← All ideas
  • Published 13 Jun 2020

boolean modifiers a need for change

Broken boolean modifiers hold 3 modeling captive from meeting its true potential. People are often forced to put objects where they don't want them, and cope with lack of absolute geometric symmetry. It often forces you to model entire models point by point and face by face, by hand, often resulting in extremely sloppy tolerances. Symmetry is often offered to help this but Blender only offers bilateral symmetry, not radial, radial mirror, radial step, radial step mirror, etc.

Fixing the boolean modifiers is the overlooked key to winning the entire CGI game. I have been a round a bit and have tried lots of software . I know two programs that have effectively solved this problem, 3D COAT, with its voxels, and ZBRUSH , with its dynamesh functions (not its regular boolean system). What I know is that voxels will solve this glitch and I see several ways to mitigate the cost voxels have on the system.

One potential fix is to coat the objects being used in the boolean modifier with a virtual layer of voxels that will only be used in the modifier, similar to how virtual particles mediate interactions between real particles in quantum physics. The idea is that by giving the surface a slight thickness, and treating it as a voxel object, there will not be any rounding errors or points that intersect, which is typically what causes boolean modifiers to crash. After performing the operation, the center of each of these voxels can be used to stitch back together and build a valid surface using calculus.

A second possibility involves using voxels to deal with the interaction by voxelizing both objects in converting them to volumes, entirely. However, in order to do this with little memory loss, a fractal compression system such as sparse voxel octtree or something better that the Blender team designs. to ensure that the inside of the object contains far less voxels than the outside. It may be useful for you to investigate spherical close packing in a 3D plane as there are irregular polygons that can tessellate 3D space far better than cubes.

One last suggestion, is that Blender switch to a voxel work flow entirely. I know that there are people working on such a system in eastern Europe, (Czechoslovakia?) It is called Atomontage ( atomontage.com ). I think that this program, if it can be pulled off, will surpass the rest of the entire 3D industry.

I am posting this as a person with autism. Many people with autism have issues with things that only work some times. I also have OCD. I am experiencing one giant glitch with boolean modifiers. This glitch is not just a problem with Blender, but exists in most 3D software.

100%

Development Status: Open

Upvote 2 Downvote 0
crystalsunshine686
13 Jun 2020

4 Comments

Lillya

The boolean modifier in Blender is planned to receive a major redesign, hopefully making it much more stable and solving the cases where it fails, like with co-planar faces. This boolean code redesign has been going for a long time now, so hopefully it's implemented soon.

I think the second option of using voxels is also planned to be added in the future, with a dedicated voxel remesher modifier, but we'll have to wait to know exactly how that will work.

1
dantabdantab37

Meanwhile have you tested volumemesh modeller addon (which is not mine). And about circular symmetry don't you use an empty and array modifier to it (very easy to use) . Boolean in blender is sometimes a mess, because of vertex proximities. a displace modifier with a very small amount solves often the problem. And with a very high count of polys, a decimate modifier keeps the job done. It depends on the complexity of the scene, imho...

thinsoldier

I'm pretty sure the advanced volume based boolean stuff I saw in C4D recently is using openVDB. In theory it should eventually be possible to do the same in Blender.

crystalsunshine686 Author

also excuse the grammar I have several disability's