Fundamentals · PixInsight 1.9.x

SCNR

SCNR (Subtractive Chromatic Noise Reduction) limits one colour channel, usually green, to the level of the other two. This corrects greenish colour errors – but can also distort real signal if the object actually contains green components.

Purpose

TaskApplication
remove a greenish colour error in the backgroundgreen, Average Neutral, with a mask if needed
correct greenish starsin the star image
shape the SHO paletteremove green so that gold and blue emerge
remove magenta fringesin the inverted image

Principle

SCNR changes only pixels whose share of the selected colour lies above a protection limit; all other pixels stay unchanged.

Protection methodLimit for green
Average Neutralmean of R and B
Maximum Neutralmaximum of R and B – stronger effect

Parameters

ParameterEffectTypical choice
Color to removecolour to limitGreen
Protection methodprotection limit (see above)Average Neutral
Amountstrength0.3–1.0 depending on the application

Operation

  1. Select the image, set the parameters, apply.
  2. Compare before/after with Ctrl+Z / Ctrl+Y at an affected spot (1:1 or 2:1).
  3. Readout on a greenish pixel: after SCNR, G at most at the level of the protection limit.

Typical applications

Green cast in the background

  • Object without green components (e.g. a pure Ha nebula): Amount 0.5–0.8 on the whole image, before raising contrast and saturation – these would otherwise make greenish blotches clearly visible. Check yellowish areas (lots of red and green) to see whether they tip towards orange; then Amount 0.5–0.7.
  • Object with OIII: only for a real green cast in the background, Amount 0.3–0.5, with an that protects the nebula. Compare before/after at an OIII spot: the teal must be kept.

Star image

Greenish stars are colour errors (channel misalignment, noise, Bayer interpolation); natural stars practically never appear green. In a star-only image SCNR is harmless because it contains no nebula with real green. Amount 0.5–0.8, 1.0 if a green cast remains.

Shaping the SHO palette

In an SHO image the strongest signal (Ha) lies in the green channel. SCNR removes the green excess; areas with SII (red) turn yellow/gold, areas with OIII (blue) turn cyan/blue. The typical gold-blue palette is thus created by removing green, not by saturation.

  • Amount 1.0: classic gold/blue look, hardly any green.
  • Amount 0.5–0.7: some green remains, more natural transitions.
  • Several passes with a lower Amount are easier to control than one at 1.0. Maximum Neutral acts more strongly.

Removing magenta

Magenta cannot be selected directly. In the inverted image magenta becomes green:

  1. Image → Invert.
  2. SCNR (Green), Amount 0.5–1.0.
  3. Image → Invert again.

Example: magenta fringes around stars in an SHO image, because stars, as broadband emitters, appear stronger in SII (red) and OIII (blue) than in Ha (green).

What a good result looks like

Typical errors

ObservationCauseRemedy
OIII areas blue or grey instead of tealSCNR applied to the nebulaSCNR only in the star image or masked in the background
Yellowish areas tip towards orangeAmount too highAmount 0.5–0.7
Star stays greenishAmount too low or wrong windowAmount 1.0; check the target image
SHO image flat after SCNRsaturation not yet raisedraise saturation afterwards via

When to use it – and when not

Good forBetter another tool
green cast without real green signaloverall colour balance → or BackgroundNeutralization
gold/blue palette in the SHO imagefine hue shifts →
magenta fringes via inversion
Used in

BackgroundNeutralization · HaLRGB workflow · MorphologicalTransformation · Narrowband workflow (HOO/SHO) · NoiseXTerminator (NXT) · OSC workflow with a quad band-pass filter · PixelMath · RangeSelection (Range Mask)