Scene walkthrough – Polygons following particles

Posted on January 10, 2012 by xsisupport

In this video, I do a walk through of a scene posted by Guillaume Laforge on the XSI mailing list. In the scene, Guillaume uses a point cloud to drive the polygons of a mesh, so that the polygons follow that transformations (pos and ori) of the particles. ICE modeling is used to “break up” the mesh into polygons. Includes a description of how vector subtraction is used to locate points relative to a polygon center.

http://vimeo.com/34804805

Tip: Use String to Array to quickly populate arrays

Posted on January 9, 2012 by xsisupport

String to Array is handy for testing when you need some sample data. Why bother adding ports to a Build Array node and typing in values, when you can just paste in a string. Depending on where you get the raw array data, sometimes all you need to do is a little search/replace magic in your favorite text editor, and then you’ve got a string you can paste into String to Array.

Here’s a couple of examples:

Note that the type of the array elements is defined by where you plug in String to Array. So if you want to put the array into an attribute, you’ve got to define the type of the attribute first.

A new attribute has no type:

Use a constant node to set the attribute type:

Now plug in String to Array:

Screenshots of the week

Posted on January 8, 2012 by xsisupport

Unique items to index
by Martin Chatterjee

RGBA Smooth Keyer
by rray

Deformer to instance
Edy Susanto

Blend particle colors of different point clouds
by Rob Chapman

Avoiding divide-by-zero errors with a divisor test
by Daniel Brassard

formula |cos(x) + cos(z)| ^ (1/2) in Parametric Sandbox ICE representation
by Daniel Brassard

Intro to scene references and Get Data daisy chaining in ICE

Posted on January 7, 2012 by xsisupport

Introductory-level video that shows you how to daisy chain Get Data nodes when you’re getting data from a scene.

Daisy chaining (connecting Out Name to In Name) is a way to separate out and isolate the different scene references that combine together to pull in scene data.

http://vimeo.com/34702959

Friday Flashback #51

Posted on January 6, 2012 by xsisupport

A look back at the last six years of the XSI mailing list, from 2006 to 2011.

12,354 different topics
81,458 posts in total
9,784 topics with at least one reply or follow-up
1,241 different posters

Word cloud for Subject lines of all posts (2006 to 2011):

To put the relative size of the words in context, here’s some word counts:

ICE = 7685
python = 1686
rendering = 1339, render = 2461
scripting = 1427, script = 1112
maya = 1265
lagoa = 597

ICE brain teaser

Posted on January 5, 2012 by xsisupport

Early this week, this ICE “brain teaser” was posted on the XSI mailing list.

Given an array like this:
8,8,8,2,2,2,2,5,5,5,5,9,9,9,9,1,1,1,1,1,1
how can can I –without using any loops–convert it to an array like this
1,1,1,2,2,2,2,3,3,3,3,4,4,4,4,5,5,5,5,5,5

Martin Chatterjee came up with a nice solution using arrays.

As an exercise, I tried to do it using a [non-simulated] point cloud. Basically, the idea is to use the particle self.ID data set to index into the array without using a Repeat node.

Here’s what I ended up with. First, I set a boolean flag on each point to indicate whether the corresponding array element is the same as the previous, and then I do a cumulative sum of array elements.

My first try at this ended up as an ICE tree version of this algorithm, plugged into the On Creation port of Add Point.

var a = [8,8,8,2,2,2,2,5,5,5,5,9,9,9,9,1,1,1,1,1,1];
var a1 = new Array(a.length);

var val = -1;
var ix = -1;
for (var i=0;i<a.length;i++)
{
	if ( a[i] != val )
	{
		val = a[i];
		ix++;
	}
	a1[i] = ix;
}
LogMessage( a );
LogMessage( a1 );
// INFO : 8,8,8,2,2,2,2,5,5,5,5,9,9,9,9,1,1,1,1,1,1
// INFO : 0,0,0,1,1,1,1,2,2,2,2,3,3,3,3,4,4,4,4,4,4

But that didn’t work, because as I found out, I wasn’t able to carry attribute values over from one point to the next. In the example below, you can see that self.tmp is local to each point (which means it starts off at zero for each point). Instead of incrementing self.tmp each time, I end up setting it to 1 every time.

Rigid body dynamic (RBD) precision and subframe sampling

Posted on January 4, 2012 by xsisupport

About precision and subframe sampling for rigid body dynamics…here’s my understanding of how these two are related. Note that nobody confirmed this when I asked around, but then again nobody gainsaid it either 😉

The precision determines how frequently the simulation is calculated. For example, if Precision=120, then there is at most 1/120 of a second between calculations.

The subframe sampling determines the number of samples per frame.

So, if you are have this:

30 fps
Precision = 120 per second
Subframe sampling =2 per frame

Then the RBD simulation is updated/calculated 4 times per frame (120/30), and ICE takes 2 samples of those calculated values per frame.

Using ICE to invert weightmap values

Posted on January 3, 2012 by xsisupport

Here’s a more general ICE tree for inverting weightmaps: it handles weightmaps that are not in the 0-1 value.
When I first worked it out in my head I saw it like this:

But that’s just the same as using a Rescale:

Here’s a weightmap before and after being inverted by the above. In this example, the weightmap has a value range of -1 to 3.
Before:

After (inverted):