Welcome to od|forum

Register now to gain access to all of our features. Once registered and logged in, you will be able to contribute to this site by submitting your own content or replying to existing content. You'll be able to customize your profile, receive reputation points as a reward for submitting content, while also communicating with other members via your own private inbox, plus much more! This message will be removed once you have signed in.

jim c

Members
  • Content count

    544
  • Joined

  • Last visited

  • Days Won

    1

Everything posted by jim c

  1. Thanks for doing this, the video looks like it covers all sorts of really cool techniques. Can't say how nice this is to see.
  2. I've got this growing tree model, and I wanted to animate the growth sequence starting where it pushes through the dirt. Using the GrainSource SOP and the PopGrains DOP I can get some decent results, so I now have some clumpy particles. Once I have those what's the best approach to rendering them? Would it make sense to use them as the source for a voronoi fracture sop and have the particles drive the various pieces? Or is there a better approach?
  3. And here's the result now, a couple of captures of different frames
  4. I've got a nice little lsystem that is producing a tree. The goal is to animate this process. So I'm going from this to this over around 360 steps. So far so good. The only real problem I'm having is how to best generate UVs while it's "growing". To add nice detailing to the geo I use the UV data to apply some deforms. My problem has been that to get this to work nicely I hand built the UV map in Modo using the last frame of the tree's geo like so: Then I used the UVLayout SOP to layout all the islands horizontally along U, and then gradually increased the V values as they run along the new U coord the to "twist" the UV, which then makes it easier to deform it to this: I don't need to do this to all the geo, just the lower trunk portions, the higher branches don't matter they'll be hidden by all the leaves anyways. So, the question is, how to apply UVs like this for the rest of the previous intermediate frames? I've tried to use a wire deformer to get the final geo into a similar position as the growing geo and then use an attribute transfer SOP for the UVs, but that doesn't work so well. Another idea I had was to use a foreach block to isolate the individual branch pieces and then somehow regenerate uvs for each piece, but while I can get the foreach to work I haven;t figured how to re-apply the UVs. The geo us being created with PolyWire Sop which I think I need to stick with as it gives the best results at joining everything together. I've cached out various pieces, so I have access to the underlying lsystem wires (prior to it being fed to the PolyWire sop) if that helps. Any ideas would be most appreciated! Thanks
  5. OK so I more or less figure out how to get this to work. First problem was that when I switched to the UV viewport, the UV's didn't show up so it was difficult to tell what was going on (they were there however, I could export the model and see the uvs in Modo). Next step was to delete all the branch geo as it was being created. Using a delete Sop with an expression that checked against the gen attribute (that's supplied by the Lsystem) pretty much did that. Next I used the UVLayout Sop to layout out the various UV islands that the PolyWire SOP creates more or less horizontally. This is done with UV at the vertex level NOT at the point level. After this was done I used an attr promote sop to calc the max UV value and store in a detail attr. Created a Attr Wrangle Sop to adjust the uv.y value so its gradually increasing as a function of uv.x/maxuv.x value. At this point we've got twisting UVs, not matter what frame we're on! The rest was re-suing the existing technique I'd worked out for the static hand built uv's I'd made in Modo. It's not perfect, when you promote the vertices UVs to point UVs there's some distortion that happens at certain areas, but I can live with that, maybe some clever camera angles to avoid focusing on those spots.
  6. if the grid is flat, then you could use the normal() expression, pass in the grid's path, pick any prim, 0 should be fine, and then use that in the Knife SOP's direction parameter.
  7. Couple of things you can do. Look up "Box Modeling" and you'll see how a simple box can be divided up and manipulated into more complex forms, like a head. To start with that drop down a Box SOP. You can then go to the Polygon shelf and use the cut lines or edges (I can't recall exactly what it's called), but it lets you interactively place new edge loops in your geo. That lets you add more detail to the mesh, where you want it and form there you can use the edit sop (typically in Point mode) to manipulate (transform/rotate/scale) points. If you just model one side, along the x-axis for example, then you can delete the other side, and use the Mirror SOP to mirror over your geo. You can use a series of metaballs all merged together and fed into a VDB Volume and then converted into a mesh (Convert VDB SOP) as another way to set up some core forms, and then use the retopo tools to give proper edge loops. There's the NURBs tools as well, but those are probably more hassle then they are worth.
  8. Thanks, I wasn't thinking along those line, it's a project for my wife, she wanted a magic growing tree Unfortunately it's taken forever to get done. Id be happty to write up the basics of how it works though (most of the core is coming up with some nice L-system code).
  9. I have a chunk of geometry, lets say a tube, that's off in space, and it's been rotated around. I'd like to "normalize" it, to translate it to the origin, and then rotate it such that's point straight up in the Y axis, perform some operations on it, and then put it back in it's original position and orientation. I know how to normalize the translation, but I'm stuck with the rotation stuff. Is there a sop that does this or some vex that would point me in the right direction?
  10. Wow! That's some super cool stuff - the animation features look really, really cool, very curious to see how these work!!! Congrats! The ocean on the twisting strip looks really cool, that's a great demo. Compiled SOPs!? Hardware accelerated SOPs? some really interesting stuff in the release!
  11. I'm trying to generated noise that flows along the topology of the geometry. But I'm missing something obvious about how to do that. The red arrows indicate how I'd like the noise to flow. The geo is generated by using the polywire SOP. The only other thing I can think of is using COP to generate a texture map with noise and then read the data from that.
  12. Thanks guys, both of these are real interesting ideas.
  13. I'm looking to upgrade my Nvidia Quadro 4000 (which I have never been that thrilled with). Since I have a 2011 Mac Pro, I'm trying to avoid a card that will require me to do things like adding a 2nd PSU (which appears to be a requirement for NVidia cards like the 780), so I'm looking at the ATI 7950. This seems to be a decent compromise, but I'm curious to know if any of you guys have used it with recent versions of Houdini, say H14 or H15 and how did it perform. I found a pretty positive review by someone on the sidefx site for H12, but nothing more recent.
  14. So I broke down and ordered a 970 from macvidcards - fingers crossed, hopefully this will be a nice improvement over the thermal disaster that is the Quadro 4000!
  15. If I have geometry that I'm passing into a CopySop that is changing over time (specifically a basic leaf shape that may have it's shape changed based on stamped values from the copy sop that in turn pull this from attributes on the template points) does it make sense to even use the Packed Geometry option in the CopySOP? IN the sidefx docs I came across this: " This is important to remember when copy-stamping packed geometry. If every instance of your packed geometry is unique, you get no memory or performance benefits. In fact this will use more memory than "real" geometry would, because each packed primitive has its own overhead. " (https://www.sidefx.com/docs/houdini/model/packed#how-to) Since the leaves are different for each template pt, am I sort of defeating the purpose packing them? Or would I use some other workflow? Total number of template points (when the tree is fully grown) is around 110K-150K pts.
  16. Maybe take a look at Peter Quints tutorials on Vimeo - some really good stuff there. 3DBuzz has (or had) a very nice set of stuff.
  17. Thanks TitanX is out of the budget According to the NVidia site (http://www.geforce.com/hardware/desktop-gpus/geforce-gtx-titan-x/specifications) its rated at 250W, but you're running it on a MacPro with no mods to your PSU?
  18. Hamp, thanks for tip. I did look there but it seemed the cards I kept looking at were over the power limit! But the one you mentioned seems to be fine (according to nvidia website it uses 145W). Will definitely look into this guy. How has it performed heatwise? One of the other reasons I want to ditch the Quadro is it's an absolute thermal pig and run's hot as hell. Atom, yes I think the PSU does have that much, apparently (this is what I found online, so take it with a grain of salt) it's an issue with how the Motherboard/LogicBoard are designed by Apple such that the maximum power draw for PCI devices is 225W.
  19. I've got a tree that's generated via the L System SOP. I'm extracting the end points of the branches to grow leaves on them, after running the points through a SOP Solver to add some custom pt attribute data used in growing and animating the leaves. I notice that the point ids kind of jump all over the place as the lsystem grows over time. Is there a way to get a stable set of points such that the oldest point is point 0, and the next is point 1, and so on? Is it just a matter of sorting the points by age?
  20. Thanks. I sort of found a work around (I was able to use a bounding sphere to select a small group of points and then since the point I wanted was closest to the ground, just sorted by Y). But ultimately it would be nicer to have a general solution.
  21. Thanks, but here's what I'm talking about - I want to track one specific point as the lsystem grows. See these screenshots from your file: So that's frame 1 above, and that's the point I want to track - I've resorted by lage Now as the frames changes (frame 14 above), the point switches as well, now the point I wanted to track (circled in red) gets missed and instead I'm left with a different point. And then a few frames later it switches again: This is what I'd like to prevent. I think what I'm after is something like how a particle system treats id, so that they are stable as the system grows?
  22. I've made a wings rig with a lag effect. The rig is made up of 12 wings (6 per side), with a curve for each wing. The curve is controlled via 3 nulls for each curve (basically a 3 degree NURBS curve). Each wing has 6 bones which follow the curve as deformers for the geometry. This all works fine. To add some lag, I used the shelf tool's Lag effect applied to the 3rd control point for each curve (this is the point at the end of the curve). This causes the end point to lag behind and gives a sort of "flapping" effect. Looks nice so far. However the performance is crap, it really slows the rig down a LOT. The performance monitor flags a couple of nodes in the CHOP network as being prime offenders but I'm not sure how to tackle this. The shelf Lag effect creates a little network each time, and I've basically got 12 sets of the same thing. Is it possible that I can combine these somehow? Or is there better way to simulate the effect when you have a number of items you want to lag? Both of the red highlighted objects have their "Compute" parm set to Full Transform.
  23. Thanks Edward. I looked into that and it seems I can bring most of the parts into one node, but at the end of the network that the shelf Lag tool creates, it seems set up to push the data back into one node with using pretransforms or something like that. I'm not sure how that works or how to get it to apply to a whole bunch of different objects. Is the network that the shelf tool generates the only way to use the lag CHOP? Is there a simpler setup? Thanks again!
  24. Is it possible to modify bones within a Digital Asset using the Bones Shelf tool if the DA is locked? I made a simple test and it didn't seem to work, but perhaps there was a step I missed? I'd like to have a rigging tool that has a template rig that you can "pose" with the bones tool, and then click a button and it generates a custom DA for you based on the template bone positions. Using the Bone Shelf tool allows for more intuitive manipulations than just the translate or rotate handles.
  25. I'm trying to add long(ish) hair to a female character using the Fur tools. I can get the hair to work, I can get the various length/density/cut maps to work, and after fiddling with the Stiffness attributes I can get the fur to sort of resemble hair, but it's not very easy to control. So, I figured that guides would be the answer, except for the life of me, I don't understand how we're supposed to do this. 1) is it the case that you can feed in a a geo object that has the guides into the fur object created with the fur tools? Because if this is the case then I don't see how this can possibly work - it certainly doesn't for me. 2) If the above is incorrect, then is the understanding that we have to edit the fur asset that's created by using the fur tools? I.e. we need to go to the guide sub node object and then edit that? 3) if #2 is correct, then how to work with a reasonable set of guides that can be manipulated? According to the docs (and what I got to work with a simple test that uses the fur SOP, not the fur asset that's auto created for you) you need to feed in one guide per point on the skin geo. However if you have a head that's got 5K pts, but you're using density/length maps to control the hair, then you end up with LOADS of superfluous guides. If, on the other hand, you use a group to limit the skin surface, from this will work great for limiting your guides BUT it then seems (from what I can tell) that when you add dynamics the hair wires fall through the skin that is NOT in the group! So I'm completely confused as to what to do here. Is there something obvious that I'm missing here? What I'm ultimately wanting to do is to be able to control the initial state of the hair before it gets fed through the simulation to help get to decent initial pose. Perhaps there's another way to do this that I'm not aware of. I've been through most of the sidefx docs on this and watched the Fur Masterclass a couple of times. If it matters I'm using Houdini 10.