strange behaviour of homogeneous scattering volume

Use this forum for general user support and related questions.
Forum rules
Please upload a testscene that allows developers to reproduce the problem, and attach some images.
CodeHD
Posts: 408
Joined: Tue Dec 11, 2018 12:38 pm
Location: Germany

Re: strange behaviour of homogeneous scattering volume

Post by CodeHD » Tue Nov 24, 2020 10:08 am

lighting_freak wrote:
Tue Nov 24, 2020 9:39 am
On this pictures you may see that the sample plate is almost specular:
Exactly: almost ;)

Your choice of normal map seems to be quite a good match. Attached is your scene back with three changes:
  • Clear Volume instead of Homogeneous. (The effect really comes form the surface alone. You can add some absorption color but there is no volume scattering in your material)
  • Increased the normalmap's Height setting a bit.
  • Switched to Path + GPU. No need for BiDir here.
Below is the rendering result. A little more Height will probably better match your picture.
render_satinice.png
Attachments
Musterplatte_satinice_CodeHD.zip
(1.41 MiB) Downloaded 21 times

CodeHD
Posts: 408
Joined: Tue Dec 11, 2018 12:38 pm
Location: Germany

Re: strange behaviour of homogeneous scattering volume

Post by CodeHD » Tue Nov 24, 2020 6:33 pm

Little addendum: I realised the sample is a bit darker than the background because I left the path depth at the default value of 6, so that a small Fresnel-reflection part is not traced. (Gamma correction then amplifies it.)

So just increase the total and specular path depth a little and the result will be brighter, as it should be without volume absorption or other losses.

lighting_freak
Posts: 219
Joined: Thu Jan 18, 2018 6:02 pm

Re: strange behaviour of homogeneous scattering volume

Post by lighting_freak » Thu Nov 26, 2020 8:26 pm

Hello all,

finally the sample part that should use the DF23 material. I hope the download-link works:
https://www.dropbox.com/t/4OvpV1lyUnVKTr5R
Material definiton in that sample is different, compared to the one that we made during the last days.

I continue the investigations on the material - the goal is still a valid volume definition. The DF23 material is a diffuse one, 100% sure. My current target procedure is to find a volume & surface setting that fits the visual appearance (somehow in daylight and lit through condition) to the real sample. This volume will find it's way into that model and for the surfaces I have to find a new way.

Do you have any idea how to model a VDI grain structure that creates a special light pattern while it shines through the surface?
I'll keep you updated for the sample plaque progress.

BR
OS - Windows 7 X64
CPU - Intel CORE i7
GPU1 - Variants of notebook card from nVidia
GPU2 - Variants of notebook onboard card from Intel
Lux - Latest possible relaease

CodeHD
Posts: 408
Joined: Tue Dec 11, 2018 12:38 pm
Location: Germany

Re: strange behaviour of homogeneous scattering volume

Post by CodeHD » Thu Nov 26, 2020 8:50 pm

Looking back at the data sheets, you might be right that there should be a diffuse volume component to the DF range, however it really looks small if any in you images... Do you have any more explicit data on what DF"23" and "8N" actually mean? I didn't find anything on their website.

What is a "VDI" grain structure? I don't know that abbreviation.

lighting_freak
Posts: 219
Joined: Thu Jan 18, 2018 6:02 pm

Re: strange behaviour of homogeneous scattering volume

Post by lighting_freak » Fri Nov 27, 2020 7:07 am

HI,

sure I can try to describe the meaning:
Let's start with "8N" - it is the clear basic PMMA material.
The "DF_XX" is the density of light diffusing elements within this basic acrylic. The higher this number the more diffuse the volume is.

There are some images that show some key facts to the pure (not satiniced, but polished surface) material (sorry that they are in German):
This is how it works:
principle.jpg
This is how diffusion is defined and measured:
measurement.jpg
They measure the angle that shows a light intensity fall-off to 50% of its highest intensity at normal view.
The diffusion angle is given as "half width half max - HWHM".

This are the results for vayiing DF materials:
scattering.jpg
I hope this helps to understand the volume.

The VDI is a description of how to make a reproduceable grained structure.
It's a definition of a surface roughness.
VDI.jpg
For my test setup I'd like to have the DF23 volume combined with the VDI 36 surface.

With best regards
OS - Windows 7 X64
CPU - Intel CORE i7
GPU1 - Variants of notebook card from nVidia
GPU2 - Variants of notebook onboard card from Intel
Lux - Latest possible relaease

CodeHD
Posts: 408
Joined: Tue Dec 11, 2018 12:38 pm
Location: Germany

Re: strange behaviour of homogeneous scattering volume

Post by CodeHD » Fri Nov 27, 2020 12:57 pm

Ok thanks for the additional info. Really don't know why I resented the idea that VDI would be the organisation in this context :lol:

You would certainly have to play with the anisotropy of the volume scattering to get this effect. I could imagine one could even recreate this test setup, and write a render-script to find good values. Unfortunately, that still needs an additional info: the thickness of the material for which this material property is defined.

Regarding the VDI grain: That seems to define a certain integral of the surface shape (I have not looked up all details), which can surely be computed for any surface. An accurate representation of your material will probably require more knowledge though, since multiple shapes coudl fulfill the same roughness criteria. How big the impact on your case would really be, again no idea...

lighting_freak
Posts: 219
Joined: Thu Jan 18, 2018 6:02 pm

Re: strange behaviour of homogeneous scattering volume

Post by lighting_freak » Mon Nov 30, 2020 2:06 pm

Hello all,
CodeHD wrote:
Fri Nov 27, 2020 12:57 pm
You would certainly have to play with the anisotropy of the volume scattering to get this effect.
I don't believe that we have to use anisotrophy settings. The material is a molding compound, no directional dependency is expected if you do some real parts from it.
CodeHD wrote:
Fri Nov 27, 2020 12:57 pm
the thickness of the material for which this material property is defined
I forgot to mention that the scattering cone that is mentioned in the pictures above is measured at a 3mm thickness. I could get it for other thicknesses as well.
CodeHD wrote:
Fri Nov 27, 2020 12:57 pm
An accurate representation of your material will probably require more knowledge though,
I can measure the scattering cone of the sample plaque, but how can I transfer this information into luxcorerender (keywords BRDF/BSDF/BTDF)?
CodeHD wrote:
Fri Nov 27, 2020 12:57 pm
since multiple shapes coudl fulfill the same roughness criteria.
You're right, this defintion is purely a mechanical one and different shapes do really create different light scattering... I know that this will require deeper investigation afterwards but at first a valid way to insert those data would be nice.

This is my current status:

Normal view:
still not blurry enough
still not blurry enough
Slight tilted view for reflection:
good specular reflex
good specular reflex
OS - Windows 7 X64
CPU - Intel CORE i7
GPU1 - Variants of notebook card from nVidia
GPU2 - Variants of notebook onboard card from Intel
Lux - Latest possible relaease

CodeHD
Posts: 408
Joined: Tue Dec 11, 2018 12:38 pm
Location: Germany

Re: strange behaviour of homogeneous scattering volume

Post by CodeHD » Thu Dec 03, 2020 2:25 pm

lighting_freak wrote:
Mon Nov 30, 2020 2:06 pm
I don't believe that we have to use anisotrophy settings. The material is a molding compound, no directional dependency is expected if you do some real parts from it.
If you want to model the ~20 degrees diffusion angles, then you do. With anisotropy = 0, scattering is equal in all directions and you have no HWHM by definition.
lighting_freak wrote:
Mon Nov 30, 2020 2:06 pm
I forgot to mention that the scattering cone that is mentioned in the pictures above is measured at a 3mm thickness. I could get it for other thicknesses as well.
One thickness should be enough.
lighting_freak wrote:
Mon Nov 30, 2020 2:06 pm
I can measure the scattering cone of the sample plaque, but how can I transfer this information into luxcorerender (keywords BRDF/BSDF/BTDF)?
Your full material is a combination of a volumetric scattering and a rough or "satinated" surface.
As said before, I would start with finding good values for the scattering scale and anisotropy. Note that for a good comparison with a real material, you should minimize other differences. Most importantly, compare it with a sample that has a polished surface, not rough.
Finding the right surface model may be a bit more tricky. For on-axis, the trick with anormal map might give you good results, but I fear at large angles you may see the influence from micorfacet-multiscattering. I'm not sure if any LuxCore matieral at the moment supports that.

Post Reply