Attribute Normal Distribution Plots in CHOPs?

[Zach]

Hey guys,

So, I'm looking for a good way to visualize, and maybe manipulate (through ramp parameter vops, i guess), normal distributions of attributes for a group of points or particles. I'm not as CHOPs savvy as I'd like to be, but this seems like something I should be able to visualize in the motion view (although it would really be more appropriate in the details view) so I was wondering if you guys could help guide the way.

If you're not familiar with the term "normal distributions," I'm talking about a way of visualizing the distribution of a particular attribute's values across a group of points in a "bell-curve"-like manner that illustrates the relationship between the individual values and the mean and standard deviations for the population (err, group of points). The idea is, if you were to create an attribute "foo" and assign it a value of "1+fit01(rand($PT),-0.5,0.5)", you'd get random values between 0.5 and 1.5, and they would average out to "1". A normal distribution plot would show how far away a point's value is from one (or, how far it deviates from the mean). Theoretically, this distribution would take the shape of a bell-curve, with 68% of the points lying within one standard deviation (I guess values between 0.83 - 1.17), 95% falling within two standard deviations (values between 0.525 - 1.475), and so forth. I'm not explaining this well, but for more information, here's a link that is less complex than what you find on wolfram or wikipedia: http://www.mathsisfun.com/data/standard-normal-distribution.html

This would be a nice visual way of interactively tweaking the effect of multiple noises, or for quickly identifying extreme values as they pop up (they'll skew the entire distribution); Massive apparently permits the user to manipulate normal distributions of behaviors or other crowd properties; you could use them to quantitatively determine the effect a force has on a system... whatever.

It seems like something I could probably whip up in python (and may have to), but I imagine performance would take a pretty large hit. Although maybe not, the math isn't super complex.

Anyway, has anyone done this kind of thing? Is this the kind of thing that you'd find useful?

[lukeiamyourfather]

If I understand you correctly this could be done with an Attribute Promote SOP (to get the minimum and maximum values of everything) and a VOP SOP (to remap and visualize the information). It would be pretty straightforward and wouldn't have much of an impact on performance. There might be some CHOP trickery to do this in a more elegant way but I'm not as versed in CHOP. If that's the case hopefully a CHOP guru will chime in.

[stevegh]

Where is this CHOP guru?