SpectrophotometricColorCalibration (SPCC)
SpectrophotometricColorCalibration (SPCC) uses Gaia spectra to compute how each star should appear through the filters and sensor used, compares this with the image and derives the channel factors from it. It also neutralises the background.
Purpose
| Task | Setting |
|---|---|
| natural colours in broadband and OSC images | sensor and filter curves, white reference |
| neutralise the background | Background Neutralization with a region of interest |
WhySPCC rather than a manual white balance?
SPCC compares thousands of stars photometrically with Gaia DR3 spectra. That is objective and reproducible; a manual balance follows visual habits.
Principle
- For each star: expected colour ratio (from the Gaia spectrum, filter and sensor curves) versus measured colour ratio in the image.
- A line is fitted through all stars (linear fit). At the position of the white reference (e.g. Average Spiral Galaxy) this gives the white point.
- To make the white reference appear neutral, each channel is scaled by the inverse of its white-point
y(G as reference) and normalised to the largest factor → white balance factors. - Prerequisites: linear image, astrometric solution, gradients already corrected. A gradient would distort background and white balance (different background colour in each corner). The photometry needs unaltered star profiles – hence before AI-based deconvolution for linear data – corrects star shapes, makes stars smaller and sharpens nebula structure. Open fundamentals →.
Parameters
| Parameter | Setting | Why |
|---|---|---|
| QE curve | the camera’s sensor | its spectral sensitivity enters the expected star colours |
| Filter R/G/B | the filters actually used; for OSC the sensor’s colour filter curves | SPCC computes how each star appears through exactly these filters |
| Additional filter | e.g. a light-pollution or multi-band filter, if listed | otherwise its effect is not modelled |
| Narrowband mode | off for broadband and OSC | |
| White reference | Average Spiral Galaxy | neutral reference: an average galaxy type appears white |
| Background Neutralization | enabled, region of interest on a nebula-free preview | see below |
| Lower / Upper limit | default −2.80 / +2.00 (sigma relative to the median) | lower the upper limit in nebula-filled fields (about +1.0 to +1.5) |
Operation
- ImageSolver, if the astrometric solution does not come from Chains calibration, cosmetic correction, weighting, registration, normalisation and integration into a reproducible pipeline – from individual subs to aligned master lights. Open fundamentals →.
- Preview (Alt+N) for the background: area without visible nebula, as few bright stars as possible, 300–400 px on a side, at least 50–100 px from the edge. Dark nebulae are unsuitable (darker than the true sky).
- In SPCC under Background Neutralization enable Region of Interest, take it From Preview.
- Apply; check the plots and the log.
Whya region of interest?
Without it the background is determined from the whole image. In fields with extended nebula, nebula components flow in and SPCC partly neutralises the nebula – faint outer parts turn grey instead of coloured. The principle is the same as with Gleicht die Kanäle so an, dass der Hintergrund neutral grau wird – ohne die Farben des Objekts zu kalibrieren. Für Falschfarben und als Nachkorrektur. Open fundamentals →.
Typical applications
Interpreting the results
R/G and B/G plots:
| Element | Meaning |
|---|---|
| green dot | one star |
| x-axis (Catalog) | expected colour ratio from Gaia spectrum and filter curves |
| y-axis (Image) | colour ratio measured in the image |
| blue line | linear fit through all stars |
| vertical black line | colour ratio of the white reference in the catalogue – x of the white point |
| horizontal black line | value of the fit line there – y of the white point |
| Feature | Ideal | A deviation means |
|---|---|---|
| point cloud | dense, linear, not curved | curvature or split-off groups → filter or sensor curves do not match the system |
| intercept | close to 0 | spectral deviation (e.g. unmodelled additional filter) or background offset |
| slope | any, but stable across the cloud | – |
| σ | small (broadband about ≤ 0.03–0.06) | noise, filter influence, faint stars |
| white point | in the dense part of the cloud | at the edge or outside → white balance poorly supported |
| outliers | single points | double stars, variables – harmless |
Example (broadband, mono RGB): 2,784 stars; R/G slope 0.916, σ 0.029; B/G slope 1.344, σ 0.058; factors 1.000 / 0.787 / 0.710. Both clouds linear and dense, white points central.
Multi-band filters on colour cameras
A quad or dual-band filter cuts out parts of the stellar spectrum. If it is not available as an additional filter, SPCC calculates with the plain sensor curves; measured and computed colours are then not strictly linear: slopes clearly different from 1, large intercept, increased scatter.
Example (OSC with quad band-pass): 5,077 stars; R/G slope 1.374, σ 0.107; B/G slope 0.645, intercept +0.521. White points still in the dense area → white balance plausible. The colours are a good approximation, not an exact calibration; the visual check is what counts.
What a good result looks like
Typical errors
| Observation | Cause | Remedy |
|---|---|---|
| Faint nebula grey | Background Neutralization without region of interest, or upper limit too high | region of interest on a nebula-free preview, lower the upper limit |
| Cast in one corner | gradient not corrected | gradient correction first |
| Curved point cloud | wrong filter or sensor curves | check the curves |
| Large scatter, slope ≠ 1 | unmodelled additional filter | expected; check visually |
| SPCC does not start | astrometric solution missing | ImageSolver |
When to use it – and when not
| Good for | Better another tool |
|---|---|
| broadband RGB and OSC, also with a light-pollution filter | false colour (HOO, SHO) → Gleicht die Kanäle so an, dass der Hintergrund neutral grau wird – ohne die Farben des Objekts zu kalibrieren. Für Falschfarben und als Nachkorrektur. Open fundamentals → |
| objective, reproducible colours | flux calibration for MGC → SPFC (see Corrects gradients using a reference map of the sky (MARS) and can therefore distinguish real large-scale signal from gradients. For broadband data; requires SPFC. Open fundamentals →) |
BackgroundNeutralization · CurvesTransformation · HaLRGB workflow · MultiscaleGradientCorrection (MGC) · Narrowband workflow (HOO/SHO) · OSC workflow with a quad band-pass filter · SCNR · WeightedBatchPreprocessing (WBPP)