[Atomic-S (OP)]

White is typically defined by something such as mean noon sunlight. That works pretty well for many purposes, but is slightly less than objective, because it is tied to radiation at a specific temperature without giving a solid reason why that temperature and not some other should be chosen. This difficulty can be avoided by turning to Fourier theory, in which a spectrum such that the power per unit frequency interval is evenly distributed across all frequencies, has special mathematical significance. Light having this property (within the visual band) could be said to be absolute white. I am wondering what its color temperature is -- that is, the temperature of the blackbody whose output most closely matches it in color.


_{Mod edit:  I've formatted the subject as a heading.  Please do this in future to help keep the forum tiday and easy to navigate.  Thanks!}

[Pmb]

White is typically defined by something such as mean noon sunlight. That works pretty well for many purposes, but is slightly less than objective, because it is tied to radiation at a specific temperature without giving a solid reason why that temperature and not some other should be chosen.

Because its tied to the sun whose color temperature is about 5780K. Our eyes have developed so that when sunlight, which has a black body distribution, shines on a surface with particular absorption properties we see “white”. At least that’s what I recall.

This difficulty can be avoided by turning to Fourier theory, in which a spectrum such that the power per unit frequency interval is evenly distributed across all frequencies, has special mathematical significance. Light having this property (within the visual band) could be said to be absolute white. I am wondering what its color temperature is -- that is, the temperature of the blackbody whose output most closely matches it in color.

The color temperature of a light source is the temperature of an ideal blackbody that radiates light of comparable hue to that of the light source. I can’t see it beeing defined according to your definition.

[evan_au]

Human vision is very adaptive - we can still recognise a white object at noon and at sunset, or under incandescent and fluorescent lights, even though the ambient light is very different in each case.

The retina has mechanisms to adapt to different local illumination of each colour, making room for various colour illusions.

So given this wide variability in human perception, I am not sure that any new definition of white would make a difference to what we see.

Digital Camera designers try to mimic this automatic adaptation in the human visual system: http://en.wikipedia.org/wiki/White_balance

[Pmb]

So given this wide variability in human perception, I am not sure that any new definition of white would make a difference to what we see.

And that's why they don't change it. There's no need for a change.

In anycase I've never seen white defined as in the OP. This has always been my understanding of the term - http://en.wikipedia.org/wiki/White

[JP]

So given this wide variability in human perception, I am not sure that any new definition of white would make a difference to what we see.

And that's why they don't change it. There's no need for a change.

Those are good points as to why we don't define white in this way, but you can still ask the question, even if it's not going to produce interesting results for human vision.  I don't know the answer, but you could compute it by looking at the (normalized) variance of Planck's law over the spectral range of interest.  This would tell you how much the intensity varies across that range, scaled by the average intensity across that range. 

You could also ask what a pulse looks like that has a flat spectrum over some range of frequencies.  In this case, the field of the pulse would look like a sinc function (since it would be the Fourier transform of a constant value over some finite window).  No black body would produce this since it continuously emits over all frequencies and the spectrum isn't flat.