorry but I'm new in that!!
Could anyone explain how do we do that? cache the particles on the disk?
Tx,
Thalie
- Allegro
- Member
- 696 posts
- Joined: 3月 2006
- Offline
If you take anoter look at that file, you'll see that the sphere_object1 runs a particle simulation, copies spheres to the particles, and then uses these as collision objects in the simulation.
The only step from your question is that they weren't written to disk before sending them to dops.
If you wish to write your particles to disk (to prevent having to resimulate them more than once), simply append a rop geometry output node to the popnet and use it to save your files to disk. Then use a file node to read the cache back in.
The only step from your question is that they weren't written to disk before sending them to dops.
If you wish to write your particles to disk (to prevent having to resimulate them more than once), simply append a rop geometry output node to the popnet and use it to save your files to disk. Then use a file node to read the cache back in.
Stephen Tucker
VFXTD
VFXTD
- thalie
- Member
- 14 posts
- Joined: 2月 2008
- Offline
- tomschwarz
- Member
- 4 posts
- Joined: 4月 2009
- Offline
Hi,
I want to make an effect similar to Coen's exploding ball but I am doing it to a model of a soldier.
Unfortunately because of the way the soldier was modelled I am unable to use the Houdini shatter RBD - this causes crazy distortion without shattering. I'm no expert so I don't know why this is. As such I have used a tool I found on Odforce to shatter the soldier object - this gave me the shattering I was after (its a platic toy soldier). Here's a link to the tool http://forums.odforce.net/index.php?app=core&module=attach§ion=attach&attach_id=7932. [forums.odforce.net]
I've run into difficulty now though because when I attach the pyro nothing happens.
I don't know how to separate the geometry and I think this could be the problem. I have saved it out as .bgeo and loaded it into an empty scene but the pyro still doesn't work on it.
How can I seperate the geometry after it has been shattered?
I want to make an effect similar to Coen's exploding ball but I am doing it to a model of a soldier.
Unfortunately because of the way the soldier was modelled I am unable to use the Houdini shatter RBD - this causes crazy distortion without shattering. I'm no expert so I don't know why this is. As such I have used a tool I found on Odforce to shatter the soldier object - this gave me the shattering I was after (its a platic toy soldier). Here's a link to the tool http://forums.odforce.net/index.php?app=core&module=attach§ion=attach&attach_id=7932. [forums.odforce.net]
I've run into difficulty now though because when I attach the pyro nothing happens.
I don't know how to separate the geometry and I think this could be the problem. I have saved it out as .bgeo and loaded it into an empty scene but the pyro still doesn't work on it.
How can I seperate the geometry after it has been shattered?
- probbins
- Member
- 1145 posts
- Joined: 7月 2005
- Offline
- tomschwarz
- Member
- 4 posts
- Joined: 4月 2009
- Offline
- old_school
- スタッフ
- 2540 posts
- Joined: 7月 2005
- Offline
Here's yet another way. (fyi Keith just wanted a straight forward exploding sphere from points so that is what I gave him. The multi-pass approach seemed too “M-like” for his tastes. You know. Tweak particles. Cache. Tweak RBD. No good. Go back. Tweak particles. Cache. and on and on…)
Just blow the crap out of the rbd fractured object with the particles themselves. Everything is live and you just tweak away.
Simple and straight forward.
Attached is a file that should be quite reusable and could make the basis for a new shelf on exploding RBD with particles directly. Lots of work on conditioning the shell to get auto pieces going. You will want to match the break clip grid polygon sizes to roughly match the resolution of the critter you wish to cut up.
I use the sdf calculated from the bounded animated rbd object to generate the explosive particle cloud. You could just create a rough sdf of your critter as well and scatter in that sdf volume. I use the volume gradient of the sdf in a VOP SOP to determine the initial velocity with what every sick minded IED builder would love to have: “up direction addition” for that extra upward lift. Less energy in to the ground.
Everything else is just a simple well timed particle blowing rbd chunks around explosion.
Main controls are the mass and initial velocity of the particles. Oh and the number of input scatter points as well along with the duration of the particle birth.
The particles are colliding with the pieces in a two-way setup. You could just have the particles collide with the rbd pieces for that extra immediate kick.
Note: Particle collisions currently are not taken in to account with glue objects so I keyed the glue strength to time with the particle explosion. It could though. Glue is just an arbitrary rule set up that isn't that arbitrary actually since it is built up for you in a particular way…
Just blow the crap out of the rbd fractured object with the particles themselves. Everything is live and you just tweak away.
Simple and straight forward.
Attached is a file that should be quite reusable and could make the basis for a new shelf on exploding RBD with particles directly. Lots of work on conditioning the shell to get auto pieces going. You will want to match the break clip grid polygon sizes to roughly match the resolution of the critter you wish to cut up.
I use the sdf calculated from the bounded animated rbd object to generate the explosive particle cloud. You could just create a rough sdf of your critter as well and scatter in that sdf volume. I use the volume gradient of the sdf in a VOP SOP to determine the initial velocity with what every sick minded IED builder would love to have: “up direction addition” for that extra upward lift. Less energy in to the ground.
Everything else is just a simple well timed particle blowing rbd chunks around explosion.
Main controls are the mass and initial velocity of the particles. Oh and the number of input scatter points as well along with the duration of the particle birth.
The particles are colliding with the pieces in a two-way setup. You could just have the particles collide with the rbd pieces for that extra immediate kick.
Note: Particle collisions currently are not taken in to account with glue objects so I keyed the glue strength to time with the particle explosion. It could though. Glue is just an arbitrary rule set up that isn't that arbitrary actually since it is built up for you in a particular way…
There's at least one school like the old school!
-
- Quick Links