Jump to content

Whatsinaname

Members
  • Content count

    96
  • Joined

  • Last visited

  • Days Won

    1

Whatsinaname last won the day on October 26 2014

Whatsinaname had the most liked content!

Community Reputation

3 Neutral

About Whatsinaname

  • Rank
    Peon

Personal Information

  • Name
    Ernest
  • Location
    Hole in Ground
  1. Isolating collided paricles in FLIP

    I'm trying to isolate those areas of a FLIP sim, that have been in contact or collided with some geometry. My problem is I've got something that works for the particles as long as they are close to the surface, but as I intend to create some dripping and surface tension effects. At the moment, I am assigning an attribute to the points which are colliding, which allows me to isolate the fluid on the surface, but of course that gets overwritten each frame. There must be a more elegant solution for this, but I cannot figure it out at the moment, so any ideas or feedback pointing me towards a better solution is welcome. collideStick_v001.hipnc
  2. Instance and point colors/groups

    Telling from your description, it feels like there is something going wrong with your packed promitive workflow. Just watch the following video, as far as I remember, they are more or less exactly explaining your problem and what you can do about it:
  3. Heightfields causing OpenCL issues on Farm

    This is very good to know, Thanks a lot. It also seems like OpenCL issues are affecting Flipbooks (= Flipbook crashing) on some constellations (speaking about Xeons running Linux).
  4. Heightfields causing OpenCL issues on Farm

    I presume the machines on the farms are mostly Xeons, running on Linux.
  5. In H16.0.573 I created a terrain using Heightfields without caching the end result, as everything computed fairly quick. Unfortunately, that scene didn't seem to work on the farm due to an OpenCL issue, clCreateProgramWithSource() failed. The machines on the farm do not have graphics cards, so OpenCL won't work in this environment. Therefore I debugged my scene and it seems like some Heightfield SOPs are using OpenCL. I was too lazy to inspect the individual nodes, but it was great if it was easier to see if a certain node will try using OpenCL, as this won't work on many farms.
  6. Change emission rate in FLIP simulation?

    Look inside the Dopnet, there should be a source volume node, which has a Scale Source Volume parameter. Increasing that one should do the job.
  7. Flat tank sim looks like in slow motion

    Are you using stickyness? If so, try deactivating it and try again.
  8. ConvertVDB- no density channel

    Well, if it is working in your test scene, most certainly there is a glitch in your original scene. go through all your nodes and check how the data is changing. But without seeing that original scene of yours it is almost impossible to tell you want went wrong.
  9. Q about disappearing fluid

    Actually, Particle Seperation should be your major control regarding resolution. Radius scale can be used to so some minor changes, but I usually try to leave it as is.
  10. Motion blur with SDF liquid fluid

    Justr out of interest...what do you mean by SDF liquid, can you clarify that? Signed Distance Field liquid makes no sense to me.
  11. Heya, It seems like I'm getting nuts with this one... I have a fast moving fluid and I''m trying to get some detail into the leading edges using disturbance, which simply won't work or at least not in a satisfying manner. In 99% of the cases, Disturbance kicked in at my fluid's tail, where there is little density left, moving a lot more slowly than the leading edge where i want the effect to happen. The result is low density areas being pushed too heavily resulting in beautiful streaky artefacts and no disturbance whatsoever in the leading edge. I'm applying my disturbance directly to the velocity field and first I tried using temperature as control field. In order to be doing so, I measured my max temperature, which is highest in the first frames at the fluid's leading edge, and remapped the control field to apply some really nuts amount of Disturbance to just that area. Result: Nothing. So I played with that curve, but all I came up with was the tail being disturbed and the leading edge being a boring smokepuff. Therefore, I added in a speed field, writing just the length of my velocities into it and used that in order to control Disturbance. Although it seems to be working slightly better than temperature, I still cannot grab my leading edge. It feels I'm still missing anything and don't know what it is. Maybe also worth mentioning I have a min/max substep setting inside my solver of 1/10 and additionally applied a gas repeat microsolver to the pyro solver to make the whole thingy move faster. Deactivating the Gas Repeat seems not to have any influence on Disturbance in my scene. Any ideas? EDIT: It might have something to do with the scale of my emitter. It is rather small (sphere- ish, scaled to 0.1), scaling it up to 1 works better, but my Disturbance block size is also small enough to work on a smaller smoke plume. Strange. disturbance_check_v001.hiplc
  12. Pyro Smoke Artifacts

    Strange. I am getting these artifacts, too. But without using any disturbance (just density, vel and temperature with a tad of dissipation). It also feel like these areas are not behaving correctly, as the smoke just stays there while the rest of that smoke is getting pushed upwards by density. It was great to know what exactly is causing these streaks, not only that it happens when there is too much dissipation (which I already knew), but what is happening under the hood that is leading to those artifacts. Any ideas?
  13. I've got a simiar issue here. I was wondering if there was any more efficient way of copystamping instances as I created a scene where I've got millions of points and instanced some packed primitives onto these points. Rendering a frame is taking ages, as well as baking the instance part. My setup is slightly different than the one posted by Matt, but it basically it does more or less the same (the main difference is I am not using any alembics and I'm randomising my instances using a stamp on a switch node's input value). The thing is, it feels rather slow. Does anybody have an idea how to make it any more efficient? Here's a basic example of what I'm doing using a less complex setup: packedInstances_v01.hiplc
  14. Heya, Crating a wide ocean surface using Houdini Ocean Waves, I encountered an issue where my ocean surface is being intersected by the underlying ocean volume (see image). I was wondering what might be causing this, maybe my wave displacement is too extreme (even though it doesn't feel that strong), or is there maybe a way toimprove the overall volume sampling for the ocean volume shader in order to get rid of these issues? Thanks for any suggestions.
  15. To put it plain and simple, I'd go for a cloud solution and as soon as I got more and more jobs coming in, I might find it makes more sense to switch to my own renderfarm. Maybe get yourself one (or two if you can afford it) extra machines for flexibility reasons. Especially when starting a new business you never know how it is going to develop in the long run, so there is no point in investing into a farm you probably do not need. You can buy your own machines as soon as your business makes enough money you actually can afford a farm. As long as you cannot, a cloud makes more sense to me.
×