Gianni Ritschard

The point density is not crazy high considering the shape of the mesh is very long and thin, that leads to bumps like this.
0.002 is not saying much about the particle density, with a small enough object/ liquid that could be low-res as well.

If the strings don't need to be all simulated different you could also think about simulating only 1 or 2 of them and then copying them and deforming the mesh slightly each time btw.

You could turn off the adaptivity in the particlefluidsurface and then either blur the normals or quadremesh it and transfer the velocities back from the original mesh or particles (see hip file attached).

Hope that helps

Definitely not enough particles I would say.


You can imagine the meshing placing a little sphere on every point and smoothing the result, which gets bumpy with too little particles.

You could either lower the particle separation(might change the look of the sim a lot though), or use reseeding like in the vimeo video that was mentioned, to get more particles without changing the look (and sim time) too much.

tamte
you can potentially also play with FLIP Solver/Particle Motion/Droplets/Detect Droplets
you can either kill them directly within threshold
or can also help with those isolated particles not influencing the velocity field and then after sim you will have droplet float attribute you could possibly use to remove those

I know it's an older thread, but what is particle density referring to here?
Is it the number of particles in the voxel?
I did some tests, but couldn't find that out for sure.