cracking using particles and primitives

[Omy]

hello everybody.

I have animated the curve based on the expression and also i have setup a pop and primitive setup.

I want the grid to be deformed based on the animation of the curve

i tried to do something,but i dint get the result.

i have attached the hip.please help me out.

Edited November 23, 2010 by Omy

[sadhu]

where is the scene file?

[Omy]

where is the scene file?

particleicebreak.hipnc

[Emanuele Berti]

Hi Omy,

I took your scene and I just changed it a bit, the main difference is about the attributeTransfer SOP, now it uses an attributeTransfer POP.

Hope that helps.

Cheers,

Emanuele

particleicebreak_fix.hipnc

[Omy]

Hi Omy,

I took your scene and I just changed it a bit, the main difference is about the attributeTransfer SOP, now it uses an attributeTransfer POP.

Hope that helps.

Cheers,

Emanuele

hi emanuele,

thank you very much

can you do the cracking in dops and metaballs and magnet force and show me how is it done.please

cheers,

[tettoffensive]

I'm trying to do something similar to this, but I'm projecting the l-system onto a sphere with a ray sop, but since the l-system is planar, the projection isn't spherical and gets more compressed toward the edges. Does anyone have any ideas for how I could project the l-system geometry a little better onto a sphere?

Thanks

[Omy]

I'm trying to do something similar to this, but I'm projecting the l-system onto a sphere with a ray sop, but since the l-system is planar, the projection isn't spherical and gets more compressed toward the edges. Does anyone have any ideas for how I could project the l-system geometry a little better onto a sphere?

Thanks

i think u myt give a try using nurbs sphere-project-profile.It does similair to raysop.

[pclaes]

I find creeping geometry onto an approximate surface (modified nurbs sphere/cylinder) and then raying on the actual surface works fairly well. You can use several "projections" to avoid pinching at the poles. You can group the edges by $BBX or $BBY before you perform the creep (the 2D geometry should be alligned to the xy (0,0;1,1) grid) and fuse the shapes based on those groups, this way you can get rid of seams.

[Omy]

I find creeping geometry onto an approximate surface (modified nurbs sphere/cylinder) and then raying on the actual surface works fairly well. You can use several "projections" to avoid pinching at the poles. You can group the edges by $BBX or $BBY before you perform the creep (the 2D geometry should be alligned to the xy (0,0;1,1) grid) and fuse the shapes based on those groups, this way you can get rid of seams.

is boolean operations are the only way to get the cracks? is there any other better methods than the cookie?

[pclaes]

Hi Omy,

there is no easy way to do physically accurate cracking in 3d. It is a hard problem, but there are ways to fake it or procedurally model it. Cookies do tend to give crappy geometry and either not work, or need a lot of extra filtering. There are ways to do it with the clip sop (as part of a foreach) to break geometry. You could also reverse engineer some of the voronoi breaking patterns to get cracks you desire. Read up on how voronoi actually works and you should be able to scatter points in such a way that the edges of the different cells will give you a fracture line. By scaling some of those cells inwards you can create cracks.

Some of those techniques are discussed in a tutorial dvd I made a while ago:

http://www.3dbuzz.com/xcart/product.php?productid=58&cat=0&page=1&featured

Breaking objects in 3d remains a hard problem. Cracking can also be faked to a degree, because a crack does not need to break all the way. Which means you can cheat it to a degree with displacements and animated textures.

There was a file posted a while back that used an L-system together with an add sop, a clip sop and a polywire.

Do a search here on the forum for "crack". It has been done a few times already.

*) I would consider going with voronoi fracturing. So you actually do break your object. Then you identify where the crack needs to go, you can propagate some kind of attribute from one point along it's edges in a certain direction to the next. Or simply trigger it with an animated curve (either carve the curve or an animated L-system).

*) I would make the scatter of the points adaptive in the volume, so you have a lot of points on the surface and few in the volume. (You can do this with an sdf if you want).

*) I would also identify those edges that are triggered and add detail to them (either by subdividing them and adding noise, or at rendertime with a displacement shader). This would be an extra layer of detail. but the main direction should come from the voronoi.

*) If you really think about what a crack is: an opening of space inside of a solid material... where has that material gone that used to be in that space? Most of the time it has simply moved aside/pulled apart, if it's a wall it may have fallen out. So if you want to recreate that, displacement seems not a bad choice. Combine it with a bit of particle dust where the fracture is and that should give you a good start.

But with all the advice people have already given you, you should be able to manage, I mean there are scene files all over the place in regards to fracturing and cracks. Also have a look at Johners' example files that come with the voronoi fracturing.

A very interesting read on voronoi fracturing (before it became part of houdini 11):

cracking with L-system procedurally modeled:

Macha's file here:

The force propagation that creates a realistic crack pattern is still quite difficult. You can try, but it is almost like an aggregation system in 3d and then using that system to break up the model.

In regards to aggregation in houdini, read this: http://nccastaff.bournemouth.ac.uk/jmacey/MastersProjects/MSc09/Cashmore/MichaelCashmore-Thesis.pdf

It could provide you with really nice patterns.

The "problem" with all of this is that it really is not beginner material. This is intermediate to advanced level effects, and it requires a certain understanding of 3d, math and Houdini. There are a lot of scene files that can help you along the way, but ultimately it will take time to put a realistic crack together yourself. It is good to push yourself and try difficult things as that is how you learn. You just chose an effect that is quite difficult if you want to do it right. There are a lot of hacks that look good enough though .

Good luck!

[Omy]

Hi Omy,

there is no easy way to do physically accurate cracking in 3d. It is a hard problem, but there are ways to fake it or procedurally model it. Cookies do tend to give crappy geometry and either not work, or need a lot of extra filtering. There are ways to do it with the clip sop (as part of a foreach) to break geometry. You could also reverse engineer some of the voronoi breaking patterns to get cracks you desire. Read up on how voronoi actually works and you should be able to scatter points in such a way that the edges of the different cells will give you a fracture line. By scaling some of those cells inwards you can create cracks.

Some of those techniques are discussed in a tutorial dvd I made a while ago:

http://www.3dbuzz.com/xcart/product.php?productid=58&cat=0&page=1&featured

Breaking objects in 3d remains a hard problem. Cracking can also be faked to a degree, because a crack does not need to break all the way. Which means you can cheat it to a degree with displacements and animated textures.

There was a file posted a while back that used an L-system together with an add sop, a clip sop and a polywire.

Do a search here on the forum for "crack". It has been done a few times already.

*) I would consider going with voronoi fracturing. So you actually do break your object. Then you identify where the crack needs to go, you can propagate some kind of attribute from one point along it's edges in a certain direction to the next. Or simply trigger it with an animated curve (either carve the curve or an animated L-system).

*) I would make the scatter of the points adaptive in the volume, so you have a lot of points on the surface and few in the volume. (You can do this with an sdf if you want).

*) I would also identify those edges that are triggered and add detail to them (either by subdividing them and adding noise, or at rendertime with a displacement shader). This would be an extra layer of detail. but the main direction should come from the voronoi.

*) If you really think about what a crack is: an opening of space inside of a solid material... where has that material gone that used to be in that space? Most of the time it has simply moved aside/pulled apart, if it's a wall it may have fallen out. So if you want to recreate that, displacement seems not a bad choice. Combine it with a bit of particle dust where the fracture is and that should give you a good start.

But with all the advice people have already given you, you should be able to manage, I mean there are scene files all over the place in regards to fracturing and cracks. Also have a look at Johners' example files that come with the voronoi fracturing.

A very interesting read on voronoi fracturing (before it became part of houdini 11):

cracking with L-system procedurally modeled:

Macha's file here:

The force propagation that creates a realistic crack pattern is still quite difficult. You can try, but it is almost like an aggregation system in 3d and then using that system to break up the model.

In regards to aggregation in houdini, read this: http://nccastaff.bournemouth.ac.uk/jmacey/MastersProjects/MSc09/Cashmore/MichaelCashmore-Thesis.pdf

It could provide you with really nice patterns.

The "problem" with all of this is that it really is not beginner material. This is intermediate to advanced level effects, and it requires a certain understanding of 3d, math and Houdini. There are a lot of scene files that can help you along the way, but ultimately it will take time to put a realistic crack together yourself. It is good to push yourself and try difficult things as that is how you learn. You just chose an effect that is quite difficult if you want to do it right. There are a lot of hacks that look good enough though .

Good luck!

Awesome.Thank you very much for the detailed answers Cheers