1

With black body radiation, light of different wavelengths are emitted in various proportions depending on the temperature of the emitter.

Graphs of spectral radiance vs wavelength for different colour temperatures look like these:

enter image description here

It is said that the colour of the radiation (as a whole) changes in a predictable manner. Samples of colours at various temperatures:

  • 1000K Red
  • 1500K Reddish orange
  • 2000K Yellowish orange
  • 2800K Yellow
  • 3500K Yellowish white
  • etc

Others claim that the visible light appears yellow at 5000K.

Leaving aside the precise temperature, the consensus is that "[a]s its temperature increases further it becomes bright red, orange, yellow, white, and ultimately blue-white".

However, if we take a look at the graph, it seems that even when the 'yellow' temperature is reached, the black body should emit quite a bit of red and orange wavelengths in addition to yellow.

It might be that the more precise phrasing is that "it first turns red, then orange (red+yellow light), and finally white (red+yellow+blue looks white to the eye)".

But standard high school introductory material and the Wikipedia article seem to suggest that the perceived colour can be a (pure) yellow at some point (2800K in the linked article above or 5000K from the high school material).

Is this just a matter of perception (we see various wavelengths but our brain combines the RGB data from the cones in the eyes into 'yellow'), or is there some other explanation for why we can (in theory) see a pure yellow light emitted from black body radiation of a suitable temperature?

Related questions that do not answer my question:

That question addresses black body radiation generically but not the specific issue of how a 'pure' colour can be perceived when the visible emissions are of a collection of wavelengths (e.g. red + orange + yellow, together perceived as a pure yellow).

That question addresses colour perception through an interplay between emitted and absorbed/reflected light.

Qmechanic
  • 201,751
Lawrence
  • 113
  • 3
    Pure yellow or whatever colour doesn't fit with BB emission anyway. There is no way to use "pure" the way you did in the question. All the rest is physiology. Rather than yellow, what we can't perceive from a BB is green. Somehow we need higher purity to see green (likely because it is in the middle). – Alchimista Oct 17 '21 at 12:03
  • 1
    I'm afraid that I can't quote a source, but I once read that the eye cannot usually distinguish an impure colour (wide spread of wavelengths) from a pure colour (very narrow spread) centred on the 'right' wavelength in the wide spread. – Philip Wood Oct 17 '21 at 12:05

2 Answers2

2

Your "more precise phrasing" is indeed exactly right.

niels nielsen
  • 92,630
1

You can plot the track of a blackbody on a colour chart. Here it is (from the wikipedia page on black bodies). Yes the range of colours goes from reddish to blue-ish white, through a yellowish/orangeish region. There is no "pure colour" and blackbody radiation is emitted over a wide (infinite) range of wavelengths at any temperature.

I don't think there is a consensus on what names to give to the appearance of blackbody radiation, not least because the colour that is perceived by the eye will also depend on the intensity of the light (i.e. photons per unit area) received at the back of the retina - the difference between photopic and scotopic vision. However, if the blackbody surface is big/close enough such that it is resolved by the eye, then this number will be roughly constant and I guess is what the colour chart claims to represent in that case.

Planckian curve on colour chart

ProfRob
  • 130,455
  • For a given temperature, the intensity of a black-body is fixed. The color percieved by the eye can only depend on the temperature. – Alfred Jan 12 '22 at 21:55
  • @Alfred not really: chromatic adaptation will influence perception. – Ruslan Jan 12 '22 at 23:30
  • @Alfred the relevant phenomena are photopic, mesopic and scotopic vision. These do indeed depend on the flux received at the eye. Your statement is only true for looking ar a blackbody "wall" where indeed that flux would be constant no matter how far away the wall was. In the real world, blackbodies do not have infinite size, so when they are far away the flux received at the eye diminishes. As a result, the colour perceived changes. That is why faint stars appear white and brighter stars may appear coloured. https://physics.stackexchange.com/a/169986/43351 – ProfRob Jan 13 '22 at 10:00
  • @Alfred i.e. it is the specific intensity (Watts per square metre per Hertz per steradian) that is fixed by the Planck function, not the flux (Watts per square metre per Hertz) at the eye. – ProfRob Jan 13 '22 at 10:04
  • @Profrob OK, I was wrong but not for the reason you give. The color of any part of the sun is the same, and looking at just a part of the sun through a small opening about 1m away will burn a smaller area of your retina, but will burn it as thoroughly as if you were looking at the whole sun. What would indeed change the perception is to be able to contract your pupil even more to reduce the solid angle of the flux reaching each neuron of your retina. So not the flux received by unit of area at your eye, but the solid angle reaching your retina from the opening of your pupil. – Alfred Jan 13 '22 at 10:36
  • @Profrob But yes, for a black body much cooler than the sun, reddish, I suppose it may look a different red in each eye if you put atropin in one eye and not in the other. But changing its size, or its distance, will not change the percieved color. – Alfred Jan 13 '22 at 10:38
  • @Probrob And this is why watching a solar eclipse without protection is so dangerous. In normal situations, you can look a the sun for a few seconds. It leaves an after-image, but this is reversible, because the pupil contracts. In an eclipse, the solid angle of sun you see is much smaller, the pupil opens in the dark. Even looking for a very short time (as you do when you look at the whole sun, moving rapidly away) can cause a permanent damage. – Alfred Jan 13 '22 at 10:52
  • @Alfred If you put the Sun 1000 light years away, it would appear white to the naked eye. The numbers of photons received per second are insufficient to trigger the colour response cells. There are many stars out there you can look at tonight to gauge the truth of this. Their spectra are still, in terms of shape, that (very roughly) of a blackbody. – ProfRob Jan 13 '22 at 11:01
  • @Profrob I am comparing incident solid angles that are "measurable" by your retina, whole sun vs a "square of sun" of side, say 1/10th of a degree : same percieved color, if the opening of the pupil is determined by the "ambiant" daylight. Once one gets to extremely small incident solid angle, like stars, so that the flux that reach each cell of the retina becomes extremely small, (even though the pupil is fully dilated) then indeed, cones cannot react at all, only rods. No color perceived at all. – Alfred Jan 13 '22 at 11:04
  • @Alfred well, yes. Forgive me for being an astronomer - these are the only "blackbodies" I work with. The difference is that the image size on the retina is fixed even whilst the distance is increasing. You are considering situations where the image size scales with the solid angle of the source. There are two regimes. I will ponder how to change my wording. – ProfRob Jan 13 '22 at 11:10
  • @Profrob The argument of yours I wanted to contradict was the "infinite wall" of black body. A blackbody seen at a solid angle of 1/10 of a steradian or 1/10,000 of a steradian (very, very roughly, the sun) or 1/1000,000 of a steradian (still reasonable for a person not myopic, or with lenses well adapted to one's myopia) will look the same colour. Much less than that is a different problem. – Alfred Jan 13 '22 at 11:11
  • @Alfred - yes, I agree with these numbers. But it is still the case that what matters is the number of photons per unit area at the back of the retina (as opposed to at the eye perhaps). – ProfRob Jan 13 '22 at 11:17
  • @ProfRob To that I agree: the perception of colour would indeed depend on number of photons per unit area at the back of the retina, I was wrong. – Alfred Jan 13 '22 at 15:45
  • @ProfRob When scotopic vision dominates, or even becomes important, color perception changes. But in this wiki page https://en.wikipedia.org/wiki/Photopic_vision there is a single curve for the luminosity function in terms of wavelength, for a huge range of luminance level, $10$ to $10^8 cd/m^2$. Unless the black-body is seen under such a small solid angle that scotopic vision plays a role, the colour perceived should depend neither on this solid angle, nor on the solid angle of the pupil as seen from the part of retina that receives the light, but only on the black-body's temperature. – Alfred Jan 18 '22 at 22:50
  • @ProfRob Then again, the curve might depend on luminance, even within the huge range of luminance where photopic vision dominates, wiki is not all-knowing. One way to decide would be to look at, say, a red-hot piece of metal in a rather bright environment with atropin (or, rather, the new kinds of safer eyedrops ophthalmologists put in your eyes to look at your retinas) in only one eye to check whether the perceived colour is the same in both eyes or not. If not, it would make the relationship between SI units nonlinear, for a given wavelength, cds won't be proportional to Watts ! – Alfred Jan 18 '22 at 23:12
  • @Alfred all but the brightest stars look white. If you look at them through a telescope they become coloured. – ProfRob Jan 19 '22 at 00:22
  • @ProfRob This is the transition between scotopic to photopic vision. I never denied that. The point is, when one is within the huge range of photopic vision, from the luminance where the scotopic contribution becomes negligible up to the point of burning one's retina, between $10$ and $10^8 cd/m^2$, for a given wavelength, are candelas and Watts proportional or not ? If yes, the perceived colour of a black-body depends only on temperature within this range. If not, the SI definition of a candela is meaningless, which, between you and me, is quire possible, but it is a problem ! – Alfred Jan 19 '22 at 05:58