unfold

[ssmeier]

This web site may be of interest:
https://papercraft-maker.com/

This is the thread that the web site is mentioned:
http://www.papermodelers.com/forum/softw...ative.html

-Steve

[micheus]

Cool! It's free and generate PDF files.

It seems to be a good option for the 2nd option I talked about in the previous post.

[engrenage]

micheus, that video you've shared is impressive! (and so is the cost of the software)

1) right, I get it now. that was just an example because I have yet to see hydroforming of 3D shapes, in my case it's somehow "too basic and coarse" to be of any real use.

2) why does your illustration in #post14 show that the columns of each color sets are detached? there still is the problem of vector output...

I wrote another reply which was somehow lost, so I'll repost its content.. basically it's aout how I managed to create the first unfold illustration I posted for my egg.

- select one face only
- view -> align to selection
- file -> export selected -> cartoon edges > [OK]

loop over for all wanted faces, import the resulting SVGs in [inkscape, ...] to rotate/assemble them.

I had some scaling issues with a few parts, probably because I inadvertently changed the zoom level at some point.

maybe it's easier to automate this rather than use the auto-UV method*? in this case, what I would suggest is let the user select a number of edges. then, recurse over the faces attached to one edge, then if not all folds have been processed create a new group and so on.

---

ssmeier, this is better than nothing but I still see some issues.

first and most important, I notice there is some co-planarity problem with pentagonal faces that were subdivided (and shouldn't have been). I suppose this is due to rounding errors somewhere along the line, so even if I don't cut the superfluous edges I will have problems to assemble afterwards.

Couldn't figure out a way to change the size of the tabs or remove them completely.. not a major issue I suppose though.

again, it goes by selecting the cuts, but _any_ real-life problem will have more cuts than folds so it (mathematically) makes more sense to select where to fold (as opposed to where to cut).

regarding the difficulty of selecting an edge, selecting both faces touching that edge is also an option (edge selected? excellent, fold this. face selected? never mind, prompt to select another face and figure out which edge connects these two faces.

I am not in favor of on-line browser-based software.

a software that unfolds should imperatively output the folds and cuts on different layers or groups. For papercraft this is not an issue, but when it comes to laser/waterjet/milling these need to be treated differently because of the depth dimension (folds will just be engraved a few micrometers, while cuts will go all the way through).


Will now send an e-mail to phuocdh to lete him know about those bugs.


* auto-UV method has the advantage of being able to apply a texture and preview how the result's gonna be. in the case of CNC machining, the texture would need to be vectorial anyway and that can be added later on.

[micheus]

Ok engrenage.

I'm going to think about something. But, it probably will take some time.

[micheus]

2) why does your illustration in #post14 show that the columns of each color sets are detached? there still is the problem of vector output...

That was intentional. Just to illustrate I unwrapped the regions in "strips". Bellow you will find it unwrapped in a different way.

I wrote another reply which was somehow lost, so I'll repost its content..

It's in #post 16. I'm not sure it was this one, but for some reason the forum engine marked one of your posts for moderation. I approved it a few days ago.

maybe it's easier to automate this rather than use the auto-UV method*?

Not for me now. That will require I to look in a part of the code I'm not used to play around. I still have my own project to "finish" and just a couple of free time to code.

basically it's aout how I managed to create the first unfold illustration I posted for my egg.

- select one face only
- view -> align to selection
- file -> export selected -> cartoon edges > [OK]

loop over for all wanted faces, import the resulting SVGs in [inkscape, ...] to rotate/assemble them.

Please don't do that!

In the video bellow I try to show you how to do basically the same process inside Wings3D in an easier way. It will require you patch two files.
I also tried to cover some possible options you may need to use to cut and "re-cut" an island/group.
There is no audio (sorry), but I tried to be slow then you can follow the steps and options being used. Shotcuts I used: [E] edge, [B ] body, [L] edge loop, [D] repeat last command, [SPACE] deselect, [A] Aim camera and [Shit+A] Frame camera.



Attached is a .tar file including two modified files is used in the video: wpc_autouv.beam and wpc_hlines.beam. Use the File->Install Plug-in or Patch to update them.

It's an unofficial version of the current plugins and valid only for v2.2.6.1 - for which I have tested.

If you decide to remove them, you will need to access the user folder, like:
"C:\Users\<user_name>\AppData\Local\wings3d\2.2.6.1\plugins"
and delete them manually.

[engrenage]

micheus: cool! I will try this ASAP and let you know of the results.

---

micheus, that was #post16 indeed. sice you're listed as as admin, I thought you had read the blacklisted message. never mind, the info got through and this is what matters.

"cloth" is totally the way to go, although I I have not been aware of a way to specify the direction of the fibers (angle your cloth fibers at 45° and the result will be _totally_ different) ; in practice the cloth will _never_ be equi-directional. also, there are many many different types of "woving" and this will radically mess up the way deformation is computed. I am not looking for a way to go as far into that, meshes or fibers are not (at this point) relevant so the surface should be considered as homogenous in nature.

---

but still.

this is also too-far-fetched anyway.

unfolding and - if possible - smoothing out vertices as suggested in #post12 is most probably the neatest way to go in this case. the rest is just-- cosmetics*?

* not withstanding the fact that it is _definitely_ easier to select edges that will be folded instead of cut because their number differ in a favorable manner.

[micheus]

Ok. While I wait you try that change in the code - which will make the process easier than you are doing now - let's keep thinking about this different way to open the mesh you have been trying to make me understand...

Back to the "egg" example, the edge approach seems to not be the case for it, but the current way used to unfold in Wings3D (like in the video).

So, thinking in something less obvious, when opening a mesh...
Do you already have something in real life you made in metal using the current approach?

Since the cloth approach seems to not really be the way to reach the result you expect, that could be a way I visualise the 3D model, the opened mesh and the object assembled.

By selecting the edge to produce the cut and unfold faces, which situation of those would you expect as result?

unfolding and - if possible - smoothing out vertices as suggested in #post12 is most probably the neatest way to go in this case. the rest is just-- cosmetics*?

That reference is about to deform an object while the unfold is about to open a mesh of an object. Again we are talking about two things.

I can imagine to add a new way to open the UV mesh as something useful somehow; But, deform the object making assumptions about where to smooth and in a non "physical" way means to create a targeted tool - not useful for the current context of CG apps which uses the cloth simulation for that - even the material not really been a fabric (it can be rubber or something more rigid by controlling some parameters involved in the algorithm).

I cant say I would have time (very sparse) to work in something like this (it may be also beyond my ability ).

[engrenage]

micheus, definitely the second (lowermost) option! although it seems to me the faces alongside the red (then yellow) line must be co-planar, or else we wouldn't have a solid group (like in the top example).

having co-planar faves at this point of the discussion and not being 100% certain these actually are co-planar or not is confusing to me.

considering this, none of the above options is correct

[engrenage]

what did I say.

None of your two illustrations match the expected behaviour.

With a single fold as you illustrated, _all_ other edges must be cut. So the result would be as such:




plus, of course, quite a number of detached, single quads.


When I wrote "select the edges to fold", I mean strictly this. All other edges must be cut. No "trying-to-be-smart-feature-that-fails-and-gets-in-the-way", just fold, if not, cut.

[micheus]

Ok. I was thinking you were still looking for some automation.

In this case, what you need is already available.
If you proceed this way - selecting only the edges to keep and cutting everything else, you just need to use Selection-Invert:
a) in Geometry window, context menu set the invert the selection, set the edges to Hard Edges and when you use the UV Mapping command all the edges will be already Marked for cut and finish with Continue-Unfold or;
b) Select the object and use the context menu UV Mapping and in the UV Segmenting window mark the edges like you did, Selection-Invert and use the context menu Mark edges for cut and finish with Continue-Unfold.

Try this approach.