Additive vs Subtractive Colour Experiment at Home

Ask a child what red and green make. If they have ever used a paintbox, they will say brown, and they will be right about paint. Then overlap a red torch beam and a green torch beam on a sheet of white paper in a dark room, and the overlap is unmistakably yellow. Not brown. Not greenish. Yellow, as clear as a lemon. This additive vs subtractive colour experiment is built entirely around that contradiction, because the surprise is the teaching. Most children, and most adults, hold exactly one theory of colour mixing, the paint one, and it is the wrong one for everything on a screen.

Key Takeaways

• You overlap red, green and blue light on white paper, then mix the matching paints, and record the difference.
• Light adds: red + green = yellow, green + blue = cyan, red + blue = magenta, all three = white.
• Paint subtracts: red + green = dark brown, because each pigment removes what the other reflects.
• The prediction is the point. Ask what red and green make before you turn anything on, and write the answer down.
• Best for ages 7 to 13. Three phones showing full-screen red, green and blue are the most reliable light sources.

What You’ll Need

Item	How much	Where to get it	If you don’t have it
Red, green and blue light sources	3	Three phones or tablets showing a full-screen pure colour at maximum brightness	Three torches with party gel filters, or two layers of red, green and blue cellophane taped over torch heads
White paper or card	2 sheets	n/a	A white plate, or a white wall
Dark room	1	Curtains drawn, or night-time	A large cardboard box on its side to shade the paper
Red, green, blue and white poster paint	A teaspoon of each	Craft shop, school supplies	Food colouring diluted on a white plate, which actually behaves more like printer ink
Cyan, magenta and yellow food colouring	A few drops	Supermarket; a basic set has yellow, red, blue, green	Printer ink is the ideal but not worth buying for this
White saucers or a palette	3	n/a	A sheet of baking parchment on a plate
Magnifier, or a drop of water	1	n/a	Used to see the subpixels on a screen
Phone camera	1	n/a	Photograph each overlap, since the eye adapts and memory is unreliable

Time: prep 10 min, experiment 25 min, cleanup 10 min Cost: nothing if you have poster paint and a phone Mess level: medium, because of the paint

Do the Experiment: An Additive vs Subtractive Colour Experiment

Step 1 - Collect the prediction, in writing

Before any lights or paint come out, ask three questions and record the answers: what do red and green make, what do all three colours make together, and does it matter whether we use light or paint? Almost everyone answers brown, muddy black, and no. Keep the paper. You will come back to it.

Step 2 - Set up three clean light sources

Open a plain white image editor or a solid-colour app on three phones and set each to full-screen pure red, pure green and pure blue at maximum brightness. If you are using torches and cellophane, use two layers of cellophane each; one layer leaks far too much of the other colours and the result goes pale. Darken the room properly.

Step 3 - Overlap them in pairs on white paper

Lay the white paper flat and shine two sources onto it from about 20 cm, at a slight angle, so each makes a patch and the patches partly overlap. Work through red + green, then green + blue, then red + blue.

What you should see: Red + green gives a clear yellow in the overlap. Green + blue gives cyan, a pale swimming-pool blue. Red + blue gives magenta, a vivid pink-purple. The overlaps are always brighter than either single patch, which is the clue to the mechanism.

Step 4 - Add the third source

Bring all three patches together so they overlap in one place.

What you should see: White, or a very pale grey-white. It will not be perfect, because phone and cellophane colours are not spectrally pure, but it is unmistakably lighter and less coloured than any of the three sources alone.

Step 5 - Now mix the paints

On a white saucer, mix a blob of red poster paint with an equal blob of green. Then mix all three of red, green and blue.

What you should see: Red + green gives a dark olive brown. All three give something close to black-brown. Every mixture is darker than either ingredient, which is the exact opposite of what the lights did.

Step 6 - Mix cyan and yellow to show the other system works too

Put a drop of yellow food colouring and a drop of blue in a teaspoon of water and stir. Then try magenta or red with yellow.

What you should see: Yellow + blue gives green, which most children already know. Magenta or red + yellow gives orange or red. These are subtractive mixtures that behave predictably, unlike red + green, and they are how a printer makes every colour from three inks plus black.

Step 7 - Look at the screen up close

Put a single drop of water on a white area of a switched-on phone screen, or use a magnifier, and look closely.

What you should see: Tiny red, green and blue stripes or dots. The white you have been looking at your whole life was never white; it is three colours your eye cannot resolve, added together. This is the single most convincing moment in the activity.

What’s Actually Going On

There are two completely different processes, and they only look like the same thing because we use one word, “mixing”, for both.

Additive mixing happens when two light sources land on the same spot. The spectral power at each wavelength simply adds. Shine red light at around 630 nm and green light at around 530 nm onto the same patch of paper and the paper is reflecting both; there is no light near 580 nm anywhere in the room. Yet you see yellow.

The reason is in your eye, not in the light. The human retina has three types of cone, with peak sensitivities near 440 nm, 545 nm and 565 nm; Andrew Stockman and Lindsay Sharpe derived the middle- and long-wavelength cone sensitivities from measurements in observers of known genotype in Vision Research in 2000. Each cone reports a single number: how much total light it absorbed, weighted by its own sensitivity curve. That is three numbers for an entire spectrum, which means huge amounts of information are thrown away. Red light plus green light happens to produce the same three cone signals as a single yellow wavelength would. Your brain cannot tell the difference, and so there is no difference, perceptually. Two different spectra that look identical are called metamers, and they are the reason screens work at all.

Subtractive mixing is multiplication, not addition. A red pigment absorbs most wavelengths and reflects the long ones. A green pigment absorbs most wavelengths and reflects the middle ones. Mix them and a photon has to get past both: whatever the red pigment reflects, the green one absorbs, and vice versa. The reflectance curves multiply, so the product is small nearly everywhere and the mixture is dark. Add a third pigment and it gets darker still. The same logic applies to stacking coloured filters, which is why two layers of cellophane in different colours transmit almost nothing.

Then there is the part that creates most of the confusion. There are three primary systems in common use and school art class teaches the weakest one. Screens use red, green, blue, which are additive primaries. Printers use cyan, magenta, yellow plus black, which are subtractive primaries chosen because each one removes exactly one of the RGB bands. Art classes use red, yellow, blue, a historical convention that cannot produce a clean cyan or magenta and so has a visibly smaller gamut. If your child has been told that red, yellow and blue are “the primary colours”, they have been told something that is true only within one tradition of paint. Cyan and yellow making green is a reliable subtractive result; blue and yellow making green works mainly because artists’ “blue” is usually cyan-ish.

Two practical notes on getting it to work. Yuka Hirai and colleagues published a subtractive mixing procedure in Physics Education in 2025 using an LED mini-torch with filters printed on a colour printer, which is a cheap way to get cleaner filters than craft cellophane. And the classic quantitative treatment of subtractive mixing by R. de Wilde, P. Forrez and J. Hellemans in Physics Education in 1980 is worth reading if you want to measure transmittance rather than just look.

How to Explain It to Your Kid

Ages 5–8: Lights pile up, paints take away

Use their hands. Two torches on one spot is like two people both shouting; it gets louder. Two paints on one spot is like two people both covering their ears; less gets through. So light mixing makes things brighter and paint mixing makes things darker. Then let them do all the overlaps themselves and name each one out loud. Do not introduce cones.

Ages 9–12: Your eye only has three detectors

Tell them their eye has exactly three kinds of colour detector and that all the colours they have ever seen are three numbers. That is a shocking fact at this age and it lands. Then explain the yellow: red light and green light push those three detectors in the same pattern as real yellow light would, so the brain reports yellow. Then do the water-drop-on-the-screen step. This connects directly to how a digital camera sensor records colour with just three filters.

Ages 13+: Metamerism and gamut

Introduce metamerism properly: different spectra, identical percept, with information irretrievably lost at the cone stage. Then ask the hard question: if a screen only emits three wavelengths, which real-world colours can it not reproduce? Have them look up a chromaticity diagram and the idea of gamut. Finish with the practical consequence: this is why a photograph of a sunset never quite matches the sunset, and why printers and screens disagree about the same image file.

The question to ask: “If I shine cyan light and red light on the same spot, what colour will the overlap be, and will it be brighter or darker than each one alone? Decide, then we test it.”

Why Teaching This Early Pays Off

This is the clearest household example of a child holding a model that is coherent, useful and wrong outside its domain. Vosniadou and Brewer’s 1992 study of children’s mental models of the earth, in Cognitive Psychology, found that children build internally consistent models and then patch them when faced with contradicting evidence rather than replacing them. That is exactly what happens here if you are careless: told that red and green light make yellow, a child will often conclude that “the torches are weird” and keep the paint model intact.

The fix is to make both models correct within their domains rather than declaring one wrong. Light adds. Pigment multiplies. Both are right, and a child who holds two mechanisms and knows which applies has a far more useful understanding than one who has memorised a colour wheel.

The payoff is unusually concrete. Every screen a child will ever use is additive, every printed page is subtractive, and the gap between them is why things never print the way they look. It also pre-loads the trichromatic basis of colour vision, which is the foundation of everything from colour blindness to camera white balance. One afternoon will not teach colour science, but it reliably kills one specific, widespread misconception, and that is a good return for twenty-five minutes.

Before You Start: What Parents Should Know

Best age	7 to 13. Below 7 the light overlaps are still enjoyable and worth doing without the explanation
Supervision	Light. Paint and food colouring need an apron and a wipeable surface
Real hazards	None beyond stained clothing. Food colouring stains hands and worktops; poster paint stains fabric. Do not shine a torch into anyone’s eyes, and do not substitute a laser pointer for a torch under any circumstances
Don’t	Don’t use a bright white LED torch with a single thin layer of cellophane and expect clean colours. One layer transmits too much of everything and the overlaps come out pale grey instead of yellow

When It Doesn’t Work

• The overlap looks white or pale, not yellow. Your filters leak. Use two layers of cellophane, or switch to phone screens showing pure colours, which are far cleaner. Also check the room is genuinely dark.
• Red + green looks orange rather than yellow. The red is stronger than the green. Move the green source closer, or dim the red, until the two single patches look about equally bright. Balance is what makes the yellow convincing.
• All three together give grey, not white. Normal, and still useful. Perfect white needs balanced, spectrally pure primaries. Point out that it is lighter than any one patch, which is the part that matters.
• The paints make a nice colour rather than mud. You probably used a cyan-ish blue with yellow, which is a well-behaved subtractive pair. Use red and green specifically; that is the pair that goes muddy.

Take It Further

• Make a colour-matching game. Put a yellow object and a red-plus-green light patch side by side, photograph them, and see whether anyone can tell which is which in the photo. They often cannot, which is metamerism demonstrated rather than described.
• Measure a filter. Shine one torch through one, two and three layers of the same cellophane onto a phone light sensor and record the reading from a free light-meter app each time. Transmittance multiplies, so the readings should fall roughly geometrically, which is subtractive mixing as a number.
• Pull a colour apart instead of building one. Run a black marker through paper chromatography and compare: there you separate the dyes a manufacturer mixed subtractively, and the magnifying step pairs with using a water drop as a lens to inspect screen subpixels properly.

What to Watch For Over the Next 3 Months

• Week 4: They say “that’s additive” or “that’s the paint kind” about something unprompted: stage lighting, a projector, a printed poster.
• Month 2: They notice that a photo of a screen looks wrong, or that a printed picture is duller than the same image on a tablet, and can say why.
• Month 3: If they ask why there is a separate black ink in a printer when cyan, magenta and yellow should make black, they have found a real engineering answer: three inks give a muddy brown-black, and black ink is cheaper and sharper for text.

Frequently Asked Questions

Why do red and green light make yellow when red and green paint make brown?

Because light sources add their spectral power while pigments multiply what is left after absorption. Red and green light together stimulate your three cone types in the same pattern a single yellow wavelength would. Red and green pigment each absorb what the other reflects, so almost nothing comes back.

Are the primary colours red, yellow and blue or red, green and blue?

Both, in different systems. Red, green and blue are the additive primaries for light and screens. Cyan, magenta and yellow are the subtractive primaries for ink. Red, yellow and blue is an older painters’ convention with a smaller range than cyan, magenta and yellow.

Can I do this with coloured cellophane from a craft shop?

Yes, with two layers of each colour. A single layer transmits too much of the other wavelengths and the overlaps come out pale. Phone or tablet screens showing a full-screen pure colour are cleaner and need no filters at all.

Is there really no yellow light in the mixture?

Correct. Measure the spectrum and you would find a peak near 630 nm and another near 530 nm, with nothing in between. The yellow exists only as a perceptual response. That is what makes it a good demonstration of how much of colour lives in the observer.

Why does my printer need black ink if cyan, magenta and yellow make black?

Because real inks are not perfect absorbers, so the three together give a muddy dark brown rather than true black. Black ink is also cheaper per page and gives sharper text, since one ink layer beats three overlapping ones.

---

About the author

Ricky Flores is the founder of HiWave Makers and an electrical engineer with 15+ years of experience building consumer technology at Apple, Samsung, and Texas Instruments. He writes about how kids learn to build, think, and create in a tech-saturated world. Read more at hiwavemakers.com.

---

Sources

1. Stockman, A., & Sharpe, L. T. (2000). “The spectral sensitivities of the middle- and long-wavelength-sensitive cones derived from measurements in observers of known genotype.” Vision Research, 40(13), 1711–1737. https://doi.org/10.1016/S0042-6989(00)00021-3
2. The Physics Classroom. “Color Addition.” https://www.physicsclassroom.com/class/light/Lesson-2/Color-Addition
3. The Physics Classroom. “Color Subtraction.” https://www.physicsclassroom.com/class/light/Lesson-2/Color-Subtraction
4. Georgia State University. “Color Perception.” HyperPhysics. https://hyperphysics.gsu.edu/hbase/vision/colper.html
5. Hirai, Y., Morita, C., & Kamata, M. (2025). “A subtractive colour mixing experiment using a mini-torch with LED and filters handmade with a colour printer.” Physics Education, 60, 035002. https://doi.org/10.1088/1361-6552/adb557
6. de Wilde, R., Forrez, P., & Hellemans, J. (1980). “A quantitative demonstration of subtractive colour mixing.” Physics Education, 15(3), 152–154. https://doi.org/10.1088/0031-9120/15/3/407
7. Vosniadou, S., & Brewer, W. F. (1992). “Mental models of the earth: A study of conceptual change in childhood.” Cognitive Psychology, 24(4), 535–585. https://doi.org/10.1016/0010-0285(92)90018-W