VRScan Tech and Accessability

[MegaPixel]

Hello,
I was wondering what the folks at Chaosgroup were using behind the scenes to create these material scans? Looking at the sample size limitations stated in the FAQ, I'm going to guess it's a desktop type scanner with a specific bed size, maybe something similar to one of the links below? If it is proprietary hardware/software, I understand not being given the answer. If we were to pursue somewhat larger surface area scans, like stone walls where patterns might represent 3 meters or more up and down, I'm guessing this would break the limits of the current tech? I'm starting to see some other material scan sites pop up that seem to be capturing rather large surface areas and I'm wondering if they are using a more portable/scaleable solution or if the scans are based off of 3d modeled objects(like Quixel Megascans)? I'm sure the tech isn't cheap at the moment but I can definitely see a desire in the near future for more artists to get involved with making and sharing their own PBR scans, considering the vast near infinite variations in materials out there. With the recent advancements in realtime tech, it seems like material development is becoming the bigger bottleneck than it was before.

Some quick search results for Physically Based Material Scanners:
www.vizoo3d.com
www.xrite.com
www.optis-world.com

Thanks! Dave

[kosso_olli]

We have extensively tested the VrScans and found out that it is by far the superior product to the three alternatives you mentioned. The only one that comes remotely close is the X-Rite system, but as of now there is no connection to 3dsMax, Maya or V-Ray. Their AxF-Format is pretty useless for our DCC software. So we tested it in Vred, but the results were worse compared to the VrScans. We have sent some materials samples to Chaos, and the scanned materials are extremely close to the sample. The amount of details is just crazy!
I don't know where the journey goes, but in some podcast Dinko said that their goal is to make the scanner available to anyone. I appreciate that, would buy that immediately!

[svetlozar.draganov]

I sent the thread to Dinko, he'll provide more details on the subject.

[dinko.dimitrov]

Hi Dave,

You have started a great topic.
Chaosgroup is using proprietary hardware and software to scan and build digital materials.
Our purpose is to create a hardware that digitalizes physical world materials in the very best possible way. This means that we not just want to create a hardware that will scan and produce a digital copy of a given material or something similar to it but an exact copy of that material as the result of this is a photorealism, accuracy, and ease of use.
We want once a material is scanned the digital copy to behave in the same way in any environment and under any resolution without any other adjustments.
Of course, there is always a tradeoff and currently, this is the scanned sample size which with the new hardware version is 30x40 cm (doubled than some months ago).
We are also looking to increase the scanning samples size double or triple in the future but larger the scanned sample is bigger the scanned material size is.
Ultimately we want to make the scanning hardware publicly available and we are targeting next year, i.e. 2018 to start offering the harware.
I will be glad if you can drop a line to vrscans@chaosgroup.com and to have the chance to talk a bit about your work.

Thanks,
Dinko

[dinko.dimitrov]

We have extensively tested the VrScans and found out that it is by far the superior product to the three alternatives you mentioned. The only one that comes remotely close is the X-Rite system, but as of now there is no connection to 3dsMax, Maya or V-Ray. Their AxF-Format is pretty useless for our DCC software. So we tested it in Vred, but the results were worse compared to the VrScans. We have sent some materials samples to Chaos, and the scanned materials are extremely close to the sample. The amount of details is just crazy!
I don't know where the journey goes, but in some podcast Dinko said that their goal is to make the scanner available to anyone. I appreciate that, would buy that immediately!

Thank you for the great feedback Oliver!

[^Lele^]

For the size Dinko mentioned, you get a resolution of 60 microns, or twice the width of a wool fibre, per pixel.
Ollie is very right in his perceptual analysis: x-Rite provides a BTF rendition of the materials, but *not* for the portable scanner, just for the room-sized one they own and operate (much like it is with us now. minus the room-sized hardware.), while OPTIS provides for the full BRDF/BSDF again only with the big sized OMS4, while the portable scanner produces a simpler result (And both can only be used within that family of tools exclusively.).

Every other solution extracts textures off the scan, to plug them into a simple principled BRDF (xrite portable, RTT/Dassault).
Granted it increases the flexibility of some aspects, but it also misses the point entirely as far as shading (if not color) goes by doing so: we scan *because* there is no principling with textures and a BRDF a number of materials.

As for the final results, the shader needs light distributed on it, so if it doesn't render with V-Ray, it doesn't get the V-Ray "lighting distribution", and so even a purportedly identical or better scanned solution may render poorly because of render engine limitations (say, GI and glossy bounces, GI methods, AA filters, IS methods, and so on and so forth.), or it may render a version of the scan in realtime, although of course cutting corners to do so, either in the material description or lighting approach (we aren't still on that front either...).