BrettjoAstro · Field Guide
Narrowband, RGB Stars
& the Milky Way
Why narrowband nebulae need broadband star frames — and how the Seestar S30 Pro's Milky Way sessions fit into your capture workflow.
The filter only passes a sliver of light
A narrowband filter like the Optolong L-Ultimate passes only two very narrow windows of the electromagnetic spectrum — hydrogen-alpha (Hα) at around 656nm, and doubly-ionised oxygen (O III) at around 501nm. Everything else — including most of the light that defines the colour and brightness of stars — is blocked.
This is great for nebulae — an emission nebula like M27 (the Dumbbell) or NGC 7380 (the Wizard) radiates almost entirely in those two lines, so the filter dramatically boosts signal-to-noise in a light-polluted or moonlit sky.
But for stars, it is a problem. The filter renders most stars unnaturally dim, colourless, or with a false colour bias towards red (Hα) and cyan (O III). Stars shot through a narrowband filter look wrong and cannot be colour-calibrated to real stellar temperatures.
Stars need their own broadband exposure
The solution used by almost every serious narrowband imager is to capture a short set of broadband RGB exposures on the same target — no narrowband filter, just clear glass (or a broadband luminance filter). These frames capture the true colour and brightness of every star in the field.
Broadband RGB records the blackbody colour of each star — hot blue-white O-type, warm yellow G-type, deep red M-type — as it would appear to the eye.
PixInsight's spectrophotometric colour calibration (SPCC) works on broadband star data. Narrowband-only frames cannot be SPCC'd to a meaningful white reference.
After StarXTerminator removes stars from your narrowband nebula, the RGB star frames are blended back in via ScreenStars or manual layering — restoring natural colour without introducing filter artefacts.
Stars are bright. You do not need five hours of RGB. Typically 20–40 minutes of broadband subs at shorter exposure lengths is sufficient to capture clean stars across the field.
RGB stars sit at project level, not rig level
An important nuance in your workflow: the requirement for RGB star frames applies to a project (a target, like M27), not to a specific telescope. Whether the narrowband data came from the FRA400 or the 107PHQ, you still need a set of broadband RGB stars for that target.
This matters for AstroLog. If the RGB stars session is linked only to a telescope record, the logic breaks when you switch rigs. The RGB stars requirement is a project-level dependency.
| Session type | Telescope | Filter | Project dependency | Required? |
|---|---|---|---|---|
| Narrowband nebula | FRA400 or 107PHQ | L-Ultimate (Hα / O III) | Defines the project | YES |
| RGB stars | Either (same FOV preferred) | None / broadband | Companion to narrowband project | YES |
| Milky Way | Seestar S30 Pro (wide field) | None | Separate Milky Way project instance | WHEN SHOT |
A different imaging mode entirely
The Seestar S30 Pro is a wide-field smart scope. When summer darkness is limited — as it is now from Shinfield at 51°N — it is well-suited to Milky Way panoramas while your primary rig is working on a narrowband DSO target.
Milky Way captures differ fundamentally from DSO imaging in ways that affect how they should be logged:
No fixed RA / Dec target
A DSO session points at a named object with precise coordinates. A Milky Way shot is a wide-field composition — it may be framed by a landscape feature, a rising arc, or a galactic centre alignment. There is no single target RA/Dec to log.
No filter — broadband by default
Milky Way imaging captures the full broadband colour of star clouds, dust lanes, and emission regions together. No narrowband filter is used, so the RGB star colour problem does not apply — the stars are captured correctly in the same frames as the Milky Way itself.
FITS metadata differs
Your Python script on the Mac already has a specific FITS importer for Milky Way instances. This means the FITS headers likely differ from DSO frames — possibly no OBJECT keyword, or a custom value. AstroLog needs to handle this gracefully rather than treating a missing OBJECT as an error.
Running both simultaneously
Polar align the AM5N. Load the target into ASIAIR. Confirm the L-Ultimate filter is in the train. Begin narrowband acquisition (Hα and O III channels, 3–5 min subs, dithering every 2–3 frames).
While the primary rig runs unattended, set up the Seestar S30 Pro on a separate mount position. Frame the Milky Way arc or galactic centre. Begin capture — the S30 Pro runs autonomously.
Before the end of the session, remove the L-Ultimate from the primary rig (or switch to a clear filter slot). Capture 20–40 minutes of broadband RGB subs on the same narrowband target. These are the RGB star frames.
Your Mac-side Python script ingests the FITS files. The DSO narrowband frames go into the project. The Milky Way FITS files are handled by the dedicated importer you added. Both feed into AstroLog as separate session records under the same night log.
Narrowband data: stack Hα and O III separately. Apply SPCC. Use StarXTerminator to remove stars. Process nebula. Recombine with the RGB star stack using ScreenStars. Milky Way data: stack independently, stretch with GHS, colour-grade separately.