How do you get existing objects into a form that can be 3D printed?

For example, something like a chess set that already exists. Is there a way to scan a small 3D object and have it be replicated or does everything have to be modeled inside a program?

You would have to get a 3D scanner. There are apps you can download, like 123D Catch, that create a 3D model by several pictures that are assembled, however I'm not sure that the quality wis as good and I'm not sure that the wireframe is directly printable, most likely you would have to import it into a program that translates it into something a 3D printer can use.

TazMan2000

You can get a physical object into 3D digital space through 3D object scanning

https://en.wikipedia.org/wiki/3D_scanner

or photogrammetry

https://en.wikipedia.org/wiki/Photogrammetry

Both can work very well, depending on what it is you are trying to scan and what your needs are.

Here's an example of some photogrammetry I'm doing for work on a particular Star Wars helmet. 71 different photos of the Navy helmet that it will be based on processed with Agisoft Photoscan and then brought into modo for clean up:



This is work in progress so I still have some steps in ZBrush and then back in modo. But hopefully this will give you an idea of the kinds of steps one would go through...

Dan

Thanks for the responses. It sounds like even with a 3D scanner there is still work that would have to be done in the computer. I've heard of ZBrush before but never used it. My skills in modeling are all outside of the computer.

Tsunemori said:

Thanks for the responses. It sounds like even with a 3D scanner there is still work that would have to be done in the computer. I've heard of ZBrush before but never used it. My skills in modeling are all outside of the computer.

Click to expand...

Yes, no matter the method, scanning or photogrammetry, there is always some work to be done via the software to clean things up and get the final 3D model into a form that you can do something with ("There is no royal road to geometry"). That said, the time taken to learn one's way around various 3D applications, like modo/ZBrush/etc., is well worth the time and will add a whole new set of very useful modeling skills...

Dan

3DImpact said:

Yes, no matter the method, scanning or photogrammetry, there is always some work to be done via the software to clean things up and get the final 3D model into a form that you can do something with ("There is no royal road to geometry"). That said, the time taken to learn one's way around various 3D applications, like modo/ZBrush/etc., is well worth the time and will add a whole new set of very useful modeling skills...

Dan

Click to expand...

Thanks again for the responses. I really appreciate it. This is almost a completely foreign world for me. I've done graphic design before (photoshop type stuff) but no 3D stuff. What's the learning curve for something like ZBrush? What kind of computer would I need to even run something like that?

Tsunemori said:

Thanks again for the responses. I really appreciate it. This is almost a completely foreign world for me. I've done graphic design before (photoshop type stuff) but no 3D stuff. What's the learning curve for something like ZBrush? What kind of computer would I need to even run something like that?

Click to expand...

I started from scratch in 3D the day after I walked out of the theater after the opening night of Avatar in 2009. I knew then and there that I wanted to be able to do that kind of 3D modeling. I too had only very good Photoshop skills up to that point. The learning curve can be steep, no way around it, but some of the applications are rather easy to learn. Depending on your learning style there is a lot of good tutorial material out there - I learn best by watching video tutorials/demonstrations, but everyone has their own way to learn. Most of the major applications have very good user forums. I spent a few years taking online courses at FXPHD. I jumped right into the deep end of the pool and started out with everything at once - modo, ZBrush, Vue, World Machine, SpeedTree. It took a good solid year until I felt fairly functional. Like any new skill, though, you'll always be learning and improving. Example - I went from starting fresh in early 2010 to delivering VFX sequences to national network television documentary productions by 2014 (and I did all that as a hobby, mind you - not even remotely related to my day job scientific research).

Now, of course the next step for you will be to identify the application(s) you want to learn to do the things you want to do. There are almost always multiple applications that can solve a particular digital modeling problem - you'll have to research and learn those that work for your needs and budget. Some are completely free and very capable (e.g., Blender). Others that are not particularly cheap may have 'learning' versions or free 30-day trials.

Most of these applications will run well on any reasonably current/up-to-date home computer. If there's one area I'd not skimp on it would be RAM. The more the better - 64 GB would not be too much. A good fast multi-core CPU will speed your render times if you're doing that kind of thing. Most of the web sites for the various applications will usually have a list of minimum machine requirements somewhere.

Good luck and enjoy the journey!

Dan

A Microsoft Kinect can scan things pretty ok.

I use a home made light tent to photograph the object in, then create a 3D model using photogrammetry (in a similar workflow to Dan), but using RealityCapture.
The results are quite amazing. Just using an average PC (but RC does require an Nvidia graphic card).

I'm researching getting a 3D printer so I am interested in this myself.

I've seen videos of people using an XBox One Kinect... has anyone done that?

Depends how accurate you want the result to be I guess. From what I've seen, the result using an Xbox is pretty average and quite low resolution.

even with the Xbox One Kinect? That has a lot better resolution than the old one

What I want to be able to do, and this may be unrealistic I don't know.

First this is strictly for my own use, I'm not starting a business or anything... I want to be able to scan parts and the resize thing and print in a different scale. My 'dream' print would be to be able to scan a 1701-D and resize it and print it in either 1000 scale or in a much more challenging 537 scale! I worked out the size dimensions for 537 and yes it would be HUGE! I also want to print a 1/18 scale BSG Viper, and then do custom prints for Phasers and blasters.

If you scan and 3d print in a smaller scale, then you could, in theory use a Kinct, but as soon as you want to upscale, then the inaccuracy of the kinnect or likewise scanners will cause trouble.

I wanted to do the same thing, but with scale models, and print them large scale, so I bought an Einscan-S, a bit more expensive, but the result is clear

whb64 said:

What I want to be able to do, and this may be unrealistic I don't know.

First this is strictly for my own use, I'm not starting a business or anything... I want to be able to scan parts and the resize thing and print in a different scale. My 'dream' print would be to be able to scan a 1701-D and resize it and print it in either 1000 scale or in a much more challenging 537 scale! I worked out the size dimensions for 537 and yes it would be HUGE! I also want to print a 1/18 scale BSG Viper, and then do custom prints for Phasers and blasters.

Click to expand...

I would love to see a 1:18 TOS Viper!

First, there are two distinctly different "forms" of 3D modeling. "Hard surface" modeling is for things like spaceships, vehicles and buildings. The other approach is more organic for sculpting things like creatures and figures. These approaches can be quite different and often require different methods and software.

I like to imagine "hard surface" modeling being like building something out of wood, paper, metal, or plastic but in 3D. In contrast, free-form organic modeling is like sculpting something from soft clay. I am strictly a "hard surface" modeler so I can't comment on organic modeling with something like Z-Brush (although I would love to jump into that if I get the chance).

3D scans and photogrammetry "solves" are really only useful as a 3D "template" for reference. The resulting scans have way too much information in them to be useful for 3D printing. It would be like scanning a photo in Photoshop at 10,000 ppi when all you needed was 300 ppi in order to make a print. The ultra-high-res scan contains way too much info and "clogs up" your system just trying to work with it.

For something complex, if you *can* get 3D scan data you could then use it as a template to build something from scratch using the scan as reference. You would then have a "clean" model that's suitable for output in a format that you can print such as SLA.

Hope that helps!

Tsunemori said:

Thanks again for the responses. I really appreciate it. This is almost a completely foreign world for me. I've done graphic design before (photoshop type stuff) but no 3D stuff. What's the learning curve for something like ZBrush? What kind of computer would I need to even run something like that?

Click to expand...

As far as scanning the hardware pricing really hasn't come down much. I have a Structured light Einscan pro. Great detail but it only handles smaller objects.
You can get good results with photogrammetry. No matter what you scan with you will end up having to clean it up and remodel it anyway to get a good print from it.
That process is usually called retopology.
I think of a scan as just really good 3d reference.

If you can sculpt in the real world Zbrush is about the closest computer analogy you will get.
Meaning the tools are fairly similar. But it is a bit of a learning curve.
It's not like Photoshop, meaning you load a file, do stuff to it, then save.
In Zbrush You can start with nothing and, shape it etc, Epxort it out, add new parts, mesh parts together etc.
Think of it more like a tool box than a linear A to B program.
For the computer any current PC will run it. Meaning a I7 Chipset. You will want as much RAM as you can get, 16 to 24 Gig at least.
A fairly good graphics card. Doesn't have to be a Nvidia 1080 card. But a good gaming card will work.
I would stick to a Nvidia cards as opposed to an AMD chipset card.
One thing you will defiantly need is a Wacom tablet.

http://pixologic.com/zclassroom/
http://docs.pixologic.com/
http://www.zbrushcentral.com/zbcinfinite.php#/page/1

Something that I've done in the past (~20 years ago) was to take one or more plastic models of what you want to recreate on a computer, and carefully cut it into thin slices with a band saw. Place the parts cut edge down on a photo scanner and use the images as reference. Don't do this if you plan on selling any part as it leaves you open to copyright infringement.

star-art said:

First, there are two distinctly different "forms" of 3D modeling. "Hard surface" modeling is for things like spaceships, vehicles and buildings. The other approach is more organic for sculpting things like creatures and figures. These approaches can be quite different and often require different methods and software.

I like to imagine "hard surface" modeling being like building something out of wood, paper, metal, or plastic but in 3D. In contrast, free-form organic modeling is like sculpting something from soft clay. I am strictly a "hard surface" modeler so I can't comment on organic modeling with something like Z-Brush (although I would love to jump into that if I get the chance).

3D scans and photogrammetry "solves" are really only useful as a 3D "template" for reference. The resulting scans have way too much information in them to be useful for 3D printing. It would be like scanning a photo in Photoshop at 10,000 ppi when all you needed was 300 ppi in order to make a print. The ultra-high-res scan contains way too much info and "clogs up" your system just trying to work with it.

For something complex, if you *can* get 3D scan data you could then use it as a template to build something from scratch using the scan as reference. You would then have a "clean" model that's suitable for output in a format that you can print such as SLA.

Hope that helps!

Click to expand...

Thanks for the info. My project isn't exactly organic looking. It's basically cubes, like a pixelated 3D object, that's why I'm trying to figure out if it would be easier to scan in a prototype or build it in a program. I can't imagine a purposefully blocky shape would be too complicated to build from scratch, but I wasn't sure how much a scanner would cut down on time in the computer, or if it wouldn't be worth it for that kind of shape.