dpz Posted February 23, 2011 Share Posted February 23, 2011 hi. stuff like bounce and friction works fine with an attribcreate and per_point values. but i stumbled over the task to vary the weight of a rather simple object like a tube .. per_point values in sops for mass and/or density won't work i guess. so i'd really like to know how to vary the weight for a tube for example. (let's say one end of the tube is heavier 'cause some metal lives in there). thx i advance. cheers, -chris. Quote Link to comment Share on other sites More sharing options...
old school Posted February 23, 2011 Share Posted February 23, 2011 Build a composite representative RBD object using Glue to mock the object you are trying to physically replicate. Each piece in to a group merged together and set up as an RBD Glue object with -1 for the glue strength (infinite). In DOPs, the RBD Glue Object DOP will create a wrapper RBD object to receive the entire glue hierarchy transform, rotate and pivot channels. You can fetch those channels via the DOP Import SOP or CHOPs to drive your original geometry. In DOPs, you can use the RBD State DOP to override the mass of each component to your liking. This is quicker than you think. See the tip below. Tip: When working with a DOP network and your viewer set to DOPs, you can use the select arrow (hotkey is "s") to select the RBD object you want to modify then tab-type RBD State and hit enter. This creates a new RBD State DOP with that object(s) already set up for you in the Group parameter field. Hit s then q to repeat and select the other pieces in turn. ---- You can also manipulate the pivot point of the RBD object to wherever you want. composite_object_with_varying_densities_v001.hip Quote Link to comment Share on other sites More sharing options...
dpz Posted February 23, 2011 Author Share Posted February 23, 2011 thx for this interesting example file. but have a look at the 2 quick tests in the flipbook (attachment ). 1. example: shouldn't the heavier side of the object fall quicker? 2. example: shouldn't the lighter part be more attracted by the magnet force? .. it seems that the different values onyl affect the sim after the ground collision. cheers. 358435468.mov Quote Link to comment Share on other sites More sharing options...
old school Posted February 23, 2011 Share Posted February 23, 2011 thx for this interesting example file. but have a look at the 2 quick tests in the flipbook (attachment ). 1. example: shouldn't the heavier side of the object fall quicker? Nope. Neglecting wind resistance, no. I just dropped a 4' long round stick with a 1/2lb transformer duct taped to one end off my 30' high atrium and she fell completely flat after 4 tries. Newton was spot on when he said that gravity affects all objects the same no matter what the mass. But I'll tell you this. After the 4th try, the plastic case on the transformer let loose big time! That's f=m*a at work for you. Dynamic fracturing at it's best. 2. example: shouldn't the lighter part be more attracted by the magnet force? .. it seems that the different values onyl affect the sim after the ground collision. Now you didn't mention "that" in the first post. You did mention varying density but not magnet forces. I did a couple simple tests by manipulating the Pivot channels and disabling "Compute Center of Mass" and you can move the pivot point about and that does indeed shift the center of mass and you can achieve a similar effect as the composite RBD file I attached before. As for the Magnet Force DOP, by changing the Sampling Mode, you can change how the RBD objects interact with the metaball field(s). The default setting for RBD objects and metaballs is "Point" or a single continuous force applied to the RBD objects. It basically ignores the geometry and the sdf densities. Changing this to Circle changes the force lookup to include the surface of the RBD object along with the pivot cog's. Finally Sphere causes the sdf to be used and can be much slower but you should now see the various masses of the objects cause different effects. Decreasing the resolution of the SDF collision volumes will increase performance so set them just high enough to get good collisions without tunneling. See the attached example file this time using pivot offset on one cone and the other at it's default. You should be able to do these kind of one-off tests to play with various theories ad nauseum. And that is why tubes, spheres, grids and cubes rule in Houdini. rbd_custom_sdf_v001.hip Quote Link to comment Share on other sites More sharing options...
dpz Posted February 23, 2011 Author Share Posted February 23, 2011 well, i have to admit i didn't mention that. but this wasn't supposed to go in different directions in the first place at all. i just stumbled over issues while playing with it. whatever .. i have to stop playing now. but thx again for this stuff you posted. definitely learned something. and stop smashing furnishings please. cheers, -chris. Quote Link to comment Share on other sites More sharing options...
Macha Posted February 24, 2011 Share Posted February 24, 2011 (edited) Interesting example, now I finally understand the sampling mode settings too. I was wondering what would happen if we attached a metaball force to each object. I tried with a sopsolver but what happens is that the sopsolver only processes one object. Does anybody know why? rbd_custom_sdf_v001.hip Edited February 24, 2011 by Macha Quote Link to comment Share on other sites More sharing options...
Recommended Posts
Join the conversation
You can post now and register later. If you have an account, sign in now to post with your account.
Note: Your post will require moderator approval before it will be visible.