Page 1 of 2

How to measure "real" material for rendering

Posted: Fri Apr 29, 2016 10:12 am
by spectral
Hi,

I have several materials here and I need accurate values for rendering.

Here are the materials that I want to analyse : https://i.materialise.com/3d-printing-m ... ifications

Is there some hardware/software that allow me to produce accurate analysis ?

Thx

Re: How to measure "real" material for rendering

Posted: Fri Apr 29, 2016 7:18 pm
by atlas
Measured BRDFs are an active field of research, I think a lot of the setups are experimental. I'd look at some of the papers on the subject, and maybe contact the authors directly.

Re: How to measure "real" material for rendering

Posted: Fri Apr 29, 2016 9:37 pm
by ypoissant
It depends on the accuracy of the measure you want. Start by looking at "Gonioreflectometers". It's only a good start. It can get much more sophisticated than that.

Re: How to measure "real" material for rendering

Posted: Mon May 02, 2016 8:43 pm
by dr_eck
To measure BRDF and/or BTDF, what you need is a scatterometer. There are several options available. Here's link to a page on my website that may help: http://eckop.com/optical-scatter-2/optical-scatter If you want really precise results, accurate over at least 5 orders of magnitude, the ScatterScope 4 is the way to go. If you can settle for 3-4 orders of magnitude and want results in a fraction of a second, a conoscopic scatterometer can't be beat. It's also a little bit less expensive.

Re: How to measure "real" material for rendering

Posted: Wed May 04, 2016 9:50 pm
by ypoissant
Also, I find a lot of today customers expect spatially varying BRDFs. Especially those customers dealing with textiles materials. SVBRDF produce huge data files though.

Re: How to measure "real" material for rendering

Posted: Fri May 06, 2016 7:52 pm
by spectral
This hardware looks interesting but completely out of our budget ! Just impressed !

Re: How to measure "real" material for rendering

Posted: Mon May 09, 2016 6:49 pm
by dr_eck
spectral wrote:This hardware looks interesting but completely out of our budget ! Just impressed !
Yes, the price is very high. If I could sell 100 systems, the price could be reduced by 90%, but the market just does not seem to be there. Maybe I should do a Kickstarter to really find out.

Re: How to measure "real" material for rendering

Posted: Tue May 17, 2016 5:46 am
by joulsoun
We're very close to snagging an imaging sphere, so I'm curious about how a conoscopic scatterometer deals with light sources? If we want to measure the reflectance with a number of light source angle increments.
We also have non-optically smooth surfaces that we measure, things that are really rough where we need a very big aperture to get a smooth measurement. Any reflections on if this is suitable for a conoscopic scatterometer? We mainly use measurements with SPEOS, but also to fit to material models for other applications.

Re: How to measure "real" material for rendering

Posted: Tue May 17, 2016 6:14 pm
by dr_eck
joulsoun wrote:We're very close to snagging an imaging sphere, so I'm curious about how a conoscopic scatterometer deals with light sources? If we want to measure the reflectance with a number of light source angle increments.
We also have non-optically smooth surfaces that we measure, things that are really rough where we need a very big aperture to get a smooth measurement. Any reflections on if this is suitable for a conoscopic scatterometer? We mainly use measurements with SPEOS, but also to fit to material models for other applications.
Hmm. Given the nature of this forum, I don't want to turn this post into a sales pitch, but it will be difficult. Please PM me if you want more details.

Our conoscopic scatterometers allow the incident angle for BRDF measurements to be varied from normal incidence almost up to the acceptance angle of the scatterometer. The only limitation is that it can only be varied in one plane, so you would have to rotate your sample to get additional planes. BTDF measurements are done with an optional sample table that offers similar capabilities. Scattering is, of course, measured over a portion of a hemisphere, not just in the plane of incidence.

Rough surfaces are no problem for conoscopic scatterometers up to a point. Our 80 degree scatterometer has a 1.5mm diameter input aperture. If your surfaces have a "period" greater than 0.5mm or so and you want to average, it would be necessary to take multiple measurements and average them.

Our software is set up to output either ABC coefficients or pictures. If SPEOS needs a different input, the software can be modified to give you data in a format compatible with SPEOS or your other software.

Re: How to measure "real" material for rendering

Posted: Thu May 19, 2016 8:24 am
by joulsoun
Thanks for the answer! Got you on the aperture size, but still not sure if I understand if it is possible to change the inclination angle of the light source in a conoscopic device. Looking at the schematic on your page, it looks like the beamsplitter always directs the light source straight at the sample?