I need smart people to tell me which program I should use to materialize low res polygon models.

cod8smith

I am working on a way to materialize Low resolution 3d poly-models and i need programs that can possibly take a 3d model, fill it up with triangles or other straight lines (or show the ones already in it), and calculate the corners in these triangles. I'd would like to adjust the resolution or something, so I can kinda make polygon art.

I know stl. files are already constructed out of geometric shapes and i need to get the information above out of them.

I don't have a lot of experience on 3d modeling programs (only solidworks), so feel free to educate me on existing programs.

Feel free to ask more questions if I am unclear, because I don't know shit about this yet XD

Eric Chadwick

What is the intended output going to be? How exactly are you going to present the finished results?

It would help to embed reference image(s) of what you want the models to look like.

Where are the models from? Are you making them yourself, purchasing them, ripping them, what?

poopipe

as above.  the answer is probably Houdini though

cod8smith

I want the model to kinda look like this, but then open. My intention is to create an open structure and physically build it. For this i need to create a model like the one below (I'd preferably scan something using a 3d scanner or take an existing model and transform it) and calculate the x/y/z values of all the corners.

Now there might not be a program available that can already do this, but something might come close.

As I said, the models can be everything.

And thanks for the quick reaction XD

poopipe

ok - so you want to take a scan/CAD model or similar highres model,  turn it into a super low res one and then build a real world physical model out of it right? 

Houdini is free, spend a few days learning about poly reduction with it and you'll be able to get what you want in terms of the model.

the dimensions/corner stuff I'm confused about.  any 3d app must know where the vertices are or it won't work. 

can you be more specific about that part?

cod8smith

well, to create a physical structure, i want to make the connection points. And i need to know the corners of each triangle. I think this has to be in the code somewhere, but i need this to find what degrees the lines are relative to each other.

This way i can produce the connection points and build the model as a puzzle.

Thanks for the advice, i will check houdini out :-)

musashidan

cod8smith said:

well, to create a physical structure, i want to make the connection points. And i need to know the corners of each triangle. I think this has to be in the code somewhere, but i need this to find what degrees the lines are relative to each other.

This way i can produce the connection points and build the model as a puzzle.

For this part you could 3D print the model to scale(of the physical model size that you intend to build) and then use a scale ruler and protractor to find edge lengths/triangle angles.

Neox

hm shouldn't Pepakura be kinda like what you want? It turns 3d models into paper models you can print out and build yourself.

cod8smith

Pepakura does what i want to do with a model, but I still need to retrieve the angles of the folding lines leaving all the vertices. I plan on doing this for more than one model, so a successful program or way of retrieving this info is crucial.

poopipe

All the information you need has to be there or it is impossible to render the object in a viewport,  it's a question of getting it out in a form you can use. 

Eg.  It's not difficult  to work out the angle between two triangles and output it to a list but making the list human readable is a problem -  not least because tris and vertices aren't necessarily stored in a sensible order. 

This is quite an interesting problem,  I'll be curious to see how you deal with it

Eric Chadwick

Handing someone a puzzle like this would be very very difficult to solve. A box filled with balls and sticks, the balls having somewhat random-looking holes, and the sticks being seemingly-random lengths. Even if you color coded them by height or something, it would be near impossible to piece together.

pior

I'd encourage you to try and clarify what you want to achieve as much as possible. For instance the follwoing sentence doesn't make sense :

"but I still need to retrieve the angles of the folding lines leaving all the vertices." 

Now I think I kind of understand what you are getting at ... but still, start by explaining your needs to the absolute best of your ability - even if that means taking the time to make diagrams.

RyanB

Use whatever tool you want to create a low poly model
Break the model into individual triangles
UV map the entire model with zero distortion, there are ways to do this in every 3d package
Pack the UVs but give yourself some space between them
    You can cut each triangle out of the UV map and it should have the exact surface area
Print your UV map onto whatever thin material you want, cut out tris, glue them together
Also need to number the triangle edges so you know where they go

This was just off the top of my head, some scripts in whatever 3D package could easily automate the whole process.

RyanB

For the verts, you can access whatever info you need with a bit of scripting or knowing where to look.

Houdini verts

Blender verts

Max and Maya are similar, you just use a bit of scripting to print them out to whatever file you want.

Eric Chadwick

Apparently the OP would like to make a puzzle as a lattice, like TinkerToy. 

cod8smith

wow thanks for the advice. And yes I would like to produce a puzzle like that, but I don't want to give to much away.

I've been trying to dive into this, but my experience with modelling programs does not go further then Solidworks atm. However Houdini and blender look promising.

For now, getting the data, even if i have to read and transfer it by hand will make the concept possible. I can work on refining it later. I just need to know where the holes in the balls need to be, and what the length of the sticks need to be.

But to sum up the idea: Take a 3d model, turn it into a real life wire-frame structure, that is build with balls and sticks kinda. Ill find solutions for finding a way to puzzle a box of balls and sticks into Ollie here for instance.

RyanB

cod8smith said:

I just need to know where the holes in the balls need to be, and what the length of the sticks need to be.

PHASE 1:
Write a script:
Place a sphere at each vert. 
For each edge of the model, place a stick (extrude a circle along the edge).  
Chop off an amount from the end of each extruded stick so it intersects the spheres by 1 cm or 3/8 inch (this amount will have to be experimented with to find right amount).
Do a boolean operation to cut out the holes from the spheres where the sticks intersect.

Assign vertex colours or some other identifier to each sphere and stick end.  This would create a weird rainbow pattern over the model.

Could be vertex number instead.  You could also put a unique identifier on the end  of each stick and the hole it goes into.  For example, 21-B would identify vert 21 and stick B.  Then stamp 21-B next to the hole and at the end of the stick.

Print a 3D prototype.

PHASE 2:
?

PHASE 3:
Profit

throttlekitty

ok, so here's the basic data you have easy access to in a 3d app. You can also check all this by loading an obj file in a text editor.

Vertex/Edge/Face ID: an index number assigned to each component.

Vertex location: A XYZ value that defines each points location in space.

Faces are constructed by listing the vertex IDs that make it up, clockwise. (I believe all programs will start the numbering with the lowest number in the vertex id index.)

So if you selected that vertex, you may have something like 13:12.011, 7.560, 130.113. (vertex id: x, y, z). Selecting a face would return something like 13,52,176,192.

Do you plan on building these by hand like with wood, or on a 3d printer? The 3d printed route would make this easy, it's just a matter of doing what RyanB suggests, a bit of automation and making sure everything is physically sound. If by hand or something, my suggestion would be to write a script that:
*Takes an input for the size of the spheres and the hole depth.
*Measures the angles at the vertex of each face.
*Does the math to calculate the coordinates where holes should be placed, and the length of each stick.
*Print this into a text file in a way that makes sense as instructions: The ID for each sphere, the coordinates for holes (and their IDs), and what sticks go in which holes.

Marking the spheres for drilling sounds fun I used to have a cool globe with a kind of ruler on it, i can't remember the name and that's bugging me.

Prime8

at least the length of each edge can be displayed very easily in Blender, if I remember correctly it's an attribute of the edge data as well. Could be extracted together with the edge ID... I don't wanna think about how to match that to the actuall model 
For the vertices... I'm sure there is no such thing as the angle between vertex and edge out of the box in any application. you would have to create some relation using normal or global space for each vertex and then figure out the angle for two axis. Not sure how you get the holes in tge balls with information though.

RyanB made a good suggestion, I would probably approach it that way too.