Everything here was shot from a back deck in the Central Iowa suburbs, under a Bortle 8.6 sky. The Bortle scale runs from 1 (middle of nowhere, the Milky Way bright enough to cast shadows) to 9 (downtown). At 8.6 the Milky Way is gone and, on a good night, the stars overhead can be counted by eye. The site name rounds up, because it feels like a 9. Running total: 52 targets, 243 hours of exposure, 26 nights and 21,931 individual frames.
Not one of these objects is visible from that yard. That's the fun part.
The gear
Two small all-in-one smart telescopes: a ZWO Seestar S30 Pro and a DwarfLab Dwarf 3. Camera, optics and mount in one box each, and both run on equatorial mounts, so they track the sky the way a proper mount does and the frame doesn't slowly rotate through the night. Polar-aligning on a wooden deck is a hobby of its own: the boards flex, and walking away from the scope at the wrong moment is enough to knock it off.
Shooting
- Lots of short frames, not one long one. Each scope takes exposures of 10 to 60 seconds all night long, and they get stacked afterwards. A plane, a satellite or a gust of wind costs a few frames instead of the whole night.
- Filters are what make this possible. Streetlights and city glow cover the whole spectrum, but emission nebulae glow at a couple of very specific colors: hydrogen-alpha (deep red) and oxygen-III (teal). A dual-band filter lets just those through and throws out most of the light pollution. Galaxies and star clusters shine in every color, so a narrow filter would throw their light out too. Those are shot broadband and simply need a lot more time.
- Some nights the Seestar is driven from a computer with N.I.N.A. instead of its phone app, taking raw 60-second frames. Every frame is kept, the bad ones (clouds, trailing, the neighbor's floodlight) get tossed, and each night is calibrated properly with darks and flats.
Many nights per target
One night is almost never enough under this sky. Most objects get several nights, sometimes weeks apart, all stacked into one image. The hours on each card are the real total that went in, counted straight from the capture files. The bar underneath is the goal for that kind of object: 1.5 hours for a bright cluster, 2 for a planetary nebula, 3 for a bright nebula, 4 for a galaxy, 5 or more for faint targets and mosaics. Past the goal, doubling the time buys only about 40% more signal, so that's where a target counts as done. The really big ones are mosaics: several overlapping fields stitched together.
Processing
- PixInsight does nearly all of it. The early images were done in Siril (a few of those are still here), and PixInsight won the side-by-side comparison. Stacking, color calibration, pulling out the gradient the city glow smears across every frame, then RC-Astro's BlurXTerminator, NoiseXTerminator and StarXTerminator to sharpen, quiet the noise and lift the stars off so the nebula can be worked on by itself.
- Stretching is where the picture shows up. A raw stack looks almost black: the faint stuff is a tiny fraction of the brightest stars. A careful stretch brings it up without turning every star into a blob.
- HOO vs Natural. Emission nebulae get two versions. Natural keeps the colors roughly the way the filter caught them. HOO maps hydrogen to red and oxygen to teal, which shows where each gas sits and how they mix. HOO is the favorite, so it opens first. 28 targets have both; the toggle above the image switches.