Accumulating vector data?

[perryharovas]

Hi everyone,

I am trying to make additive point position data.
I am trying to basically make each point position add to the previous point position, and accumulate.

The reason I want to do this is I want access to the path the particle travels to make a wetmap.

I am using Pointcloud Lookup in the Rendertree, and it is correctly seeing the data I want, but the data changes every frame.
I want to make it see each frame of data as something that accumulates with each new frame.

I am REALLY close on making this wetmap thing happen, it just needs this one problem solved!
I have tried many things, but can't seem to get it to accumulate the vector data.

I was able to get a scalar value to accumulate (as the Accumulate Proximity Map compound does with Weightmaps, but I have had no luck with PointPositions.
Any help would be very much appreciated!

Thanks,

Perry

[TwinSnakes007]

Use strands and read the strandpositions for all the points that the particle has traveled.

[xsisupport]

If you want the path the particles follow, then it could be as simple as this. That's like the strandpositions suggestion.

But I'm not "adding" points, so I'm not sure whether this is what you want.

[perryharovas]

Thanks guys!

Steve, that is almost what I need, but it doesn't work.
I want, ultimately, the renderer to know the point position and at each frame, it should know all the previous
point positions up to and including the current frame. In this way, if I can continue to write to a custom property that
keeps accumulating the point positions, hopefully Pointcloud Lookup will see the data and be able to use it.
I have basically been able to create a wetmap for any frame (without UV's, and without weightmaps) that does exactly what I want with the
sole exception that it doesn't consider the previous positions of the lagoa particles, just their current one...

TwinSnakes007, thanks also. I tried using strand trails (because I didn't want to mess up the sim I already have), but it doesn't seem to work either.

[xsisupport]

Doesn't Point Cloud Lookup output a vector, not an array of vectors?

[perryharovas]

I don't know, but if so, then obviously that would be why it isn't working...

Argh! Can you think of any alternatives to Point Cloud Lookup, Steve?

Thank you so much for you help!

Perry

[Bullit]

Isn't there a node that is looks to other frames instead of current one?

[Daniel Brassard]

Do you know if the pointposition array will be the same size at each frame?

Some though, cache the pointposition then read it at the next frame and add (append) to the current point cloud position and save the cache file. If you have a lot of particles, this cache file will become huge!

my 2 cents!

[perryharovas]

Isn't there a node that is looks to other frames instead of current one?

Yes, but it is just the previous frame, and isn't added to the point position of the current frame (or any before the previous frames).

[perryharovas]

Do you know if the pointposition array will be the same size at each frame?

Some though, cache the pointposition then read it at the next frame and add (append) to the current point cloud position and save the cache file. If you have a lot of particles, this cache file will become huge!

my 2 cents!

Thanks Daniel.

It won't be the same for each frame, it is a pouring water sim, with continuous collisions and some water dropping off the geometry and some falling on it for the first time.
Thanks though!

[grahamef]

I think the simplest approach is to accumulate "wetness" values in a weightmap on the geometry. However that may be too coarse depending on the resolution you need.

Another approach would be to copy the water particles in contact with the surface to another pointcloud every frame and use that instead.

[perryharovas]

Interesting idea to copy the contact particles to another cloud.
The weightmap approach IS too dense because I need control over the resolution
and the model may change VERY often. This client is known for many lastminute changes to entire geometries,
but also want fast turnarounds (and the dense mesh approach makes both extremely difficult).

Thanks for your input!

Perry

[talent103]

did you ever find a good wetmap solution perry?