View Full Version : rain drops
11-19-2003, 02:41 AM
I'm trying to make it rain on my T. rex for a class project (its a dynamics class), and I can make it rain just fine, but when the that rain hits the T. rex, that's where the problem starts. I want it to hit him realistically, with drops rolling off and everything. I tried using a particle collision event, that makes a particle wherever a rain drop hits, but when I apply gravity and make it collide with the rex, it ignores the the surface its on, goes inside the T. rex and collects as if he's a bucket. does anyone know of a good way to go about doing this? this has been driving me insane. thanks in advance
11-19-2003, 04:35 AM
Try adjusting the collision offset. I don't know where that attribute is located off the top of my head but it may be deciding that when the collision event is happening that it is inside the geometry.
11-19-2003, 07:01 AM
I just now tried that out, it's actually called trace depth, apparently collision offset is for cloth, anyway, no matter what I do to the number it still falls through. I've tried writing an expression to start the particle a little bit above where the original particle died, but so far nothing has worked, I think because I'm not experienced in the ways of expression writing. any help woud be much appreciated.
11-19-2003, 07:04 AM
Did you make sure the normals are facing outwards? Sounds like if the particles are collecting inside then the normals may be pointing inside your body... though I don't know why it would register the collision even at the outside of the skin though. Hmm, lemme know on the normals. Also, if it's NURBS, you might try rebuilding each of the surfaces and reparametarizing each patch, that sometimes solves some problems.
11-19-2003, 07:23 AM
yeah, the normals are facing outwards, and its a polygon surface. also, it seems a few of the particles, a few few, are going along the outside, but then they go back inside, and collect with the others
11-19-2003, 07:27 AM
Hmmm, try uping the tesselation factor on your collision node on the poly surface to something higher. try something really high like 10,000 or 20,000 or maybe even higher if you want to give it a go, and see if that helps it.
I'm just throwing out any ideas I'm having, hopefully something works, because logically I think what you're doing should be fine!
11-19-2003, 07:29 AM
If the tesselation thing doesn't work, try making a crude standin for just a part, like his back, and set it to non-renderable and view mode as wireframe (can get it to be like that in the attr editor for the geo). See if the rain will collide properly with the standin, cuz that ought to tell us if it's a geometry problem or the particles problem.
11-19-2003, 08:25 AM
I think I got it, I just tried a new expression,
particle position = birthPosition + <<0,.1,0>>;
this puts the newly born particles just above the surface , they now fall onto the surface instead of through it
11-19-2003, 08:30 AM
Let me clarify your question.
You have a rain particle, collide with the T-rex skin, on Collision you create a collision event, emit new particles and maybe kill original particles, you connect the new particles with gravity.
The result is your new particle go into the skin and collected inside as the skin is a bucket.
Maybe what is happening you set your geoconnector resillience value to a very low number or even 0, so it doesn't bounce.
Maya got confuse when she calculated the collision event because the new particle is actually at the skin (it can be outside or inside, but probably she consider it is inside).
To get away with this situation, set resillience attribute back to 1.
But now the newly generated particles will move away from the skin, so you need to modify the collision event, set inherit velocity attribute to a less than 1.0, but not too small, just enough to tell Maya it is the direction is to outside of the skin.
You need to balance the inherit velocity, conserve and resilience in some how not to let the bounce to much but at the same time retain the particle outside the skin.
11-19-2003, 08:42 PM
well, here's what I've got so far, using that creation expression, i just need to figure out how to make the drops stick to the skin a bit better, as friction doesn't do it right (it's already at .9 here). thanks for all replies, and thanks in advance for future replies. if anyone can offer help with the drop stickiness, don't hesitate to comment.
11-22-2003, 12:16 AM
ok, I've been trying to use goals, inside the T. rex at certain points are locators, about 10 of 'em, that the particles want to go to. I was going to try and make it so that per each particle the goal weight would depend on the distance between itself and the locator. is goalpp per goal weight or for every goal weight? is it possible to this for each goal weight? is this even a good idea? thanks in advance
11-22-2003, 03:21 AM
yeah, its not following the underside, like real water would, this has been driving me insane, I tried using goals, with the locators, but it didn't work, because goal weight is not per particle, and goalPP is for all goals, if there was a per particle per object one then it would work, but no. now I'm trying to make the actual body of the T. rex as a transform goal, they stick more, but they move in weird directions across the surface, not straight down but at diagonals. if any one knows a good way to solve this problem, well... I'd appreciate it a whole lot. thanks in advance
11-22-2003, 08:59 PM
Originally posted by PilotLtKettch
... but it didn't work, because goal weight is not per particle, and goalPP is for all goals, if there was a per particle per object one then it would work, but no.
goalPP it`s a Per Particle attribute ... so it can have different value for each particle ... but it can`t be set via channel box ... try runtime/creation expressions
11-22-2003, 09:33 PM
Yeap, for goalPP write expressions or use the component editor to set them manually (this assumes a set amount of particles). This is the method to get controlled motion on Softbodies.
11-22-2003, 09:52 PM
whoops, I meant it is per particle and not per goal object
11-22-2003, 10:24 PM
You should also be able to write some expressions to change per object goal weights on the fly. I did something similar in a thread about Forming Smoke into Text a couple weeks back. The thread shouldnt be too buried, check that out for some inspiration on how to turn that into what you need perhaps.
11-22-2003, 10:59 PM
you can make dino skin a goal set to 0 until the drop hit it and then set it to 1 and the dropt will folow the skin topology.... (but lots of expressions need to be made to get a realistic movement)
11-23-2003, 01:05 AM
ok, I think I've got it, i create a newton field, select both the field and the rex, go fields>use selected as source, and then turn on under special effects, apply per vertex, I'll post results as soon as possible
11-23-2003, 09:43 AM
ok, new problem cropped up, without apply to vertex, the field is placed in the average of all of the vertices, apply per vertex applies the force equally to every vertex, so, places with more vertices get more combined force (rain drops get sucked to the eye). is there a way to fix this? adjust the magnitude per vertex? an expression, painting it like weights? thanks in advance and thanks for previous replies
11-25-2003, 10:14 AM
I have make a rain with SW-tube particles and I
have put 3D Motion Blur.I am not happy with it,
did someone know how to make a realistic CG rain?
man I can't download this, is link broke
11-25-2003, 05:09 PM
Motion blur doesn't work with particles. You can make realistic rain just with Streak type particles rendered in the hardware renderer. There IS a workaround for making motion blur on particles but you will have to do a search on this forum for it, I actually posted a question specifically about motion blur a couple weeks back and someone posted an excellent workaround.
11-26-2003, 10:33 PM
here is my rain sample
if i`ll spend some time developing it ever further I can make each drop trajectory fully controllable by expressions
01-16-2006, 06:00 PM
This thread has been automatically closed as it remained inactive for 12 months. If you wish to continue the discussion, please create a new thread in the appropriate forum.
vBulletin v3.0.5, Copyright ©2000-2013, Jelsoft Enterprises Ltd.