Announcement

Collapse
No announcement yet.

Simulation Crash with Particle Skinner?

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • Simulation Crash with Particle Skinner?

    Hi there,

    I´m trying to improving an existing animation with some phoenix love, but I´m having trouble getting it to work.
    The animation consists of a sphere smashing into another one, tearing it apart.
    I´m using an mParticles Pflow and I have skinned the torn mesh via Particle skinner.
    Now I set a Phx source up to emit smoke and particles from the torn mesh and animated the discharge from 0 to 1o and back to 0 over 3 frames.
    But whenever I´m trying to simulate it, it crashes max.
    I tryed emitting from a sphere inside the torn sphere instead, hiding the particle skinned mesh and now it simulates, but of course without the interaction with the torn mesh.
    So I guess it´s an issue with the particle skinner.
    Can you confirm this? Does phoenix just not like meshes with changing vertex count?
    Unfortunately i also don´t have a mesher like frost at hand, so I´d have to make it work without any other plugins.
    I´m currently on nightly build 25297.

  • #2
    Hmm, it could be the way the skinned particles are created. Is it possible to attach a scene that crashes? Hopefully it can be fixed in a couple of days....
    Svetlin Nikolov, Ex Phoenix team lead

    Comment


    • #3
      the changing vertex count will cancel the velocity calculation, but except that no crash should happen. nevertheless, would be good to see the scene
      ______________________________________________
      VRScans developer

      Comment


      • #4
        Hi there,

        so...I stripped down a scene for support and as I was checking if everything was crashing as it supposed to be...nothing crashed anymore.
        Guess it must have been something else in the scene, that was interfering.
        I am having some other issues though, maybe you can help me with these aswell:

        1. How do I get really thick smoke? I want it to look like liquid in water. Would it help to discharge more smoke in the beginning? Or increase the smoke in the PHX Source? Or could I even render the smoke as mesh and maybe put an SSS shader on it?
        Because when I set preview to mesh, It showed me only a solid block from the simulators bounding box, instead of only the smoke as a mesh.
        I´m actually going for a look similar to this vimeo video:
        https://vimeo.com/107876630
        This is actually simulated as liquid with phoenix, so I guess that would be the best solution. I just really like the high detailed smoke look in huge scale sims (nukes and such) and would like to recreate it for my purpose.

        2. Whats the word on render elements for atmospheric renderings? Has there been any development? Because I really can´t live without DoF in post and without Zbuffer I´m out of options right now for smoke/fire.

        3. I´ve never used this before, how do I get the adaptive grid to work with liquids? Do I simply have to lower the threshold? or should I just do a lores sim and experiment with the grid size to get it right for the hires sim?
        Last edited by ben_hamburg; 23-06-2015, 05:20 AM.

        Comment


        • #5
          can you give an image how the geometry looks like? in the most cases the breaking of objects is not a problem, they are prepared in broken state, but not rendered and the simulation is driven by the pieces from the very beginning.
          ______________________________________________
          VRScans developer

          Comment


          • #6
            Click image for larger version

Name:	Torn_Mesh_Phoenix_Liquid.jpg
Views:	1
Size:	99.1 KB
ID:	856388

            In this case the mesh is not prefragmented, the tearing happens via the particle skinner "on the fly". In my original scene I had several skinned meshes that were ripped apart via an mParticles Glue setup.
            In the stripped down version I deleted all but one mesh, so maybe there was just too much going on in the background.

            But as you can see from the screenshot, the solution seems to be working now.
            I´m just gonna sim in a stripped down scene and then import the phoenix grid into the original scene to render.

            Right now I´m using the liquid solution, emitting from a sphere inside the torn mesh for one frame, that makes some nice splatter effect.

            I´d still like to know how I can achieve the effect of very dense smoke, any advice on that?

            Comment


            • #7
              About the very dense smoke - when you switch to mesh preview/mesh render mode, make sure your surface channel is set to smoke, and the surface level is a value between 0 and 1 (or more, if you emit smoke above 1). Without using a mesh render mode, you could indeed discharge smoke from the PHXSource with a higher value than 1, but then you have to make sure you render using the Spherical sampler, as the Linear might cause visible grid artifacts.
              Svetlin Nikolov, Ex Phoenix team lead

              Comment


              • #8
                Ah, right, the surface channel...never had to use this before.

                But just to clarify:

                Those big scale smoke sims just produce a lot of smoke because of the temperature I guess, right?
                So if I want to recreate the look with just smoke (and no fire)...

                1. I´d discharge with a higher value than 1
                2. I´d make sure me scene scale is setup (or the phoenix scale settings) to match those values I´d get with a nuke explosion or something like that.
                Last edited by ben_hamburg; 24-06-2015, 02:17 AM.

                Comment


                • #9
                  Yup, these will help get a very thick smoke. Only one thing - smoke density is not related to the temperature, especially if not using fuel and burning - the smoke is just as much as you discharge from the source If course, render-wise you can still regulate the visible density using the Simple Smoke factor.
                  Svetlin Nikolov, Ex Phoenix team lead

                  Comment

                  Working...
                  X