"Thomas de Groot" <tDOTdegroot@interDOTnlANOTHERDOTnet> wrote:
> This looks terribly good indeed.

Thanks Thomas :)

I can't seem to stop toying with it now (I'm sure you can all understand that).

So, after many more tweaks I've got it looking like I wanted - more subtle. But
I've noticed that whatever texture I use it seems to sort of fade away,
"overpowered" by the SSLT. I've tried all kinds of different SSLT parameters
and diffuse settings(even the ones from the Jenson, et. al. 2001 SIGGRAPH
paper).

Am I missing something simple here? Anyone have some tips on retaining my
texture colors/details and still getting the cool SSLT scattering?

-Rob

Post a reply to this message

Attachments:
Download 'sslt.lucy.test2.jpg' (460 KB)

Preview of image 'sslt.lucy.test2.jpg'

"Robert McGregor" <rob### [at] mcgregorfineartcom> wrote:
> "Thomas de Groot" <tDOTdegroot@interDOTnlANOTHERDOTnet> wrote:
> > This looks terribly good indeed.
>
> Thanks Thomas :)
>
> I can't seem to stop toying with it now (I'm sure you can all understand that).
>
> So, after many more tweaks I've got it looking like I wanted - more subtle. But
> I've noticed that whatever texture I use it seems to sort of fade away,
> "overpowered" by the SSLT. I've tried all kinds of different SSLT parameters
> and diffuse settings(even the ones from the Jenson, et. al. 2001 SIGGRAPH
> paper).
>
> Am I missing something simple here? Anyone have some tips on retaining my
> texture colors/details and still getting the cool SSLT scattering?
>
> -Rob

Okay, here's an example of what I mean: this is the chest of the figure from my
"Rebirth" IRTC entry. The left is straight "plastic" old-school radiosity; the
right is SSLT using the Jensen skin parameters and averaged diffuse component
(since POV-Ray doesn't currently do rgb diffuse). What happened to his nipples,
etc?? It becomes a sort of embryonic skin...

Post a reply to this message

Attachments:
Download 'sslt_skin_compare.jpg' (219 KB)

Preview of image 'sslt_skin_compare.jpg'

> Okay, here's an example of what I mean: this is the chest of the figure from my
> "Rebirth" IRTC entry. The left is straight "plastic" old-school radiosity; the
> right is SSLT using the Jensen skin parameters and averaged diffuse component
> (since POV-Ray doesn't currently do rgb diffuse). What happened to his nipples,
> etc?? It becomes a sort of embryonic skin...

Wow, I'm starting to feel like AlphaQuad here, but I can't stop - I'm on a roll.

Okay, so I used a three-pass technique to get something pretty nice just now. I
set up brighter lights on each of the two previous render passes (normal
radiosity and SSLT radiosity) and ran them both through a third pass via my
ambient occlusion compositing macro (to multiply them) and got the third panel,
which is much more like real skin. This is what I would expect from the direct
SSLT if it had some sort of damping filter built in to average the effect. Just
a thought. At least I know I can get decent results this way from now on...

Cheers,
Rob

Post a reply to this message

Attachments:
Download 'sslt_skin_compare2.jpg' (324 KB)

Preview of image 'sslt_skin_compare2.jpg'

Robert McGregor schrieb:
> I can't seem to stop toying with it now (I'm sure you can all understand that).

Sure! And I'm pretty envious that you got Lucy in POV format... MeshLab 
crashes on me when I try to load her.


> So, after many more tweaks I've got it looking like I wanted - more subtle. But
> I've noticed that whatever texture I use it seems to sort of fade away,
> "overpowered" by the SSLT. I've tried all kinds of different SSLT parameters
> and diffuse settings(even the ones from the Jenson, et. al. 2001 SIGGRAPH
> paper).
> 
> Am I missing something simple here? Anyone have some tips on retaining my
> texture colors/details and still getting the cool SSLT scattering?

You're missing something indeed: The "safety notes" that came with the 
change description :-P

SSLT is still very, very alpha stage, and just hasn't had the time to be 
integrated with features such as textures, photons or what-have-yous. 
The time will come, I promise :-)

You can try layered textures though, with a filtering layer over the 
SSLT one. I guess that should work.

Post a reply to this message

clipka <ano### [at] anonymousorg> wrote:
> Robert McGregor schrieb:
> > I can't seem to stop toying with it now (I'm sure you can all understand that).
>
> Sure! And I'm pretty envious that you got Lucy in POV format... MeshLab
> crashes on me when I try to load her.

Well, she's pretty big - my PoseRay generated include file is 40.7 MB. I have a
zip file of her that's 11.4 MB if you want it.

> > Am I missing something simple here? Anyone have some tips on retaining my
> > texture colors/details and still getting the cool SSLT scattering?
>
> You're missing something indeed: The "safety notes" that came with the
> change description :-P
>
> SSLT is still very, very alpha stage, and just hasn't had the time to be
> integrated with features such as textures, photons or what-have-yous.
> The time will come, I promise :-)
>
> You can try layered textures though, with a filtering layer over the
> SSLT one. I guess that should work.

Aha, the "safety notes!" And I've tried some averaged textures but that was sooo
slow. Layering sounds like a better option; I suppose I would've thought of that
myself evertually. Thanks Christoph :)

-Rob

Post a reply to this message

Robert McGregor schrieb:
> Well, she's pretty big - my PoseRay generated include file is 40.7 MB. I have a
> zip file of her that's 11.4 MB if you want it.

I'd love to - after all, she's such a gorgeous beauty for showing off 
SSLT, ain't she?

You're sure that doesn't go against the licensing?

Post a reply to this message

clipka <ano### [at] anonymousorg> wrote:
> Robert McGregor schrieb:
> > Well, she's pretty big - my PoseRay generated include file is 40.7 MB. I have a
> > zip file of her that's 11.4 MB if you want it.
>
> I'd love to - after all, she's such a gorgeous beauty for showing off
> SSLT, ain't she?
>
> You're sure that doesn't go against the licensing?

Sure, no problem - here's what it says on the Stanford Model Repository web site
(http://graphics.stanford.edu/data/3Dscanrep/):

"You are welcome to use the data and models for research purposes. You are also
welcome to mirror or redistribute them for free. Finally, you may publish
images made using these models, or the images on this web site, in a scholarly
article or book - as long as credit is given to the Stanford Computer Graphics
Laboratory. However, such models or images are not to be used for commercial
purposes, nor should they appear in a product for sale (with the exception of
scholarly journals or books), without our permission."

I just uploaded her to my web site at:
http://www.mcgregorfineart.com/zip/Lucy_POV_geom.zip

And make sure you post some shots so I can see how she's *supposed* to look <g>

Cheers,
Rob

Post a reply to this message

clipka <ano### [at] anonymousorg> wrote:
> SSLT is still very, very alpha stage, and just hasn't had the time to be
> integrated with features such as textures, photons or what-have-yous.
> The time will come, I promise :-)
>
> You can try layered textures though, with a filtering layer over the
> SSLT one. I guess that should work.

Okay, I've played with layering, averaging, etc. for hours now with not great
results :(

While I love the overall effect it is quite difficult to tame (at least for me,
right now). Anyway, since it's still in the alpha stage here is a suggestion of
what would make sense for me SDL-wise: localize the effect per-material.

If the feature was localized per material with a simple scale of SSLT
blend-amount (0-1, like basic reflection) it would become very easy to control
and blend with the pigment. Likewise, a localized samples param (like media)
and scattering distance would give full control of the overall quality (or lack
thereof, useful for quick tests) of the effect, something like:

material {
   texture {
      pigment { MyPigment }
      finish {
         diffuse ff
         subsurface {
            <RgbScatteringCoefficient>,
            <RgbAbsorptionCoefficient>,
            ScatteringDistance
            PigmentBlendAmount // 0-1
            Samples nn
         }
      }
   }
   interior {ior ff}
}

Of course I haven't seen your code and don't know if this would even work well
with what you've written so far, but it sure would make the feature much more
predictable and controllable IMO.

Just food for thought, trying to think of general use scenarios and how I would
like to be able to use the feature in my own scenes.

Post a reply to this message

Robert McGregor schrieb:
> Just food for thought, trying to think of general use scenarios and how I would
> like to be able to use the feature in my own scenes.

That's why the feature has been released "into the wild" already, so 
your feedback is much appreciated.

Post a reply to this message

This object has it's own magic. It looks great! Well-done!

Post a reply to this message