Jump to content


  • Content count

  • Donations

    0.00 CAD 
  • Joined

  • Last visited

Community Reputation

0 Neutral

About shadowst17

  • Rank

Personal Information

  • Name
    Dominic C
  1. Hey guys, I've got a digital asset that has a parameter to select primitives that's hooked upto a base group parameter inside the assset. I want to make it dive into the asset and select primitives from a certain node. Not the node that's plugged into the asset at the top. Anyone know a way to do this? Any help would be appreciated
  2. Hi, I've got a group node inside a subnet. I've promoted the base group parameter. I'd like to be able to use the selection tool by the group parameter to select primitives from the geomtry that's input into the Group node within the subnet and not the geometry that is input into the subnet. Any help would be great.
  3. Hello, So I have agents that are climbing using a locomotive clip. Each hand pulling the rest of the body up and the other etc. At the moment the way I have it having the pivot set to the hand joint for the first pull up. Then I'm using xyzdist to find the closest prim on the deforming object. Then I use primuv to have the agent hook onto that. This only works for the initial pull up, once it tries to pull it's self up with the other hand it just stays on the same location. I need a way to switch the pivot of the agent to another and keep the already moved up position of the previous frame when the other hand is locked in place. A more advanced way would be able to tell when the other hand is moving etc but i;d be fine with just inputting that data for now. Anyone got some ideas how I can achieve this. I've been trying to get this to work with solvers and getting the position of the other joint on the frame before it needs to change but using that to offset the next hand is not working. I think I was probably on the right track but my comprehension of the solver node is spotty and just gives me a headache.
  4. Quick question, How do I go about creating a colour parameter on a node? There doesn't seem to be a ParmTemplate for it and there's no documentation on this type of parameter for what I can find.
  5. Hello, I have a few questions about creating FLIP Splash particles and what is considered the typical approach to them. One issue I have is that I have a FLIP tank that's quite small. It works great for the actual FLIP(object impacting the water) and the waves tend to dissipate by the bounds. Of course Splash tends to go much further than that. What is the usual solution to this? Is it really just having your original FLIP tank bigger? That seems quite inefficient when you know you don't need that resolution of FLIP at that distance A common question i've noticed but no straight answers. Spray & foam always looks very "ballish" when settled when motion blur isn't taking effect. Is the solution really to have hundreds of millions of particles? I see that Houdini shelf likes to convert them into a volume but that volume always looks too wispy than no matter what density.
  6. I have a few problems with my crowd sim and was wondering if anyone had any solutions. Link to the crowd sim video: A lot of my agents tend to get stuck on the ragdolls as the POP Steer Separate with 'Force Separate on'. I tried using a group that only contained points that weren't in the state of ragdoll but that does'nt work. It seems to run over all agents no matter what I do. Some of the agents end up going through the sandbags. Would be great if I could stop that but another solution might be to have the agents reach a select few positions by the sandbag and transition to a climb animation? I have a certain amount of agents made to seek out certain points and then crouch. I'd like to have them look roughly in a certain direction. I've tried messing with the Orient attribute on the certain Crowd State but it does some weird things. I've found using multiple POP Steer Obstacles Bug out and cause the agents to go "NaN" once both POP Steer Obstacles are influencing the agent. Is this a bug or am I going to have try and consolidate multiple objects into one for the nodes? Preferably I would have liked to have one just for the boats as the values would be significantly different to the sandbags. This did work at one point but randomly it started to cause the NaN issue, I've backtracked a lot to find the problem and still can't solve it. Only thing I haven't done is redo the entire DOP Network. Can I manipulate the weights of certain POP Steers individually for each particle? Say at one point I want half of the agents to have 0.5 rather than 1.0 weight for the POP Steer Wander for example.
  7. Crowd Questions

    Hello i'm trying to learn how to use crowd and I think i've got the basics down but I have a few questions that I can't seem to find solutions to online. How would I go about orienting a certain crowd state to be looking(whole agent) in a certain direction? I was under the impression it was the @orient attribute but no matter what value I feed into that just causes my agents to turn in some weird directions. Sometimes I have agents that are supposed to be standing still and some running. In some cases they get stuck going forward because they're blocked by a few standing still agents. What methods can I go about to fix this issue and have my agents realise they need to turn around or go around? Similar to question 2 I have certain objects I don't want my agents to go through no matter what. Steer Force is not absolute so sometimes my agents will force their way through said objects if they can't turn around sooner due to particle steer separate. Just to confirm this because it seems to be the case with any tests I do. Crowd Triggers through Custom VEX are static and to have them trigger dynamically, I have to use a Particle Lookup instead? I have a feeling questions 2-3 are solved through fuzzy logic but I can't find any sort of tutorial on learning it and the examples given in Houdini are'nt enough for me to get my head around. Any links to good tutorials for it would be appreciated.
  8. Hello, I created this water simulation of water hitting some rocks on the shore. I'd appreciate any suggestions on improvements. My main issue I feel is the white water, it looks like sand and any advice on how to go about making that area in particular better would be greatly appreciated. Shore V1.0
  9. Hello, I was wondering how I would go about having the alpha effected by the refraction with an OceanSurface shader. I need to have the alpha effected so I can comp in a ship I rendered separately in the middle using a matte within the ocean.
  10. Thanks for the reply. The mesh had used boolean at some point so adding a poly doctor afterwards seems to have fixed it. Thought a clean node would have fixed that so never occurred to me to try poly doctor instead.
  11. Hello, How can I go about exporting vertex UV's from Houdini to Maya through Alembic? Only way I've been able to get it to work was by promoting the UVs to point but that drastically changes the UV's. FBX exports them but only if the mesh is static.
  12. Hello, I'm trying to distribute my wave layer tank FLIP to about 8 computers and they seem to cause this(seem image below) when they are brought back in where the sections that meet go crazy. (i'll add a file when I get a chance if the image does'nt instantly spark an obvious solution.)
  13. Hi guys i'm sure there is a really simple answer to this but I as wondering what was the best way to render out Whitewater? I've tried the method that Houdini sets up by default and it comes out looking far more like fog than what i'd consider foam. I've tried rendering them as particles but to get the density that looks good it uses up way too much RAM (only have 24 GB). Meshing them with a particle fluid surface node also look laughably silly.
  14. Hi guys, I've got a simple simulation of a skull that slowly comes out of water and back in that's simulated in a flat FLIP tank and I was wondering how I could go about distorting/blending the final result into a sphere.
  15. Hello, I'm trying to create a slow motion effect using .sim files with cache substep data. Problem is this method seems to be incredibly unpredictable. When save checkpoints is disabled the simulation is correct, when enabled it randomly 90% of the time will simulate completely differently to when it's disabled. Sometimes pressing reset simulation or clearing all in cache manager will fix this but it's not 100% of the time with maybe a 3% chance of fixing the problem. Is there a solid fix for this or is this bug that still hasn't been fixed? I've tried three different versions of houdini. I guess if this is unfixable I could try the FPS method instead but I've found sometimes keyframes get a bit out of sync doing that method.