A very minor point; "Variable reflection", in Section 6.7.3.4 of the 
documentation contains the sentence:

"Remember that in real life many opaque objects have a thin layer of
transparent glaze on their surface, and it is the glaze (which
-does- have an IOR) that is reflective."

I read that as suggesting that opaque materials do not have an index of 
refraction. Maybe the segment in brackets could be removed, since I
think the emphasis on the glaze having an IOR gives the impression that
the object beneath the glaze does not. Or is that just me? :-)

Thanks,

Tom York

Post a reply to this message

"Tom York" <tom### [at] compsocmanacuk> wrote in message
news:3cf51e5a$1@news.povray.org...
> A very minor point; "Variable reflection", in Section 6.7.3.4
>
> "Remember that in real life many opaque objects have a thin layer of
> transparent glaze on their surface, and it is the glaze (which
> -does- have an IOR) that is reflective."
>
> I read that as suggesting that opaque materials do not have an index of
> refraction. Maybe the segment in brackets could be removed, since I
> think the emphasis on the glaze having an IOR gives the impression that
> the object beneath the glaze does not. Or is that just me? :-)

Opaque, in this context, is probably meant to be about no transparency at
all for the sake of POV-Ray. Therefore, the statement should be right. The
point of attention is about refraction of only a surface, however I don't
know if this itself is misleading since 'fresnel' is used for bodies of
water, etc., too.
But anyway, the paragraph there is actually trying to put emphasis on the
fact you may also want fresnel reflection from the ior of a solid objects
surface. I think it was discussed before that once you have any IOR you also
have some sort of transparency anyhow. The only thing I can think of about
this matter would be that the infinitely thin transparent surface is
presumed to be a part of a finite non-transparent object, the two being
apart yet the IOR applied to all. Either to be dealt with on a per object
basis or you must create a CSG pair of objects to act as the "glaze".

While on the subject... a sentence in that last paragraph has what a person
might think could be a keyword but is not:
So with a fresnel >>reflection_type<< an interior { ior IOR } statement is
required, even with opaque pigments.

bob h

Post a reply to this message

In article <3cf5a5f2$1@news.povray.org>, bob h wrote:

>Opaque, in this context, is probably meant to be about no transparency at
>all for the sake of POV-Ray. 

Yeah, that's what I thought it meant. But:

>Therefore, the statement should be right. 

Surely a material doesn't lose its refractive index just because it's
opaque. If I have a very thin slice of lead, and a thick block of lead,
both have the same refractive index, even though one sample absorbs more
light than the other.

>The point of attention is about refraction of only a surface, however I
>don't know if this itself is misleading since 'fresnel' is used for
>bodies of water, etc., too.

I don't think the IOR could be a property of the surface only,
because it depends on the speed of light in the bulk material.
Obviously, refraction and reflection are surface phenomena only because
that's where the IOR changes.

>But anyway, the paragraph there is actually trying to put emphasis on the
>fact you may also want fresnel reflection from the ior of a solid objects
>surface. 

Yep, fair enough.

>I think it was discussed before that once you have any IOR you also
>have some sort of transparency anyhow. 

I don't think that can be true; metals are not (usually) transparent but
the Fresnel reflectance model can be applied to them. Can you point me
at that discussion, if it was in these newsgroups? My search must have 
been more fumble-fingered than I thought :-)

>The only thing I can think of about
>this matter would be that the infinitely thin transparent surface is
>presumed to be a part of a finite non-transparent object, the two being
>apart yet the IOR applied to all. Either to be dealt with on a per object
>basis or you must create a CSG pair of objects to act as the "glaze".

Is that really necessary when dealing with the Fresnel model? I can
understand the "glaze" doing the reflecting for ceramic tiles, etc, but
metals, simple plastics and the like don't have to have a glaze to
reflect. I could make an opaque material with an IOR and apply the
Fresnel model to it in POV, couldn't I? It'd render faster and be more
realistic (unless you were really dealing with ceramic tiles).

>While on the subject... a sentence in that last paragraph has what a person
>might think could be a keyword but is not:
>So with a fresnel >>reflection_type<< an interior { ior IOR } statement is
>required, even with opaque pigments.
>
>bob h
>

I agree, a stray underscore.

Thanks,

Tom

Post a reply to this message

in news:3cf5f533$1@news.povray.org Tom York wrote:

> I could make an opaque material with an IOR and apply the
> Fresnel model to it in POV, couldn't I?

Yes

> It'd render faster and be more
> realistic (unless you were really dealing with ceramic tiles).

No

In 'daily life' it is assumed that transparent materials have an ior and 
opaques have not. From a physical point of view this is not true, all 
materials have an ior. The thing is that the refractive index is a 
complex number. For transparent materials the real part of it has much 
bigger influence than than the imaginary part, so the latter is usualy 
neglected in calculations. For example for metals the opposite is true. 
The real part of the refractive index has "no" influence.

POV-Ray only uses the real part of the ior in its model, so adding an 
ior to a metal would result in an unrealistic reflection.


Ingo

Post a reply to this message

In article <Xns### [at] povrayorg>, ingo wrote:

>... The thing is that the refractive index is a 
>complex number. For transparent materials the real part of it has much 
>bigger influence than than the imaginary part, so the latter is usualy 
>neglected in calculations. For example for metals the opposite is true. 
>The real part of the refractive index has "no" influence.

>POV-Ray only uses the real part of the ior in its model, so adding an 
>ior to a metal would result in an unrealistic reflection.

For dielectrics (opaque or transparent) though, the position is the
opposite from that of metals, isn't it? The refractive index will be
dominated by the real component (no free charge carriers) and POV's
implementation of the Fresnel model is perfectly adequate from a
realism, er, POV. No glaze is necessary, except as far as the modelled
object might have one.

Thanks,

Tom

Post a reply to this message

in news:3cf63854@news.povray.org Tom York wrote:

> For dielectrics (opaque or transparent) though, the position is the
> opposite from that of metals, isn't it?

yes,

Ingo

Post a reply to this message