[Radiance-general] Re: rsensor file format

[Greg Ward]

Hi Axel,

I'll try not to duplicate what Jennifer said too much -- who, by the  
way, could not have read my response first (nor I hers) because it was  
composed on the plane and not sent until later...

> > The sensor file is normalized to 1.0 in arbitrary "response" units.
> > Think of it as a multiplier on the incoming radiance value in each  
> direction.
>
> So I shouldn't think of it as an 'inverse' IES file, then. I've just  
> run some simple examples with an ideal cosine data file, and the  
> results do match the one I get from rtrace. Multiplying each value  
> in the matrix with 2.0 gives me twice the illuminance.

Yes, it's not really comparable to luminaire data.

> I guess that the real question should be 'what do these sensors  
> actually measure?'. It's not lux, because for this they would have  
> to be cosine and v(lamda) corrected, which one could expect from a  
> 500 Pound lux meter, but not from a 5 Pound sensor. I guess it all  
> boils down to the commissioning of the sensors. The only thing we  
> can say with certainty is that they measure 'some sort of radiation  
> thing', which is different in different directions.

Yes -- the spectral response is unknown, and may or may not be  
important to the sensor output.  The rsensor program reports the  
sensor-weighted sum of RGB radiance values over the hemisphere, and if  
you have an idea of how to multiply the RGB results based on the  
spectral sensitivity, you can get a more accurate output.

> So is it correct to assume that SPOT only uses the relative response  
> from the photocells, and not absolute W/m2 or lux values?

I think Jennifer answered this pretty well, and she knows more about  
it than I do.

> Here is my own cunningly crafted cosine response. Just one column of  
> azimuth values at zero degrees wouldn't work:
>
> degrees	0	90	180	270
> 0	1.0000	1.0000	1.0000	1.0000
> 10	0.9848	0.9848	0.9848	0.9848
> 20	0.9397	0.9397	0.9397	0.9397
> 30	0.8660	0.8660	0.8660	0.8660
> 40	0.7660	0.7660	0.7660	0.7660
> 50	0.6428	0.6428	0.6428	0.6428
> 60	0.5000	0.5000	0.5000	0.5000
> 70	0.3420	0.3420	0.3420	0.3420
> 80	0.1736	0.1736	0.1736	0.1736
> 90	0.0000	0.0000	0.0000	0.0000
>
> Does rsensor use a linear interpolation to fill in in-between rows?

Yes, effectively.  The rsensor program actually inverts the sensor  
distribution to derive Monte Carlo ray samples, so it's a bit tricky,  
but amounts to the same.

> I am still puzzled that the response is not proportional to the  
> volume of the shape, or at least the area of the curve. Guess that's  
> just the way it is, and there isn't really any strict photometric or  
> radiometric reason behind this...

Radiation as we use it normally is not a volumetric quantity.  It can  
be area-based, but since we report things as a proportion of area  
(e.g., watts/meter^2), this factors out again.  I'm not really sure  
where you're confused, since there are so many things that are  
confusing about lighting units.  I spent months in the beginning  
trying to get my head around them.

Cheers,
-Greg