Decide how much total integration you want, then see how many sub-exposures it takes, how many clear nights you'll need, and how much disk space — so you can plan a deep-sky campaign realistically.
Total integration time is the biggest lever on a deep-sky image. Random noise falls with the square root of the number of frames, so quadrupling your integration roughly halves the noise. That is why a faint galaxy that looks grainy after one hour can look smooth after eight — you are averaging away the noise that hides the faint detail.
Realistic starting points for a clean result from a reasonably dark site (cities and faster optics shift these):
| Target | Typical total integration |
|---|---|
| Bright broadband nebula or cluster | 2–4 hours |
| Spiral galaxy (broadband LRGB) | 6–12 hours |
| Emission nebula (narrowband) | 12–30 hours |
| Faint IFN / very small galaxies | 20–40+ hours |
Light pollution adds a bright, noisy background that competes with your target, so under heavy light pollution you need far more total integration to reach the same signal-to-noise — often three to five times as much as a dark rural site. Narrowband filters help enormously because they block most of the unwanted sky glow, which is why they are the city imager’s best friend.
Modern deep-sky images are stacks of many sub-exposures totalling hours of integration time. Faint targets reward more data — signal-to-noise improves with the square root of total integration — so knowing how many subs and nights a goal takes helps you decide whether a target is realistic from your site.
Longer subs cut overhead and read noise but risk losing frames to clouds, satellites or guiding glitches, and can saturate bright stars. The sweet spot is "sky-limited" — long enough that sky background swamps read noise. Our field-of-view and guiding tools help you get the gear side right first.
It depends on the target and sky. Bright objects look good in 1–3 hours; faint nebulae and galaxies often want 10–30+ hours, and narrowband from light-polluted skies can run higher still. Because SNR grows with the square root of time, doubling your data only improves noise by ~1.4×.
The stretch of real darkness when your target is high enough to image well (typically above ~30°). It shrinks for low-declination targets, near twilight, and around the full Moon. Be conservative — clouds rarely cooperate for a full night.
Every sub costs download time plus dither-and-settle. With 300 s subs and 20 s overhead you lose ~6%; with 30 s subs and the same overhead you lose ~40% of the night. Longer subs (when guiding allows) reclaim that time.
Longer subs, less aggressive dithering, a faster (lower f-ratio) system, or simply a lower integration target. A dual-rig or wider sensor also gathers more sky per frame.
A planning estimate. Real yield depends on weather, target altitude, the Moon and how strict your frame culling is.
It depends on the target's brightness and your sky. Bright nebulae can look good in 1-2 hours, while faint galaxies and narrowband targets often need 10-30 or more hours of total integration to build a clean signal.
Once each sub-exposure is long enough to overcome read noise, total integration time matters most. Many shorter subs are more forgiving of tracking errors and satellite trails, while fewer long subs risk clipped stars and lost frames.
Yes. Random noise falls with the square root of the number of frames, so roughly quadrupling your total integration time halves the noise and reveals fainter detail.
Multiply your sub-exposure length by the number of subs, then add overheads such as dithering and download time. This calculator does the maths and shows how many nights the target will take.