woodenduck

FEM is for when you have all the time in the world. Vellum is for when you have to deliver. ;D

Ha, typical! As soon as I post on OdForce I find the solution. So the "Rest Orientation Frame" parm on the Vellum Constraint Sop appears to have been the culprit. In my case this needed to be set to the start frame of the sim. Not the default F1. Cheers, WD

My Vellum hair immediately shrinks in length when I try to sim it. I've tried stiffness and bend at outrageous settings. As well as constraint iterations up as far as 25000. Substeps, dampening, mass, thickness, changing restlength. Pretty much everything I can think of and it makes no difference at all to the simulation. It's like I'm not even changing any parameters. I've resampled all my hair curves to have 4, 8 and 12 segments. This also provides very little in the way of changes to the sim. Unfortunately I can't post a hip file. But if anyone has any suggestions I would be keen to hear them. Thanks, WD

Just need to change your expression from opinput() to opinputpath(). The first returns just the name of the operator. The second returns the path to that operator. This particular parameter requires a path.

Can you post a hip file or screen shot?

DOP network - can be a path. Object Mask - Should be just the name of the dopobject. Not a path.

pop wrangle plugged in to "particle velocity" on the flip solver... v@v *= fit(@Frame, 1, 24, 0, 1); This will multiply your velocity on a ramp from 0-1 between frames 1 and 24.

if you attach a null to the camera at obj level and scale it to 0.1, does that line up?

So you cached the abc then you read it back in to render? Is the geometry packed? Do you have the same problem if you unpack followed by normal sop?

Glad you got it working, and of course whatever gets the job done is great. However, I would encourage you to investigate using convex hulls for all collisions. You will see vast improvements in speed and stability. They can be a little trickier to set up, but the rewards more than worth it.

The blue collision object you are visualising... is that created using a static object dop? If so, are you visualising the "Collisions/RBD solver/Volume" tab rather than "Collisions/Bullet Data" tab??? I think you may not be seeing what the solver is seeing.

If 'v' exists on the incoming points, the flip object should pick that up automatically. So however you set 'v' in sops, should be what you see in dops. Otherwise, you can grab the 'v' and plug it in to the 'particle velocity' (inout 2) on the flip solver. pop wrangle. Inputs tab. Input1 = First context geometry. v@v = point(0, "v", @ptnum);

I usually point anyone looking to get started with VEX at Matt Estela's excellent wiki... http://www.tokeru.com/cgwiki/index.php?title=HoudiniVex

in a wrangle. Plug static points in to input1 and animated points in to input2 @P = point(1, "P", @ptnum);

If you post a hip file, I'm sure someone will help you out.