Jump to content


  • Content count

  • Donations

    0.00 CAD 
  • Joined

  • Last visited

Community Reputation

19 Good

1 Follower

About WLVL

  • Rank

Personal Information

  • Name

Recent Profile Visitors

1,774 profile views
  1. Glad you found this useful. Double check that you are copying bbox max, bbox min, bbox max1, bbox min1, values correctly to the material instance parameters in unreal
  2. Is moving to Australia reasonable?

    You need to have at least 2 years of experience in the field you are being hired in, to be granted a full working visa (high skilled professional). If you go to Australia with the Working-Holiday Visa, you can open your self-trader account (ABN, australian business number) and freelance. The market in Sydney is not as dynamic as in London for example, and being at the beginning of your career won't help. So, if you decide to come you will have to accept the fact that your work might not be continuos, and have to do odd jobs in between VFX gigs. Note that Iloura Sydney ( rebranded to Method Sydney ) closed down in September 2018. Iloura Melbourne is now Method Melbourne. Mill Film has just opened in Adelaide and planning to expand. a list of the main vfx houses in Oceania here: http://vfxvoice.com/vfx-down-under-the-now-and-the-new-in-australia-and-new-zealand/
  3. Hey Houdinians, I have made a check list gathering infos from the webinars/tuts done by the great Houdini GameDev Toolset Team. Please let me know if something is missing or wrong, I hope this is helpful for some of you. Rigid Body Dynamics Vertex Animation Texture - RBD VAT optimal for high # of pieces. alternatives (bones, straight fbx) in HOUDINI: 1- rbd material fracture SOP proxy to rbd bullet sim 2A- dopimportSOP(fetch geometry from DOP network) OUT to VAT ROP 2B- this step if you need to add edge/interior details or manipulate geo static highres full geo add uvs (autouvSOP) limit number of poly, polyreduceSOP (precision errors in ENGINE if npoints/nprims is too high) hou material: shop_materialpath prim attribute (unreal will replicate houdini's) or assign unreal mat: unreal_material prim attr (in UNREAL/content menu//right click on material/copy ref path) pack fragments (assemble SOP) add rest point attribute (VAT ROP needs rest) dopimportSOP (create point to represent objects) use transform pieces to apply motion to high res geo copy template attributes ON (VAT ROP needs orient + pivot) OUT to VAT TOP 3- VAT ROP method to RBD will export a mesh fbx, a pos texture, a rotation texture. textures size min will be: [ n of packed prim * n of frames ] in UNREAL: import VAT passes (pos.exr, rot.exr, mesh.fbx) pos/rot texture: details/compression_setting/vertex displacement (HDR default not needed, exr is normalized) details/texture/filter/nearest mesh: details/import_setting/mesh/vertex_color_import_option/Replace (pivots are stored here) details/import_setting/mesh/Remove Degenerates/OFF details/LOD0/build_setting/Remove Degenerates/OFF details/LOD0/build_setting/full precision uv/ON create material details/material/tangent space normal OFF (we write data in world space) details/material/num_custom_UVs , set to 3. copy/paste material code from houdini VAT ROP/Rigid Vertex Animation UE Code connect outputs (base cd, normals, trans offset) connect (uv1,2, ...) (pass data from vertex shader to texture shader) create material instance from material copy bbox max, bbox min, bbox max1, bbox min1, n frames, from houdini VAT ROP set playback speed to 1 (it's a multiplier ) (framerate fps / total number of frame animation) assign pos and rot texture to translation and rotation textures assign material instance to rbd mesh.fbx --- tested on Houdini 17.0.416, Game Dev Toolset 1.96 rbd_VAT_workflow_v01.hiplc
  4. Hello! exporting rbd sim to unreal using VAT. Issue: active pieces are floating above ground in Unreal. non active pieces are placed correctly. (see images) It seems like the y value of the texture is somehow capped ? how can I debug this? Everything worked fine in the simulation iteration before. I changed a few constraints properties and added a fracture level to the geo. 800 fragments, 240 frames using Game Dev Toolset 1.96 using Houdini Indie 17.0.416 any help appreciated thanks!! solution: bbox_max1, bbox_min1 copied values were wrong (I was using same value of bbox_max, bbox_min, old method depracated)
  5. Hi Odforce! I have a few fetch ROP in series (dependent one on another: sim, mesh and mantra render). The Fetches are sourcing the filecacheSOP/render node. This happen also if using Geometry ROPs chains. Valid Frame Range: Save Frame Range Only (Strict) Rendering locally on my machine When I try to render from the last node of the chain, the last node frame range is applied to all. What am I missing? Thanks for helping.
  6. Hi F1! I am back on this How should I modify the above code to create only one single primitive? At the moment is creating one new primitive each time there is a branching. Is it possible to just keep adding the points to the first created primitive? I have tried to modify the conditions, addprim or primpoint function but didn't find a solution. Thanks in advance
  7. Hello! I am trying to fuse multiple primitives into a single one. This is a branched structure. I get some odd results where I get a single primitive but the ends are also connected. Suggestions? workarounds? Hip file attached with some examples and trials. Thanks! [ polypath SOP, join SOP, fuse multiple primitives, fuse multiple curves / polygon primitive into a single primitive ] fuse_nprim.hip
  8. “I'm having a hard time; I'm preparing my next mistake.” B.Brecht

  9. ok i found that using PolyCut instead of the Fuse/unique solved the problem. So F1 code + separation post solver. Convert line doesn't keep attributes, but is definitely useful in other situation. thanks again
  10. I found the piece of code that i was looking for to modify the algorithm in the creation stage. I asked for a "post" solution at first because i knew some side fx would come along, but now i got everything under control. My code generates one prim at each iteration, the advantage of this approach is that i can separate the various branches in chunks after the creation stage. That's what i do using a FuseSOP set to Unique, acting only on a group of points ("primchunk_node" group). This primchunks are used later for the "flip/unfold effect". F1 code is much better optimized since it creates only one primitive per branch. But in this way i can't separate them afterwards, or at least not with the FuseSOP/Unique approach. I modified the code to create one prim also for each primchunk_node. The only downside is that while tweaking the unfold effect, if i want to change the primchunk distribution i have to re-run the solver. Thanks for helping once again! ///Original - Primchunk selection after solver int new_pt = addpoint(0, newP); int new_prim = addprim(0, "polyline"); addvertex(0, new_prim, @ptnum); addvertex(0, new_prim, new_pt); ///F1's - Primchunk selection in solver int new_pt = addpoint(0, newP); if (neighbourcount(0, @ptnum) == 1 && !@group_primchunk_node ) { // Add new point to existing polyline. int new_prim = pointprims(0, @ptnum)[0]; addvertex(0, new_prim, @ptnum);//addded addvertex(0, new_prim, new_pt); } else { // Create new generation. Seed case + branching nodes + primchunk nodes int new_prim = addprim(0, "polyline"); addvertex(0, new_prim, @ptnum); addvertex(0, new_prim, new_pt); int old_prim = pointprims(0, @ptnum)[0]; int new_branch = prim(0, "branch", old_prim) + 1; setprimattrib(0, "branch", new_prim, new_branch); Attached you find the project im working on, I started from a already done space colonization algorithm. I customized it and worked mainly on the post solver effects to randomize growth speed and other attributes, and on the unfold effect. I will publish a cleaner network, when ill be done! If you have any question feel free to ask space_colonizer_v11.hip
  11. ps. sorry for the mistyping, editing the post doesn't seem to work for me.
  12. ... i can't modify the algorhitm at this time, i will need a "post" solution. so maurits's option are not an option unluckily. mestela's technique is nice but it doesn't support the branched structure, for what i experienced. i'd like to have one prim for each primchunk ( = my id attribute in the example file).
  13. thanks for helping. Maybe i oversimplified the example hip. attached you will find the geometry i am working with. It's a geo created throught a space colonization algorithm, for no branched_structure.193.bgeo.sc
  14. Hello Odforce, i have several curves made of 1 prim every 2 points. Each curve is identified by an id attribute. I'd like to have only one primitive per curve. I tried the join SOP for each id without success. Example hip attached. thanks for helping multiprim_reduce.hip