Jump to content


  • Content count

  • Donations

    0.00 CAD 
  • Joined

  • Last visited

Community Reputation

0 Neutral

About bentway23

  • Rank

Personal Information

  • Name
    Mark Reynolds
  • Location
  1. So, Jesper--if I understand you correctly it's something like the screenshot (pointwrangle vP=v@uv;), then the grouping and smoothing, all after the Remesh? I'm afraid the "set your Position back to original position" has me lost, though. Using the vertexsplit did so much good work. These aren't hero trees, so the small distortion will probably be usable. And, martinkindl, ideally, yes: the problem is that the hi-res was too high-poly to sim, and the point deform from the reduced model to the high-poly one wasn't working--there were too many errant polygons flying off into space. I think this is because it's a library tree model, so there are lots of spindly branches that are four triangles for a large amount of space--I'm not having this same problem with our made-for-this-shot models. The downres'd one, though, still has enough detail to use as the finished model, so we're rolling with that. Thanks!
  2. Okay, I found that "uv" has to be specified, so it's a LOT closer, although a bit (as the thread you referenced said) wobbly. However, right now I'm not using the attribute transfer at all--is there a way to incorporate that to do the final bit of smoothing/restoration/
  3. I'm not quite following how to apply the vertexsplit. I have the setup on the attached screenshot, which obviously is wrong since it's not working, with the vertexsplit applied in the pipe right before the remeshing, and then using that as the source for the transferattributes. I did confirm with the uvquickshade that up through/including the vertexsplit the UVs are intact and lovely.
  4. This is a bit fx, a bit modeling, a bit rendering, so I'm throwing it here-- I am running sims on a tree model. The original tree was too hi-res to simulate, so I had to run Clean and Remesh to reduce its poly count a little bit. It still proved too oddly-modeled to run a point deform on the original based on the sim, but the sim'd model is still high-poly enough so we're going to use it as the final. What I need to do is transfer the UVs, in world space, from the original model to the down-res'd sim'd model. I tried using Maya's Transfer Attributes, which does exactly what I need, but choked on the source model's size. How can I do this in Houdini? (I know there will be some changes, but those should be mostly in the little tiny branches that are okay to lose, and this is not a hero tree.) I assume it will be something like storing the original models' point positions and then somehow applying those points to the end result of the sim'd/cached model, at which point I could do a straight attribute transfer (and although I get this theory I don't know the steps required), or is there a better way (and if not, how would I do my suggested way)? Thanks so much!
  5. The subject pretty much says it all--setting up a local HQueue farm, the server is a Mac, the farm computer is Linux. The shared folder seems to be successfully registered with both computers, and the client shows up in the server's HQueue broser console. However, when I submit a job, under the job details for the child job, it fails immediately, and I get the error: /bin/bash: ./houdini_setup: No such file or directory Any pointers? I've dug around the Linux machine's directories and can find no file or folder with that name. On the Mac I found it buried way in the subdirectories under the applications directory. (The Linux machine is only running Houdini Engine.) Thanks!
  6. **I'm amending this--originally it was just the below question, about making POP particles collide with FLIP sims, but I just was toying with multiple POP emitters/particles in the same scene, and they don't seem to want to play together easier--so my more fundamental (surely stupider) question is: how does one make different particle systems (POP, FLIP) interact? For instance, in the supersimple screenshot, I would want the orange particles (which have initial velocity on x) to collide with the blue ones. I would think, since they're part of the same POP solver, this would just happen. I would think wrong.** I have a puddle of bood, a viscous FLIP fluid. I have POP particle rain (a la https://www.sidefx.com/tutorials/rain-systems-and-wetmaps/). How can I make the rain (the particles--I assume not the instanced geo) collide with/splash the FLIP fluids? I understand how to use POP forces on FLIP sims, but when I put a POPnet and FLIP sim in the same scene, they just ignore each other. Since static/RBD and particle (FLIP or POP) solvers all play so well together with minimal extra effort, I'm sure getting POP and FLIP sims to play together is something simple and dumb. Thanks!
  7. The first of what will probably be 2 - 3 stupid questions today--I'm doing this tutorial to learn about FLIP Fluids: https://app.pluralsight.com/player?course=introduction-fluid-simulations-houdini-2078&author=john-moncrief&name=introduction-fluid-simulations-houdini-2078-m2&clip=4&mode=live The thing is, when sourcing from an object, his create tab (and he's using H14) has a bunch of tabs that H17 doesn't have, primarily force-related (noise, vortex, &c.). This was done just using the shelf buttons. Where did those tabs go? I've hunted around but can't find them living elsewhere. Thanks!
  8. Sim still cooks even after cache

    Thanks! It's even more ornery than that--I actually had to dive into my DOP network and disconnect my FEM solver for it to stop trying to re-solve. I'm used to Maya's "enable" switch, for which there doesn't seem to be a corollary in Houdini. (This is not a criticism of Houdini, just an observation.) Fortunately, disconnecting and reconnecting things is simple enough.
  9. Can't select points in DOPnet

    Alrighty, so I guess I need a touch of remedial help here. When I click the backtick, it says "restart selecting," but I can't actually select anything. No matter what I click on, nothing happens. (Also, it's not saying the "select solid object" thing it usually says--perhaps I have to tell it that I'm working on the "constrained object" field first?)
  10. Hello! So this weekend I cached a FEM simulation and it went swimmingly--sort of. Okay, I didn't do a .sim cache, but a .bgeo.sc cache. The problem is that the file node I have set up in the project to read the cache doesn't read what was written--it keeps recooking all the finite elements, even though it is not hooked up to anything except for the cache directory. (I even tried switching the file cache node I used to write it to "load from disk," and it still wants to recook everything. I pulled the cache into an empty file and it loads beautifully, so I know it actually cached fine. The next step in the project is to pull it in and use it to deform the original hi-res geo. How can I tell the project to stop cooking and just read what it wrote? Thanks for any help, as always-- **Addendum** - creating a whole fresh new file read node worked fine. I guess at some point Houdini nodes get tired and forget where they are? At any rate, I would love to know what I clicked that made it so forgetful, so I know how to fix it when I screw up again. **Addendum to the Addendum** - the new file node worked fine, until I went out of that subnet and back in, and then it started doing the cooking again. So, still not fixed, and the file nodes are still forgetting themselves.
  11. That did it--thanks!
  12. Hello-- I've been working with using distance falloff from reference objects to control various attributes (mainly FEM stiffnesses). I had been using straight 0 - 1 Fits in a pointVop, but wanting more control, I've added a ramp to my setup. My way-simple test setup isn't working, though (attached)--theoretically it's measuring distance from a sphere to the recolored grid, but now my radial black-to-white falloff is not reading any translation in my reference sphere, either through scale or translation--EXCEPT on one tiny little strip of the grid, where it looks like it's repeating and not quite working, but at least some colors change when I move the reference object. You can see on the screenshot attached, too--it looks like the falloff is working, but moving the sphere doesn't move the gradient--and the upper right edge shows the spots where it IS updating, just weirdly. My main, big project has done this a tiny bit--by and large the falloff has worked, but there have been some stray vertices that just won't behave. What am I doing wrong? Thanks for any help! falloff_issue.hip
  13. Can't select points in DOPnet

    I know you can create nodes that way--that's the way I usually do it, which is why I wasn't getting the "select this" options that appear when you create it from the shelf. If you create a constraint node that way, how do you nudge it so it then lets you choose what is affected, the way bringing it in from the shelf does? Or is that just why god made shelves?
  14. Can't select points in DOPnet

    You are my favorite person of the day. Is there a way to do that if you're not using the shelf tool? Or is that exactly why you use the shelf tool? (Shelf works great, this is a future-reference question.)
  15. Can't select points in DOPnet

    I have a question that is so dumb that I can't find an answer in Google. I'm working with FEM objects. I need to select points in the DOPnet to make constraints. I can't select points, just objects. Thus: can't do pinning and stick things to things. I've done this a million times on my home computer, but I'm using a different one now (with H17 installed mere days ago), so I'm sure it's some big obvious setting. Thanks and duh.