Am 10.10.2011 19:23, schrieb Cousin Ricky:
> I used CIE.inc from LightsysIV to display black body colors at various
> temperatures.  As the temperature increased, I expected the color to become
> yellower before it turned white.  Instead, I could not discern any change in
> hue.
>
> I figured that the change in hue might be too subtle for me to detect, so I
> #debugged the hues.  Too my surprise, after becoming slightly yellower, the hue
> doubled back on itself and got redder, then onto magenta, violet, and blue.
>
> Is this right?

Yes.

> If so, why do candle flames (glowing soot particles) look
> yellower than glowing coals?

Because the flame of a candle is far away from being a perfect blackbody 
radiator.


> Could it be a whitepoint issue?

As the whitepoint of sRGB (the default color order system from CIE.inc) 

this temperature. And only *close* because the sRGB whitepoint is 
defined as a D-illuminant (and is implemented as the Daylight macro 
within CIE.inc) and this is not quite identical to blackbody.


> Floating point
> accuracy?  That warning that splines are an experimental feature?
>
> The attached image shows my results, with a fully-saturated swatch to the lower
> right of each sample.

Looks good to me.
-Ive

Post a reply to this message