Re: [Phys-l] sun's true color or lack thereof

[John Denker <jsd@av8n.com>]

On 08/24/2008 06:21 PM, Anthony Lapinski wrote:
> Does anyone know the actual color of the Sun? The Sun certainly appears
> yellowish, and then Wein's Law can be used to calculate its surface
> temperature (about 6000 °C). Textbooks confirm this result.
>
> However, I read an article in Astronomy magazine that the Sun is actually
> pure white. Our daytime sky is blue since our atmosphere scatters blue
> light the most. This leaves sunlight with less blue and hence a yellowish
> tint. The article said the color of snow closely matches the Sun's actual
> color because snow reflects the sky's blueness as well as sunlight.

1) The color of the sun is a perceptual issue, outside the realm of hard 
 physics.  There is no "true color", just various notions of perceived color.
 
2) Note that the chromaticity of the _white point_ of any decent imaging 
 system is an important design parameter, often a configurable parameter.

3) White points that mimic _daylight_ are commonly chosen, and probably 
 should be your first choice for many applications.  But beware that others
 may choose differently.  

4) The D60 and D65 white points are particularly common, and correspond to 
 black body temperatures of 6000 and 6500 K respectively.

5) Note that item (3) carefully talked about "daylight" ... which does not 
 necessarily answer the original question.

6) If you ask about the apparent color of the sun's disk, in the evening
 it is red ... not white or yellow.

7) The sun's disk at mid-day is yellowish-white, for the same physical
 reason that it is red in the evening.  Much of the blue component has
 been scattered out.  This means the disk is not a black body source.  
 But keep in mind that all this physics has only secondary importance.

8) Much more importantly, the sun's disk is not the only source of light 
 in the scene.  It is not "the" _illuminant_.

9) If you add up the light from the disk plus the light from the blue
 sky, that's more-or-less what folks consider _daylight_.

10) The whiteness of white clouds or snow does not answer the question,
 not at all.  White things look white under a wide range of illuminants.
 The relevant perceptual issues here go by the names of "color constancy"
 and "discounting the illuminant".