Add Hα to RGB While Preserving Natural Star Colors

Hα data can make hydrogen-emission structures easier to distinguish in a broadband image. RGB supplies the broadband color context, including reflection nebulosity, continuum light, and stellar colors. Adding Hα changes that balance; it does not make the result a more precise natural-color measurement.

This lesson uses a restrained linear NBRGBCombination trial on starless inputs, then restores stars extracted from calibrated broadband RGB. Keeping Hα stars out of that restoration helps avoid replacing broadband star colors with narrowband profiles.

It is an optional branch that spans stages: prepare and calibrate RGB first, combine the starless data while linear, stretch the object and RGB stars separately, then restore stars during finishing. Do not follow it as an instruction to mix Hα into uncalibrated RGB immediately after ChannelCombination.

Start with calibrated, integrated Hα and broadband RGB data of the same field. Register them to one reference and apply a common crop. Check stars at 1:1 in the center and corners before removing them. Preserve the original integrations and geometry.

Complete appropriate linear gradient correction on each branch. Use a genuine matching MARS reference for Hα where available, or the DBE workflow when suitable. Do not apply broadband response settings to an emission-line master.

Complete broadband RGB plate solving and SPCC with the actual camera/filter responses before extracting its stars. Keep it linear. Save Target_RGB_calibrated_linear.xisf and Target_Ha_prepared_linear.xisf. Apply any intended BXT work at its appropriate earlier stage rather than adding it automatically later.

For this exercise, use a genuine Hα integration. A dual-band color integration requires a validated extraction workflow first; an arbitrary red-channel extraction is not guaranteed to be pure Hα.

On a duplicate of the linear calibrated RGB, use StarXTerminator with Generate Star Image enabled and Unscreen Stars off. Save the RGB starless and stars outputs separately. Follow the star-separation lesson for model, license, and extraction checks.

Remove stars from a duplicate of the linear Hα too. Preserve that extraction for inspection, but do not add its stars to the RGB restoration. Inspect both starless inputs for holes, leftover halos, removed compact emission, and mismatched edges.

Keep all three working branches aligned: RGB starless, Hα starless, and RGB stars. The main combination uses the first two. Preserve the inherited RGB-stars STF rather than independently auto-stretching that output to exaggerate residual background.

Locate NBRGBCombination in the installed scripts, commonly under Script → Utilities. It is a different tool from ChannelCombination, NarrowbandNormalization, and ImageBlend. The linked worked examples use older interfaces; consult the installed help before setting response bandwidths or scale controls.

Keep a starless RGB-only baseline. Give it and the HαRGB trial comparable test stretches on duplicates, with STF off afterward. Judge the same sky area and presentation brightness rather than accepting whichever auto-stretch looks more dramatic.

Inspect emission structures alongside reflection regions, dust, galaxy continuum, faint outskirts, and smooth background. Reject a trial that introduces a broad red wash, magenta patches, harsh noise, or loss of subtle broadband color.

If emission dominates, return to the original inputs and reduce its contribution. Do not compensate with aggressive global desaturation. The preferred balance can remain subtle; acquiring more Hα improves its information content, whereas increasing a scale factor does not.

Hα contains continuum as well as line emission, and combining datasets with different bandpasses involves assumptions. Do not interpret the display blend as continuum-subtracted flux or abundance measurements. Specialized continuum-subtraction workflows require separate calibration and validation.

Save the accepted HαRGB starless result while linear, then stretch a duplicate with MAS or GHS. Preserve the RGB-only comparison. Perform intended object contrast and palette refinements before star restoration.

Stretch the saved RGB stars gently using the star-separation lesson. Avoid clipping star cores or raising their residual background. Do not apply the nebula’s strong Hα enhancement or palette treatment to that stars branch.

In ImageBlend, choose the stretched HαRGB starless object as Base and stretched RGB stars as Blend, using Screen mode and neutral additional controls for the baseline. Create a new image and inspect the composite before changing star strength.

Compare stars with an RGB-only restoration using the same stars and blend settings. Screen interacts with the brightness and color of the object beneath each star, so it does not guarantee unchanged RGB pixel colors. Check bright nebular areas as well as dark sky, and keep the enhancement restrained where it causes an objectionable tint.

Stars look red or doubled: verify the Blend is RGB stars only, the Base is starless, and the geometry remains aligned. Do not add Hα stars or a second full RGB field.

Halos or holes appear: inspect the original starless branches and optical profiles. More blend opacity cannot reliably repair a flawed extraction.

The object becomes too red: compare the RGB-only baseline, filter setup, and contribution settings. Do not run broadband SPCC on the Hα-enhanced composite merely to force it back to the original broadband balance; calibrate the broadband source before the enhancement.

The result is noisy: review Hα exposure depth, extraction residuals, and stretch strength. A high contribution can make weak narrowband noise dominant.

Save prepared linear inputs, RGB stars, accepted linear HαRGB, stretched branches, and final composite. Record acquisition filters, bandwidth entries, scaling, extraction, stretches, and restoration settings. Label the result HαRGB or Hα-enhanced RGB so the added emission contribution is clear.

This is an original beginner practice sequence. Linked authors provide combination and extraction guidance; their examples do not establish current defaults. Installed script help governs control names and parameter meanings, and natural-looking RGB stars still require careful preparation and comparison.