Astro
DEENFR

Learn · Sky geometry

Projection — what only looks that way in the sky

August 30, 2026 · 8 min read

An astrophoto is a map with one fundamental defect: it has no depth axis. Everything that happens to lie along the same line of sight is stamped onto the same spot — the star next door right on top of the nebula behind it, the foreground cluster right inside the background cloud. Anyone who wants to read images of the sky has to know this defect, because it produces the finest illusions in this gallery.

The image is a line of sight

Above, seen from the side: a foreground star, a star cluster and a nebula lie at completely different distances along the same line of sight. Below, the camera image: all three appear in the same place, the depth is gone.seen from the sidecameraforeground stara few hundred light-yearsopen clustera few thousand light-yearsnebulafarther back againwhat the camera seesall three seemingly in the same place —the depth axis is entirelylost in the image
The camera integrates everything along the ray. What stands “in the middle of things” in an image may in truth lie thousands of light-years in front of it or behind it — adjacency in an image proves nothing at all.

Constellations connect stars by their directions in the sky. Many of these stars are far apart in space, although some smaller subgroups can be physically related. The familiar pattern alone does not establish a shared origin. Constellations primarily organise directions on the sky.

Size lies as well

The image's second omission: it knows only angles. Whether something is large and far or small and near cannot be read off an angle — a thumbnail at arm's length covers the Moon, and the Moon would cover the Crescent Nebula:

A cone of view from the eye: the Moon at 384,000 kilometers and the Crescent Nebula at 5,000 light-years appear under a similar angle — the angle in the sky says nothing about size or distance.the same angle in the skyat a distance differing by a good hundred billion timeseyeMoon384,000 km · 3,476 km acrossNGC 68885,000 light-years · 25 light-years across
Moon and Crescent Nebula appear under a similar angle — yet light would not even need a second to cross between Earth and Moon, while the nebula measures 25 light-years across. The sky shows angles; sizes have to be computed.

A few things worth knowing about this:

  • Every image is also a mosaic of times. The light of the foreground star is a few hundred years old, that of the nebula behind it a few thousand, that of M 31 two and a half million. A single photograph mixes epochs — what stands side by side is not simultaneous.
  • Andromeda is six full Moons wide. In the sky M 31 measures a good three degrees — the eye sees only the bright core of that, and many therefore take the galaxy to be tiny. The apparent contradiction “huge in the sky, yet invisible” is pure surface brightness: plenty of angle, little light per angle.
  • How the depth is won back. The missing axis is measured, not seen: parallax for the nearby stars, standard candles such as the type Ia supernovae from the supernova article for the distance, plus clues such as reddening by dust. Only these measurements turn the flat image into a space.
  • Sometimes the physics helps right inside the image. Silhouettes prove a foreground (dust in front of glowing gas), ionization rims prove a togetherness (the star has to be close enough to drive the front). An image that has been read holds more depth than it shows.
AI-generated illustration: a cone of view runs through space; inside it, at different depths, a bright foreground star, an open star cluster and far at the back a red nebula.

The cone of view as a scene: star, cluster and nebula at completely different depths — seen from the pointed end they collapse into a single sight. This image is AI-generated — not a photograph but an illustration.

In our own sky

The gallery is a small collection of projection cases — here are the best of them:

52 Cygni in the Witch's Broom — perfect superposition, false kinship1 object — show
The classic: 52 Cygni sits seemingly in the middle of the Witch's Broom — a perfect overlay, a false kinship. The star stands a few hundred light-years away, the supernova remnant some 2,400 behind it. No photograph in the world could show the difference.My photograph · NGC 6992 · 0.82° × 1.40° · view gallery entry
Dust front at the southwestern rim · Stock 7 in front of it only in projection1 object — show
Stock 7 in front of the dust wall of the Heart Nebula: the cluster lies in front of the front only in projection — they belong to different distance layers. Without a distance measurement, the obvious reading “the cluster shaped the wall” would be entirely plausible and entirely wrong.My photograph · IC 1805 · 30.2′ × 31.1′ · view gallery entry
NGC 1027 and the northeastern front · a foreground cluster over a young W4 rim1 object — show
The same pattern once more: NGC 1027 seems to hang on the young northeastern rim of W4, but it is a foreground cluster. Strikingly often, foreground clusters “decorate” nebulae they have nothing to do with — simply because the plane of the Milky Way is full of both.My photograph · IC 1805 · 1.05° × 0.74° · view gallery entry
The Wizard as a whole — a figure of light, dust and viewing direction4 objects — show
The Wizard owes its entire figure to the viewing direction: light, dust edges and line of sight have to stand exactly as they stand for a “face” to appear. Seen from the side there would be no wizard — only an ordinary bubble with rims.My photograph · NGC7380 · 39.4′ × 38.2′ · view gallery entry
IC 1318 and LDN 889 — the Butterfly is a silhouette3 objects — show

And the useful side of projection: the “Butterfly” exists only because the dark cloud LDN 889 happens to lie in front of the glowing surface — the silhouette proves the order. Sometimes the illusion itself is the depth gauge.

My photograph · IC 1318 Panorama · 1.90° × 1.60° · view gallery entry
NGC 891 – one galaxy on edge, others in the distance8 objects — show

NGC 891 extends this principle beyond the Milky Way: foreground stars, the large edge-on galaxy and much more distant small galaxies occupy the same image. Additional catalogue data, not angular separation, establish which belong to an association. Read more in Galaxy associations and the cosmic web.

My photograph · NGC 891 · view gallery entry

Further reading

← All articles