Jump to content


  • Content count

  • Donations

    0.00 CAD 
  • Joined

  • Last visited

  • Days Won


bunker last won the day on September 3 2017

bunker had the most liked content!

Community Reputation

150 Excellent

About bunker

  • Rank

Personal Information

  • Name
    Julien D
  • Location
  • Interests
    DOPs, volumes, fire, smoke, FLIP, Bullet, particles, VEX, Python, lighting, rendering, comp, coffee, travel, photography, Linux, ...

Recent Profile Visitors

8,701 profile views
  1. How to apply Turbulence to Pressure field?

    to visualize the pressure field go to smokeobject/smokeconfigureobject/pressure_data and plug a scalarfieldvisualization DOP or you can create your own field with a gasmatchfield, copy the data and use the same scalarfieldvisualization
  2. you probably want this which should be default IMO: You can save it as default in >Edit->Preferences->3D Viewports.
  3. Color drops (flip)

    by COPs I think you mean CHOPs You should have a look at the ripple solver before going full FLIP, it's much faster for what you need.
  4. Convert Slice images to Volume

    yes, that works great.
  5. Wax/chocolate tornado

    You can kill FLIP particles just like in POPs. On the FLIP solver, turn on reap particles and age particles. On the source, you can set the life expectancy and life variance.
  6. Convert Slice images to Volume

    cool, I didn't know about the "COP2 Network SOP" How do you load more than one image at a time though?
  7. Convert Slice images to Volume

    it works brocoli.zip - Broccoli MRI scan gif : https://imgur.com/gallery/KwekIF7 - convert GIF to PNG sequence : https://ezgif.com/split/
  8. Convert Slice images to Volume

    here is a working example, I'm creating the images with a noise in COPs bunker_imageseq2vol_001.hipnc
  9. Convert Slice images to Volume

    That should be fairly simple do achieve as long as you have a your images in an ordered list (img_1.tif, img_2.tif, img_3.tif, ...) in a volume wrangle, use the texture function to read the image RGB values > luminance to convert to greyscale and set the volume density using that value. getbbox, to convert x,z to u,v coordinates and the y coordinate as your image index. something like this: // number of images int lastImageIndex=30; // get bounds vector bmin,bmax; getbbox(0,bmin,bmax); // create u,v,imageIndex(y coordinate) float u = fit(@P.x,bmin.x,bmax.x,0,1); float v = fit(@P.z,bmin.z,bmax.z,0,1); int imageIndex = int(fit(@P.y,bmin.y,bmax.y,0,lastImageIndex)); // create imageName, sample RBD, convert to greyscale and write to density string imageFileName = “/path/image_”+itoa(imageIndex)+”.tif”; vector color = texture(imageFileName, u, v, "pixelblur", 2.0); float greyscale = luminance(color); @density = greyscale; hope that makes sense.
  10. Swaping Low res pieces with high res

    the Assemble SOP is a digital asset so you can have a look, it's using a Pack SOP inside. I fixed the issue by creating the name prim attribute inside the forloop but left the pack SOP after the loop. Something got changed since H17.0.416, I couldn't see that at home since I have a slightly older build.
  11. solver-initializing attributes- basic question

    v@v is the initial velocity so you need to initialize v outside the solver SOP gravity is a constant so it doesn't matter if it's set on every frame inside the solver SOP or not Solver_fake_physics_demo2.hipnc
  12. Hscript problem

    hscript: fit($F,1,10,0,1) vex: f@attribValue=fit(@Frame,1,10,0,1);
  13. Houdini 18 Wishlist

    what if the self buttons were creating DOP setups inside one SOP network instead of having a Geometry node, a DOP network for simulation and another Geometry node to import the data and save to disk. It makes much more sense to see the data flow from top to bottom in one network without having to jump to different levels for no reason. maybe it's just me... grains.hipnc
  14. Swaping Low res pieces with high res

    the DopImport node has an Object Mask, that's where your RBD object name should be set, in your case HIGH_RES_PIECES or LOW_RES_PIECES, so *_RES_PIECES will work. if you simulate with hires pieces, the collision shapes (convex hulls) are created automatically in DOPs, you can see them if you toggle on "Show Guide Geometry" (blue shapes). That might work OK in some cases but you don't have any control over the collision shapes/details. I ran 2000 frames simulations for lowres and hires, both are very similar ( lowres: 56s hires: 58s). Those times are for simulation + writing the points to disk only. Displaying objects in DOPs or in SOPs is a completely separate thing, that's much slower. Also you shouldn't turn on "Show Guide Geometry" while simulating and displaying the objects inside DOPs, it makes everything much slower too. I feel like I should make some tutorials Setup3_03.hipnc
  15. Swaping Low res pieces with high res

    Just needed to promote the "name" attribute from point to primitive Setup3.hipnc there seems to be a small bug (or feature?) in the connectAdjacentPieces, the connection type modes are inverted. maybe someone else knows what's going on? I usually create my constraints in a wrangle, it's a lot more clear IMO. hope that helps