Why James Webb Space Telescope Images Are in False Colour: What Infrared Light Actually Shows

The James Webb Space Telescope does not carry an optical camera. Its primary instruments, NIRCam, NIRSpec, MIRI, detect near-infrared and mid-infrared radiation, wavelengths that sit well beyond the red end of the visible spectrum. Human eyes top out at roughly 700 nanometres. Webb's NIRCam works from about 600 nanometres to 5 micrometres. MIRI goes further, from 5 to 28 micrometres. At those wavelengths, the detector registers photons as numerical data, intensity values at each pixel, not as colour. There is no colour to capture, because colour is a property of human vision, not of light itself.
This is why the raw output from Webb looks nothing like the released images. It arrives as grayscale frames, one per filter, each representing a narrow band of infrared wavelengths. The science team at the Space Telescope Science Institute (STScI) in Baltimore then begins the process of turning those frames into something a human can read.The method is called false colour mapping, and it has a long history in astronomy, Hubble images used it too, though Webb's infrared range makes the translation more dramatic. Each filter's grayscale frame gets assigned a colour from the visible spectrum, typically following a rule: shorter wavelengths get blue tones, longer wavelengths get red tones. Stack the layers, adjust the brightness and contrast so that faint structures become visible without washing out bright ones, and the final image emerges.
The people doing this work are not graphic designers making things pretty. Alyssa Pagan and Joseph DePasquale at STScI, along with citizen scientist Judy Schmidt who has processed Hubble and Webb data for years, make decisions grounded in what the data actually contains. Which filter should map to blue? How much to stretch the brightness scale so that a dim gas cloud and a blazing star core can appear in the same frame? These are scientific choices. The colours are chosen to maximise information, not aesthetics, though the two sometimes coincide.Take the Carina Nebula image released by NASA in July 2022 as part of Webb's first science images. The orange-red cliffs at the bottom of the image represent dense clouds of gas and dust, glowing in mid-infrared. The blue regions above them are hotter, more ionised gas. The points of light scattered across the top, stars previously hidden behind dust, appear because Webb's infrared wavelengths pass through dust that blocks visible light entirely. A Hubble image of the same region shows a wall of obscuring cloud. Webb shows what was behind it.
In the Stephan's Quintet image from the same release, the red arc sweeping across the frame marks a shockwave where one galaxy is crashing into a gas cloud at around 800 kilometres per second. That shockwave is invisible in optical light. In Webb's mid-infrared data, mapped to red, it dominates the frame. The colour is not decoration. It is the data.If you mapped Webb's infrared filters directly to visible colours, treating 1 micrometre as red, 2 micrometres as deeper red, and so on, the resulting image would be almost entirely dark red and black. The wavelength differences between filters are too large, and the human eye's sensitivity to those near-red tones is too compressed, to distinguish anything. You would lose the structural information that makes the images scientifically valuable.

Optical telescopes like Hubble can produce something closer to true colour because their filters overlap with the visible spectrum. Even then, Hubble's released images are processed and colour-balanced. The idea of a raw, unprocessed, true-colour space image is largely a myth even for visible-light telescopes. Sensors don't see the way eyes do. Processing is not manipulation, it is translation.
ISRO's own imaging missions use similar principles. Chandrayaan-2's orbiter camera produces data in multiple spectral bands, and the false-colour composites generated from those bands are what reveal mineral distribution across the lunar surface. The science is in the colour choices, not despite them.A Webb image is real in the most important sense: every photon that contributed to it actually arrived at the detector from the object shown. The distances are real. The structures are real. The chemical signatures encoded in the brightness values are real, NIRSpec can identify specific molecules by the wavelengths they absorb or emit, and those identifications are what drive the colour choices in many images. What is constructed is the visual representation, because there is no other way to represent data the human eye cannot access.

The colours in a Webb image are closer to a musical transcription than to a photograph. A piano arrangement of a symphony is not the symphony, but it carries the same notes, the same relationships, the same structure. You are hearing the real piece through a different instrument. Webb's false colour images work the same way. The universe sent the signal. The scientists chose the instrument to play it back on.
What this means is that the most visually dramatic space images ever produced are also among the most information-dense. The gold and crimson of the Pillars of Creation in Webb's 2022 image encode dust temperature, molecular hydrogen emission, and the locations of newly forming stars. Strip the colour choices away and you have grayscale frames that carry the same data but reveal none of it to a human reader. The image is not a picture of space. It is a map of what space is made of, rendered in the only language eyes understand.