Assigning a per-object random value with ICE

Posted on January 31, 2013 by xsisupport

Suppose you want to use an ICE attribute to drive some behavior in the render tree. For example, you may want to introduce some randomization to a procedural texture. Here’s one way to go about it.

First, create an ICE attribute on each object. You’ll use this object as the seed for the Random Value node (you need a different seed for each object, otherwise you’ll get the same “random” number for each object).

si = Application
i = 0
for o in si.Selection:
	a = o.ActivePrimitive.Geometry.AddICEAttribute("_seed1", 2, 1, 1  )
	a.DataArray = [ (i) ]
	i = i + 1

Now apply a simple ICE compound to each object.

Seed and random scalar value for each object

All the compound does is feed the seed into a Random Value node, and store the random value in another ICE attribute.

Deleting particles by ID

Posted on January 29, 2013 by xsisupport

Let’s say you wanted to delete certain points for some reason; maybe they were misbehaving or something. You could just stick the IDs into an array and feed that into Find in Array, like so:

This is an answer to a question on si-community, which timed out on me when I clicked Submit. So now it’s my post for today. After struggling with Maya nParticles for an hour, I can’t think hard anymore.

Using an ICE attribute to drive a procedural texture

Posted on January 28, 2013 by xsisupport

Here’s a simple example of using an ICE attribute to drive a procedural texture. In this case, I’m using a random integer to drive the number of repeats of a checkerboard:

To set it up, I ran this script to programmatically add an ICE attribute. Unfortunately, I found I had to add an ICE tree to get the attribute to show up in the render tree.

import random

si = Application
    
for o in si.Selection:
	# Long, Single, Singleton
	a = o.ActivePrimitive.Geometry.AddICEAttribute("_random", 2, 1, 1  )
	a.DataArray = [ (random.randint(1,8))]
	si.ApplyOp("ICETree", o, "siNode", "", "", 0)

Then in the render tree I used Integer Attribute node to get the attribute value. This same material is applied to every cube in my example.

And if you wanted to update the ICE attribute later, eg increase the range of random values, you could do something like this:

import random

si = Application
    
for o in si.Selection:
	a = o.ActivePrimitive.Geometry.ICEAttributes("_random")
	a.DataArray = [ (random.randint(1,12))]

Screenshots of the week

Posted on January 27, 2013 by xsisupport

sparta-like particle editing in softimage

Point position from group
by Vincent Ullmann

lambert shading
by Mr.Core

Creating strands between two different objects
by Ola Madsen

wet fur effect

Create hexa tesselation
by msktkhs
“Be further divided into two right-angled triangle is divided into six triangles hexagons, X direction will take time to get to that direction cos60 * 4, Y is I will be next to the point sin60 * 2 too (laughs)”

Nodes for whirlpool spiral
by msktkhs

Volume rendering with ICE

bend paper

Screenshots of the week

Posted on January 20, 2013 by xsisupport

Creation Trees and Softimage and Arnold
https://vimeo.com/57690501

whirlpool
by tekano

Fake SSS
by Tekano

ICE sun position

Mandelbrot
by grahamef

Bend wrapper
by csaez

Accumulating point positions in an array
by xsisupport

Instance on curve
by julca

CrowdFx meets Momentum
by XAPKOHHEH

Frame match from animation clip
by Mochi-Mochi

Using modulo to delete points from a cached simulation

Posted on January 14, 2013 by xsisupport

hat tip to tekano bob who did all the work; I’m just pointing out some things about his ICE tree 🙂

First, he’s using the Element Index in case the ID was not cached (if you cache with the Cache Manager, then by default the ID attribute is not cached).

When I saw that, I thought maybe I could use First Valid to use either the ID, if available, or the Element Index:

But that gives wacky results, because Delete Point automagically restores IDs to all the points in the cached simulation. So you have no choice but to use the Element Index.

Another interesting thing I’d like to point out is how you can use modulo to delete two-thirds of all points (instead of just deleting every third point). Now, modulo by 3 has three possible results: 0, 1, and 2. If you test for modulo != 0 or modulo > 0, then you’ll be deleting two-thirds of all points: