Preserve Bright-Core Detail with HDRComposition

Long exposures can record faint surroundings while saturating a bright core. Short exposures can retain that core’s intensity differences. HDRComposition combines compatible linear integrations from those exposure ranges into a new linear master.

This is different from HDRMultiscaleTransform. HDRComposition uses additional captured measurements; HDRMT later compresses brightness differences to help present detail already present. Neither can reconstruct a core clipped in every input.

Use this optional workflow when the data supports it, not simply because the object is bright. Begin with two exposure groups from the same camera setup and filter. A single master with a merely aggressive display stretch may need a gentler stretch instead.

Calibrate and integrate short and long exposures separately in WBPP, keeping appropriate matching calibration files. Preserve filter and exposure grouping rather than widening the exposure tolerance just to make one master. Include an intermediate group only when it provides useful overlap between the ranges.

For a first monochrome exercise, use one filter at a time. Do not combine an Hα short master with an O III long master as an exposure bracket. For broadband one-shot-color data, use compatible linear RGB integrations with the same acquisition filter configuration.

Inspect the shortest group for retained core structure and acceptable signal. A single very noisy short exposure is a weak basis for replacing a large bright region. Compare acquisition settings, focus, and star profiles across groups before processing.

Keep all inputs linear with stars present. Avoid independent stretches, creative color changes, star removal, or attempts to match every histogram before composition. Save the original exposure-group integrations.

Choose one common registration reference and align the exposure-group masters where needed. Check stars at the center and corners. Then apply the same preparation crop to every input, keeping only shared usable coverage.

Confirm matching dimensions, pixel positions, channel count, and orientation. Similar framing and equal dimensions are not enough if stars are offset. Different focus or halo profiles can cause seams even with accurate centroids.

Save the prepared files with clear names such as Target_Ha_short_linear.xisf and Target_Ha_long_linear.xisf. This exercise uses saved files in HDRComposition’s input list, not a blend of stretched open views.

Find HDRComposition through All Processes, usually under ImageIntegration. Reset a reused instance and consult its installed help; the references include older interfaces.

Examine the masks alongside the aligned input masters. Confirm which mask values indicate contributions in your version, then check that replacement regions cover the intended bright areas and their transitions. A mask is a diagnostic, not proof that the blend is correct.

Use STF to inspect the linear HDR result. Automatic display stretches may differ substantially between the source and composite because the dynamic range has changed. A darker first display does not establish that faint data was lost; inspect corresponding regions and readouts, then make comparable test stretches on duplicates.

Check retained core structure, intermediate-brightness regions, faint outskirts, and bright stars at 1:1. Look for seams, dark rings, halos, color shifts, abrupt noise changes, and doubled stars. A pleasing core alone is insufficient if the transition damages the surrounding object.

High-precision output reduces numerical limitations; it does not recover missing measurements or make mismatched inputs compatible. Keep a long-exposure-only baseline for comparison.

If the replacement misses a clipped region or reaches too far into weak short-exposure signal, review the threshold and actual input values. If boundaries are visible, inspect registration and profiles before modifying mask growth or smoothness.

Repeat each trial from the same prepared input files. Change one parameter, save the output and masks, and compare the same regions. Excessive smoothing can spread an unsuitable transition rather than fix its cause.

If exposure evaluation fails or produces an implausible result, check linearity, compatible filter responses, shared unsaturated signal, clipping, background problems, and input geometry. Do not make the files look equally bright with independent permanent stretches to force the process to accept them.

Strongly different backgrounds may require a carefully controlled linear preparation strategy. Reassess those inputs on separate copies and document the changes instead of accumulating unrelated corrections. If the bracket lacks useful signal overlap, additional acquisition may be the better solution.

For monochrome data, repeat the validated procedure separately for any other filter with its own exposure bracket. Preserve a common grid across all outputs. A filter with an adequate single exposure range does not require a fabricated short/long pair.

Save the accepted HDR master while still linear. Continue with appropriate gradient correction, channel matching and combination, and broadband color calibration as applicable. Recheck astrometry and regenerate SPFC metadata on the actual current master before MGC; inherited input metadata is not proof of validity for the composite.

Then stretch a duplicate with MAS or GHS, keeping the HDR linear checkpoint intact. HDRMT can be tested later if the presentation still needs core compression, but it is not a compulsory second HDR step.

Record input paths, filters, exposure groups, registration reference, common crop, exposure-evaluation results, composition parameters, and why the chosen transition was accepted.

The core remains featureless: inspect the shortest source for saturation and actual structure. Composition cannot replace clipped pixels with information that was never captured.

Rings or seams appear: check star alignment, focus and halo differences, transition masks, and where the short master becomes noisy. Adjusting mask controls cannot repair every acquisition mismatch.

The result looks unexpectedly dark: inspect its linear state and STF before altering pixels. Use a deliberate test stretch to judge the available dynamic range.

Color changes near the core: verify compatible RGB preparation, source clipping by channel, and transition quality. Do not fix a spatial discontinuity by globally increasing saturation.

No useful improvement: compare against a carefully stretched long-only master. Keep the simpler result when the bracket does not add reliable core information.

This is an original beginner sequence. The official demonstration explains process controls, while the worked examples provide composition concepts. Installed documentation governs current behavior; no single threshold or exposure ratio fits every target.