Absolutely at wits end

I am in a bit of a conundrum and I’m not very far off from pulling my hair out.

One of the things I do as an artist is make models for 3D printing. Most of the time my tasks are a snap. Between Blender, Photoshop, Zbrush, Netfabb and like TWENTY THOUSAND other programs, there is no shortage of tools at my disposal to aid in efficiently putting out good, printable 3D models. However, there is one extremely vexing problem I frequently encounter -hollowing out very complex models such as this: http://bobbie-the-jean.deviantart.com/art/Temp-post-2-415262765

There are MANY ways to hollow out simple models but for more complex models, things become more difficult. The methods I know of for hollowing out complex models frequently result in some rather irritating issues. Here are the problems that arise from the various methods of hollowing:

1.) Solidify modifier: This often results in messy, intersecting geometry, especially if the thickness is ramped up. It is entirely possible I’m just doing it wrong but when I try to use a solidify modifier, it ends up ballooning the geometry in strange ways. I also want a nice, uniform 3mm thickness.

2.) Extrude: I thought perhaps I could select all faces, extrude in, and BOOM! Done. I was always able to do this really easily in Maya but I absolutely cannot get it to work in Blender. I have no idea what I’m doing wrong. Another issue with this method is that Blender doesn’t like high poly counts.

3.) Piece by piece: I could build the internal thickness piece by piece but that would take FOREVER. When I say forever, I mean months. There are so many little nooks and crannies, maintaining a uniform thickness by hand would require A LOT of eyeballing, and generally, doing things like this by hand would be extremely inefficient.

Does anyone have any suggestions? I would really love an easy, efficient way to get a good, uniform thickness with as little intersecting geometry as possible, if there is a way, that is. I realize I’m probably not going to get a magical “easy-fix” solution but at this point, I’m just hoping there is something more efficient than doing it by hand.

What 3D printing software are you using? Many modern ones can calculate what you’re asking for from a .stl file at print time automatically.

You could try solidify with even thickness and high quality normals turned on, but I assume you’ve already given that a try. My only other suggestion is to do an extrude along face normals (E > Alt + S), but this could give strange results as well. Honestly, I’ve never had to do this manually so there may be better options that I’m unaware of.

I’m sending my prints off to Shapeways. I tried solidify with even thickness and high quality normals. Maybe I’ll fiddle around with that some more but I don’t have high hopes for that option. I’ll try your extrusion suggestion and see how that works.

Thanks for the reply. :slight_smile:

Well unfortunately, that didn’t work. :\ Same problems as before. I have no real control over thickness without causing seriously weird geometry issues.

That kind of task is for marching cubes algorithm or a like. If you can find this toolamongst your 20k i’d suggest try it out ;).

I’d simply use the technique of a clay modeler and scoop out a hollow inside the piece without reference to the surface irregularities. Is there some overriding reason why the interior of the piece must be absolutely 3mm from the surface of the piece? When you use an algorithm that gets rid of intersecting meshes that probably won’t be the case anyway.

Just take the lower edge, extrude it in and up to create a smooth hollow that stays roughly 3mm from the exterior surface. Alt+B cross section views can help making sure the hollow stays inside the outer mesh, and I suspect that it wouldn’t take more than five or ten minutes for the piece in your DeviantArt temp post.

Save your hair. Ignore the complexities of the exterior of your model.

The Solidify modifier uses a “brute force” approach to create the thickness. It’s meant to be fast rather than pretty and, ultimately, it’s just a tool with no big button labelled “Read my mind and make everything awesome”. You must work if you want some good result.

I see 2 ways: Either you re-work the mesh created by the Solidify modifier or else you give it something less convoluted to work on.

I prefer the 1st solution. First, make sure you have several (flashy) materials and set the Solidify modifier to use them so that, after applying, you can easily select the outside, the inside or the rim by their material. Hide the outside (the original mesh) and the rim and work on the inside only. The first thing you can do is to select the vertices in the problematic areas and use [CTRL NUMPAD_+] to select around and, then [W] --> Smooth. Repeatedly. Until the fold/overlap/intersection disappears. If that’s not strong or fast enough, delete the problematic vertices, select the edge of the hole and re-build faces with [CTRL F]. (Triangles aren’t a problem for 3D printing.) And if you need more drastic measures, you can simply separate the inside into a new object and use the Decimate or Remesh modifiers, apply and re-join.

For the second solution, make a copy of you mesh and do everything that I suggest in the 1st solution on it. When this copy looks like a melted version of the original one, then you can use the Solidify modifier on it and apply. Select the inside (thanks to its distinct material) and separate it to join it with the original mesh. Use [W] --> Bridge Edge Loops to close the rim. (Recent versions of Blender work even if the number of vertices don’t match.)

(Do I need to remind you to keep your normals under control?)

As for having a uniform thickness everywhere, that’s a dream that you can buy only with the money from the Tooth Fairy. :wink:

EDIT: Orinoco summed it up right: “Save your hair. Ignore the complexities of the exterior of your model.”

The problem is that 3D printing is currently VERY expensive and cost scales by cubic centimeter. So I need to eliminate as much unnecessary geometry as possible to keep costs down. For the material I want to print this in (sandstone) the minimum wall is 2mm. 3mm would be sturdier though. That’s why I need the thickness to be uniform. If you doubt the importance of this, understand that I’ve brought a 450$ model down to half the cost by cutting away unnecessary material. So this is very important. I can’t just halfarse this. Getting a nice, uniform 3mm thickness will be the difference between people being able to afford my model and very much not being able to afford it.

I could do it by hand but it would be a great deal more efficient By calculation. I’m not averse to doing work, that much should be obvious. I didn’t just fart that model out of my bum. :stuck_out_tongue: But I need efficiency And I need consistency. This is a problem I encounter frequently. See my problem?

You have summed it up nicely. You can save time or you can save money. Or a little of both. In any case it is geometrically impossible to have a solid be completely inside another one and also be a uniform distance from it at all points. (Edit: for usual purposes. Clearly it’s possible for convex smooth surfaces… )
Use the solidify mod and apply it, then select interior faces to find problem spots and use smooth vertex to get rid of them. The solidify will make sure no parts of the mesh are too close to the outside surface. Using the normal extrude will entail much the same procedure, the advantage being that you can do it in batches.

Like you said, you have the skills and the patience and the first snowbirds are probably showing up already so you might as well stay inside and plug away:D

Time for a graphic, “Blender, why do you mock me?” :smiley:

you do have the mesh analysis that can help you find thickness and you can manualy try to correct some of the worst faces!

would like to find the script how it is done

happy bl

I was using the aforementioned method before (correcting by hand) but I am really hoping there is a more efficient way to do this. :\ The problem with correcting by hand is that on a model this complex, there can literally be thousands of places where the model is way too thin and/or way too thick. This can result in high costs and unstable structure.

Blender’s mesh analysis is buggy (for me) but it can be helpful sometimes. I realize I cannot have a perfectly uniform thickness all the way around but it would be nice if I could get it close enough and correct the trouble parts by hand. The problem with solidify is that pushed too far, it causes this weird ballooning effect. Not in front of work computer ATM but I will show pics of what I mean.

unitl we get it in blender
did you try whit matlab i think
it has some tools to do constant thickness

happy bl

I will be trying that when I get home. Here’s hoping. Last time I tried it I couldn’t get it to work but on the other hand, I never gave it a fair shake. So I’ll be trying that again. :slight_smile:

I understand why you are doing what you are doing and the costs involved. I’m merely suggesting that if a quick and dirty by hand solution can reduce your printing cost from $450 to $225, and an ‘ideal’ solution would reduce your printing cost from $450 to $217, then how much is that ideal solution really worth?

Unless that model in your DeviantArt temp file is 200 cm high, you aren’t going to get a uniform 3 mm thickness on the inside. Not doing it by hand, not using some fancy algorithm.

You might also consider making a cheap plastic model and using that to make an actual latex mold, then casting the piece in the latex mold. Might get your costs down well below $200, even if you didn’t hollow out the initial casting at all. That would have the added advantage of being able to reuse the latex mold for more copies, if it sells.

That wasn’t a quick and dirty solution though. :frowning: It literally took me weeks and it was it unfathomably tedious and frustrating. O___O; I’m going to try this Meshlab idea but you’re probably right. I’m probably going to end up having to do this by hand…again. It really is a nightmare. I’m not a lazy artist. I just love efficiency and I hate busywork. I’d rather spend my time being creative than on tedium. But the fact is, in 3D, tedium is sometimes unavoidable. I still hope there might be some quick fix… Not because I’m looking for easy fixes for the sake of easy fixes. I frequently need to hollow out very complex models and it takes a LOT of time that I could be spending on other things.

I wish I could explain… It’s almost like having to resculpt the entire model all over again in negative while making sure to maintain as consistent a thickness as possible. O_____o

It’s almost like having to resculpt the entire model all over again in negative while making sure to maintain as consistent a thickness as possible.

Exactly!:smiley: It is essentially retopology - from the inside. Just like retopo there is no substitute for doing it by hand. And like retopo if you don’t like doing it it’s going to be horribly tedious.

No explanations necessary, we get it. But dude, if it takes you weeks to save $225 in material cost when you could save $200 in material cost by being less of a perfectionist, and do it in five minutes by extruding a simple cylinder up the middle of your sculpture and making a few minor adjustments… You have to ask yourself about the value of your time.

Shrinkwrap a mesh around your object. Scale the shrinkwrap mesh down a bit, move it inside. Attach the mesh at the bottom to turn it into a hollow space inside the object. OK, not perfect, but it only takes about ten minutes.

Question really is, how much better can you do by trying to make a meticulous model of the outside of the object on the inside, such that the wall thickness is as close to 3 mm as you can get it. And is the difference really worth the time you spend?

This is really a failure on my part to communicate the complexity of the situation so let me explain:

I have had a lot of people tell me that they would buy that dragon jar if I made it into a 3D print. I have some experience with 3D printing so I figured why not? The model’s already done, mostly, I just need to hollow it out. However, from experience working on other complex models, I knew this would be NO easy task. Yet I could stand to make a lot of money here. The best part is that the dragon jar will sit in the store collecting money while I continue to work on other things. Great plan! Right? No more student loan and medical debt for me! Yaye… not so much. Here’s the problem- 3D print costs scale by volume. So if I make a model that is 10cm, it might cost 30 bucks. If I scale that same model up to 20cm, the price can skyrocket into the hundreds.

In order to keep costs down, I need to cut away as much unnecessary geometry as possible. A uniform 2mm or 3mm thickness would accomplish exactly that while still allowing for structure integrity. I realize some parts could not support that… like… his little nubblies would have to be solid. That’s fine. But in order to keep costs down, the walls would need to be as uniformly thick as possible.

It’s not really about being a perfectionist so much as making this a feasible business venture. Let’s face facts here- in this economy, people can’t really afford to drop 500$ on something so frivolous. Don’t get me wrong, I’m not insulting my work. I worked very hard on that model and I’m proud of it but the fact is, it’s not a necessity. It’s a luxury item and a very niche-interest one at that. Most people wouldn’t even spend 200$. So if I want to actually make any money on this, I would need to keep the costs as low as possible. And trust me, scooping out what looks like only a little material can lead to hundreds of $ in cost reduction. But doing it by hand can also cause structural integrity problems.

I too wonder if this is worth the effort which is the only reason I’m not investing as much time in 3D printing as I’d like to. If I could, I would quit my dayjob and just do 3D printing. However, I just don’t think the tech is quite there yet for that. It’s too young and currently, way too expensive for my liking.

On the sort-of-not-really-upside, this is the only really big issue I have with 3D printing. Aside from not being able to efficiently hollow out 3D models and create a relatively uniform thickness… omg, it is 1:13 am. XC I need to go to bed before I die. I think I’ve made my point? I’m going to try Meshlab method. If that doesn’t work, I’ll try shrink wrap. If that doesn’t work, I guess I’ll just have to do it by hand. But for now, I need to bed. Dying here. Falls face down on keyboard.

Keep us posted. I would be curious to know if everyone is indeed wrong and you can do this automatically, and if it really makes sufficient difference in the volume to warrant the extra time. Diminishing returns… Good luck.