Jump to content

redpaw

Members
  • Content count

    32
  • Donations

    0.00 CAD 
  • Joined

  • Last visited

Everything posted by redpaw

  1. Digital Assets and the Tab menu

    Any idea how to query that "Tab Submenu path" string on an HDA using python? cant' seem to find that bit of data anywhere?
  2. Hi, I was wondering if there is any way that I am not aware of to cache or precalculate the collision data that gets calculated at the beginning of a packed Bullet simulation. I've got a scene with approx 11,000 rbd objects from a shatter as a packed fractured bullet rbd, and it takes approx 5 minutes to get thru the first frame where it is I assume, building the collision convex hulls for the geo. I would assume there's some way to cache that and re-use it? but I don't see anything. The startup time on a sim is killing me. Any ideas? thx -redpaw
  3. fracture animating object

    Knowing the way houdini users are, I know there must be some way to pull this off, I just don't know how complicated it could get. I have a walking figure geo cached. Is there a way to pipe that animated cache into the fracture system, so that it breaks off pieces and they turn into RBD's but the main chunk still stays animated? I'm thinking this may end up being some sort of cyclic nightmare where it may be impossible to plug the result of a fracture back into its own input. I think it should at least be possible tho to latticeSOP the pre-broken geom as non solving RBD objects until they "kick off" and then when that happens, turn them into breakable objects so that when they fall and collide, they can sub-crack recursively. if anybody has some tips as to the workflow I'd have to try out to make this work, that'd be awesome. thanks in advance -johnc redpawFX
  4. I've got another bit of weirdness I'm hoping someone can help me with. Attached is a test file I've put together. What I'm trying to do is auto fracture at sim time with multiple iterations, and have the texture UVs stay "sticky" to the interior surfaces once they are applied. Ive accomplished this by inserting a UVtexture node right after the inside group creation point in the voroni fracture sop within the rigid body solver OTL. I tried a bunch of different ways to do this, but this was the most sucessfull except for one issue. It works with native houdini geo, or stuff you completely UV in houdini, but if you are trying to use an already uv'd OBJ from a file as the input, it seems to need the "uv" parameters explicitly set, like with a point sop the only way I have found to get the fracture time UV application to work is if i put a point SOP on the object before the dopImport However, the point sop, (with add texture on) ends up squishing the UV's at the borders and ruining the good uv layout in some cases. this is illustrated in the test file. I'm sure there's a way around this.. anyone? thanks in advance -johnc shatteringTexWeirdness.rar
  5. I've got a particle sim I'm meshing with openVDB, I would like to export this with motion blur to an alembic cache to render in maya/arnold. simple example, not using particles, but yeilding the same results.. make a torus, and add rotation animation xform node. convert to VDB using vdbFromPolygon convert back to polygons with convertVDB node. add a trail sop to compute velocity. export using alembic rop. If I skip the convert to vdb and back and just export the rotating torus, I get subframe motion based on vertex velocity in the alembic file without exporting extra motion blur steps in the alembic file... however with the openVDB stuff I end up with static geo inbetween frames. even though it seems to be creating proper vertex velocity. I can export this as a realflow mesh and it will pass the velocitiy correctly so is this a problem with the houdini alembic exporter? or user error on my part? I'm writing out 20+ mil poly meshes so I'd rather not have to subsample them and write out the data 2-3x just to get motion blur.. at that point I might as well go back to realflow meshes instead. any help would be appreciated. thanks in advance -johnc redpawfx
  6. Sorry, its been about a half a year since I read this topic.. it may not be a perfect solution, but basically, what I did was inside the solver itself where it splits and generates the new fracture geometry, it holds the inside/outside groups... I inserted a uv texture node on the inside section so on creation, it applies the tex coords.. main issue is that the projection is not perfect , there will be some stretching etc.. but if you just need tex coords to be able to map a 2d noise or something it should work pretty well. I'll have to dive into houdini later and see if I can't export something to share, but this hopefully might put you on the right track. -redpaw
  7. HOT in MR for Maya

    I'm trying to get this to compile using the latest source, slightly modified for proper linux compiles, but I'm running across a problem. I get a full compile on the deformer without errors, however it seems that libImath.so is needed in addition to libblitz.so and libfftw.so when I compile it, it links and seems ok, but if I run ldd on the plugin, it shows me that its looking for libblitz and libfftw3 but libImath is nowhere in the list.. thus when I load it into maya I am getting an undefined symbol error, I know its finding the .so when linking, because if I change the name it errors out at compile time but for the life of me I cannot figure out why I can't get it to work. I'm compiling in fedora 15 with a custom gcc version for maya (4.1.2), could that be the problem? anybody with a little more experience debugging stuff? help! thanks -johnc
  8. is there RealFlow-houdini11?

    I just got the official word that the latest and greatest houdini-11.0.658 plugin is being uploaded to nextlimits servers today. Hopefully this should be for all platforms, but at least Linux 64 -johnc
  9. New autodesk product features announced...

    As a mainly maya user concentrating on FX stuff, This update makes me just want to use houdini more. Especially if all the dynamicss plugins aren't available for linux. Oh well.. we'll have to wait and see.. -redpaw
  10. Hi guys, quick question. Is there a way to set up and use an attribute to flag a piece of geometry for only a certain number of frames? in code, I'm thinking like add an attribute called separated, set it to 0, then modify it with some expression to set separated to 1. then for every frame after that, if separated is already 1, it should always stay as 1, even if the original modification expression would evaluate it back to 0 again. basically a one way switch attribute. I know the nature of sops may not handle this, so is this a case for a sop solver within dops? if it would be possible to do it in sops tho that would be ideal. if anyone has any ideas that'd be awesome. thanks in advance -johnc
  11. Yes!, actually this afternoon I got back to this problem and after watching a video or two from peter Quint on houdini UV stuff, he showed exactly the same problem I was having, and the solution was to be using vertex uv's, so I gave it a shot and after a little fooling around with the placement of the nodes again, I got it working now. so to change the file I submitted earlier I added a vertex sop to the input of the dopImport, put another vertex sop before the texture sop in the voroni fracture otl, and had to remove the switch sop after the texture I put in since it didn't work anymore. I was waiting to try it out on my actual non test file before I posted again to see if it fixed it in all cases, but it seems that it has for now. thanks ! -johnc
  12. I would like to use a simple sop solver setup to scale the pieces of a fracture object when they register an impact right now I'm trying to create an impact group with dopnumrecords(".", $OBJNAME", "Impacts", "Impacts")>=1 to activate the group and then pipe that into a multi solver with a sop solver that scales down everyting in the group "impact" but it doesent seem to work, the same setup works if I set it up with a single RBD object, and dopnumrecords(".", "realNameOfsingleRBDobject", "Impacts", "Impacts")>=1 but it seems to have to be different with a fracture/glue object and / or using $OBJNAME instead of the real string name of the object any help would be greatly appreciated thanks -johnc
  13. dopnumrecords impacts to sopSolver issue

    Hmmm, well I am still having a bit of a problem with this, I'm closer, but I still think I'm doing something wrong here. now it registers the dop impact and sets the group for true for all the objects and the sop solver then scales every object at the same time. what I want is for each object to scale as it enters the group, because it registers its own impact. I've attached a simple demo file, if anybody can point me in the right direction that'd be awesome. thanks -johnc sopSolverTest.hipnc
  14. fracture animating object

    AWESOME, I knew there had to be a way to do it. :-) I will begin dissection.... now. thanks so much -johnc
  15. fracture animating object

    unfortunately, not yet, I'm still working on some ideas, but am currently working around the problem. there are some videos I've seen from Igor Zanic that sorta does what I want to try to achieve. Igor? you out there? -j
  16. fluid (smoke) emission rate..

    Ok, so racking my brain here. but short of any custom stuff, is there no way to set the density amount per frame from either a source, or a object.. like say I want to have a large sphere emit a density of .1 into a smoke box? or a fluid source object , sputtering a noisy density amount into a smoke container? either I'm missing something on the default shelf tools, or its something i have to code up myself.. any hints? thanks johnc
  17. fluid (smoke) emission rate..

    Hmmm, I looked at this again this morning, and it seems the pyro setup uses a multiplier on a subnode inside the pyro solver tool itself, it has a "source density" control which basically multiplies up the density of the source .. its the gas calculate node called "mix in source density" I still find it weird that this is the only way to emit a certain amount of density into a fluid grid like this.. I guess I'm gonna start working on writing my own "emitter" setup to do what I want. anybody else got any ideas.. I will probably be referencing Peter Claes's work as well I'd imagine. -johnc
  18. trying to create a nearest neighbor vex sop.. I've put together this little snippit of code.. and it compiles.. but doesent return proper Point ID's to the array. I'm just scattering points on a surface and trying to get the closest 1-5 points to each point to use in a copy sop. sop neighborhoodGroup(string fileName = ""; int maxSearchPoints=5; float searchRadius = 1;) { float n[] = {0.0}; resize (n,maxSearchPoints); int handle = pcopen(fileName, "P", P, searchRadius, maxSearchPoints); int i = 0; while (pciterate(handle)) { int pointID = -1; pcimport(handle, "PT", pointID); n[i] = (float)pointID; i++; } addattribute("nearest", n, maxSearchPoints); } I'm passing the string op:`opinputpath(".",0) to the fileName parameter, which works in another file / thread I found on odforce. Not sure if its possible to ge the point number from the pcimport command? is there any (other) way to do this? I tried getneighbor but that relies on connectivity and I don't have connected points. its just a cloud. thanks in advance redpaw
  19. figured it out.. pcimport(handle, "point.number", pointID); does exactly what I need.
  20. Is there a way to override the file open and save dialogs with a custom one ? to support naming conventions etc? alternatively we can write one in python , but how do I go about re-directing the menu items to point to the new command to build the window etc.. any help? thanks -johnc
  21. Power Noise volume shader

    Not sure if I understand this correctly, I'm trying to write my own version of this same thing. Using volumes (iso offset spheres) I'm trying to copy them to a particle simulation but I am not sure if what you guys have been saying is correct or not.. Is there a way, using the copy sop, to generate copies of the volumes and then have the volume shader stick to the transforms of each copy using object space? I've tried a few different things to no avail. I would much rather have the visualizaton of the volumes directly in the viewport instead of working with the instance Object. Does the instance object work differently and actually pass the proper object space transforms for rotation and translation . I'm using a vex shader that pulls most of its noise generation from the powerNoise thread but no matter what I have tried, using wo_space(P) or (P) results in objects swimming thru a static texture. I have even tried to promote the particle position into the shader and then subtract it from ws_space(P) but that did not work either.. (although I may not have had it working correctly) is it just not possible? any help would be greatly appreciated. -john redpawFx
  22. I have a bunch of Dynamics dops.. (glue objects) which I'm parsing thru and exporting each curve to a chan file. Once I run up a simulation, I want to make and update the dynamic chops to use the cache.. but its INCREDIBLY slow. I have about 300 objects in the scene and I have tried 2 different approaches.. the first would be preferred as I don't need to have a separate dynamic dop for each output.. where I just switch the object its pointed to, hit the update, and then write out the chan data (tx rx and pivot only) but in between "hitting" update (either in the interface or using an hscript) and actually even displaying the curve in the interface (switching from obj to chops for example..) the actual update time is over 10-30 seconds per object... So I'm trying different approach.. where I build a separate dynamics node for each object and then update them all at once which still takes quite a long time.. maybe a little faster than the one by one approach but compounded by doing them all at once (we're talking MINUTES here.. to update an interface??? from data in memory? ) I haven't actually gotten to the point of being able to fully test the time on this second approach yet.. as I'm building the nodes and updating and exporting along the way and thats taken over 3 hours so far either way the actual writing out of the chan file is pretty quick.. but its taking me upwards of 4 hours to write out 300 objects worth of chan files..... what is causing this slowdown... the cache exists already... and I don't have the re-sim button on in the dops node.. anybody got any hints? thanks -john
  23. Dynamic Chop Update For Exporting

    Ah... memory cache size bit me .. damn it would be nice to have a handy way to calculate that before kicking off a sim. general rule of thumb I guess is if it seems to be playing back in the interface too slow after cached then its swapping and you should increase the cache.. problem solved.. yay! -j
  24. I have a simulation that uses a rbdGlue object to break up an object. it contains objects labeled as shard_# grouped up properly in a sop network, but now I want to make a semi generic script/control to look at the glue object given, get the number of objects inside of it, and export a separate chan file for each piece one at a time. can anybody give me a quick hint as to what command I can use, script wise, to get a list of the objects contained in a RBDFractured/glue object? I'm not really concerned with the idea of writing a loop to change the input of the dynamics CHOP each time and then write out the chan file.. I just need to be able to get even the count of how many objects are being passed in so I can populate my 'for each' loop with proper names. the only other solution is to build a separate output CHOP chain for each object separately and thats a LOT of nodes when I could do it with a simple loop script.. anybody? thanks in advance -redpaw
  25. AAAAhhh.. I can't believe I've never seen any of these pages in the hdox before..... sigh.. changed my settings to hdox window left/right instead of up/down... maybe now I won't forget all this other stuff is here.. thanks
×