POV-Ray : Newsgroups : povray.off-topic : Physically based rendering Server Time
25 Dec 2024 12:36:52 EST (-0500)
  Physically based rendering (Message 54 to 63 of 63)  
<<< Previous 10 Messages Goto Initial 10 Messages
From: clipka
Subject: Re: Physically based rendering
Date: 2 Jul 2015 12:16:38
Message: <559563e6$1@news.povray.org>
Am 02.07.2015 um 10:25 schrieb scott:
>> - Always use "fresnel on" for reflections (don't forget to specify
>> "ior"), or "metallic on" where applicable, and always use 0.0 for the
>> minimum reflection (or leave it out, specifying the maximum reflection
>> only).
>
> If you are using fresnel, is it valid (from a PBR point of view) to even
> scale the results using a maximum reflection other than 1?

Absolutely - if, for instance, your surface is littered with 
non-specular material at a microscopic level. Fine rust on iron steel, 
for instance, or a thin coating of dust.

In those cases it might be better, however, to use an average of two 
materials.


Post a reply to this message

From: Nekar Xenos
Subject: Re: Physically based rendering
Date: 2 Jul 2015 23:09:11
Message: <op.x06gc9hjufxv4h@xena>
On Thu, 02 Jul 2015 18:16:23 +0200, clipka <ano### [at] anonymousorg> wrote:

> Am 02.07.2015 um 10:25 schrieb scott:
>>> - Always use "fresnel on" for reflections (don't forget to specify
>>> "ior"), or "metallic on" where applicable, and always use 0.0 for the
>>> minimum reflection (or leave it out, specifying the maximum reflection
>>> only).
>>
>> If you are using fresnel, is it valid (from a PBR point of view) to even
>> scale the results using a maximum reflection other than 1?
>
> Absolutely - if, for instance, your surface is littered with  
> non-specular material at a microscopic level. Fine rust on iron steel,  
> for instance, or a thin coating of dust.
>
> In those cases it might be better, however, to use an average of two  
> materials.
>

What about plastic? It reflects, but it's not highly reflective. I don't  
think I would go higher than 0.5 maximum reflection on plastic.
-- 
-Nekar Xenos-


Post a reply to this message

From: scott
Subject: Re: Physically based rendering
Date: 3 Jul 2015 04:34:39
Message: <5596491f$1@news.povray.org>
>>> If you are using fresnel, is it valid (from a PBR point of view) to even
>>> scale the results using a maximum reflection other than 1?
>>
>> Absolutely - if, for instance, your surface is littered with
>> non-specular material at a microscopic level. Fine rust on iron steel,
>> for instance, or a thin coating of dust.

But are you sure simply scaling down the Fresnel reflection term in 
those cases gives a physically correct result for those materials?

>> In those cases it might be better, however, to use an average of two
>> materials.

Yes that's my concern, that if you are using the reflection scaling to 
"fake" a different effect (roughness, absorbtion or scattering etc) then 
it *might* not be physically correct to just scale it.

In other words, if you created a geometrically microscopic rough surface 
(or added physical dust, sub-surface scattering or whatever) and 
rendered it with max reflection 1, would it be possible to get the same 
result by rendering a perfectly smooth, clean surface with max 
reflection set to some lower value?

> What about plastic? It reflects, but it's not highly reflective. I don't
> think I would go higher than 0.5 maximum reflection on plastic.

Agreed completely, but does using the Fresnel equations to calculate 
reflection (which gives an absolute value) and scaling the result to 
simulate some other effect (surface roughness, sub-surface scattering 
etc) give a physically correct result? Or is it just a "fudge factor" to 
give something that looks correct?


Post a reply to this message

From: Thomas de Groot
Subject: Re: Physically based rendering
Date: 3 Jul 2015 07:00:22
Message: <55966b46$1@news.povray.org>
On 3-7-2015 10:34, scott wrote:
>>>> If you are using fresnel, is it valid (from a PBR point of view) to
>>>> even
>>>> scale the results using a maximum reflection other than 1?
>>>
>>> Absolutely - if, for instance, your surface is littered with
>>> non-specular material at a microscopic level. Fine rust on iron steel,
>>> for instance, or a thin coating of dust.
>
> But are you sure simply scaling down the Fresnel reflection term in
> those cases gives a physically correct result for those materials?
>
>>> In those cases it might be better, however, to use an average of two
>>> materials.
>
> Yes that's my concern, that if you are using the reflection scaling to
> "fake" a different effect (roughness, absorbtion or scattering etc) then
> it *might* not be physically correct to just scale it.
>
> In other words, if you created a geometrically microscopic rough surface
> (or added physical dust, sub-surface scattering or whatever) and
> rendered it with max reflection 1, would it be possible to get the same
> result by rendering a perfectly smooth, clean surface with max
> reflection set to some lower value?
>
>> What about plastic? It reflects, but it's not highly reflective. I don't
>> think I would go higher than 0.5 maximum reflection on plastic.
>
> Agreed completely, but does using the Fresnel equations to calculate
> reflection (which gives an absolute value) and scaling the result to
> simulate some other effect (surface roughness, sub-surface scattering
> etc) give a physically correct result? Or is it just a "fudge factor" to
> give something that looks correct?
>

I don't know, but in all these cases different from the /simple/ 
reflective surfaces, we need some good working examples demonstrating 
the properties. I confess that this is too difficult for me - and 
probably most people - to do and expect them from our own gurus :-) I 
have also difficulties visualising an averaged texture or a layered 
texture staying within the properties boundaries.

-- 
Thomas


Post a reply to this message

From: scott
Subject: Re: Physically based rendering
Date: 3 Jul 2015 09:13:46
Message: <55968a8a$1@news.povray.org>
> I don't know, but in all these cases different from the /simple/
> reflective surfaces, we need some good working examples demonstrating
> the properties. I confess that this is too difficult for me - and
> probably most people - to do and expect them from our own gurus :-) I
> have also difficulties visualising an averaged texture or a layered
> texture staying within the properties boundaries.

What really would be ideal would be some macros to use for various 
groups of materials (glasses, metals, plastics, paints etc) with 
parameters designed such that no matter what values you give you are 
guaranteed a physically correct material.

I'm not a regular user of POV anymore so it's all too easy for me to get 
one number "wrong" and then I don't get a physically correct material. 
If there were some macros-for-dummies then this whole problem would be 
avoided and would surely lead to much higher quality work from all 
non-experts.


Post a reply to this message

From: clipka
Subject: Re: Physically based rendering
Date: 3 Jul 2015 09:36:08
Message: <55968fc8@news.povray.org>
Am 03.07.2015 um 05:08 schrieb Nekar Xenos:

>>> If you are using fresnel, is it valid (from a PBR point of view) to even
>>> scale the results using a maximum reflection other than 1?
>>
>> Absolutely - if, for instance, your surface is littered with
>> non-specular material at a microscopic level. Fine rust on iron steel,
>> for instance, or a thin coating of dust.
>>
>> In those cases it might be better, however, to use an average of two
>> materials.
>>
>
> What about plastic? It reflects, but it's not highly reflective. I don't
> think I would go higher than 0.5 maximum reflection on plastic.

For plastics - and actually any material that's neither metallic nor 
coated - I think heavily blurred reflections and a well-chosen ior are 
the key, not reducing the reflection maximum.

With a high ior, the maximum reflection only occurs when viewing the 
material almost edge-on. And with blurred reflections, even there the 
maximum isn't truly achieved, as the effective reflection brightness is 
tuned down by averaging in reflections at a less shallow effective angle.

BTW, the ior of most plastics is in the range from 1.5 to 1.6.


Post a reply to this message

From: clipka
Subject: Re: Physically based rendering
Date: 3 Jul 2015 10:01:08
Message: <559695a4$1@news.povray.org>
Am 03.07.2015 um 10:34 schrieb scott:
>>>> If you are using fresnel, is it valid (from a PBR point of view) to
>>>> even
>>>> scale the results using a maximum reflection other than 1?
>>>
>>> Absolutely - if, for instance, your surface is littered with
>>> non-specular material at a microscopic level. Fine rust on iron steel,
>>> for instance, or a thin coating of dust.
>
> But are you sure simply scaling down the Fresnel reflection term in
> those cases gives a physically correct result for those materials?

If it's soot, then yes (provided you also tune down the diffuse term); 
any other stuff mixed in, and things get more complicated, hence my 
recommendation:

>>> In those cases it might be better, however, to use an average of two
>>> materials.


>> What about plastic? It reflects, but it's not highly reflective. I don't
>> think I would go higher than 0.5 maximum reflection on plastic.
>
> Agreed completely, but does using the Fresnel equations to calculate
> reflection (which gives an absolute value) and scaling the result to
> simulate some other effect (surface roughness, sub-surface scattering
> etc) give a physically correct result? Or is it just a "fudge factor" to
> give something that looks correct?

For surface roughness, "roughness" is the POV-Ray parameter to tweak; 
this also tunes down the brightness of individual pixels in highlights 
(provided you use the "albedo" keyword) and reflections (provided you 
also use the "roughness" keyword in the reflection block), by spreading 
the highlights over a larger area (so much that you might not even 
notice the specular reflection anymore), but just tuning down the 
maximum brightness obviously doesn't get you the same effect, as it 
leaves reflections crisp and sharp.

Sub-surface scattering has nothing to do with specular reflections - to 
the contrary: It is the result of all the light that is /not/ reflected 
in a specular manner (which is to say, at the object's surface). As a 
matter of fact, sub-surface scattering (in a very simple case) is what 
the traditional "diffuse" mechanism models.


Post a reply to this message

From: Nekar Xenos
Subject: Re: Physically based rendering
Date: 4 Jul 2015 00:24:13
Message: <op.x08eiaqmufxv4h@xena>
On Fri, 03 Jul 2015 15:35:51 +0200, clipka <ano### [at] anonymousorg> wrote:

> Am 03.07.2015 um 05:08 schrieb Nekar Xenos:
>
>>>> If you are using fresnel, is it valid (from a PBR point of view) to  
>>>> even
>>>> scale the results using a maximum reflection other than 1?
>>>
>>> Absolutely - if, for instance, your surface is littered with
>>> non-specular material at a microscopic level. Fine rust on iron steel,
>>> for instance, or a thin coating of dust.
>>>
>>> In those cases it might be better, however, to use an average of two
>>> materials.
>>>
>>
>> What about plastic? It reflects, but it's not highly reflective. I don't
>> think I would go higher than 0.5 maximum reflection on plastic.
>
> For plastics - and actually any material that's neither metallic nor  
> coated - I think heavily blurred reflections and a well-chosen ior are  
> the key, not reducing the reflection maximum.
>
> With a high ior, the maximum reflection only occurs when viewing the  
> material almost edge-on. And with blurred reflections, even there the  
> maximum isn't truly achieved, as the effective reflection brightness is  
> tuned down by averaging in reflections at a less shallow effective angle.
>
> BTW, the ior of most plastics is in the range from 1.5 to 1.6.
>

So a shiny plastic has a high ior I assume. Then it makes sense.

-- 
-Nekar Xenos-


Post a reply to this message

From: scott
Subject: Re: Physically based rendering
Date: 6 Jul 2015 03:11:28
Message: <559a2a20$1@news.povray.org>
> So a shiny plastic has a high ior I assume. Then it makes sense.

What makes a plastic shiny or not is usually the geometric finish on the 
surface, not anything to do with the material itself. It all depends on 
the surface finish of the mould tool that was used to make the part. It 
can be anything from high polished (to give a high gloss finish) to a 
very fine texture (to give varying degrees of matt finish) to a large 
patterned texture for something like a leather effect on car trim.


Post a reply to this message

From: Kenneth
Subject: Re: Physically based rendering
Date: 11 Jul 2015 06:00:01
Message: <web.55a0e86c18c52d49a46a090d0@news.povray.org>
Thomas de Groot <tho### [at] degrootorg> wrote:

>
> Now, that is a comprehensive tutorial about do's and dont's. Thanks
> indeed Christoph.
>

I second that! So much wonderful/useful information--and in only one post!!


Post a reply to this message

<<< Previous 10 Messages Goto Initial 10 Messages

Copyright 2003-2023 Persistence of Vision Raytracer Pty. Ltd.