absorption and scattering coefficients for liquids

[OnoSendai]

Jensen has a nice paper up on his site with a table full of absorption and scattering coefficients for many liquids - wine, coffee, milk etc..

http://graphics.ucsd.edu/~henrik/papers ... roperties/

The values are in units of 10^-2 mm^-1,
so i think you need to multiply them by 10 to get to the m^-1 that indigo uses now.

EDIT: the scattering coefficient can be used as is, once u have multiplied by 10. To get the absorption coefficient, subtract the scattering coefficient from the extinction coefficient (and multiply by 10).

Note that i have only implemented uniform scattering, so non-uniform scatters will not look 100% correct yet.

I haven't tried this data yet, if u get any results, pls post them!

nik

[fused]

uh thank you. was wondering about the absorption and scattering of coffee... have been modeling my cup

will try that when my cpu got some spare time

[Phil]

Hi,

super link really interesting!
Right now I have a lot of crash tryzingHi nik,

super link really interesting!

I am just playing with this and right now I have a lot of crash trying to render a glass of Merlot wine!
I do not know where the problems are coming from, I also use the 07 exporter, and I will investigate this more in detail.
At this point I have some questions:

@all:

1- Does anybody here achieve some good results with this?

2- What about the units in Blender?
I offen import obj files from CAD program, and in this case the units are in mm, so should I scale the model?

3- Does the sss function work with all lights (map, sunlight, emitters)?

Thanks in advance, regards

Phil

[OnoSendai]

Hi, apparently there are crashes with SSS and sunlight, so if ur using sunlight, try using a mesh emitter instead (it will render faster as well)

[Zom-B]

I would love to test some milk or wine too,
but I feel somehow stupid with all these terms I don't understand

perhaps somebody could tell me how to translate the values from the Table into a Indigo material...

so what exactly are:
- Extinction Coefficient
- Scattering Coefficient
- Average Cosine
- % RMS Error
in Indigo language???

The first three are RGB values but I only see two possibilities for RGB in Indigo:
<scattering_coefficient_spectrum>
<absorption_coefficient_spectrum>

[IanC]

Ono, the values for milk are about 0.9124 for the scattering and 0.9126 for the extinction.

That gives scattering of about 10, which seems fine, but absorption of 0.001 if you take one away from the other. This seems really small, am I just not understanding it, or is that right?

[ryjo]

Here is a render of milk inside glass:



edit: The medium def:

Code: Select all

	<medium>
		<name>milk_scattering_medium</name>
		
		<ior>1.35</ior>
		<cauchy_b_coeff>0.0</cauchy_b_coeff>
		<precedence>10</precedence>

		<absorption_coefficient_spectrum> 
			<rgb> 
				<rgb>0.001 0.003 0.013</rgb> 
			</rgb> 
		</absorption_coefficient_spectrum> 
		
		<subsurface_scattering> 
			<scattering_coefficient_spectrum> 
				<rgb> 
					<rgb>11.873 13.293 14.589</rgb> 
				</rgb> 
			</scattering_coefficient_spectrum> 
          
			<phase_function> 
				<uniform/> 
			</phase_function> 
		</subsurface_scattering>
	</medium>

[Zom-B]

OK... I read Nick's first post again, and think I now understand how to get
<scattering_coefficient_spectrum> & <absorption_coefficient_spectrum> set
So far so good

So I tried to get a glass Merlot & Chardonnay rendered using the following mediums:

Chardonnay

Code: Select all

   <medium>
      <name>wine</name>
      <ior>1.33</ior>
      <cauchy_b_coeff>0.003</cauchy_b_coeff>
         <absorption_coefficient_spectrum>
      	   <rgb>
               <rgb>0.7586 1.6429 1.9196</rgb>
           </rgb>
         </absorption_coefficient_spectrum>
      <precedence>4</precedence>
      
      <subsurface_scattering>
         <scattering_coefficient_spectrum>
           <rgb>
               <rgb>0.053 0 0</rgb>
           </rgb>
         </scattering_coefficient_spectrum>
         <phase_function>
            <uniform><value>0</value></uniform>
         </phase_function>
      </subsurface_scattering> 
   </medium>

Merlot

Code: Select all

   <medium>
      <name>wine</name>
      <ior>1.33</ior>
      <cauchy_b_coeff>0.0</cauchy_b_coeff>
         <absorption_coefficient_spectrum>
      	   <rgb>
               <rgb>0.1547 0.1701 0.3443</rgb>
           </rgb>
         </absorption_coefficient_spectrum>
      <precedence>10</precedence>
      
      <subsurface_scattering>
         <scattering_coefficient_spectrum>
           <rgb>
               <rgb>0.030 0.047 0.069</rgb>
           </rgb>
         </scattering_coefficient_spectrum>
         <phase_function>
            <uniform><value>0</value></uniform>
         </phase_function>
      </subsurface_scattering> 
   </medium>

Here are some quick Renderings:





doesn't look very promissing at the moment... I realy hope I did something wrong

[ryjo]

ZomB,

I think your absorption_coefficient_spectrum is too low. This is what I get for Merlot:

absorption_coefficient_spectrum: 7.586 16.429 19.196
scattering_coefficient_spectrum: 0.053 0 0

I haven't tested this though...

[IanC]

ryjo, thanks!

Thats a beauty of a render, btw.

[Zom-B]

ZomB,

I think your absorption_coefficient_spectrum is too low. This is what I get for Merlot:

absorption_coefficient_spectrum: 7.586 16.429 19.196
scattering_coefficient_spectrum: 0.053 0 0

I haven't tested this though...

hmmm... you did multiply the scattering_coefficient_spectrum AND the scattering_coefficient_spectrum by 10.

Nicks first post tells only about multiplying scattering_coefficient_spectrum
OK I' try to Render Some Wine Over Night the way you say

[OnoSendai]

Both the coefficients should be mutliplied by 10, i've added a clarification to the first post now.

See the glass test post: http://www.indigorenderer.com/joomla/fo ... .php?t=637 for information on how to model the glass correctly for indigo.

You also need to make sure your glasses are the correct size, bearing in mind that indigo units are meters.

[Zom-B]

OK, here the result after 7h.... Merlot & Chardonnay

[Camox]

@ZomB
Nevertheless, this already looks much better.
Have you provided with Blender or with c4d and xml edited ?

[Zom-B]

Hi, apparently there are crashes with SSS and sunlight, so if ur using sunlight, try using a mesh emitter instead (it will render faster as well)

Sorry to tell you dude, but it seems to crash with env_maps too, and also if you try to illuminate the Scene with a background Spectrum thing without other lightsources

@ZomB
Nevertheless, this already looks much better.
Have you provided with Blender or with c4d and xml edited ?

At the moment I still use the old Exporter from BlackLizard, and edit the XML by Hand... after understanding the new medium system its quite easy
Learning Blender is still on my list, but I'm working on a big C4D (School) Project atm,
after this is done I move to Blender...

[/quote]