Vincent Le Chevalier <gal### [at] libertyallsurfspamfr> wrote:
> >   Unless I have understood something completely incorrectly, I think that
> > "unbiased rendering" simply means that light is not assumed to be coming
> > from a specific direction, but the entire space is sampled for possible
> > incoming light.
> > 

> But that means that the images should be different, no? And I mean, more 
> than just noisy...

  It means that the scene has global illumination. It doesn't make surfaces
more opaque.

> >   (Besides, I think the only difference pure unbiased rendering would do
> > is to make the image have global illumination, ie. what povray calls
> > "radiosity", especially if big area lights are used instead of point
> > lights.)
> > 

> Well that's a big difference

  Not that much of a difference. It doesn't make surfaces more opaque,
for instance.

>, and I don't know if it's really possible 
> or even desirable to make a single path if you're not looking for an 
> unbiased result.

  I don't believe unbiasing makes surfaces more opaque, so I don't see
how that is any relelvant.

> I was thinking more of your reflection 1 transmit 1 example. Energy 
> conservation is still one of the basic properties of BRDFs. There is no 
> way you can represent that with a BRDF. Of course all BRDFs are only 
> approximations, but that does not mean that they do not have 
> constraints. I guess my use of "physically accurate" is the problem here...

  A BRDF can take light coming from several directions and calculate how
it affects the outgoing light towards a certain direction. I don't believe
BRDFs have the requirement that these different light amounts must be
*averaged* instead of being added together.

> A black object can have a bright highlight, of course. Or rather, a 
> bright reflection. In that case, you would lower diffuse_amount, to be 
> able to set a higher reflection_amount.

  How is the program supposed to guess that you really wanted a lower
diffuse amount?

  Besides, the black object was just an example. Take a red object, with
a relatively high diffuse, and a white light source. The surface can have
white highlights even though the surface is red.

  If you average the diffuse and the specular reflection, the latter
would get a red tint.

-- 
                                                          - Warp

Post a reply to this message

>   I don't believe unbiasing makes surfaces more opaque
> 

It could very well be the case. The unbiased renderer cannot allow any 
non-conservative material. If it did, it would be unable to reach a 
stable solution, because it attempts to simulate the physical system. If 
you had just one non-conservative material in real-life, the scene would 
just end up saturated with light...

So the biased renderer can handle surfaces that are perfectly 
transparent yet perfectly reflective, because it samples a limited class 
of rays that make it work. The unbiased renderer must sacrifice either 
transparency or reflectivity in order for its solution not to explode.


>   A BRDF can take light coming from several directions and calculate how
> it affects the outgoing light towards a certain direction. I don't believe
> BRDFs have the requirement that these different light amounts must be
> *averaged* instead of being added together.
> 

If you do not make an average, as more rays are traced the brightness of 
the image rises, which is not realistic. So regardless of lighting 
theory, you must do an average. It is exactly what happens in anti 
aliasing, I don't see the problem. That average, I think, stems from 
Monte-Carlo integration.

>> A black object can have a bright highlight, of course. Or rather, a 
>> bright reflection. In that case, you would lower diffuse_amount, to be 
>> able to set a higher reflection_amount.
> 
>   How is the program supposed to guess that you really wanted a lower
> diffuse amount?
> 

It does not guess, you tell it so. Because as I said, colors should be 
normalized. If you want black you have to lower the diffuse amount.

>   Besides, the black object was just an example. Take a red object, with
> a relatively high diffuse, and a white light source. The surface can have
> white highlights even though the surface is red.
> 
>   If you average the diffuse and the specular reflection, the latter
> would get a red tint.
> 

If you add them up, there will be a red tint as well. It's just that you 
clamp the components of the color after the fact, ending up with 
something white. If no clamping happens, the red tint should still be there.

-- 
Vincent

Post a reply to this message

> It does not guess, you tell it so. Because as I said, colors should be 
> normalized. If you want black you have to lower the diffuse amount.

Are we talking about a biased raytracer like POV-Ray, or a different 
algorithm? POV-Ray definitely doesn't require normalized colors (3.7 
beta and Mega-POV both support HDRI output).

Post a reply to this message

>> It does not guess, you tell it so. Because as I said, colors should be 
>> normalized. If you want black you have to lower the diffuse amount.
> 
> Are we talking about a biased raytracer like POV-Ray, or a different 
> algorithm? POV-Ray definitely doesn't require normalized colors (3.7 
> beta and Mega-POV both support HDRI output).

No I was just trying to point out a problem with mixing 
reflection_amount & co with colors in the tentative pseudocode. POV-Ray 
certainly has no need to normalize colors.

Thinking about it the easiest way would be to define all the amounts as 
colors with all components between 0 and 1, and run the algorithm for 
each color, rather than trying to treat all colors simultaneously...

-- 
Vincent

Post a reply to this message

"Warp" wrote:
>
>  If you average the diffuse and the specular reflection, the latter
> would get a red tint.

Sure, there will be a red tint no matter if you add them or average them but 
in both cases the red tint is so small that it's insignificant.

Remember that the specular reflection of a light source is *many* times 
brighter than the diffuse reflection of that same light source in that spot 
where the highlight appears. So the contribution of the red diffuse 
reflection is insignificant compared to the specular reflection - exactly 
like if you had added them together.

Remember to think in terms of high dynamic range and not brightness values 
clamped to a [0-1] range...

Rune

Post a reply to this message

Vincent Le Chevalier wrote:
> OK, let's say more physically correct then ;-) I don't know, something 
> that conserves energy, for a start :-)

Fair enough. Hey, does it get a prism right, defracting different colors 
by different amounts? That seems pretty fundamental and possible even 
without doing QED calculations.

-- 
   Darren New / San Diego, CA, USA (PST)
     Remember the good old days, when we
     used to complain about cryptography
     being export-restricted?

Post a reply to this message

Darren New wrote:
>  Hey, does it get a prism right, defracting different colors 
> by different amounts? That seems pretty fundamental and possible even 
> without doing QED calculations.
> 

I think you can manage that in much the same way POV does at the moment, 
yes... It's just more costly in terms of rays, of course.

-- 
Vincent

Post a reply to this message

Vincent Le Chevalier wrote:
> I think you can manage that in much the same way POV does at the moment, 
> yes... It's just more costly in terms of rays, of course.

Oh! I didn't even know POV handled it at all. :-)

-- 
   Darren New / San Diego, CA, USA (PST)
     Remember the good old days, when we
     used to complain about cryptography
     being export-restricted?

Post a reply to this message

Darren New wrote:
> Vincent Le Chevalier wrote:
>> I think you can manage that in much the same way POV does at the 
>> moment, yes... It's just more costly in terms of rays, of course.
> 
> Oh! I didn't even know POV handled it at all. :-)
> 
Actually you made me doubt whether it was in Megapov or POV-Ray, but 
there it is, in mainstream POV:

http://www.povray.org/documentation/view/3.6.1/415/


-- 
Vincent

Post a reply to this message

Vincent Le Chevalier wrote:
> Actually you made me doubt whether it was in Megapov or POV-Ray, but 
> there it is, in mainstream POV:

Kewl. Faked, but kewl. ;-)  When changing the IOR in an unbiased render 
automatically makes a spectrum, let me know. ;-)

-- 
   Darren New / San Diego, CA, USA (PST)
     Remember the good old days, when we
     used to complain about cryptography
     being export-restricted?

Post a reply to this message