[Radiance-general] color space and efficacy in color.h

[Gregory J. Ward]

Hi Christian,

Sorry for the slow reply.  I was hoping someone else would field  
this, then I forgot about it.

Radiance does not use the NTSC color space.  By default, it uses the  
primaries you found after the NTSC ones in color.h.  There is a  
history to the WHTEFFICACY value, which is explained (more or less)  
in the old Radiance Digests.  See:

	http://radsite.lbl.gov/radiance/digests_html/v2n0.html#LUMFACTOR

The bottom line is that the peak efficacy of 683 lumens/watt is only  
the right value at 555 nm, and a poor approximation to the efficacy  
over the visible spectrum.  In the end it doesn't matter what factor  
you use, as you divide by it going from photometric units to radiance  
and multiply again coming out.  In other words, the factor goes away.

If you want to change the notion of what RGB colors Radiance uses,  
there is no formal way to do this.  You just do it.  You specify your  
sources and materials in your desired RGB space, and Radiance doesn't  
really care what color space it's working in.  To work properly with  
various programs that properly interpret Radiance RGBE images, you  
should probably add a line of the form:

PRIMARIES= .603 .352 .289 .590 .146 .066 .319 .348

where these numbers correspond to xr yr xg yg xb yb xw yw,  
respectively.  To add a line to a Radiance image, you can use the  
vinfo script.

I hope this is helpful.
-Greg

> From: Christian Fusenig <christian_fusenig at gmx.de>
> Date: February 11, 2006 11:37:58 AM PST
>
> Dear all,
>
> I want to visualize 3D designs of TV-Sets and display them on  
> calibrated broadcast class A monitors with EBU phosphors.
> Is it possible to tell Radiance not to work with the WHTEFFICACY of  
> 179 lm/w but  the MAXEFFICACY factor of 683 lm/w?
> Furthermore i want to use RGB color space with the 'nominal CRT  
> Primaries' which are the same as the EBU phosphor coordinates  in  
> 'color.h'.
> I saw the data in the header file but i dont know how to change it.  
> I guess this change will affect other programs like falsecolor to.
>
> Whats the advantage working with 179lm/w anyway?  In every book i  
> see the factor of 683 lm/w!
>
> -----
> #ifdef  NTSC
> #define  CIE_x_r        0.670        /* standard NTSC primaries */
> ....
> #else
> #define  CIE_x_r        0.640        /* nominal CRT primaries  =  
> EBU color space*/
> #define  CIE_y_r        0.330
> #define  CIE_x_g        0.290
> #define  CIE_y_g        0.600
> #define  CIE_x_b        0.150
> #define  CIE_y_b        0.060
> #define  CIE_x_w        0.3333        /* use true white */
> #define  CIE_y_w        0.3333
> #endif
> -----
> Does Radiance use the NTSC color space by default ?
>
> I want to create an octree containing the data of a tungsten light  
> source i measured.
> My goal is to measure the right illuminance under the lamp. I  
> achieved it with this command:
>
> echo "1 0 0 0 0 1" |rtrace -I -h lux_test.oct \
>            | rcalc -e '$1=179*(0.26510582*$1+0.67010582*$2 
> +0.0647883598*$3)'
>
> But i want to create an octree where i can use this command:
> echo "1 0 0 0 0 1" |rtrace -I -h lux_test.oct \
>            | rcalc -e '$1=683*(0.299*$1+0.587*$2+0.114*$3)'
> These factors are the EBU equivalents for brightness in the Y channel.
>
> I hope i made myself understandable, its critical that i can  
> measure the correct colors and the correct brightness in Radiance  
> to make sure the display will show the exact colors of the design!
> Thanks for help so far,
>
> Christian