WhyGee

Still waiting for sidefx to fix the bug. Ill email tomorrow to see if i can get an ETA. Hopefully this will make houdini's cloth usable again

Hi Atom, Thanks for the scene. The 2 grids in my sample hip were just to show the problem, the actual scene is far more complex. Support has confirmed that this is a bug. In the documentation they explicitly advice against adding thickness to cloth by extrusion and to use the collision radius instead. This does explain the poor collision behavior in houdini's cloth and the extreme substeps required to get something somewhat useful but it's never really close to the speed and accuracy of other solvers. I've just tried in Houdini 16.0.633, same behavior. A bit surprised that noone flagged this bug before.

so sidefx marked this is a bug. But if anyone has a workaround meanwhile, I'd really appreciate it. Thanks

Sorry about that, here's a sample file. I've tested on the latest production and daily build and seems like it's completely ignoring the collision radius. Thanks! cloth_collision_test2.hiplc

Hi, Has anyone else noticed that the cloth collision radius has no effect on actual collision distance? No effect on cloth to cloth or cloth to other collider types. I would appreciate a confirmation and of course a workaround if anyone has managed to find one. Thanks

Is there a way to have part of a gas repeat network solve every Nth pass without doing nested gas repeats? I've tried reading the solve pass data as an expression on an "enable solver" but it doesn't seem to be evaluating every pass but rather every simulation step. Anyone have a working approach for something like that? Thanks /Y

(if any of requests below already exist I apologize ahead of time ) - an easy way to rename multiple takes or at least a branch of take children. - a fast cloth solver even if at the expense of physical accuracy. something like Marvelous Designer's engine would be amazing. capable of handling strong forces without a ridiculous number of substeps/collision steps. And of course, you guys do an incredible job, This is by far the best software on the market!

Thanks that sounds promising! the seeded points inherit the nearest particle age from what I've encountered but now that I might have grouping of newly added points that shouldn't be a problem to reset. Thanks again!

Hi all, I was wondering if there is a simple way to group newly seeded particles during a flip sim. The reason I would like this option is that I have a sim where I'm not meshing the particles but rendering them directly and the seeded particles pop into existence in an unpleasing way, so I was hoping to somehow scale them up gradually over a few frames. Thanks in advance

Hi, Is it possible to have an RBD object act as a volume collider, for example against a flip sim but also as a surface collider when colliding with cloth (in the same dop network of course) If someone could provide an explanation or a simple scene setup, it would be of great help. Thanks in advance! YG

unfortunately no, that was just a simple example to explain the problem. The scene itself is more complicated and the cloth is moving around. I need some of it to pass thru wherever it may encounter a collider. I'm looking for something like the wire solver "nocollide" attribute if I remember correctly. Thanks!

Hi Atom, Not sure I understand what you mean, but maybe I wasn't clear. Imagine a table cloth falling onto a table top. I'd like to be able to have half of the cloth collide with the table while the other will pass through it. I want it to be controllable on the cloth and not delete part of the collider itself if that's what you meant? Thanks! YG

Hi, I was wondering if it's possible to disable collisions on a per point level? Thanks YG

Hi, I've searched through the forum and couldn't find an answer so hopefully this isn't a recurring question that I've missed. Attached is a simple scene with a rigid body flying onto a piece of cloth. When the rbd is set to volume collision, it works fine, when it's set to surface, the collision is ignored. Is this a limitation or have I done something wrong? I have done other tests where cloth falls onto an inactive rbd surface collider which worked. Once it has become active and started moving, the collision was ignored either completely or was not correct. The reason I insist on surface collision because in the actual project, there is also other type of dynamic interactions along with particles and flip involved and I need surface collision information on the particles. Perhaps a workaround would be if it's possible to have the same rbd sim interact as a volume for some dynamic objects such as cloth and as a surface for particles and flip. (the flip does affect the rbd, so splitting the sim is not desired) Thanks in advance, Yaniv active_rbd_surface_cloth_collision_test2.hipnc

I agree this effect should be much more straight forward. Thanks for the hip, I'll have a look now to see how it behaves in my scene. The previous file tends to get worse when the sim resolution is finer so I hope your setup will work. Cheers Yaniv