DrVertice; May I ask then how you would have a stack system handle a multi-layered boolean system involving dozens of objects (with heavy utilization of the Union, Intersect, and Difference operations at the same time)?
Now add dozens of different bevel operations on top of that (for the purpose of having different settings for each one), how would a stack handle it without becoming unwieldy?
You must understand, ace, that I want to make proposals that can be made now with actual tools. I don’t want to make “hype” about tools. Only search the easy way to have the feature.
I think that you agree with me that Blender developers don’t have time to make a lot of things, and modeling is a part that they don’t consider a lot. It appear to be secondary. With this problems over the table I prefer to have a restrained debate. If we would had a exclusive developer to make modeling tools, ok… but we don’t have it.
@cgstrive: should work when you enable “Store Edge Bevel Weight” in Properties>MeshData>Geometry Data? [iirc you have to do it on the cutout meshes as well as on the ‘main’ object – not sure though and not in front of blender atm…] And thanx for the windows build!
Ace is on to something.
Funny that in 3dsmax MaxCreationGraph does exactly what you feel is a failure of a concept - use nodes to create modifiers. Also speaking Houdini you assemble nodes into subcomps with own controls that roughly equate to modifiers. That’s what a lot of the tools like Flux are about! They are in essence a set of lengthy nodal instructions packed into few nodes that act like modifiers to be far more efficient at work than doing things low level. That’s what would be ideal for Blender aswell as you cannot write .py/develop plugins for Modifiers. It’s a lot of work though and if we continue debating it here then we need to once again fragment the thread and start a new, one that has been spoken of 100 times. Devs know best when/how it can and should happen, anything we say is just noise subtracting from the Bevel thread.
Well the proposals here are good for now (HowardT or someone else codes an auto-merge for beveling and vertex-group generation in the boolean modifier to use for the bevel modifier), but the insane examples from Houdini may still not be an easy thing to replicate (because again, your stack will grow to a huge size).
I’m fine with what is proposed (it will be enough for most cases), but we still could keep the node-based idea in mind for the future (once the BF actually gets the resources to do the ‘everything nodes’ project).
Thank You! This works great!
Win Build link for anyone interested: http://cgstrive.com/blend/2d23d_Boolean_Mod.rar
It’s patch by 2d23d that enables Crease/Bevel weight per boolean https://developer.blender.org/D2169
Actually somebody is working in the tools of the thread?
Fortunately that looks to be the case. As I said, it would indeed be nice to have this stuff in the initial 2.8 release as more advanced functionality like I described will have to wait (the way things are looking, this will already satisfy a lot of cases).
I prefer see in a 2.79 experimental build, because final 2.8 maybe will need one six months or more to be complete.
This is really nice, exactly what we need !
The only issue right now is that the bevel weight on mycube are not used anymore.
Is this thing working in the case of dense meshes too?
@Fatesailor: unfortunately not [in many real world cases] – but thats the limitation of blenders bevel (stopping at edges and not dissolving them)…
This is a problem. Because if the only needed thing is to have bevels in boolean seams of very simple, low poly meshes, that is something that can be achieved manually also. The really needed thing is to be able to make bevelling-filleting on any object. The novelty, besides, that apps like Mesh fusion have introduced is exactly this: giving to the user the opportunity to make such tasks without obstruction in all kinds of meshes.
I made a small gif to highlight issues with Bevel modifier when used with high density meshes(really hope Devs will come up with something) as well as a small issue I identified with your experimental build (Stacking more than 2 Booleans fails for me):
It’s a very powerful feature though and gives room to maneuver to avoid some artifacts that would be there if a generic bevel width would be used.
We are talking about two different tools. The boolean bevel like houdini is usefull for a lot of works and is a good tool. It’s different of a mesh fusion solution like you talk.
DcVertice you mean vdb booleans of Houdini as far as I understand. Good idea. Although I think that it is not necessary to be closed regarding other ways of achieving such results (Ilya Rodinkov’s plugin, for example, is an excellent try following a different route). Also, I am not sure about vdb’s final quality. It seems to me that it presupposes having a too dense mesh for giving really smooth results. Mesh fusion’s final quality is of very high caliber, without having to produce, necessarily, a very dense mesh. The work it does is focused on the seams area mainly. And it has the ability to produce very complex structures. Below I put a demonstration using Groboto (the app that preceded Mesh fusion). I put it in two different shadings and one mesh screenshot for showing the smoothness which produces on the seam areas and the mesh structure which underlies the shaded-rendered results:
Yeah, I ahve asked to Ilya if he could make a modifier with that pipeline. But he told me that he don’t know to code in C, I don’t think that we could see a tools like this without implication of the foundation or experience developer.
Well C4D has very robust nodal system called xpresso, that has been there for years.
Actually, there’s a lot of other things you can do as well. It’s just that no other DCC has a high level node system that’s useful so we all just think it must not be possible/useful.
Both Maya and Modo’s system is ridiculously complex and makes no sense visually. It should be:
Origianl geo>modifier>modifier>>modifier>etc.
But unfortunately in Maya all modifications going into a list on the main geo. And in modo it’s just totally bizarre. the geo node goes into the deformer and another modifer (or is that effector? I can’t remember) does into the deformer too. You also can’t visually just put something in between deformer nodes.
iirc The idea is to make Blender’s modifier nodes just like I described above. What I’m hoping for is a way to use mesh selections, vertmaps, and textures to effect the parameters of the modifiers. That would be super useful. You can do a lot of this now with the stack and the vertmap modifiers but it makes more sense to do this visually. Also, since nodes are all python, it would be cool if you could write your own.
Yes but this is a very different kind of node graph actually. Xpresso is meant for writing expressions visually. So, it’s meant to modulate parameters, not geometry. So, anywhere you would add an expression, you would use a node graph to derive it’s output. This is also similar to Modo’s schematic view. Up until they added deformers, you could only use it to write expression.
It’s still very useful but not to mesh modification Which is what we’re talking about here.
