What does it mean that when light is reflected the "wavelengths of each contributing hue are subtracted"?

[zenterix]

Homework Statement

One process that occurs for a human to be able to perceive color is called "color mixing".

Relevant Equations

Basically, we have three different types of light-sensitive receptor cells in our retinas.

Each type of cell responds to light with wavelengths in a certain range.

The distinct ranges center around yellowish red, green, and bluish violet, respectively.

A "response" from a receptor cell is some neural signal to the brain.

The combination of the responses of all our receptor cells gives rise to the perception of a rich palette of hues of color.

My question is about the origin of the light that reaches our eyes.

A TV screen, for example, emits light. Each tiny spot on the screen emits different wavelengths that reach our eyes and the combination of wavelengths is perceived as a certain hue, as described above.

The wavelengths are thus "added", in the sense that each wavelength of the object we are looking at (a pixel on the screen) reaches our optical receptor cells.

On the other hand, suppose we are looking at yellow paint on a canvas. We can see it because the yellow wavelength of white light is being reflected by the paint and reaches our eyes. All other wavelengths are absorbed by the paint.

Suppose we mix in some blue paint together with the yellow paint.

The book I am reading says that in this case the "wavelengths of each contributing hue are subtracted".

I don't quite understand this statement.

Sure, when light hits the mixture of the two paints, the blue paint absorbs yellow light and the yellow paint absorbs blue light.

But each paint still reflects its specific color, no?

If we mix red, blue, and green, we get black paint, suggesting that all the wavelengths are absorbed in totality.

Is it simply the case that yellow and blue don't let some light be reflected but red, green and blue together don't? Is it just the case that three paints have much more absorptive capacity than two paints and this is why mixing the three is black but mixing two is not?

I don't understand the meaning of the word "subtract" in all of this.

Here is a picture from the book

 

[kuruman]

Look at the part of the spot on the right that is blue. You and I see blue which means that blue reaches our eyes. The spot is illuminated with white light (all colors). If we see blue that means that all the non-blue colors are absorbed by the surface and ony blue is reflected. "Subtracted" is another way of saying "absorbed by the surface upon reflection of the incident white light".

 

[haruspex]

But each paint still reflects its specific color, no?

If we mix red, blue, and green, we get black paint, suggesting that all the wavelengths are absorbed in totality.

Green paint, e.g., absorbs red and blue, but what happens to the green light is a combination of reflection and transmission. What is under it matters. Artists apply a brilliant white layer under the paint to get brighter colours.

Mixing red, green and blue paints and applying in a thick layer will not produce black. A muddy brown, more likely. The surface will be a side-by-side mix, so some light of every colour will be reflected.

If the paint layers are thin enough to act mostly like filters, passing through their stated colour range and absorbing the rest, then you could get black. With just a blue layer, the red and green would be absorbed while the blue would be reflected off the white substrate and pass back through the filter. Adding a green filter would absorb that blue, producing black with no need for a red filter. This is why newspaper printing uses magenta, cyan and yellow as its palette. Magenta absorbs green, cyan absorbs red, so overprinting the two lets just the blue pass through both ways.

 

[zenterix]

Look at the part of the spot on the right that is blue. You and I see blue which means that blue reaches our eyes. The spot is illuminated with white light (all colors). If we see blue that means that all the non-blue colors are absorbed by the surface and ony blue is reflected. "Subtracted" is another way of saying "absorbed by the surface upon reflection of the incident white light".

That makes sense to me but only in the very basic sense of light being absorbed or reflected.

I am asking about what happens when we mix colors.

"When light is reflected, from a photograph or book, the wavelengths of each contributing hue are subtracted."

When both yellow and blue paint are mixed, both are reflected (by yellow and blue paints, respectively) but also absorbed (by blue and yellow paints, respectively). Green is totally absorbed by both paints.

We see green because yellow and blue together stimulate our eyes' receptor cells in a way that we perceive green (that is, the mix of wavelengths present are perceived as green).

What exactly is subtracted here?

The way I understand it, every color is "subtracted", ie absorbed, except yellow and blue. Thus, I guess what is meant is that what we see depends on what is absorbed (because everything else (being reflected) is what reaches our eyes. But this everything else is "added", just like when we are looking at a pixel from a screen which emits certain wavelengths that are added.

 

[DaveC426913]

What exactly is subtracted here?

You don't seem to struggle with the concept of additive light: Red and Blue add to make Magenta.
Why do you strruggle with the concept of subtractive light: White, subtracting Green makes Magenta.

Thus:

The white light, falling on a painting, contains red, blue and green light. If the paint absorbs green light, then that green light is subtracted out of the initial white light, leaving only red and blue for us to see (as magenta).

 

[zenterix]

Artists apply a brilliant white layer under the paint to get brighter colours.

The first thing I notice here is the use of the word "brighter".

Technically, brightness seems to be determined by amplitude of the light waves, though I am not sure exactly how this works. I am not sure if this is the sense that you are talking about.

If we have a white canvas under a certain color of paint, you seem to be saying that potentially a little of every color will be reflected by the canvas underneath the paint, unless the paint is applied in a thick layer.

The surface will be a side-by-side mix, so some light of every colour will be reflected.

What do you mean by side-by-side mix? You assumed we mixed red, green, and blue, so why would it be side-by-side?

If the paint layers are thin enough to act mostly like filters, passing through their stated colour range and absorbing the rest, then you could get black.

You say "passing through". Isn't the light reflected? Ie, none of it passes through?

When you say "thin layer" aren't you trying to say that light passes because the paint isn't covering the entire surface beneath it?

 

[zenterix]

You don't seem to struggle with the concept of additive light: Red and Blue add to make Magenta.
Why do you strruggle with the concept of subtractive light: White, subtracting Green makes Magenta.

Thus:

The white light, falling on a painting, contains red, blue and green light. If the paint absorbs green light, then that green light is subtracted out of the initial white light, leaving magenta.

Sure, but then all the reflected light is added.

In both cases of a canvas or a screen, what we see depends on what is added (in one case emitted, in the other reflected). But I guess it is not wrong to view it from the perspective that in the case of the canvas, what is "added" is determined by what is "subtracted".

At this point, I don't think I have any conceptual doubts. It seems to be a purely semantic thing.

 

[haruspex]

brightness seems to be determined by amplitude of the light waves

or, in the particle view, more photons.

If we have a white canvas under a certain color of paint, you seem to be saying that potentially a little of every color will be reflected by the canvas underneath the paint, unless the paint is applied in a thick layer.

Of every colour that makes it through the paint, yes.

What do you mean by side-by-side mix?

Pigments are molecules. At the surface layer of these, pigments of different colours may lie adjacent.

When you say "thin layer" aren't you trying to say that light passes because the paint isn't covering the entire surface beneath it?

If you apply a thin layer of paint you can see some of what lies beneath. This is why one often uses an undercoat in house painting. Old Masters overpainted with a glaze https://www.dutchfinepaintings.com/old-masters-painting-tricks/

 

[DaveC426913]

Sure, but then all the reflected light is added.

Look at the whole system like a black box i.e. what happens inside you don't know about. All you count s the initial state and the final state. You don't know anything about any intermediate colours, such as red and blue or any internal additions.

1. White light goes in.
2. Magenta light comes out.

Ergo: Green has been subtracted.

 

[sophiecentaur]

Homework Statement: One process that occurs for a human to be able to perceive color is called "color mixing".
Relevant Equations: Basically, we have three different types of light-sensitive receptor cells in our retinas.

The book I am reading says that in this case the "wavelengths of each contributing hue are subtracted".

I don't quite understand this statement.

IMO it's necessary to get a good understanding about Additive Colour Mixing first. The brightnesses of each of the sets of three colour phosphors on a TV display are adjusted to produce the same subjective effect that will match ('metameric') the colour you're trying to display on each pixel.

For colour printing and colour film you can't do it that way because the illuminating (white) light is reflected from or passed through absorbent pigments to produce a wanted colour. So you have to subtract colours from the original white light. Whereas colour TV uses primary sources of Red, Green and Blue light, film uses Cyan, Magenta and Yellow filters. Cyan is what you get when you subtract Red from White light, Magenta is what you get from - Green and Yellow is what you get when you get from - Blue.
You can say C=-R, M=-G and Y=-B

 

[sophiecentaur]

Homework Statement: One process that occurs for a human to be able to perceive color is called "color mixing".
Relevant Equations: Basically, we have three different types of light-sensitive receptor cells in our retinas.

The wavelengths are thus "added", in the sense that each wavelength of the object we are looking at (a pixel on the screen) reaches our optical receptor cells.

It is more correct to say that the output from each receptor is the sum of the effect of the amplitudes of wavelengths in its passband. You add the amplitudes and not the wavelengths.

 

[haruspex]

It is more correct to say that the output from each receptor is the sum of the effect of the amplitudes of wavelengths in its passband. You add the amplitudes and not the wavelengths.

I think you meant you add the outputs, but I don’t think that is quite right either.
It's probably nonlinear.

 

[sophiecentaur]

The amplitudes of each output is the sum of the power for each frequency times the value of the response at that frequency. How ever you describe it, it’s not a sum of wavelengths. You have to be precise or it means nothing. Unfortunately, the popular pages are very sloppy at times.

 

[haruspex]

The amplitudes of each output is the sum of the power for each frequency times the value of the response at that frequency.

Sure, but in post #11 you had used "amplitudes" wrt the power distribution in the incoming light and "outputs" for the amplitudes of the responses of the cells. Just felt it was a bit confusing to then be summing the "amplitudes" instead of the outputs.

 

[sophiecentaur]

Sure, but in post #11 you had used "amplitudes" wrt the power distribution in the incoming light and "outputs" for the amplitudes of the responses of the cells. Just felt it was a bit confusing to then be summing the "amplitudes" instead of the outputs.

Try again - it's hard to say this in a non-confusing way.

There are two sets of 'additions'. First there is the weighted addition of the amplitudes of the incident wavelengths in the spectrum of the light in each of the analysis filter curves. It's a correlation process between the received spectrum and the three filter characteristic, which reduces a continuous spread of values across the spectrum that's seen to just three values. When you see two objects of the same colour, they may well have totally different spectra but the tristimulus theory says that two matching objects will have the same three values. That's an extremely coarse form of spectrometer but it's all we evolved with.

To synthesise a matching colour on a display you choose three suitable phosphors (sources) 1. which you can actually manufacture annd are bright enough and 2 which form a triangle on the CIE chart which encloses most of the commonly viewed colours. An appropriately weighted addition of the lights from the three phosphors can produce a pretty good match of input and output colours.

PLUS there is another analysis that takes place inside the eye / brain which (ideally) produces the same tristimulus values as the original object colour but in an entirely different way. Dead clever but to get a reliable match, the viewing conditions need to be the same for source and display. That's usually not possible so the camera processing has to adjust the weighting of the three synthesis colours so that white / grey is the same at both ends. Not always possible of course but our eyes do try hard to make what we 'see' look real. Look at the options on a camera or in Photoshop to allow automatic or manual control to put the White Point in the right place.

 

[sophiecentaur]

Look at the part of the spot on the right

Sorry; what 'spot'? And is the spot on a TV screen or a piece of paper?
If you want to use the concept of 'negative colours', you need to specify the co ordinates of the illuminant so that you can exactly 'remove' those negative colours.
Subtractive mixing is so much harder (buckets of pigment or dye) than additive mixing (three knobs to control the amount of light from each of three phosphors.

Something else that often gets ignored about colour mixing in film (in particular) is that the consequence of trying to make highly saturated colours is that the Luminance drops and drops. In practical terms that means that the 'prism' shape of the CIE chart over different saturations on a TV screen becomes a pyramid which slopes down towards the CMY primary points as the minuses take over. Saturated colours just come out darker. With computer - aided colour printing, that can be taken care of because you have control of every pixel on the image to (at least partially) correct for this but a cinema film image has to be treated the same all over.

Visualising CMY processing is beyond me, frankly.

 

[Charles Link]

See https://www.physicsforums.com/threads/on-mixing-colors-of-light.1066329/ for a thread somewhat related to what you @zenterix are asking.

It looks to me like they have already answered things very well, but this other thread that I posted about a week or two after yours covers similar topics, and shows a couple of the basic principles behind mixing a couple of colors of light, including showing how the CIE chart works.