A glimpse into the future of POV-Ray's Subsurface Light Transport; the 
left bead uses a classic texture; the others use SSLT textures with 
increasing translucency; material definition syntax is as follows:

     texture {
       pigment {
         bozo
         color_map {
           [0.0 color rgb <1.00,0.05,0.05>]
           [0.4 color rgb <1.00,0.80,0.05>]
           [0.7 color rgb <1.00,1.00,1.00>]
           [1.0 color rgb <0.05,0.80,1.00>]
         }
         scale <0.3,1,0.3>
       }
       finish {
         diffuse 0.6
         ambient 0
         specular 1.0 roughness 0.001
         reflection { 1.0 fresnel }
         conserve_energy
         subsurface { translucency Trans }
       }
     }
     interior { ior 1.5 }

where Trans is either a float or an RGB vector.

Note how the effective color is determined by the pigment (except for 
unusually "low density" materials, where some side effects kick in); 
also note the addition of transparency where the object's "density" is 
particularly low compared to its size.

I also threw out some yet undiscovered bugs.

Please don't get too excited yet though: The changes will not make it 
into 3.7.0; I expect to rush out a patch though as soon as a 3.7.0 
version is released with a non-beta license.

Post a reply to this message

Attachments:
Download 'translucency_test.png' (251 KB)

Preview of image 'translucency_test.png'

Same beads, illuminated from behind and without the ground plane

Post a reply to this message

Attachments:
Download 'translucency_test.png' (124 KB)

Preview of image 'translucency_test.png'

On 20/02/2011 5:58 PM, clipka wrote:
> Same beads, illuminated from behind and without the ground plane

Looks good! Keep up the good work.

-- 
Regards
     Stephen

Post a reply to this message

clipka <ano### [at] anonymousorg> wrote:
> A glimpse into the future of POV-Ray's Subsurface Light Transport...

Wow, that looks really great Christoph! I've been thinking about SSLT ivory
since Stephen first showed his lamp scene; such a pain to get it right! Here are
a couple of 10cm-long test rhinos, on left with no SSLT, on right with SSLT base
texture and layered texture with pigment at filter 0.6 (trying to match a
reference photo), no radiosity. Not too bad I guess...

But your new method looks to be SO MUCH better! A few questions:

* Will radiosity work okay with it? I get lots of blocky artifacts with SSLT and
radiosity in the current implementation.

* Will we be able to set a scattering color that's different than the specified
pigment?

* Does this new version still use the scattering/absorbing coefficients and
mm_per_unit?

Regardless, really looking forward to trying it - thanks for your hard work!

Cheers,
Rob
-------------------------------------------------
www.McGregorFineArt.com

Post a reply to this message

Attachments:
Download 'rhinos.png' (560 KB)

Preview of image 'rhinos.png'

On 20/02/2011 8:16 PM, Robert McGregor wrote:
> I've been thinking about SSLT ivory
> since Stephen first showed his lamp scene; such a pain to get it right!

You're telling me ;-)
Have you tried putting your models in a scene? I'd be interested in the 
results.

> Here are
> a couple of 10cm-long test rhinos, on left with no SSLT, on right with SSLT base
> texture and layered texture with pigment at filter 0.6 (trying to match a
> reference photo), no radiosity. Not too bad I guess...

Not too bad at all, at all. I've not thought of layering over SSLT.

-- 
Regards
     Stephen

Post a reply to this message

Am 20.02.2011 21:16, schrieb Robert McGregor:
>
> * Will radiosity work okay with it? I get lots of blocky artifacts with SSLT and
> radiosity in the current implementation.

I'm not sure - that depends on what you're seeing. I found and 
eliminated a bug in the code that kind of "shattered" a surface 
according to which component of the surface normal was smallest. Other 
than that, I'm not aware of any problems with combining radiosity & SSLT 
(except that SSLT-enabled objects currently don't actually use the 
radiosity code, but an inferior substitute).

> * Will we be able to set a scattering color that's different than the specified
> pigment?

Yup - the sample images actually do use this, being more translucent for 
reds and oranges (giving them the brownish hue seen in the backlit 
scene). However, the effect of different translucency settings is not as 
straightforward as the effect of the pigment (as can be seen on the 
rightmost bead in the examples, in which the reddening effect is less 
present).

Also note that if you want an object to appear as if coated with a 
different color, you'll still need to layer textures.

> * Does this new version still use the scattering/absorbing coefficients and
> mm_per_unit?

No - the current parameterization will cease to be supported. For the 
new syntax, see the sample code I posted with the first image.

The mm_per_unit setting will still be required though, as the 
"translucency" parameter value is specified in mm instead of POV-Ray 
units, so that the same material can be used regardless of scene scale.

> Regardless, really looking forward to trying it - thanks for your hard work!

Heh - thank /you/ people for encouraging me to pick up that work again :-)

Post a reply to this message

Am 20.02.2011 21:40, schrieb Stephen:
> On 20/02/2011 8:16 PM, Robert McGregor wrote:
>> I've been thinking about SSLT ivory
>> since Stephen first showed his lamp scene; such a pain to get it right!
>
> You're telling me ;-)
> Have you tried putting your models in a scene? I'd be interested in the
> results.

I guess the new parameterization will really make it a lot easier: It'll 
boil down to first picking a suitable pigment, and then just toying 
around with a single translucency parameter until it looks convincing.

Post a reply to this message

On 20/02/2011 9:27 PM, clipka wrote:
>> Have you tried putting your models in a scene? I'd be interested in the
>> results.
>
> I guess the new parameterization will really make it a lot easier: It'll
> boil down to first picking a suitable pigment, and then just toying
> around with a single translucency parameter until it looks convincing.

Better and better :-D

-- 
Regards
     Stephen

Post a reply to this message

"clipka" <ano### [at] anonymousorg> schreef in bericht 
news:4d618629$1@news.povray.org...
>
> Heh - thank /you/ people for encouraging me to pick up that work again :-)

Hail to the Chief! Great work indeed.

Thomas

Post a reply to this message

Stephen <mcavoys_at@aoldotcom> wrote:
> On 20/02/2011 8:16 PM, Robert McGregor wrote:
> > Here are a couple of 10cm-long test rhinos, on left with no SSLT, on right
> > with SSLT base texture and layered texture with pigment at filter 0.6
> > (trying to match a reference photo), no radiosity. Not too bad I guess...
>
> Not too bad at all, at all. I've not thought of layering over SSLT.

Here's what I was using - curious to see what this looks like on your lamp:

// SSLT Units
#declare SSLT_UNIT_CM = 10;
#declare SSLT_UNIT_M  = 10 * 100;
#declare SSLT_UNIT_IN = 10 * 2.54;
#declare SSLT_UNIT_FT = 10 * 2.54 * 12;
#declare SSLT_UNIT_YD = 10 * 2.54 * 12 * 3;

#declare SSLT_Samples = <16, 8>*8;

global_settings {
   assumed_gamma 1
   mm_per_unit SSLT_UNIT_CM*10
   subsurface { samples SSLT_Samples.x, SSLT_Samples.y }
}

#include "rhino.inc"  // the model is 1 POV-unit long

object { Rhino
   texture {
      pigment { rgb 1 transmit 0.1 }
      finish {
         // ivory?
         subsurface { <0.70, 0.72, 0.57>, <0.000314, 0.00049, 0.00067> }
      }
   }
   texture {
      pigment { srgb<235,220,190>/255 filter 0.6 }
      normal { agate 0.1 scale 0.15 }
      finish {
         specular 0.7 roughness 0.01
         reflection { 0.05, 0.95 fresnel } conserve_energy
      }
   }
   interior { ior 1.54 }
}

Cheers,
Rob
-------------------------------------------------
www.McGregorFineArt.com

Post a reply to this message

On 21/02/2011 1:29 PM, Robert McGregor wrote:
> Here's what I was using - curious to see what this looks like on your lamp:

I'll post a comparison when your material has finished rendering. From 
what I can see your texture looks much better. Mine finished in 10 
minutes and after 20 minutes yours is 18% finished.

I think it needs higher subsurface { samples }

In the meantime I posted the DFX mesh in povray.binaries.misc in the 
thread Lamp with SSLT Figure, if you want to try it. (3 Megs)

-- 
Regards
     Stephen

Post a reply to this message

Am 20.02.2011 18:35, schrieb clipka:

> Please don't get too excited yet though: The changes will not make it
> into 3.7.0; I expect to rush out a patch though as soon as a 3.7.0
> version is released with a non-beta license.

It's coming, it's coming! Dev team colleagues did agree to include it 
even in the next 3.7.0 release candidate!

(jumps up & down in excitement)

I can't wait to see what you guys will make of it...

Post a reply to this message

On 21/02/2011 8:37 PM, clipka wrote:
> It's coming, it's coming! Dev team colleagues did agree to include it
> even in the next 3.7.0 release candidate!
>

Good news and well done :-D

> (jumps up & down in excitement)
>
> I can't wait to see what you guys will make of it...

With your artistic talent you shouldn't need to wait. ;-)

-- 
Regards
     Stephen

Post a reply to this message

On 21/02/2011 6:48 PM, Stephen wrote:
> I'll post a comparison when your material has finished rendering. From
> what I can see your texture looks much better. Mine finished in 10
> minutes and after 20 minutes yours is 18% finished.

I don't know what order these images will be posted but yours is the 
lighter one. I think that for this model the agate pattern is 
superfluous. But more discussion might be wasted with Clipa's news.

-- 
Regards
     Stephen

Post a reply to this message

Attachments:
Download 'v4lamp01a_inc00b_.jpg' (42 KB) Download 'v4lamp01a_inc00brobs_.jpg' (40 KB)

Preview of image 'v4lamp01a_inc00b_.jpg'

Preview of image 'v4lamp01a_inc00brobs_.jpg'

clipka <ano### [at] anonymousorg> wrote:
> Am 20.02.2011 18:35, schrieb clipka:
>
> > Please don't get too excited yet though: The changes will not make it
> > into 3.7.0; I expect to rush out a patch though as soon as a 3.7.0
> > version is released with a non-beta license.
>
> It's coming, it's coming! Dev team colleagues did agree to include it
> even in the next 3.7.0 release candidate!
>
> (jumps up & down in excitement)
>
> I can't wait to see what you guys will make of it...

Awesome! I'm really looking forward to it :)

-------------------------------------------------
www.McGregorFineArt.com

Post a reply to this message

Stephen <mcavoys_at@aoldotcom> wrote:
> On 21/02/2011 6:48 PM, Stephen wrote:

> I don't know what order these images will be posted but yours is the
> lighter one. I think that for this model the agate pattern is
> superfluous. But more discussion might be wasted with Clipa's news.

Just curious, how many POV-units is the model in your scene? After I rotated
VickiLamp and ran her through PoseRay it shows:

size dx,dy,dz=(0.192059,0.631494,0.42944)

So I just used scale 1.59 to make her 1 POV-unit tall. Assuming she's about a
foot tall as a lamp I changed my mm_per_unit to SSLT_UNIT_FT and got this result
after 13 min. (using the same samples of <16, 8>*8). The SSLT is subtle but
really makes a big difference when comparing left to right.

-------------------------------------------------
www.McGregorFineArt.com

Post a reply to this message

Attachments:
Download 'sslt_vickilamp.png' (737 KB)

Preview of image 'sslt_vickilamp.png'

On 22/02/2011 4:22 AM, Robert McGregor wrote:
> Just curious, how many POV-units is the model in your scene? After I rotated
> VickiLamp and ran her through PoseRay it shows:
>
> size dx,dy,dz=(0.192059,0.631494,0.42944)
>

Good question, I make it 16.5, after scaling. But I said to my self if 
it were standing straight it would be about 20 and the height 10 inches. 
This gives mm_per_unit = 12.7 but for some reason I've been using 
mm_per_unit  74.86


> So I just used scale 1.59 to make her 1 POV-unit tall. Assuming she's about a
> foot tall as a lamp I changed my mm_per_unit to SSLT_UNIT_FT and got this result
> after 13 min. (using the same samples of<16, 8>*8).

It took me about 40 mins to render her with your material 10 mins with mine.

Width = 800
Height = 600
Antialias = On
Sampling_Method = 1
Antialias_Threshold = 0.300
Jitter = Off
Antialias_Depth = 3

> The SSLT is subtle but
> really makes a big difference when comparing left to right.

It does indeed. :-D
I'm playing with the Stanford Bunny ATM.


-- 
Regards
     Stephen

Post a reply to this message

Am 21.02.2011 21:43, schrieb Stephen:

>> I can't wait to see what you guys will make of it...
>
> With your artistic talent you shouldn't need to wait. ;-)

Not really: I keep getting distracted by things to fix & improve in the 
code :-)

Post a reply to this message

On 22/02/2011 7:37 PM, clipka wrote:
> Am 21.02.2011 21:43, schrieb Stephen:
>
>>> I can't wait to see what you guys will make of it...
>>
>> With your artistic talent you shouldn't need to wait. ;-)
>
> Not really: I keep getting distracted by things to fix & improve in the
> code :-)

Better than being distracted by hunger and poverty. ;-)

-- 
Regards
     Stephen

Post a reply to this message

Stephen wrote:

> Better than being distracted by hunger and poverty. ;-)

poverty [pov-er-tee] - noun: The tendency to be distracted by pov

Post a reply to this message

On 23/02/2011 1:09 AM, Christian Froeschlin wrote:
> Stephen wrote:
>
>> Better than being distracted by hunger and poverty. ;-)
>
> poverty [pov-er-tee] - noun: The tendency to be distracted by pov
>
>
LOL should be an entry in the Uxbridge English Dictionary.
http://www.alspcs.com/main.html


-- 
Regards
     Stephen

Post a reply to this message

Stephen <mcavoys_at@aoldotcom> wrote:
> On 23/02/2011 1:09 AM, Christian Froeschlin wrote:
> > Stephen wrote:
> >
> >> Better than being distracted by hunger and poverty. ;-)
> >
> > poverty [pov-er-tee] - noun: The tendency to be distracted by pov
> >
> >
> LOL should be an entry in the Uxbridge English Dictionary.
> http://www.alspcs.com/main.html
>
>
> --
> Regards
>      Stephen

My 2 euro cent:

POVer tea: break in the middle of the night while POVing.

POVing: Art of using POV. Can be dangerous for mental health and transform a
        human being into a POVer. Moderate use strongly recommended.

POVer: Victim of excessive POVing.

POV-Ray: aka POV. Best ray-tracer ever ...

Post a reply to this message

On 23/02/2011 8:53 AM, Bruno Cabasson wrote:
> My 2 euro cent:
>
> POVer tea: break in the middle of the night while POVing.
>
> POVing: Art of using POV. Can be dangerous for mental health and transform a
>          human being into a POVer. Moderate use strongly recommended.
>
> POVer: Victim of excessive POVing.
>
> POV-Ray: aka POV. Best ray-tracer ever ...

Very good, you must join in the next Mornington Crescent in OT.

Render – Change the name of a folder in DOS.


-- 
Regards
     Stephen

Post a reply to this message

Coming soon: SSLT support for CSG

(Caveat: Unions of overlapping objects will probably give unexpected 
results; merge should work though.)

Post a reply to this message

Attachments:
Download 'translucency_test.png' (276 KB)

Preview of image 'translucency_test.png'

clipka <ano### [at] anonymousorg> wrote:
> Coming soon: SSLT support for CSG
>
> (Caveat: Unions of overlapping objects will probably give unexpected
> results; merge should work though.)

Great!

Post a reply to this message

On 2/20/2011 3:22 PM, clipka wrote:
> Am 20.02.2011 21:16, schrieb Robert McGregor:
>>
>> * Will radiosity work okay with it? I get lots of blocky artifacts
>> with SSLT and
>> radiosity in the current implementation.
>
> I'm not sure - that depends on what you're seeing. I found and
> eliminated a bug in the code that kind of "shattered" a surface
> according to which component of the surface normal was smallest. Other
> than that, I'm not aware of any problems with combining radiosity & SSLT
> (except that SSLT-enabled objects currently don't actually use the
> radiosity code, but an inferior substitute).
>

Definitely seeing that... attached is a pathological case.

This scene uses radiosity and photons, both of which really need to be 
turned off to get a more pleasing rendering. But the red+white checked 
pattern of the tablecloth really angers the SSLT code in this case ;)

Post a reply to this message

Attachments:
Download 'pathological radiosity and sslt.jpg' (18 KB)

Preview of image 'pathological radiosity and sslt.jpg'

High!

On 23.02.2011 09:53, Bruno Cabasson wrote:

> POVer tea: break in the middle of the night while POVing.
>
> POVing: Art of using POV. Can be dangerous for mental health and transform a
>          human being into a POVer. Moderate use strongly recommended.
>
> POVer: Victim of excessive POVing.
>
> POV-Ray: aka POV. Best ray-tracer ever ...
>
POVghanistan: phantasmagorical raytracing paradise far beyond the sands 
of ignorance, where the clicking of drag'n'drop is never heard and only 
console-hacking CLI nerds ever find their way to. Subdivision of 
Khyberspace.

POVghan: native of POVghanistan. Tends to be long-haired, shaggily 
bearded and wearing strange headgear.

See you in Khyberspace - www.khyberspace.de

Yadgar

Post a reply to this message