[Fernando Tella]

Would be nice to have a Nd list for opaque objects. I'm quite lost with that. Well, 20 for mirrors, 3 for diffuse, but I would like something a bit more accurate: natural wood, varnished wood, natural arid stone, polished stone, copper, steel, aluminium, matte plastic, reflective plastic,...

[tom]

Since Nd is a generelized approach of measured ior, it's truly experiemental. For example, a typical glass has Nd 1.5 but this is not highly precise comparing to a measured ior data. So, feel free to set Nd depending the look you need.

[Frances]

Since Nd is a generelized approach of measured ior, it's truly experiemental. For example, a typical glass has Nd 1.5 but this is not highly precise comparing to a measured ior data. So, feel free to set Nd depending the look you need.

It is very difficult to get consistent results and it is very easy to get unrealistic results with this approach. This is what is so frustrating about the new material model. Trial and error versus inputting known values for known materials. If the look we need is varnished wood, then the nd setting shouldn't be a guessing game - especially since meaningful feedback from editing takes such a long time.

And as for Attenuation - having to have a separate material for each object according to size, even if they have the same texture is inconceiveable. This is what I've gathered about that parameter from reading what Mike had to say about it.

[-Adrian]

I verymuch agree with Frances, this bugs me.

[jonathan löwe]

yep!

seems strange...or maybe we understand something wrong?

[tom]

Hi Frances,

I think I made a big confusion by saying "experimental". Well let me try to
explain with my horrible English. Of course there are known Nds for
materials like water 1.33, glass 1.5 etc and you don't need to think twice if
water is 1.33 or not. This is same for all the applications, not specific to
Maxwell. So what is experimental here? It's unknown reflective opaque
materials. For example shiny wood floor. If you know or measure the Nd
for the varnish on it and the naked wood, you can have the look you want
precisely. But personally I do not think this much precision is necessary
and it's not because Maxwell cannot do it, it depends on the information
you have about the substances you want to simulate. I would simply set
my varnish Nd 1.5 like glass and go on. About the word you used
"unrealistic", no this doesn't mean you will have unrealistic results.
Because there are substances in this range and this doesn't mean
"unrealistic", this means you are setting wrong Nd for what you want to do.
Let's say a user should know 1.33 for making water, right? Or instead,
he/she has to use a preset material. Where and how does the experiment
(experiment doesn't mean doing arbitrary things) begin? There are 2
parameters; Ref 0 and Ref 90 and there is a transition between them, it's
called fresnel. The curvature (transition) is under control of Nd value,
which is a physcially accurate and correct method of helping user
simulating decay of reflection with fresnel phenomenon. Nd=0 is ghost,
which is used for special cases. And Nd>1 turns your object reflect starting
from ref 90, ending at ref 0, which is varying for millions of objects in real
life and you don't need it 100% precise to have the realism you need. For
example you can set 1.48 or 1.6 for glass and it's still glass. It's primarily
your eye looking at a preview sphere which will decide the transition of
reflection you need it to be as you perceive these substances in real life.
But if you look for something like, "What's Nd of my coffee cup?", you
won't find an accurate answer. You have two ways to solution: 1) You can
measure it with expensive devices or you can find the Nd for sodium line in
the catalog or 2) You can simply set it with your perception like you did
the same with other applications for years because Nd is same with ior in
other applications and it's not something special with Maxwell. Here's a
simple definition of how this system works, I guess it would help.




In short, you will use the same IOR value you use with other applications.
And if you have further questions, please let me know.

Best regards,
Tom

[Frances]

Thanks for the explanation tom.

[tom]

And as for Attenuation - having to have a separate material for each object according to size.

This is a wrong assumption, too. Because attenuation of a substance is always same. For example "honey", let's say the honey sample I have has the attenuation of X and it has a color R1,G1,B1. With this **single** setting I can render a smear of it, a jar of it or a tank of it. It will produce me the correct result because with different thicknesses of honey, the result with natually change. Can we talk about a tinted transparent substance having different thinknesses but producing homojeneous same color? Physically no. However traditional render engines are able to produce this with omiting absorption and this means correlated color independent of thickness, which is an old habit of thinking non-physically. In short, you don't need to prepare different versions of same material about attenation for your scenes. If your glass is tinted cyan for 1m thick, just set it's transmittance color to cyan and put 1 m attenuation as in real life, so you'll see it will render white for 3 mm thickness. But if you tend to make glass with white transmittance (which is unreal) and expect having color when it's thick and when you can't have the color, if you attempt to make a second version for tinting the glass, this run will never end and you will produce hundreds of same thing for different sizes just because you think it designed to work this way. No. I'm making the explanation very straightforward to make a better sense, no worries here.

[Fernando Tella]

Thank you Tom for the technical info.

[tom]

Thank you and don't worry, I will soon provide you a better and detailed explanation of this material approach on the forum with full interaction. Only text and manual is not enough, you need to ask questions and have answers.

[Fernando Tella]

For previewing the attenuation I think it would be nice to have some preview scenes as the one is used in the manual (the valley thing) with different heights or scales; for example, from 0 to 1cm, o to 10cm, 0 to 1m, 0 to 10m,... so it would be easier to preview. Maybe different radious spheres are enough.

[Frances]

This is a wrong assumption, too. Because attenuation of a substance is always same.

Perhaps you and Mike should talk.

[-Adrian]

...

[Miles]

[tom]

ROFLMAO