Hi vytas_rauckis,
I have no fast solution for that, to create a natural cluster of bubbles, let me explain.
The practical part:
The soap-bubbles typically find a specific balance of all participating bubbles where the tension is equalized among all in direct contact, and so on.
This was practically used, to my knowledge, in the architecture design of Frei Otto. We studied his light-weight buildings back in the days.
An example is given here, even more targeting roof curvatures:
https://www.youtube.com/watch?v=-IW7o25NmeA
A longer introduction to this theme is here:
https://www.youtube.com/watch?v=rEXj69hCi44
As you might notice, this is not a simple “formula”.
In other words, to place spheres randomly in space and “draw” their contact areas as flat elements wouldn’t be correct. The whole “physics” in such a cluster is complex.
The CG part, I guess that mathematical precision is not the target here.
So, here is the closest I get with my first idea, and it is not really where I think it should be. But perhaps it delivers some ideas.
The VoronoiFracture gives of course the “cut” not only to its neighbors. So there is a limitation on top of that, it might extend the cut if the cluster is more uneven. I share this file, as it show some limitations if the cluster of bubbles is spreading out.
Two scene files: (You might ignore these or just check them out as a brain storming file
https://www.amazon.com/clouddrive/share/0qJiewzK8OsLRniP4jLAYlzrUPYNAjfgtaHwKCuPFVZ
The point production (XPresso) could be simplified, but given the current state of the idea above, I think it is not the prime problem.
For the texturing, please have a look here:
https://www.youtube.com/watch?v=FxjROrq8U4w&index=66&list=PLD428BE966FF3B1E9
All in all, one would need to Boole all the Spheres to get one surface and then fracture it. Again, if the set up has some “branches” of bubbles, the points might cut…
Here is a manual suggestion. (See image below), based on your lead to use Voronoi Fracture.
With the little gap the seam is nice.
Scene file: First step of exploring a possible workflow, by setting up the membranes with Splines, which was a test, but I guess the Voronoi can deliver this, with a tight sphere mesh to start with as well, but here is the set up …
All files are in this collection
https://www.amazon.com/clouddrive/share/rD38g7crCgMwnv5AfyKEP07bUGPxYtvcBKcrsjSlUb6
… and here is then the next step, which uses as well the Boole set up under a Voronoi Fracture.
This set up uses the “Current State To Object” function, and while a gap was introduced in the Voronoi Fracture to get the mesh[es], the resulting meshes receive an Extrude [M~T] with Create-Caps on, to give each of the four segments a thickness. The material thickness is below half of the initial gap. All parts from the file …_22 that were used are deleted and only the polygon based result is left for clarity. The Selection Tag (produced without the gap of the Voronoi): Current State To Object, then Edge To Spline, then Sweep.
The Sweep set up simulates the little fluid accumulation along the contact areas. Again, real bubbles pull them self to a equalized position, this was not the target here.
All the best
edit: all scene files are now in one collection. /edit