Hi! here again with more R&D. I always wanted more control about group change over time in TP, so I decided to test it. My goal was to do a random group change (the typical workflow, a random value per particle and per call from 0 to 100, and after a treshold passing particles to the new group if the random value is less than 10 for example) but with manual control about the probability a particle change from group.
I was thinking it will be complex but one more time TP shows me is the way to go, I did the setup in less than 5 minutes:
In this example the weighting is relative to the distance of every particle from a node. I define a max distance with “range” value. With normalize I convert this distance to “0″ to 1. Next expression is “1-a” to invert the value, so particles more near the object has “1″ and particles more far “0″ (this really is not so necesary). I remap this with a “value to time” plugged in “fade in” node. In this node I have animated my curve, and its this curve where I can control the weigh of the threshold, so making modifications on the timeline I can control with total precision and manually the random group change.
Add the end I think it give a lot of flexibility, combining a procedural aproach (selection of particles are still procedural), with manual refinement (possibility to change the probability that one particle are selected) using the curve.
I don-t know how I finish with a butterfly video in fact, XD. I wanted to illustrate this technique, I decided that can be cool some butterflies. After for more realistic motion I decided that butterflies that are near one moving butterfly will move too, so I added change of group based on proximity (when one butterfly is near one moving butterfly this one move too). Also I put in practice the animation tree states in geom instance that I don-t use so often, and has a lot of potential too (butterfly motion change when they move or when tehy are stopped).