Jump to content


  • Content count

  • Donations

    0.00 CAD 
  • Joined

  • Last visited

Community Reputation

1 Neutral

About davide445

  • Rank

Personal Information

  • Name
  1. I'm a bit ashamed to ask, but is something I seems not be able to solve. Having the attached mesh Blade_align.fbx, where seems the pivot is not centered and also most importantly the axes are wrongly rotated. The object has also strange orientation. I remember there was a C4D beautiful tool to automatically set pivot position in the volume center of mass, and also align the axis to the object main simmetry. In Houdini I seems not being able to achieve anthyng related to axes, I was able to just reset the pivot using the Transform node. Someone can please give me a hint.
  2. Since Houdini tet embed mesher is very good and very fast I will use it for prototyping, but need next to export the tetrahedron volumetric mesh in a simulation ready format such as gmsh .msh format. https://gmsh.info/doc/texinfo/gmsh.html#MSH-file-format I suppose this can be done directly using Houdini API and gmsh API without going through a mesh file, writing a Python reader from Houdini, then import the data directly in Gmsh using gmsh.model.mesh.addNodes and gmsh.model.mesh.addElements Another option will be converting Houdini native .geo format in the .msh format. Attached an example of the .geo format containing a tet mesh created in Houdini. Another path will be using third party converters such as this one https://people.sc.fsu.edu/~jburkardt/m_src/fem_to_gmsh_test/fem_to_gmsh_test.html https://github.com/nschloe/meshio so first exporting Houdini mesh in an intermediate format and next using the converter to create the final .msh mesh Will leave to the specialist the definition of the best path. I need an estimate to understand how complex and portabe around Houdini releases the result can be. liver3hd60kv2.geo
  3. I have created tet using tet embed in this liver mesh, want to export the mesh as solid and not triangle for specialized simulator usage. As example STL format I tried to export into is composed only from triangles, there is any format supported from Houdini that can preserve the created tet solids.
  4. Houdini 19 Wishlist

    - MPM solver - Mapbox on steroids integration
  5. Why still no MPM solver

    Will finally test on Maya since seems there are no other options, will need to start from scratch since never used it. Looking at research papers there are applications where MPM work in real-time (faster than PBD), so I suppose it's not necessary slower than other solvers, even if I suppose this can be a very special case. MPM can be used in many kind of sim for granular materials, thin objects and fibers but also soft objects with breaking parts etc and more in general any large deformation, contact and fracture problem. There are even GPU implementation of the method so I suppose something innovative can be made with it. I'm not complaining SideFX can do what it want, just trying to understand.
  6. Looking since some time how to realistically sim volumes of snow moved by objects (vehicles), with the ability to change his properties (I.e. wetness, density). I've seen lately Vellum (improved PBD solver) used for this kind, but my humble testing didn't see any way to realistically simulate other than dry snow. The "old" POP Grains (original PBD) seems also limited to a few choices. Using FLIP or SPH solver might result in more flexible results with a lot of custom work to be done, but so far the more complete solution (realism + flexibility) I have learned about is MPM. MPM is available out of the box in Bifrost, in Houdini I only found an Azmisov 2015 old third party solutions that will sure never work on 18.5, another 2015 tutorial from cmivfx using VEX either probably not working anymore. So just wondering why there is so little love for this kind of solver (that can be used for many other kind of sim other than snow) in Houdini, or if I miss some point so that maybe it's really not needed for my goal.
  7. Approacing the problem of simulate snow plowing, I find there are two kind of grains solvers: POP Grains (PBR based) used to create Wet sand, or Vellum Grains (XPBR based). Wondering what can be the best for snow, I found both of them used but didn't understand what can be better for a thick volume of snow and not only for surface effects.
  8. Looking a way to generate buildings resembling as most as possible the real one, avoiding a lot of manual work. Can define procedurally any kind of building, but in fact will be always a subset of the real one you can find on a street, the alternative being a lot of analysis just to define a shape grammar applicable for the specific location. I can stitch street view images on the 3D shapes, except it's not allowed to store them offline so no way to legally proceed on this path. Want to ask if there is any way to extract from street view data the basic elements of the facade (i.e. prevalent color, number of windows, floors) and next generate from this a facade (even as simple 2D texture to be applied to the 3D shape).
  9. Pyro Fx vs physics sim

    Maybe a noob question, was considering how to actually simulate fire dynamics in a realistic but fast way on different dynamic situations. Using Pyro Fx that produce realistic looking results, another one is using specialized tools such as Fire Dynamics Simulator based on Nvier Stocks equations for fluid dynamics. How much are they different?
  10. Smoothing DEM data

    It's an option, but was making it work in H16 release, with later H17 breaking the APIs so didn't work past that. Not a problem so far I'm not working on terrains on continuous basis, even if I will have still a lot of work to do for this specific case.
  11. Smoothing DEM data

    @j00ey art the end discovered the realtime engine accept RGBA 16 fixed format, this and using the simple heightfiled_blur node I obtain good results.
  12. Smoothing DEM data

    Testing the hiplc is working fine, but the result didn't change in the realtime engine. I was noticing another element, that maybe show how beginner still I am: the converted_heightmap.tga I created just using an Heightfiled Output attached to the Mapbox node using Image Output settings Packet Raster, Single Channel, 32f, it's in fact a RGB8 image. Trying to output using RGB format instead of Single Channel and 16b fixed Type produce an RGB16 image. What's the correct Image Output settings to obtain an R32F or R16F image?
  13. Smoothing DEM data

    @j00ey I can't smooth it in the realtime engine (except developing some new component that's not the path I want), want to smooth (or better to say refine) it first in Houdini and next just export and use it as is in the realtime engine. Here are how the original Mapbox data appear in Houdini, a problem since I need to have a vehicle running on these slopes, that for sure are not terraced in reality
  14. Smoothing DEM data

    @j00eywas trying to replicate your work into my project, but i faced a problem: you converted all to VDB and I need to output an heightmask. Normally using heightfield_output, it's not working with your output node. I can't find any way to convert back the VDB to a terrain. @nukiattached the original Mapbox height and the heightmap result of a conversion I'm using into the 3D realtime engine. maps.zip
  15. Smoothing DEM data

    @j00ey thanks will check this back in the office. @nukithe terrain was downloaded from Mapbox and next converted to heightmap using Terrain_Output node. The original height it's from 980 m to aboit 2100 m. The shown picture its not in Houdini but in our 3d realtime engine.