Rank novice with Blender … doing almost anything seems to result in very large amounts of hidden (interior) mesh geometry. For example, I created a UV sphere and chose various blocks of four faces and extruded them outwards a short distance. I was surprised to find that the original faces were still in there and the extruded “cap” consisted of new faces. I had thought it would take the hollow-shell approach.
Do Blender modelers really have to mess around with getting inside their meshes and cleaning up all this hidden stuff? I’m vaguely aware of using things like Limited Dissolve, but in my experimentation it hasn’t worked out too well.
In one tutorial I followed, the guy started out using subsurf on a cube and the final model duplicated probably 75% of the geometry. I didn’t realize this until my UV unwrap produced what appeared to be multiple copies of the same faces. I still don’t know how he got his clean UV map, I followed his very first couple of instructions (which didn’t leave any room for user error) and went straight to UV map and interior dupes were everywhere.
I feel like I must be missing something important. It’s made worse by the fact that many Blender tutorials are oriented to static renders or videos, rather than game modeling – which is probably the only place this matters much, apart from even less-common uses like 3D printing.
One quick gotcha is that many of the tools start the process and leave you in a state of needing to move the geometry. For example, when you extrude, you should just do it exactly once, and right then move the mouse. If you escape, or right click, the extruded geometry is still there, right on top of where it came from, and then if you extrude again thinking it didn’t work, you suddenly have the problem you describe. There are other tools that are similar.
The other gotcha that is similar involves the mirror modifier. If you have a face right on the plane of the mirror, it gets duplicated sometimes as well, depending on the tools you use. I’d have to see exactly which tutorial you were following, and maybe follow it in my head to catch some of the other things that happen. But Blender in general CAN work fine without these problems, as long as you complete each tool as you do it.
Thanks. This is the tutorial. Mostly I was looking for a primer on UV so it served its purpose in that regard. (I’m at that amusing stage where I’m trying to grasp just about every step in the process all at once… fairly comfortable with Unity, now I’m nailing down how to get useful stuff into Unity!)
His first step is to subsurf a cube (subdivisions = 2) and when I started backtracking, that’s where I saw duplication straight away. I took just that object with no further changes and went straight to UV and most of it is duplicated. I think that’s probably just the nature of subsurf. (That’s what I meant by no user error – it’s just a few steps, no mousing around, etc.)
Eventually I tried just extruding from a simple UV sphere (as described above) and that’s where I found the extra geometry after extrusion. I wasn’t extruding multiple times. I’d just select four adjoining faces, drag them straight out and apply. Then after deleting one of the newly-added side faces, I could easily see the interior original faces.
Or maybe people just don’t bother cleaning up the interiors?
Neither of those things should create duplicate geometry. Modifiers create “sort of” duplicates, in that it keeps the original geometry and constantly calculates the resulting geometry. But that is the point of modifiers, to be able to create changes without destroying the original, which is useful both in creation and render time(if you were using Blender’s renderers). Once you apply the model, and when you export it(or Unity does it automatically) it applies all of the modifiers top to bottom so the original is no longer there.
As far as extruding, I just tried it with a subdivided cube(single face), and also with a UV sphere(4 faces like you said), and I extrude, drag, and then delete the extruded verts, and I’m left with a square whole in the sphere. I even try doing a bit of extra clicking and re-extruding, and it still deletes what is under. The extrude process is best done in vertex or face mode though, with proper selections. It is possible you are selecting only edges or something, I don’t know. Also, just to make sure you understand properly, the extrude works in a 4 step process.
Select the faces(either via all the verts, or in face mode).
Hit ‘E’ key(or select extrude in a toolbar)
Drag the geometry out with the mouse(Don’t click at all until you are done positioning).
Left click to confirm the operation.
If you use the E key, you can have your mouse cursor at the right spot, so the dragging is much more intuitive than when using tool bars. And any random clicking will complete the extrusion too. Extrusion by definition should delete the geometry where it is coming from.
I don’t have anything set up, but if you happen to, I’m willing to offer a small bit of free help(like 10 minutes or something) if you have something handy for remote viewing of your screen. If you do this though, have handy and easily reproducible the ways you are getting duplicate geometry so we can easily work out what is happening. I’m willing to install something if needed as well. In fact, I’m considering making a bit of consulting business so this would be a way to start. And I don’t mind helping for free a bit too as I’ve gotten plenty of help from the community here and other places.
No idea what the heck was going on when I tried the tutorial and my earlier experimentation. I just now loaded up Blender again and tried both the simple UV sphere extrusion and the subsurf/unwrap and neither produced duplicated geometry this time and the unwrap looked clean. Bizarre.
I guess my general question still applies, though. Is it commonplace to spend time cleaning up hidden geometry? For example, in this excellent tutorial about using Blender’s built-in UV paint features, the little spikes on the cactus are independent meshes that intersect the cactus body. Each one has geometry inside the cactus:
Well, if you create the geometry using “box modelling” which is generally taking a form and turning it into what you want by moving and extruding(and possibly including the subsurf modifier), than everything is attached to that one thing and you are good.
Now, it is generally acceptable, moreso in modern game engines to have geometry like the cactus there. If you force the spikes to be connected to the trunk, you will get some really ugly geometry in the trunk to accomodate that. But if you keep it separate it can stay clean. It is much easier to create that way as well, possibly using tricks like particle systems to create the spikes. As long as you don’t have any faces that are too close to being co-planar(which would cause Z-Fighting) you are OK with that. Animation can be affected as well, so you may have to pay attention to it if you are animating the model in question. You also want to watch the potential waste. For something like the spikes, there really isn’t much of that, but for other things, you may need to make sure you don’t have a bunch of polys in there, not because of problems, rather because they aren’t seen and therefore are wasted when rendered. This shouldn’t be an issue though if you personally are creating the model.
There are situations where you can’t have a model be of separate pieces like that. Some older game engines and renderers required single manifold meshes(though I don’t have any examples). Also, you would get initial problems if you try to do 3d printing with the model.