Model, assumptions & sources
The optimal sub-exposure makes sky-background shot noise dominate read noise (Dr. Robin Glover, SharpCap):
t_sub = C · RN² / (LPR + dark) C = 1 / ((1+p)² − 1)
where p is the accepted fractional increase in total noise: p=5% → C≈9.8, p=10% → C≈4.8, p=2% → C≈24.8.
The sky electron rate (LPR, e⁻/pixel/s) is built from your site and gear:
LPR = P₀ · 10^(−0.4·m_sky) · W · T · A · Ω · QE · τ
P₀ = 1.0×10⁷ photons·s⁻¹·m⁻²·Å⁻¹ per arcsec² for a V=0 source (Bessell V zero point, 3640 Jy ≈ 999 ph·s⁻¹·cm⁻²·Å⁻¹) ·
m_sky = sky surface brightness (mag/arcsec²) ·
W = filter FWHM (Å, top-hat approximation) ·
T = filter peak transmission ·
A = πr² effective aperture (m²) ·
Ω = (206.265·pixelµm·bin / FLmm)² (arcsec²/px) ·
QE, τ = quantum efficiency and optical transmission.
Labelled approximations: filters are treated as top-hats; the V-band photon zero point is used across the visible band; narrowband uses sky continuum only (real light-pollution emission lines, e.g. Na/Hg, are not modelled — narrowband subs may be shorter than predicted if such a line falls in-band); Bortle→SQM mapping is approximate (SQM is authoritative); the moon model is illumination-based and rough. Star-saturation is not modelled — long subs may clip bright stars regardless.
Filter database: FWHM/bandwidth values are the manufacturers' marketed specs (well corroborated across sources — the dominant driver of exposure). Peak transmissions are the makers' stated peak/typical figures; several brands publish transmission curves rather than numbers, so treat T as ±a few %. Transmission scales exposure linearly, so small errors there matter far less than FWHM. Pick Custom… in any row to enter measured values.
One-shot colour (OSC) mode: on a Bayer sensor each colour pixel only sees its own emission line — red pixels collect Hα/SII, blue+green collect OIII/Hβ. The tool computes the sky rate (and required sub) per Bayer channel and recommends the longest one, so read noise is swamped in every channel. The Bayer response fractions (Hα→R≈1.0, OIII→G≈0.85/B≈0.80, etc.) are representative values, not a measured per-sensor matrix — treat OSC numbers as good estimates. Camera presets use the colour-version QE for OSC and the mono-version QE for mono; read noise is the gain-dependent working value (editable). Sensor full-well is quoted at gain 0 and drops at high/HCG gain.
Sources:
· Optimal sub formula & swamp factor — SharpCap Smart Histogram, SharpCap forum
· V-band photon zero point — Astronomical Magnitude Systems (Case), Zero point (photometry)
· Aluma AC4040 — Diffraction Limited · Stellarvue 130EDT — Stellarvue
· Filter bandpasses — Baader, Astrodon, Chroma, Antlia, Optolong, Astronomik, ZWO, Radian/STC/SVBONY/Askar (manufacturer + retailer specs)
· Sensor specs — ZWO & QHY product pages + Sony sensor datasheets (read noise is gain-dependent; values are representative working figures)
Built for Stellar Nomads · sky & gear specific · no data leaves your browser.