Astro
DEENFR

Learn · Method

Are the colors real? — how these images are made

August 30, 2026 · 9 min read

It is the most common question asked of every one of these images: “Does it really look like that?” The honest answer has two halves. No — no eye in the world sees a nebula this way, they are far too faint for that. And yes — the colors are not an invention but the actual colors of the spectral lines, only collected over hours instead of over the blink an eye is given.

The trick: not seeing more, but seeing longer

The eye exposes for roughly a tenth of a second, then throws the image away and starts over. A camera is allowed to add up: hundreds of 60-second frames become hours of collected light. On top of that comes the second trick from the hydrogen article: emission nebulae send out their light almost entirely in a few narrow lines. A filter that passes only those lines locks out the whole stray lighting of a suburb — and that is exactly what makes these images from a garden at the edge of the Taunus possible in the first place.

The visible spectrum from 380 to 700 nanometers. Marked are the nebular lines H-beta, [O III] and H-alpha as well as sodium street lighting. The dual-band filter passes only two narrow windows around [O III] and H-alpha; everything else is darkened.Hβ 486[O III] 496 · 501Hα 656sodium lamp 589400 nm500 nm600 nm700 nmthe two windows of the dual-band filterEverything outside — streetlight, moonlight, airglow — is locked out.
To scale: the nebular lines and the two passbands of the dual-band filter. The sodium line of the classic street lamp falls between the windows and is discarded completely. Note how narrowly Hβ sits beside the [O III] window — depending on the filter width it partly comes along.

What the camera does differently from the eye

The color sensor itself is, incidentally, color-blind: it only counts photons. The color comes from the Bayer pattern — tiny red, green and blue filters laid over the pixels like a checkerboard. And the raw summed image would still be almost black to an eye: the nebula sits in the lowest percent of the value range. Only the stretch lifts it into the visible — an amplification of brightness, not a recoloring. Hα is deep red, [O III] is blue-green; the images stretch the brightness axis, not the hue.

The imaging chain: collect photons for hours, the Bayer sensor sees only one color per pixel, stacking many frames lowers the noise with the square root of their number, stretching makes the linear signal visible.Collecthours instead of seconds,hundreds of framesBayer sensorevery pixel sees only one colorStackaveraging over all frames —noise falls as √NStretchthe linear signal becomes visible,the line colors stayAn unstretched astrophoto is almost black — the information sits in the lowest percent of the values.
The chain from photon to image. The gain from stacking is mathematics, not magic: quadruple the number of frames and the noise halves — which is why the deep exposures in this gallery are sums from several nights.

A few things worth knowing about this:

  • “False color” is something else. Famous Hubble images map three narrowband lines anew onto the red/green/blue channels (sulfur, hydrogen, oxygen — the “Hubble palette”). That is deliberate cartography, not natural color. The images in this gallery, by contrast, keep the natural assignment: Hα in the red, [O III] in the blue-green.
  • Here color is a measurement, not a matter of taste. Where red turns into blue-green, the ionization state physically changes — the color gradients in these images are maps of excitation, connecting directly to the fronts and supernova articles.
  • Galaxies need the opposite route. Starlight is a continuum across all wavelengths — a dual-band filter would nearly extinguish a galaxy. M 31 and M 33 are therefore made broadband; only their H II regions benefit from narrowband.
  • What is left out is documented honestly too. Whatever processing the images in this gallery have been through — calibrating, stacking, removing gradients, stretching — is written down in each entry's “Acquisition and processing” section. The standard: amplify what is there; paint nothing in.
  • And visually? In the telescope eyepiece, bright nebulae show a pale gray-green — the dark-adapted eye is most sensitive to [O III] and nearly blind to deep red. Anyone who has once seen the Veil visually through an [O III] filter knows: that gray-green has a magic of its own.
AI-generated illustration as a split image: on the left the same nebula pale in an orange-gray, light-polluted sky, on the right high in contrast in red and blue-green through a narrowband filter.

The same sky, two ways of imaging it: on the left the suburban reality with its dome of light, on the right the view through the two narrow windows of the narrowband filter. This image is AI-generated — not a photograph but an illustration of the principle.

Under our own sky

Four images, four statements about color:

NGC 6888 — WR 136 catches up with its own past1 object — show

Two lines, two stories: the Crescent Nebula carries Hα-red filaments and a blue-green [O III] arc — here the color literally separates the shock fronts (see the Wolf-Rayet article). Nothing about it is tinted; these are the two channels of the filter.

My photograph · IC 1318 Panorama · 21.3′ × 21.7′ · view gallery entry
W4 in context · heart bubble, dust fronts and IC 1795 at the edge5 objects — show
Chemistry over a wide field: in the Heart Nebula field the red-to-blue-green gradient shows where the radiation of the central stars excites the gas more strongly. The color map is an excitation map — readable like a weather chart.My photograph · IC 1805 · 1.95° × 1.52° · view gallery entry
IC 5070 — the Pelican as the second half of the same bubble2 objects — show

The Pelican in dual band: the drawing lives on the contrast between glowing hydrogen and dust in front of it. Without a narrowband filter this field would simply not be photographable from a suburban garden — it lies in the middle of the bright summer Milky Way.

My photograph · NGC 7000 · 1.05° × 1.59° · view gallery entry
Nucleus and inner disk6 objects — show

And the opposite case: M 33 is starlight, hence continuum — here the broadband route does the work, and the colors tell of populations (see lives of the stars): yellowish for age, bluish for youth, pink for the H II regions as narrowband islands in a broadband image.

My photograph · M 33 · 19.5′ × 17.6′ · view gallery entry

Further reading

  • Astrophotography — techniques and history (Wikipedia).
  • Bayer filter — how a color-blind chip sees colors (Wikipedia).
  • H-alpha — the line that gives the nebulae their red (Wikipedia).
  • Light pollution — the opponent the filters are up against (Wikipedia).

← All articles