"Jim Charter" <jrc### [at] msncom> wrote in message 
news:42569e69@news.povray.org...
> I think the diffuse component is irrelevent for rgb 0.  I think it is 
> basically the product of the two values, rgb and diffuse, and the angle of 
> the lightray, that are used to calculate shadow.  So if you go rgb 0 you 
> loose all capacity to get non-lighted shadows and instead have to rely on 
> specular and/or reflection to mold the surface.  So, as you say, adding a 
> little color gives "diffuse" something to work with.  The trouble is that 
> just about the time that adding color starts to make a difference, it 
> starts to look like grey, not black, material.

Very true.  Though in some cases, I've had rgb 0 objects show up like black 
holes in the middle of a picture, and giving it a very small value still 
makes it appear black, without making it look like there's black hole in the 
picture.  But if it looks ok in relation to other objects, no reason to 
change it.

>
> Yes I remember Jaime's suggestion too, though I thought he was talking 
> about specular.  I also remember that a long time ago there were often 
> threads discussing such formulas. Alas, am I now just catching up? Well 
> not quite. For a long long time I typically adjusted 
> diffuse/specular/reflection as some sort of loosely conceived division up 
> of a unit whole.  Then it came as something of a revelation to me when I 
> freed myself from that discipline.  Now, while Jaime's statement comes as 
> a sobering remainder to keep it sane, I am still not convinced that a 
> formulaic relationship between these elements always makes sense.

I'm not convinced either, but Jaime's lighting is always better than mine, 
so I don't argue.  Plus it makes my life simpler (biggest reason). ;-)

Incidentally, I looked up a demo scene of his, and he goes by the 2nd 
component of reflection (which is usually the highest), which would make it 
diffuse 0.85 in your case.  But with rgb 0, I don't think it matters, as 
you've already stated.

>
> I was introducing my daughter to raytracing just the other night.  We were 
> giving a sphere, with no finish specified, different colors just to 
> demonstrate how the color vector worked.  I was trying to get her to guess 
> how to get yellow by introducing the concept that yellow is really 
> not-blue.  So first I had to demonstrate that white is <1,1,1>. We did the 
> render.  "It doesn't look white, it looks grey", was her immediate 
> reaction. :(

Some day, when you grow up, you'll realize that there is no such thing as 
rgb 0 and rgb 1, just various shades of rgb 1*A, where A is a value between 
0 and 1, non-inclusive*.  ;-)

>
> With a low rgb, radiosity also is neutralized.  So the only way to get any 
> tones into the shadows is reflection. But even the slightest amount of 
> reflection gives a hard-shiny look.
>
> About the bottom of the shoe. ("Bottom" is the correct term.) Yes, if it 
> was made of leather it would show wear and have tan patches.  If not, then 
> it would be some homogenous material other than leather and therefore show 
> different surface characteristics.  It is an important question because it 
> may be true that the solution to the problem has to do with providing 
> massive amounts of minute surface detail.  After all, shiny, homogenous, 
> black patent leather is favored by some precisely for its "dark liquid", 
> form-denying properties.

I was just thinking that the bottom might be made of a different material. 
...or it might not...

>
> So the hunt goes on.  But I was interested to peoples reactions to gross 
> adjustments of these factors across the scene in general.
>

If you have the time/patience/computing power, making rotating animations 
can help point out flaws, especially for things such as specular.  From 
certain angles, specular reflections may appear ok, but from others, you 
realize they are too high / too rough / too etc.

Just mentioning some tricks that I've learned.

*I'm thinking of making that my sig.  :-)

-- 
Jeremy
www.beantoad.com

Post a reply to this message

Jeremy M. Praay wrote:

> I'm not convinced either, but Jaime's lighting is always better than mine, 
> so I don't argue.  

Exactly.

> Incidentally, I looked up a demo scene of his, and he goes by the 2nd 
> component of reflection (which is usually the highest), which would make it 
> diffuse 0.85 in your case.  But with rgb 0, I don't think it matters, as 
> you've already stated.
> 

I will have to study this matter further.


> 
> If you have the time/patience/computing power, making rotating animations 
> can help point out flaws, 

Sometimes brutally so.  I used that technique for my first ever IRTC 
entry.  I was inexperienced at placing objects then.

especially for things such as specular.

Gawd, not specular too.

Post a reply to this message

Jim Charter wrote:
> Jeremy M. Praay wrote:
>> Jaime recently mentioned "diffuse + reflection = 1". [...]
>
> For a long long time I typically adjusted 
> diffuse/specular/reflection as some sort of loosely conceived division 
> up of a unit whole.  

   Now that you mention it, I was just thinking about specular, and also 
phong, having the same "relation" with diffuse. Re-reading the part of 
the docs about diffuse, seems it accounts for anything that is not 
direct light reflection. As specular and phong both simulate reflection 
of the light sources, seems that also diffuse+specular or diffuse+phong 
should be 1. But we use traditionally specular or phong coupled with 
reflection, so I really don't know what to do with all this...

   Oh... I almost forgot about the original post: I prefer 22 or 23, 
both look very convincing.

--
Jaime

Post a reply to this message

21

Post a reply to this message

23 is the best (if I had to buy one pair)
24,25,30 and 31 are very "disco"
27,28 found at Chelsea flea market !


Rene
http://rene.bui.free.fr - online portfolio

Post a reply to this message

Rene Bui wrote:
> 23 is the best (if I had to buy one pair)

Sold!

> 24,25,30 and 31 are very "disco"

They pulse.

> 27,28 found at Chelsea flea market !
> 

Two for one!  They do look a little rusted over.

> 
> Rene
> http://rene.bui.free.fr - online portfolio
> 
>

Post a reply to this message

Jaime Vives Piqueres wrote:
> Jim Charter wrote:
> 
>> Jeremy M. Praay wrote:
>>

> 
>   Oh... I almost forgot about the original post: I prefer 22 or 23

Interesting.

20 shows a sort of starting point with all settings low.
22 shows an increase in roughness with specular still the same as in 20.
23 shows the increase in specular with roughness remaining the same as 
in 20.

Post a reply to this message

Jim Charter wrote:

> 
> Opinions sought.
> 
> http://www21.brinkster.com/jrcsurvey/Temp/index.html
> 
> 
> The link is to  some test renders permutating different
> combinations for background, specular, and roughness.
> 
> Constants:
> pigment { rgb 0 }
> diffuse 1
> reflection { .075 .15 fresnel metallic 0 }
> 
> Which looks best?

21, 22 or 24. 24 was my first choice. 

would like to see the final selection sometime.

Post a reply to this message

"Jim Charter" <jrc### [at] msncom> wrote in message 
news:425610d3$1@news.povray.org...


> Which looks best?

   Blimey! I remember those things at school! All of a sudden, *that girl* 
was either your height now, or was taller than you by three inches. Those 
were the days.

  23 for me. Although, I like 24 too.

I see a rainy night scene under an amber street light unfolding...  :oO

   ~Steve~

Post a reply to this message

Ross wrote:
> Jim Charter wrote:
> 
> 
>>Opinions sought.
>>
>>http://www21.brinkster.com/jrcsurvey/Temp/index.html
>>
>>
>>The link is to  some test renders permutating different
>>combinations for background, specular, and roughness.
>>
>>Constants:
>>pigment { rgb 0 }
>>diffuse 1
>>reflection { .075 .15 fresnel metallic 0 }
>>
>>Which looks best?
> 
> 
> 21, 22 or 24. 24 was my first choice. 
> 
> would like to see the final selection sometime. 
Based on the responses I am getting here: a strong favouring of 23, as 
well as my own preferences: I like some of the liveliness of 24 too, I 
will probably experiment with some interpolation such that, if

20.....23
.
.
21.....24

are vectors, then
X marks the likely settings:

20.....23
.....X
.
21.....24

Post a reply to this message