Hi there
I'm still trying to get the correct values for the Radiance reflection model,
they're close but not quite right. So for isotropic reflectance I get K as below

[ exp(-tan^2(delta/alpha^2)]
K = -----------------------------------
cos(theta_i)*cos(theta_r) *4pi*alpha^2

and use the formula to get the returned colour as
Li = Ii*projected_solid_angle(rho_diffuse/pi + rho_specular*K) where

cos(theta_i)*PI*r^2
projected_solid_angle = ------------------- dist_to_light^2

and r is the radius of my spherical light source.
when I do rview -vf view.vp -av .5 .5 .5 file.oct do I just add this -av value
for ambient light as below?
Li = light.colour*projected_solid_angle*(rho_diffuse/pi +rho_specular*K)
+ ambient;

Thanks for any help

Tarik

···

--
Tarik Rahman
PhD student, Institue of Perception, Action and Behaviour
School of Informatics
University of Edinburgh

Hi Tarik,

The formula for K should be:

exp(-tan^2(delta)/alpha^2)
K = -----------------------------------
cos(theta_i)*cos(theta_r) *4pi*alpha^2

How are you computing delta? It is the angle between the half-vector and the normal.

Also, for large disk sources, the exact calculation for solid angle is:

2pi*(1 - cos(theta_s/2))

Where theta_s is the subtended angle of the light source. This is a insignificant correction to what you have for distant sources.

-Greg

···

From: Tarik Rahman <tarik.rahman@ed.ac.uk>
Date: January 11, 2005 7:05:40 AM PST

Hi there
I'm still trying to get the correct values for the Radiance reflection model,
they're close but not quite right. So for isotropic reflectance I get K as below

[ exp(-tan^2(delta/alpha^2)]
K = -----------------------------------
cos(theta_i)*cos(theta_r) *4pi*alpha^2

and use the formula to get the returned colour as
Li = Ii*projected_solid_angle(rho_diffuse/pi + rho_specular*K) where

cos(theta_i)*PI*r^2
projected_solid_angle = ------------------- dist_to_light^2

and r is the radius of my spherical light source.
when I do rview -vf view.vp -av .5 .5 .5 file.oct do I just add this -av value
for ambient light as below?
Li = light.colour*projected_solid_angle*(rho_diffuse/pi +rho_specular*K)
+ ambient;

Thanks for any help

Tarik
--
Tarik Rahman
PhD student, Institue of Perception, Action and Behaviour
School of Informatics
University of Edinburgh

Thanks for that, made the correction for tan^2(delta) and delta is as you say
delta = normal.dotProduct(halfwayVectorUnit);
where
halfwayVector = point_to_lightUnit - point_to_camUnit;
halfwayVectorUnit = halfwayVector.normalize();

So for the ambient light of .5 .5 .5 I just add it at the end? I don't need to
multiply by the diffuse component rho_diffuse?

Tarik

Quoting Greg Ward <gregoryjward@gmail.com>:

···

Hi Tarik,

The formula for K should be:

exp(-tan^2(delta)/alpha^2)
K = -----------------------------------
cos(theta_i)*cos(theta_r) *4pi*alpha^2

How are you computing delta? It is the angle between the half-vector
and the normal.

Also, for large disk sources, the exact calculation for solid angle is:

2pi*(1 - cos(theta_s/2))

Where theta_s is the subtended angle of the light source. This is a
insignificant correction to what you have for distant sources.

-Greg

> From: Tarik Rahman <tarik.rahman@ed.ac.uk>
> Date: January 11, 2005 7:05:40 AM PST
>
> Hi there
> I'm still trying to get the correct values for the Radiance reflection
> model,
> they're close but not quite right. So for isotropic reflectance I get
> K as below
>
>
> [ exp(-tan^2(delta/alpha^2)]
> K = -----------------------------------
> cos(theta_i)*cos(theta_r) *4pi*alpha^2
>
>
> and use the formula to get the returned colour as
> Li = Ii*projected_solid_angle(rho_diffuse/pi + rho_specular*K) where
>
> cos(theta_i)*PI*r^2
> projected_solid_angle = -------------------
> dist_to_light^2
>
> and r is the radius of my spherical light source.
> when I do rview -vf view.vp -av .5 .5 .5 file.oct do I just add this
> -av value
> for ambient light as below?
> Li = light.colour*projected_solid_angle*(rho_diffuse/pi
> +rho_specular*K)
> + ambient;
>
> Thanks for any help
>
> Tarik
> --
> Tarik Rahman
> PhD student, Institue of Perception, Action and Behaviour
> School of Informatics
> University of Edinburgh

_______________________________________________

--
Tarik Rahman
PhD student, Institue of Perception, Action and Behaviour
School of Informatics
University of Edinburgh

Hi Tarik,

So for the ambient light of .5 .5 .5 I just add it at the end? I don't need to
multiply by the diffuse component rho_diffuse?

That's correct -- Radiance handles multiplication of the ambient with the various surfaces. (It would have to for any scene containing more than one surface type.)

-Greg

···

From: Tarik Rahman <tarik.rahman@ed.ac.uk>
Date: January 12, 2005 8:34:14 AM PST

Thanks for that, made the correction for tan^2(delta) and delta is as you say
delta = normal.dotProduct(halfwayVectorUnit);
where
halfwayVector = point_to_lightUnit - point_to_camUnit;
halfwayVectorUnit = halfwayVector.normalize();

So for the ambient light of .5 .5 .5 I just add it at the end? I don't need to
multiply by the diffuse component rho_diffuse?

Tarik

Oh, I know that Radiance handles the ambient term but say I wanted to write my
own code implementing the Radiance model, which terms would I need to multiply
the ambient value by?

Tarik

Quoting Greg Ward <gregoryjward@gmail.com>:

···

Hi Tarik,

> So for the ambient light of .5 .5 .5 I just add it at the end? I don't
> need to
> multiply by the diffuse component rho_diffuse?

That's correct -- Radiance handles multiplication of the ambient with
the various surfaces. (It would have to for any scene containing more
than one surface type.)

-Greg

> From: Tarik Rahman <tarik.rahman@ed.ac.uk>
> Date: January 12, 2005 8:34:14 AM PST
>
> Thanks for that, made the correction for tan^2(delta) and delta is as
> you say
> delta = normal.dotProduct(halfwayVectorUnit);
> where
> halfwayVector = point_to_lightUnit - point_to_camUnit;
> halfwayVectorUnit = halfwayVector.normalize();
>
>
> So for the ambient light of .5 .5 .5 I just add it at the end? I don't
> need to
> multiply by the diffuse component rho_diffuse?
>
> Tarik

_______________________________________________

--
Tarik Rahman
PhD student, Institue of Perception, Action and Behaviour
School of Informatics
University of Edinburgh

Hi Tarik,

This is explained pretty well in section 4 of the materials.pdf document. I assume you are reading this:

You multiply the ambient by the rho_a parameter of the reflectance model, which is the same as rho_d in most cases.

-Greg

···

From: Tarik Rahman <tarik.rahman@ed.ac.uk>
Date: January 14, 2005 4:11:13 AM PST

Oh, I know that Radiance handles the ambient term but say I wanted to write my
own code implementing the Radiance model, which terms would I need to multiply
the ambient value by?

Tarik