Hi.
I'm trying to model a pearl, well, lots of them eventually.
I found a nice reference photo on the web.

http://www.loppleman.com/pearls_white.jpg

but so far I'm having difficulty. What I've got so far is,
to paraphrase: almost, but not quite, completely unlike a pearl.

http://billhails.net/Pearl.png

That's done with lots of nested semitransparent spheres
and the render time is already feeling painful (around five
minutes for that pic,) considering I'll be doing lots of them.

I don't want to steal a ready-made texture, I'd rather gain the
experience and roll my own, but hints would be welcome.

I can post the source for discussion if anyone thinks I should
pursue my current tack, but I'm thinking of trying something
completely different.

I'll continue trying anyway.

-- 
Bill Hails
http://billhails.net/

Post a reply to this message

Bill Hails <me### [at] billhailsnet> wrote:
> Hi.
> I'm trying to model a pearl, well, lots of them eventually.

Oh, good. They'll go well with my diamonds. :)

> I found a nice reference photo on the web.
>
> http://www.loppleman.com/pearls_white.jpg
>
> but so far I'm having difficulty. What I've got so far is,
> to paraphrase: almost, but not quite, completely unlike a pearl.
>
> http://billhails.net/Pearl.png

That's not too bad, it's just a bit too glassy or metallic, IMHO. But it's
almost impossible to model shiny things properly in a bare scene. Follow
your reference image's example and give your pearl some things to reflect.
Other pearls would be ideal, but for development work, just throw in a few
metal balls and rings, and maybe use a simple lightprobe (or just an
environment sphere with no HDRI).

> That's done with lots of nested semitransparent spheres
> and the render time is already feeling painful (around five
> minutes for that pic,) considering I'll be doing lots of them.

Nested spheres never do what you want them to. I tried doing clouds with
them once... :)

> I don't want to steal a ready-made texture, I'd rather gain the
> experience and roll my own, but hints would be welcome.

Ok, here's a few ideas off the top of my head.  Try using a scattering
media, with a reflective layer below it. I think that will eliminate the
need for most of your spheres. I don't use media much at 566MHz, so you'll
have to ask the media experts about the fine art of subsurface scattering.

Vary brilliance: try .5-10, but really high values can give interesting
effects, too. Try using both phong and specular. Phong for the general
'glow' of the pearl and specular for the actual highlight spot. Just keep
the total of phong+specular less than one. Even try using a non-zero
ambient. It should be ok here, even if you are using radiosity, since
pearls always seem to have a bit of a glow about them.

I think pearl has a fairly high IOR, but I can't seem to find a value for
it... I guess it's pretty hard to measure properly with all those layers.

Good luck, Bill!

Post a reply to this message

PM 2Ring wrote:

> [...]
> 
> Ok, here's a few ideas off the top of my head.  Try using a scattering
> media, with a reflective layer below it. I think that will eliminate the
> need for most of your spheres. I don't use media much at 566MHz, so you'll
> have to ask the media experts about the fine art of subsurface scattering.
> 
> Vary brilliance: try .5-10, but really high values can give interesting
> effects, too. Try using both phong and specular. Phong for the general
> 'glow' of the pearl and specular for the actual highlight spot. Just keep
> the total of phong+specular less than one. Even try using a non-zero
> ambient. It should be ok here, even if you are using radiosity, since
> pearls always seem to have a bit of a glow about them.
> 
> I think pearl has a fairly high IOR, but I can't seem to find a value for
> it... I guess it's pretty hard to measure properly with all those layers.
> 
> Good luck, Bill!

Thanks for the suggestions. I did try Media after posting, but didn't
think to put a reflective surface behind it. Using phong + specular < 1
is a new one on me, I usually just use specular and roughness.

The other problem I've hit is that when switching to a hdri light probe,
the texture looks *completely* different, probably because the resemblance
to a perl was only coincidental in the first place :-)

Onward and upward.

-- 
Bill Hails
http://billhails.net/

Post a reply to this message

Bill Hails nous apporta ses lumieres en ce 06/05/2006 06:50:
> Hi.
> I'm trying to model a pearl, well, lots of them eventually.
> I found a nice reference photo on the web.
> 
> http://www.loppleman.com/pearls_white.jpg
> 
> but so far I'm having difficulty. What I've got so far is,
> to paraphrase: almost, but not quite, completely unlike a pearl.
> 
> http://billhails.net/Pearl.png
> 
> That's done with lots of nested semitransparent spheres
> and the render time is already feeling painful (around five
> minutes for that pic,) considering I'll be doing lots of them.
> 
> I don't want to steal a ready-made texture, I'd rather gain the
> experience and roll my own, but hints would be welcome.
> 
> I can post the source for discussion if anyone thinks I should
> pursue my current tack, but I'm thinking of trying something
> completely different.
> 
> I'll continue trying anyway.
> 
The spheres are probably to transparent. Try with lower transmit or filter.
Try using blured reflection. Maybe for the inside sphere(s).
Try adding some scathering media, use a fairly large density, probably >100.
Try adding irid to the iner sphere(s). Keep it low (<0.1).
Use an ior in the interior block, something around 2 and dispersion around 1.01 to
1.1.

To help reduce the rendering time, you may increase adc_bailout from the default of
1/256.

-- 
Alain
-------------------------------------------------

Post a reply to this message

Alain wrote:


> The spheres are probably to transparent. Try with lower transmit or
> filter. Try using blured reflection. Maybe for the inside sphere(s).
> Try adding some scathering media, use a fairly large density, probably
> >100. Try adding irid to the iner sphere(s). Keep it low (<0.1).
> Use an ior in the interior block, something around 2 and dispersion around
> 1.01 to 1.1.
> 
> To help reduce the rendering time, you may increase adc_bailout from the
> default of 1/256.
> 

Thanks for the pointers Alain.
I've switched to a HDRI lighting setup, and the previous texture
looked awful in that light, so this is a second attempt using
internal scattering media and a semi-transparent surface.

http://billhails.net/Pearl-002.png

There's no internal surface now, just the media, and It seems to me
to be even less like a Pearl than my first post. I'm going to try
putting inner semitransparent reflective spheres back in, but keeping
the media.

-- 
Bill Hails
http://billhails.net/

Post a reply to this message

Bill Hails <me### [at] billhailsnet> wrote:

> There's no internal surface now, just the media, and It seems to me
> to be even less like a Pearl than my first post. I'm going to try
> putting inner semitransparent reflective spheres back in, but keeping
> the media.

Try adding irid to the sphere, at a low value, pearls are definitely
somewhat iridescent! Experiment with transparency of the outer sphere and
density of the media within to give it a softer look, it's a bit too hard
and shiny IMO.

Post a reply to this message

Bill Hails <me### [at] billhailsnet> wrote:
> Alain wrote:
>
>
> > The spheres are probably to transparent. Try with lower transmit or
> > filter. Try using blured reflection. Maybe for the inside sphere(s).
> > Try adding some scathering media, use a fairly large density, probably
> > >100. Try adding irid to the iner sphere(s). Keep it low (<0.1).
> > Use an ior in the interior block, something around 2 and dispersion around
> > 1.01 to 1.1.
> >
> > To help reduce the rendering time, you may increase adc_bailout from the
> > default of 1/256.
> >
>
> Thanks for the pointers Alain.
> I've switched to a HDRI lighting setup, and the previous texture
> looked awful in that light, so this is a second attempt using
> internal scattering media and a semi-transparent surface.
>
> http://billhails.net/Pearl-002.png
>
> There's no internal surface now, just the media, and It seems to me
> to be even less like a Pearl than my first post. I'm going to try
> putting inner semitransparent reflective spheres back in, but keeping
> the media.
>
> --
> Bill Hails
> http://billhails.net/

Well, this thread is a year old, but I found it and figured I'd offer my
advice anyway. Pearls are made up of multiple transluscent layers, right,
due to the way they're formed. So it was clever of you to form it in the
same manner.
One thing that occurs to me is that these surfaces would be transparent when
viewed straight-on, which, if you loook at photos of pearls, explains why
they often have a darker spot in the middle.
I think combining these tricks with varying reflections will work well for
you.

Post a reply to this message