[Radiance-general] rtrace -i -I flags

Gregory J. Ward gregoryjward at gmail.com
Fri Mar 2 19:01:57 CET 2007


I guess the man page is pretty confusing.  The '-i' option sends a  
ray into the scene, just as rpict would for generating an image or  
rtrace would for generating a radiance value, but then computes the  
incident irradiance at that surface point rather than the radiance.   
In other words, it makes most surfaces (with the exception of glass  
materials) look perfectly diffuse, with a normalization factor that  
computes irradiance.

The '-I' option, which is only supported by rtrace, changes the  
notion of the input to be interpreted as an illuminance measurement  
point as Guilio said, rather than a ray origin and direction.  This  
makes it convenient for placing virtual irradiance meters anywhere in  
your scene.

-Greg

> From: "Bleicher, Thomas" <T-Bleicher at bdp.co.uk>
> Date: March 2, 2007 7:04:22 AM PST
>
> Jelle.
>
> As Giullio has already explained
>
> -i    will return the value for a 'light meter' at (x,y,z) pointing  
> in direction (dx,dy,dz)
>       Only that one point will be calculated.
>
> -I    will return the value for virtual light meter at (x,y,z)  
> pointing in direction (dx,dy,dz)
>       but it will sample the whole hemisphere around the direction  
> vector.
>       Basically rtrace introduces a virtual plane and turns the  
> 'virtual light meter' around
>       to get the value of a point on this plane. You have to  
> increase -ab by 1 to get the
>       same number of effective bounces as with '-i'.
>
> Try a simple scene with both options and low -ab values (0,1,2).
>
> If you print origin and direction of the rtrace ray as well you  
> will see that '-I' changes
> them. Quite confusing if you're not prepared for it.
>
> Thomas
>



More information about the Radiance-general mailing list