Break Glue network based on velocity

[JuriBryan]

Hey all,

I am starting to work on a new shot to learn more about destruction and I am running into a problem.

What I am trying to do is "sculpt" the destruction by creating velocity values in sops for each piece, this way I can control the direction of the initial explosion and have a lot of control over the look.

The problem is, that I still want the control of the Glue network. for example I want parts of the ground/wall to be still together after the explosion.

But the glue network only breaks connections based on impact and not velocity.

My question is now, is there a "easy" way to make the glue network react to velocity values or do I have to change the entire tool to make this work?

If that is the case I would rather through only the pieces that I "sculpt into a active sim and have the rest as static objects... but that takes a way a lot.

So if someone figured out how to do it, or has a idea on what could work... please let me know:)

all help is appreciated!!!!

Here is a quick test scene:

GlueNetwork_velocity.hipnc

Cheers and thanks for the help in advance,

Juri

[mawi]

Hey all,

I am starting to work on a new shot to learn more about destruction and I am running into a problem.

What I am trying to do is "sculpt" the destruction by creating velocity values in sops for each piece, this way I can control the direction of the initial explosion and have a lot of control over the look.

The problem is, that I still want the control of the Glue network. for example I want parts of the ground/wall to be still together after the explosion.

But the glue network only breaks connections based on impact and not velocity.

My question is now, is there a "easy" way to make the glue network react to velocity values or do I have to change the entire tool to make this work?

If that is the case I would rather through only the pieces that I "sculpt into a active sim and have the rest as static objects... but that takes a way a lot.

So if someone figured out how to do it, or has a idea on what could work... please let me know:)

all help is appreciated!!!!

Here is a quick test scene:

GlueNetwork_velocity.hipnc

Cheers and thanks for the help in advance,

Juri

This should work. (atleast in theory)

You can connect a SOP solver to 2nd input of your constraint. By default the SOP-solver contains a objMerge called relations, that is your constraint network Geometry.

Use the dopobjectid point attribute to fetch the velocity of the piece that belongs to each point.

Now blast the point if the length is higher then your threshold.

Keep in mind that the display of anything usefull while inside that sopsolver is in best case limited.

A tip is to build and test your network inside the object containing your glue pieces network and then copy it over to the sop solver.

Might be able to pop a file later on today or tomorrow...

[JuriBryan]

That is what I am trying to do now.

I created a vop sop that compares the length of the Velocity against my threshold and colours the points accordingly. After that I blast everything with a higher red colour then 0.2

This works perfectly in sops and I see the result and it is what I want, but in dops nothing changes.

It is like the points don't get blasted away, and since the display is bugged the way it is I can't even check and see what is wrong.

I even tried delete certain points in sops and feed it in as a new glue network... but that changes nothing.

Could it be that it is just a bug and I need to get the latest build?

[Hudson]

Please share your latest hip-file. I had the same problem in my latest project. I could take a look to remember how I solvet it

[JuriBryan]

here is the new hip...

if I get it to work I will create a few OTLs out of it... but only if I get it to work:)

GlueNetwork_velocity2.hipnc

[Hudson]

Hi Juri!

Actually it seems to work. not perfectly but it works.. try to unlock your glue net node inside Glue_, then reset your simulation and go back to your glue network to lock it again. Try simulating it should work. You should see some fracturing. also delete the channels for the "Overide with SOP" parm and set it to 1 or $F in your "Glue" node in DOPs. It will then update you glue net for every frame instead of just frame 1.

Let me know if it works

here is your file with the changes applied

/Hudson

GlueNetwork_velocity2.hipnc

[JuriBryan]

well yes... it is breaking.

But the problem is that I lose all of my control.

I build the velocity field of the points inside sops and apply the velocity to the pieces before they come into dops.

So I know that I will get only the pieces moving form the start that I really want to be moving.

But with the Glue network applied they either just fall down, not really what I want, or some other part breaks and moves.

The sop solver does not really do anything... could be that the fact that I use initial velocity on the Fracture pieces messes with it...

I will try and apply it as an external force rather then directly on the pieces... but up until now it is working but with out any control and not in the way I define the velocity of the pieces.

Thanks for all your help so far!

hope that we can figure it out, would love to have more control and the awesomeness that is the glue network working together:)

[JuriBryan]

Here is a file showing kinda what I am after.

It is using the field force and some directed particles to give me the motion that I want.

Turn of the glue network and you see what I want to get with this test, besides all the side pieces falling of course.

When I turn on the glue network together with the sop solver that should check for velocity movements with a threshold of 0.2 (so almost nothing) nothing is moving at all.

Any idea why this is happening?

And thanks again for all the great help!

GlueNetwork_velocity3.hipnc

[Hudson]

I understand what you want to achieve now.Well first of all your methed actually works if the Glue net is not locked. So to say a possibility would be to ROP output the glue net since it takes time to reload all the time. Also make sure that your params in the SOP solver inside DOPs are related to your parms inside Glue_ SOP. Otherwise you will be out of visualisation. I am refering to the filter expresion parm inside the group node. it was $CR>=.7 in the Glue_ and $CR>=.2 in the SOP solver :)Another aproach would be to transfer the velocity to the pieces and create a velocity_impulse node inside DOPs wich you can control by addind velocity to whole pieces or points. I am still a noob so I couldn't figure out how to achieve the interaction between the point velocity inside SOPs and add it to the velocity impulse inside DOPs.I multiplied the velocity in your point SOP inside box_object with .3 since it was insanely high and all the pieces just disapeared from the screen in 2 frames :PTo make sure it works try to apply the velocity only for a few pieces in the middle to see the result. at the moment allmost all the glue net dissolves since all pieces has a pretty high velocity.Good Luck!

GlueNetwork_velocity3_HM.hipnc

Edited May 3, 2013 by Hudson

[Hudson]

Just saw your recent post. I take one more look. For me it seems that you are really close. I was getting the glue net to dissolve prims in the file I sent you. the only problem was that almost the whole glue net disaperad exept two prims. I was thinking it was because almost all pieces had an initial vel

[Hudson]

Well, I have to give up on this one. My knowledge is not sufficient since it seems to be a matter of VOPSOPing so you can control your values a bit more. The deeper I get into this issue the mor eI understand your problem and also realize I can't help very much.

Good luck, I hope others give it a shot

[JuriBryan]

Okay I fixed it, well it works:)

I had to create a work around, importing the particles that drive the DOP sim and which are driven by the "sculpted" velocity on the pieces into a sop solver.

Then I colour them based on velocity, transfer the colour of the points on the glue network and then I delete the glue network.

This works and keeps all the control from sculpting the explosion intact and adds the control of the glue network to it.

So now I can sculpt a house out of wood, metal, concrete and glass and have it in one simulation:)

that will be fun!

cheers and thanks to everyone that helped!

here is the file showing how I solved it:

GlueNetwork_velocity4.hipnc

PS: I need to find a way to keep the force scales "normal"... more then a 1000 in scale is just a strange way to work with it.

[Hudson]

Good job man! Thank you for sharing