What you see here is one edge of something six full moons wide.
Somewhere between ten and twenty thousand years ago, a star in Cygnus ran out of fuel. Its core collapsed, the rest of it came apart, and for a few weeks it would have been one of the brightest things in the sky. People saw it. Nobody could write yet, so nobody wrote it down.
What's left is a bubble of shocked gas about 130 light years across, still tearing outward at roughly 1.5 million kilometres per hour.
Here's the part that took me a while to believe. Those ropes and ribbons aren't ropes or ribbons. The shell is thin — thinner, relative to its own size, than a soap bubble. Less than one part in fifty thousand of its radius. You only see it in the places where it curves edge-on into your line of sight. Everywhere else in this frame, the bubble is still there. You're just looking straight through it.
The red is hydrogen. The cyan is oxygen, glowing where the shock front is hottest and moving fastest, which is why it rides the leading edge of the arc rather than sitting inside it.
Five hours of exposure, one 8-inch telescope, a backyard north of Bucharest.
The long version
NGC 6992 is the brightest section of the eastern rim of the Cygnus Loop, catalogued as G74.0-8.5 and known variously as the Veil, the Network Nebula, or Caldwell 33. William Herschel found it on 5 September 1784 — the same year he catalogued NGC 7331, which was the subject of the last post here. His description of this one is unusually good: "branching nebulosity... the following part divides into several streams uniting again towards the south."
The remnant is middle-aged as these things go, with age estimates clustering around one to two times ten thousand years. X-ray spectra of the interior show a metal-rich plasma consistent with a Type II core-collapse event, so the progenitor was a high-mass star, and the remnant's unusually round shape suggests it exploded inside a cavity that the star's own wind had already cleared.
Distance has been argued over for decades. Minkowski put it at 770 parsecs in 1958 from proper motions Hubble had measured in 1937; later estimates ranged from 440 to 1400 pc. The current figure comes from Gaia parallaxes of stars sitting in front of and behind the expanding shell, giving 725 ± 15 pc, or about 2,400 light years.
The filament explanation is the standard one and it's worth stating precisely: the shock front is so thin compared with the radius of the shell that the emitting layer is only optically detectable when viewed exactly tangentially. Undulations in the surface produce several such tangent lines at once, which is why the filaments appear braided. Each one is roughly four billion miles thick — about the distance from here to Pluto.
Framing note: the Cygnus Loop measures roughly 2.8° × 3.5°. This telescope sees 1.00° × 42′. The whole object doesn't fit, and won't, even on the diagonal. The Eastern Veil on its own is a reasonable compromise.
Acquisition
| Date | Night of 17/18 September 2026 |
| Integration | 40 × 450 s, 5 hours total |
| Calibration | Matched darks, flats, bias |
Equipment
| Optics | GSO RC8 carbon, 203 mm, with TS 2″ Photoline 0.8× reducer — 1271.5 mm at f/6.3 |
| Camera | QHY8L one-shot colour CCD at −10 °C, gain 0, offset 140 |
| Filter | Optolong L-eXtreme dual-band |
| Mount | HEQ5 Pro belt-modded, on PLL C82 pier, EQMOD |
| Guiding | ASI220MM Mini on TS 50 mm, pulse-guided, 4.13″/px |
| Control | N.I.N.A. 3.2.0.9001, PHD2 2.6.14 |
| Image scale | 1.265″/px, field 1.00° × 42′ |
Processing
This is HOO from a one-shot-colour Bayer sensor with a dual-band filter, not a mono camera and a filter wheel. The L-eXtreme passes hydrogen and oxygen in narrow windows, and the two lines land in different parts of the Bayer matrix, so they can be pulled apart afterwards into what are effectively two separate monochrome images.
The same frame, without the stars
Almost none of the stars in this picture have anything to do with the nebula. They sit in front of it and behind it, scattered across thousands of light years of the Cygnus arm, and the supernova remnant just happens to lie along the same sightline. Modern processing can separate the two — a neural network trained to recognise stellar profiles pulls them out into their own layer, which is normally done so that stars and nebula can be stretched differently.
Run that separation and throw the star layer away, and this is what's left.
The Eastern Veil with the foreground and background stars removed. The small dark specks are where the brighter stars used to be.
It's a strange thing to look at. The structure is the same, but without the stars to give a sense of depth, the shell reads much more clearly as what it is — a surface, curving away from you, catching the light only along its folds.
- Siril — calibration, registration, stacking, then the dual-band data split into separate Hα and OIII channels.
- GraXpert — AI background extraction to flatten the light-pollution gradient, then stellar deconvolution and AI denoise on each channel. Stars came down from 4.7 to 2.6 pixels; background noise dropped about 40%.
- StarNet — stars separated from nebula, so the two could be treated independently. This is what fixes the cyan-star problem that HOO images usually have: the nebula gets a pure, unmixed Hα-red and OIII-cyan, while the stars get a blend of both channels that restores their natural colour.
- Siril again — generalised hyperbolic stretch per channel, HOO composition, and the two halves recombined.
- Photoshop — black point, colour balance, curves, a little saturation.
Shot from a backyard north of Bucharest, Romania.
