bobbybob

Try using the position for calculating the length. If you create an attribute vop and use the UV as the color (at sop level), you can see why you are getting this distortion.

We can use a modify data node, and get the velocity data with the dopfield function/expression, and use it as a condition

Is there a easy way to set a "speed limit", like we can do on packed, in DOP's? The velocity data is not available at pre- or post solve, and it's not possible to use a wrangle after the RBD solver (or is it?), so what would be a good solution here?

How is it possible to source in RBD objects (hero) into DOP over time?

Is it possible to create two way collision between rigid body and grains? I am trying to make a simple example where a sphere is dropped down in a container of grains.

Thank for the answer @coltonmiland @davpe I am still in doubt how to texture the terrain. Is it possible to generate masks from the heightfield to blend materials, to break up tiling patterns? DO you use very big textures for cliffs. Normally cliffs have very sharp and contrasted features, that makes it very easy to see tiling, how can we go about solving this? Again, is it normal to use some kind of LOD setup? If the texture have to cover big areas of the cliff, it would be possible to use a single texture, if the cliff is far away, but if the cliff get's closer, we are able to see the low texture resolution. If we then scale down the textures, we are getting tiling patterns? I just can't find any sources on this, and find myself pretty stock. It would be possible to create a lot of custom textures, and do a lot of mapping work, but it would be great if there was some kind of known workflow for this problem

Workflow for creating terrain I am trying to create a realistic mountain terrain, but I can't find many resources on it. There is Alex Dracott's "Generating HeightField Terrain Textures" and "GDC 2017 - Houdini 16 Terrain Tools" but it leaves me with some questions that you guys maybe have the answer to. When I look at Alex Dracott, and other artists terrains, it looks like the vertical side of the mountains is texturered very nice, but how is this possible? The vertical sides of a hightfield will always have very low resolution, because the UV's a projected down on the surface in the Y direction. One though I had, was to create the UV's based on the quads that make up the heightfield, but this will involve some kind of remeshing, because the mesh also are scaled (Stretched). Extreme example of UV stretching <stretched_uvs.PNG> Another thing is the texturing. In Alex Dracott's video, he uses cop to create colored noise from masks and applies them back as textures, but what is the practice for texturing a big terrain with textures/materials? I have been playing around with substance painter, and projection, to "draw" materials, is there another way? Is it normal to export the heightfield mask and generate materials in a more procedural way via substance designer, for these big areas, or it more common to create the material from scratch without any masks in something like substance designer? And last, is it normal to create some kind of LOD system for such a project, or using UDIMS, to get the details we want when getting near curtain areas of the terrain? Maybe I am lost, and not even near what a real world workflow looks like, but I am sure that someone on the forum has experience with this.

Is it possible to create a custom heightfield mask with my own name? It seems to me that only the predefined mask names (mask, height, water, debris...) is possible to reference, when working with heightfields? If I wanna create 3-4 different grass masks, for texture splat mapping later, it would be nice to be able to create these mask (grassMask1, grassMask2....) It's going to be used for working with them in COP, but also export them, for later use when create the final materials

Thanks for the answer @konstantin magnus, but what I really wanna know is, why we can't delete the @P attribute. Just to understand the data architecture better.

I ran in to this problem, because I have a sphere, and I need a deformed version of it, but only as a rest attribute. I know how to do this in other ways, but I thought, that to just make the deformation, create the rest, delete the P attribute and then merge it with the original sphere would be nice. But when trying to delete the P attribute with a Attribute Delete, nothing happens? Is this because P is a special attribute or something? Would be interesting to know.

I am trying to separate geo that intersect. I am trying to do this with simulation. I am able to do this via the wire solver (File example) but I need to be able to separate more custom geo then just "tubes". What I want, is to first separate the geometry (custom geo), then take that geometry and use it with the FEM solver. I am not able to make the FEM or Grain solver push geo away from each other, when it intersects like I can with the wire solver. Is there any trick or something that I have overseen? I am able to push away points from each other via vex, but it is pretty slow, because it has to iterate over a lot of points, and run more then ones because some geo that is pushed, will intersect with other geometry, and it is not that easy to setup if I don't start with lines, but real geometry, (like geo from zbrush) wire_separation.hipnc

@StepbyStepVFX, thanks for your answer. So this is actually described in the FEM docs, I feel a bit stupid Quote from documentation

When would you use FEM over Grain? The last couple of weeks I have been working with Grains, and got a good understanding of how it works. It seems that you can do pretty much everything with this tool in case of cloth and simulation of object that will deform. I haven't been using that much time on FEM, and from my perspective it looks like a tool that has a lot of overlaps with the grain tool. It could be nice to get an understanding of where the FEM and Grain tool differs and where to use which, from people that has good experience with both tools.

In this hip, there's a cloth simulation that does what I am trying to explain. The other network is a grain solver, where I am trying to create the same "surface stiffness" by using the targetP, but I don't think that is the right way. surface stiffness.hipnc

Is it possible to control the "surface stiffness" of a granular sheet, so it will act like a sheet of paper or another stiff material?