Booleans used straight are fine for many things.
And in subsurf modelling, they’re a great starting point to get your shape right from the beginning. After you have the right shape, then you can worry about topology. I’ve found fixing bad topology is usually faster than building subsurfs from the beginning, but that’s just me, your mileage may vary.
Booleans with triangulate can make some “alright” topologies sometimes (double check the hell out of this) but you won’t be working with an all quad mesh anymore. I’ve done the same with applying a subsurf modifier and it can give alright topology as well, but yeah keep in mind there is a huge difference between alright and good.
Alright topologies are good for things in the background that don’t need to deform and don’t need precision uv mapping.
I’m working on this for speedflow.
With this automatic bevel, the cleanning process is really fast.
For hard surface objects with low to medium poly count, booleans work great as long as the edges are set to sharp and have a bevel weight like in Pitiwazou’s example. For high poly modeling with booleans (like the workflow for modeling weapons in AAA studios), you’d need something like zRemesher in order to get a clean result. Unfortunately the current remesh modifier mangles the edges far too much for that workflow to be possible in Blender.
[ATTACH=CONFIG]476536[/ATTACH]
I modeled this revolver frame (and the rest of the revolver) without booleans. I did use the intersect operator heavily and I can only care so much about topology… A lot of the flat planes had some control edges to keep everything from blowing up, but plenty of nasty spider webs.
Really, the intersect operator has the same topology issues as the boolean modifier, but it is faster to just stay in edit mode, rather than needing to hop back and forth from edit to object mode.
Is it true that Boolean operations used to be a rather bad way to model in Blender, but that wasn’t because of the actual concept as opposed to the algorithms back then being just horrible.
The implementation that was in place before Bmesh for instance were infamous in creating extraneous vertices, odd faces, and other pieces of bad geometry. Now the Blender booleans are fairly clean (ie. they only create new geometry where it is needed and cleanup work now falls in the range of topology tweaks rather than general cleanup).
Now it is true that Booleans can be improved yet (ideally, there should also be a concept of bracing edges, which would be making cuts across the new Ngons depending on the number of border vertices. I would also love to see some simple retopology algorithm for the immediate area around a boolean as well, but coding that would be a bit more involved and take a bit longer than the first idea).
Houdini 16 has new (upgraded) Booleans and they are awesome.
It think in terms of precision and stability these should be considered the new technical standard.
You can animate them and they are free of flickering and glitching, the surface is very stable.
Unfortunately the whole program is a little unstable at the moment, at least under Windows.
Hmmm. This is an interesting conversion. I guess I would fall under the category of never using Bools because of the cleanup afterwords. I always just subdivide the geo and cut holes that way. But I’m mainly modeling characters where all of this is mostly unneeded. Especially when I mostly use sculpting. However, I do see a point in the future where I’m using primitives to get the general shape right and then Boolean’ing them together and Dynotopo sculpt the results to smooth the intersections. But you know, old habits die hard. I usually just make a base mesh with the skin modifier (or Z-spheres in Z-brush) and sculpt the results into what I want.
As all we know the problem with booleans is in the seams: it is difficult to have bevelled, chamfered seams in Boolean-ed objects. And this makes models done by the use of them look artificial. This is the main problem of them. If someone had a way to overcome this problem than they would be the most intuitive way of modeling hard surface objects. So the needed thing is to find a way to achieve this goal.
Groboto, Mesh fusion (a Modo tool) and Hard mesh (it is a Maya plugin) like attempts are all attempts aiming to solve this problem. And they give quite nice results. Blender too needs such tools. It is a very important need for every advanced modeling app. It is inescapable for any such app to confront this challenge in the coming years. So, there should be a focusing on it. It is a serious necessity.
Below I put a model I made with the use of Groboto (its authors are those that made Modo’s Mesh fusion too). It is a boolean model in its total shape. The so important role that the smoothly connecting seams play is openly evident. And it should be kept in mind that the seams are controllable as to their softness or sharpness… something very important.
(the render is in Cycles Blender, while the mesh screen capture is from Wings 3D)
No matter how insignificant it may seem, to limit your toolset is to limit your creativity.
Don’t put yourself in a box.
I a way with blender we can do something nice like in zbrush.
Pitiwazou can you, please, put a screenshot of the mesh of the final result…
This is the remesh modifier, so it’s not as good as Groboto or Mesh fusion.
I tested this technique but that doesn’t work with rounded objects.
I’m sure blender can do it better if a dev work on that.
I think that BF must pick a dev only for this task (because the 2.8 launch will be longer) and the people want this. Make a kickstarter or something for this type of features. I want put money for this!!!
All those are nice tries Pitiwazou but the mesh density required for achieving good results, in such a case, is too high. Imagine what would be if you had a much more complex object to model. And, if you would take the route of retopologizing it by hand the time that the whole process it would consume would be too high also. While the meshes which Meshfusion like tools produce are ready to use without any more editing.
I double the DcVertice’s thought: it is something that many people would put some money to support such a project. Imagine that you have some such plugin in, say, some fifty Euro… who would resist buying it with the amount of money paid wholeheartedly?
Yesterday I tried solutions like remesh, but the density is excesive and the solution it isn’t clean and polish. Maybe a new remesh modifier will be better.
A lot of times i told that we need kickstarter for some tools that the proffesional users needs. I cannot pay a coder to make a tool, but i can put money with more people to make this tools. This tools could be free after a few time or something, to incentive to the people to pay something. The BF always have this hole in the development. We don’t see great third partie plugins and we need that.
I know this is not as good as mesh fusion, but right now it’s the best option.
You can add a decimate to reduce the number or tris.
