some tests made during the development

[Ivaylo Katev]

some of the materials are included in the free library

[Ivaylo Katev]

... and more

[Ivaylo Katev]

the height map and the brdf of the cylinder are from two different materials, beside the fun, it can be useful in some cases

[super gnu]

forget the materials scanner.. i want that van

p.s. these materials are amazing. im still skeptical about their usefulness for large surfaces ( architectural etc) do you have any examples of say 2m x 2m or bigger?

[Ivaylo Katev]

do you have any examples of say 2m x 2m or bigger?

the currently operating devices can't handle such a size, but if there is significant interest toward large scale materials, we will construct corresponding devices

[super gnu]

i actually meant one of your existing materials tiled over a larger surface to see how it looks... Lele suggested they actually tile much better than expected.

[Ivaylo Katev]

ah yes, really the main part of the efforts were spent to the tilling, you know the annoying problem when you try to use simple texture over a large surface, the tilling is visible even the material is "uniform", there are actually always differences that make the borders visible. we spent significant efforts to equalize the borders, in the most cases the tilling is invisible

[stevesideas]

Looking very nice !

[Art48]

So do you have some images where we can see a larger surface being covered? I think it's all black magic what you guys have done there, and I don't believe it until I see it :P

[dinko.dimitrov]

You can play around with the scenes we have here:
http://docs.chaosgroup.com/display/V...VRayScannedMtl
All the materials in them a VRscans. I believe the chair model will answer your questions.

[Ivaylo Katev]

So do you have some images where we can see a larger surface being covered? I think it's all black magic what you guys have done there, and I don't believe it until I see it :P

no problem, we have some test renders with big objects, the sizes are in mm
you can see the borders of some materials, there is place for improvements, but for now we are happy with the results

[super gnu]

interesting.. i think for almost any architectural scale scene you are going to need much larger scans.. is it practical to scale up the scanner to massive sizes? do you just need an x/y motorized platform to move the scan head?

id imagine that if it takes 4 hours to scan a 200x200 sample, then 5mx5m panels are going to be quite painful.

another alternative would be to do multiple 200x200 scans and develop a method to randomly tile them so repetition is practically eliminated.
this would be a good solution for "noisy" textures, but im not sure how that would work for materials like wood.

anyway, its probably a more practical solution than a room sized scanner that may only be needed once in a while.

[Dan Andersen]

Awesome, looks very convincing.

[Ivaylo Katev]

well, currently we have no device for large materials, but if we have one, of course the pixel size will grow, to keep the same material size.