RFE? Some thoughts about fields, neighbors, proximity and attributes

POPs
1

Field POP and neighbors

I’ve been playing around a lot this week with the idea of taking neighboring points and trying to use those as affectors for a second POP. I saw @greg comment about the second input to those which is great for creating extra attributes but I’m wondering how/if that would work with a field where you create an attribute with a falloff.

Basically it’d be great if the field had a second input and could use whatever logic is going on with the neighbor POP to allow for falloffs from points or lines in the second input geometry and then have an attribute on the first input with that falloff.

For example if you had a cube made from line strips intersecting a volumetric sphere of points it’d be super cool to be able to have a field take those line strips (or even just the points) and then calculate the closest points to furthest with falloff. This functionality would work equally well on the neighbor POP actually too…rather than the neighbors as their own attributes just have each attribute have a weight which basically says “hey my closest neighbor is x distance away” and then normalize that with a falloff curve or something.

Proximity POP - group to group

It’d be great if the proximity POP could have two group inputs and have it exclusively connect from one group to the other. My example of this is that I have a light position which is a single point and I’m connecting it to a bunch of points on a 3D scan to make a fake god-ray kind of effect. Right now I have to ensure my light is far enough away that I can just use the threshold.

It’d be similarly cool to have a cube and a sphere and they only connect lines between one another.

EDIT: Maybe this could also be done in a more flexible way with incoming attributes rather than groups

null3
null31920×1080 117 KB

2

Lots to respond to here, but the next Proximity POP has 2 inputs - i think it does what you are asking.

And we were thinking Neighbor that as of today, just averages neighbors, will do a gaussian rolloff - closest points have greater weight. Though that may not be as general as what Field or some other rolloff functions may be able to provide.

Now I’ll re-read what you wrote…

3

That neighbour setup is exactly what I was thinking for the field pop so that works too!

For my second request it’s just that right now any line can connect to any line whereas I’d like to be able to say any line in group A can connect only to any other line in group B

EDIT: Thinking further on this perhaps it’d also be good if the line connection has a % chance of forming and this % is determined by an attribute. If it’s zero then that point cannot connect but if it’s 1 it will definitely connect. That’d give some nice falloffs in the line connections.

1 Like
4

Chiming in, I think what we’re adding (2n inputs and more weighing features) will cover what you want to do.

Regarding the % chance of forming I think you could also do that with the new features but the other way around, the lines created inheriting the attribute and you could use that to delete them afterwards. with the threshold aheum delete POP.

we’re also adding optionally having the distance as an attribute so you could use that to fade lines in and out for smoother transitions (somewhat already possible with GLSL POP or converting lines to line strips and using line metrics but a bit convoluted)

1 Like