Well there is digital scanning technology. In the Lord of the rings special features they describe using some sort of adapted meat scanner for
creating 3d surface data from marquettes. heres some info.

[Link to www.aranz.com]

---

that sounds exactly like what i was talking about... why hasn't it gained more popularity? surely if a person sticks his/her head in the scanner, it would save hours/days of modelling..... and i'm guessing the resulting mesh density can also (hypothetically) be controlled since it's all digitized...

---

Cogito, Ergo Sum

what you get is point cloud data which tends to be very messy and very high res. If you want to animate the meshes you pretty much need to re model them over the top using the scanned mesh as a guide. still usefull though.

heres a pic I made with a model I found here: [Link to graphics.stanford.edu]

---

---

wow that's really hi-res....not very practical i suppose unless that's to be the centre of your scene, and not modified or animated in any way....
but it's definitely interesting, and maybe some progress in that area would yield better results, and the mesh density could be controlled
i don't really understand their procedure though...will have to find out what all that jargon means....
looks exciting....do you think maybe if autodesk took it up seriously they could improve it and streamline it? iron out the chinks?

---

Cogito, Ergo Sum

You can get these models in lower resolutions but they miss the detail that comes with the super hi res point clouds and the main problem is that the topology is completely useless for animation. I think for animatable meshes you will probably always need to manually organise the topology so it works for the motion of the character. I have seen a Maya script used that conforms vertices of one mesh to an underlying mesh. but that is still quite a manual process. you can also re-topologise using z-brush so that could be used to create user friendly animatable meshes from point clouds. It would be awesome if there was some sort of scanning equipment that could yield much more user friendly meshes but i can't think how they could ever organise the topology automatically so it flowed well for animation and also get a detailed scan without using loads of polygons. One method that I saw once was some sort of pen that could be used to manually plot vertices in 3d but physically pressing onto a marquette. that way you could draw the best topology onto the physical model and then build the 3d mesh by kindof tracing over it polygon by polygon. you never know though. clever people keep coming up with amazing stuff all the time anything is possible.

---

I saw that pen method in the special features of Toy story.

---

[Link to www.3dprevis.com]

I've seen that too, I always wondered how they generate the 3d positional data from that. Anyone know?

---

Please support us and post a [Link to www.digitalartsfront.com] on your website.

--Matt

Now that you mention it no. The pen must have sort of fixed reference point, otherwise I can't imagine how it would work.

---

[Link to www.3dprevis.com]

Now that you mention it no. The pen must have sort of fixed reference point, otherwise I can't imagine how it would work.

---

[Link to www.3dprevis.com]

the pen method you describe sounds reasonable enough....if you get to specify the topology then it's awesome otherwise, as you say, it's pretty useless....
i'm just wondering, since the cross-section at any one given point of the object (the part being scanned) is 2d, then can't the cross-section be treated as something similar to a nurbs spline for instance....since we can create loft surfaces to attach splines together, can't we use a similar technique to create the in-between mesh? (between two cross-sections)..... for high-res meshes, theoretically, we'd then have to sample cross-sections more frequently, so that there is less in-between mesh to interpolate..... and for low-res meshes, increasing the distance between cross-sections....
i'm on pure theoretical ground here i know nothing of how this works
but in actual fact, i'm guessing you'll probably only use the 3d scanner if you need real detail.... i saw the pic of them scanning the LotR troll, and the quote there said that they didn't want to leave it to the animators/modellers alone, probably because somewhere along the design process, there would inevitably be changes to the mesh, and it would not look exactly the same as the physical marquette.... considering their budget and the level of their design team, that's probably a little paranoid

---

Cogito, Ergo Sum

I'm not sure how it all works either but I'm sure new methods will come up for procedures like this. there alot more substantial 3d scanners being used than the meat scanner method for creating digital doubles for actors and such. On LOR I think its just that the process was that all these incredible traditional sculptors spent months on end detailing thier designs and transfering that directly into displacement maps was going to be better than getting another set of artists to try and copy those details in zbrush. I think they did train several concept artists to use zbrush though. seeing as its such a naturalistic way to add detail.

---

I guess this sort of thing would be handy for companies working to a deadline that want to replicate something exactly, but for the average Joe hobbyist it would take all the fun and satisfaction out of creating models. Unless you create the model in clay yourself first...

---

[Link to www.3dprevis.com]

yeah that's true....although if you're working to a deadline, and, say, your scene has to be full of hi-res characters with lots of detail, it would help if you had a few clay models lying around, which could be scanned.....but i guess if you put time and effort into making clay models, you might as well do it straight into 3d

---

Cogito, Ergo Sum