PBR Material Scanners?

MeshMagnet

I was wondering if anybody could shed some light on PBR Material Scanners. I've seen a few postings around here about them, I've also heard that the Quixel guys built a custom one for the Megascans that will be available later this year.

So far, I could only find this link. It seems more practical for industrial applications and uses proprietary software:
http://www.optis-world.com/products/hardware/OMS2.html

Does anybody know if there are any other types available?
Has anyone build one from scratch?

Any info would help.

Xoliul

I think you might be underestimating the effort and money that goes into this.

MeshMagnet

I really have no idea, that's why I was asking.

Do any of you think it would be possible to use a laser 3d scanner; something similar to this?

[ame="http://www.amazon.com/Artec-3D-Spider-Scanner/dp/B00GPDS3GO/ref=pd_sim_sbs_indust_4?ie=UTF8&refRID=15X8ESXD4CZ8T50G7HPS"]Artec Spider 3D Scanner: Remote Controlled Robots: Amazon.com: Industrial & Scientific[/ame]

I understand it's not cheap by any means, (at least for a student or hobbist).

Could something like this be used to capture PBR materials? Or am I way off?

sprunghunt

MeshMagnet wrote: »

I really have no idea, that's why I was asking.

Do any of you think it would be possible to use a laser 3d scanner; something similar to this?

Artec Spider 3D Scanner: Remote Controlled Robots: Amazon.com: Industrial & Scientific

I understand it's not cheap by any means, (at least for a student or hobbist).

Could something like this be used to capture PBR materials? Or am I way off?

laser scanners only capture the geometry - which is fine if you just want to make the bump/normalmap.

However a proper material capture device usually involves a bit more than that. And the technique you probably are thinking of is called photogrammetry Here's a paper from an academic about it:

http://www.pauldebevec.com/Research/LS/

and a company which loosely describes their process:
http://ir-ltd.net/cracking-the-black-box-of-photogrammetry-scanning/

to just capture the BRDF of a material - which is not the same as PBR material capture but close. What you want is called a gonioreflectometer.



here's an paper about that:
http://research.microsoft.com/en-us/um/people/xtong/bootstrap.pdf

Xoliul

Well there's a few maps you'd want to capture, right;
Diffuse/Albedo and normal are relatively easy. You can do that with a regular camera and something like Agisoft.
However that's by no means full PBR yet, you want to capture the specular reflected light too. You can capture that with polarizing filters and combining it with an unpolarized photo, this would give you just the reflected light. However, that's where it gets really tricky; the reflected light intensity is determined by many factors:
-Light intensity, but you can control this
-Fresnel angle, this changes for every angle, hence why this stuff can't be easily photoscanned like diffuse.
-Roughness/smoothness; super tricky to capture since light rays are scattered, and make the intensity appear different.
-IOR/F-0 value; similar value for most materials, but there's a lot of materials (metals notably) where this value will differ and influence your intensity.

Basically, you have a single equation with 3 unknown values. That gets pretty tricky to solve.

And keep in mind, that scanner you linked does scanning on a single point (it's able to control the angles and measure scattered light like that). It's massively more complicated to do it for an area.