... maybe interesting for people who like to toy around with subsurface
light transport, so I thought I'd share:

https://youtu.be/ClmtcwfWI6M

Post a reply to this message

clipka <ano### [at] anonymousorg> wrote:
> ... maybe interesting for people who like to toy around with subsurface
> light transport, so I thought I'd share:
>
> https://youtu.be/ClmtcwfWI6M

Definitely one of those refractive plays on light, can be seen to occur even
without the foil added. So probably a slight air gap does same.

Recreate using SSLT? Simple enough, right? Maybe not. I tried animation frames
to look for good values of translucency and mm_per_unit.

Only able to determine a very low mm_per_unit is needed, along with high
translucency "color". At least without trying extreme numbers or changing object
sizes (only a pair of boxes 1 unit thick).

My test object blocks use pigment {rgb 1}, and a gap between them to sandwich
the pseudo foil sheet (no reflection, dark gray, yet specular highlighted).

You will probably want to pause this very short animation near the beginning to
see where it looks best, most closely matching the real video wax
brighter/darker. Light source is above, room being darker near bottom.

/* testing for correct wax SSLT properties */
//cmd:-kfi1 -kff50 -sf8 -ef8

#version 3.7;

#declare MM=25.4*clock;//4;
#declare ST=2-2*clock;//1.5;

global_settings{assumed_gamma 1.0
 subsurface{samples 200,50} mm_per_unit MM}

light_source{-z*99,1 rotate <60,-15,0>}

camera{location -z*9 look_at -y/2 angle 50
 rotate <30,30,0>}

sphere{0,1 pigment{function{abs(y)} // room
 color_map{[0 rgb 0][1 rgb 1]} scale 2 translate -y}
 scale 111}

box{-1,1 hollow scale <2,0.5,1> translate y*0.511
 // upper wax block
 pigment{rgb 1} finish{diffuse 0.5 specular 0.5
 subsurface{translucency ST}}
 interior{ior 1.4}}

box{-1,1 hollow scale <1.95,0.01,0.95>
 // foil sheet between wax blocks
 pigment{rgb 1} finish{diffuse 0.1 specular 0.9 roughness 0.01}}

box{-1,1 hollow scale <2,0.5,1> translate -y*0.511
 // lower wax block
 pigment{rgb 1} finish{diffuse 0.5 specular 0.5
 subsurface{translucency ST}}
 interior{ior 1.445}}

text{ttf "times", concat(str(MM,-2,1)," mm per unit"),0.1,0
 pigment {rgb <0.4,0.8,1>}
 scale <0.75,0.75,0.75>
 rotate <30,30,0>
 translate <-2.25,-2.5,-1>
}

text{ttf "times", concat(str(ST,-2,1)," translucency"),0.1,0
 no_shadow
 pigment {rgb <1,0.4,0.2>}
 scale <0.5,0.5,0.5>
 rotate <30,30,0>
 translate <-3,2.75,-1>
}

Post a reply to this message

Attachments:
Download 'wax_mm_ss.mp4.mpg' (128 KB)

Am 15.12.2016 um 00:51 schrieb omniverse:

> Recreate using SSLT? Simple enough, right? Maybe not. I tried animation frames
> to look for good values of translucency and mm_per_unit.

Ah, you folks make me mad! How come everyone and their uncle want to toy
around the "mm_per_unit" setting?

It is supposed to be the number of millimiters in a POV-Ray unit. The
general scale of your scene. Couldn't be easier as that, could it? It's
not some fancy knob to twiddle if you happen to be unhappy with how your
subsurface-enabled materials look.

(scuffles off, mumbling)


No, seriously: Any visual effect you can achieve by tweaking the
"mm_per_unit" setting can just as well be achieved by tweaking the
subsurface settings instead.

Also, the "mm_per_unit" setting is /not/ an SSLT parameter. Subsurface
light transport just /happens/ to be the only feature currently making
use of it -- the idea being that this allows you to re-use the very same
material definition in a scene that uses an entirely different scale.

Other existing features that could also benefit from this mechanism
include fog, media and fading interiors, but as they were invented
before "mm_per_unit" that would mean changing their syntax.

Post a reply to this message

clipka <ano### [at] anonymousorg> wrote:
> Am 15.12.2016 um 00:51 schrieb omniverse:
>
> > Recreate using SSLT? Simple enough, right? Maybe not. I tried animation frames
> > to look for good values of translucency and mm_per_unit.
>
> Ah, you folks make me mad! How come everyone and their uncle want to toy
> around the "mm_per_unit" setting?

Um... because it's there.  LOL

> No, seriously: Any visual effect you can achieve by tweaking the
> "mm_per_unit" setting can just as well be achieved by tweaking the
> subsurface settings instead.

I will have to take your word for it. I just haven't been able yet.

When I use a very large, or what I consider large, translucency number(s) the
effect again matches the real experiment but also makes the blocks too
transparent. In this case up to 20 or <20,20,20>.
I haven't tried different pigment numbers yet, and will look at the
documentation on SSLT in the meantime to see what I might have missed.

Bob

Post a reply to this message

Am 15.12.2016 um 02:50 schrieb omniverse:
> clipka <ano### [at] anonymousorg> wrote:
>> Am 15.12.2016 um 00:51 schrieb omniverse:
>>
>>> Recreate using SSLT? Simple enough, right? Maybe not. I tried animation frames
>>> to look for good values of translucency and mm_per_unit.
>>
>> Ah, you folks make me mad! How come everyone and their uncle want to toy
>> around the "mm_per_unit" setting?
> 
> Um... because it's there.  LOL
> 
>> No, seriously: Any visual effect you can achieve by tweaking the
>> "mm_per_unit" setting can just as well be achieved by tweaking the
>> subsurface settings instead.
> 
> I will have to take your word for it. I just haven't been able yet.
> 
> When I use a very large, or what I consider large, translucency number(s) the
> effect again matches the real experiment but also makes the blocks too
> transparent. In this case up to 20 or <20,20,20>.
> I haven't tried different pigment numbers yet, and will look at the
> documentation on SSLT in the meantime to see what I might have missed.

It will be difficult to match the experiment exactly, as the aluminum
foil is also reflective. The SSLT algorithm can't account for that.

Post a reply to this message

clipka <ano### [at] anonymousorg> wrote:
>
> It will be difficult to match the experiment exactly, as the aluminum
> foil is also reflective. The SSLT algorithm can't account for that.

Okay, had made it reflective and didn't notice obvious changes so that would
explain why.

However... what about photons? Wouldn't that interact, bouncing off the foil
going through the wax?

Attempt to mimic SSLT using media scattering only proved it needed two differing
parameters. Posted at images after the with/without ior pair.

Bob

Post a reply to this message

Am 15.12.2016 um 07:32 schrieb omniverse:
> clipka <ano### [at] anonymousorg> wrote:
>>
>> It will be difficult to match the experiment exactly, as the aluminum
>> foil is also reflective. The SSLT algorithm can't account for that.
> 
> Okay, had made it reflective and didn't notice obvious changes so that would
> explain why.
> 
> However... what about photons? Wouldn't that interact, bouncing off the foil
> going through the wax?

No; the photons wouldn't even reach the foil - they'd just be deposited
on the surface of the way.

To fully model the wax & aluminium foil experiment, we'd need
full-fledged monte-carlo sapling of scattering media.

Post a reply to this message