
A series of aligned photographs taken with identical capture settings can be fused in Enfuse to reduce photon-shot noise through weighted averaging. You do not need to invent different exposure values for every frame. This repeated-capture task differs from recovering a broader exposure range, and it requires checking that the scene signal remains aligned.
Key Takeaways
- Independent repeated captures can provide averaging evidence that one file cannot.
- The documented default weights are an appropriate starting point for this use.
- Hard-mask focus selection gives up the same averaging benefit.
Choose a repeatable subject
For a hypothetical still-life photograph, keep the camera, subject, lighting, focus, white balance and exposure consistent across the series. Check the original frames for moving labels, changing reflections and other altered content.
Identical settings are not proof of identical geometry. Register the inputs when needed and inspect small edges. A moving person or flickering reflection is not random sensor noise; averaging it can create a different visible problem.
The HDR motion guide explains why scene movement needs its own decision. Calling a series a noise stack does not eliminate that requirement.
Start from the repeated-image workflow
Section 8.2 of the official Enfuse 4.2 package manual specifically discusses repeated images and noise reduction. It describes the default exposure weight of 1.0 and saturation weight of 0.2 as a suitable combination; other criterion weights default to zero.
For actual aligned repeats, an illustrative invocation is:
enfuse --output=repeated-average.tif repeat-a.tif repeat-b.tif repeat-c.tif
These are separate captures, not three filenames pointing to copies of one image. Duplicating an existing noisy file supplies no new independent observation of the subject.
The manual also suggests that comparing a run with --saturation-weight=0.0 can be worthwhile. That is a comparison option, not a universal guarantee that removing saturation weighting improves every subject.
Preserve the averaging mechanism
Enfuse computes a weighted average, so this is not necessarily a uniform arithmetic mean of the files. The criteria determine their contributions.
Do not add --hard-mask merely because it helped a focus stack. That route selects a best-weighted contribution at the finest scale instead of averaging fine detail, increasing the noise tradeoff. The purpose here is to combine repeated observations of the same signal.
Nor is increasing contrast afterward equivalent to noise reduction. Inspect texture and smooth areas separately rather than judging only whether edges look stronger.
Compare noise against retained detail
For a proposed review, inspect a smooth shadow, a fine label edge and a reflection in each original and the fused result at matching scale. Look for cleaner random variation without double edges, erased texture or an averaged moving highlight.
The noise-versus-plastic-skin guide concerns a different denoising decision, but its detail-preservation check remains useful. Keep the originals when the fused result fails that comparison. No repeated-capture series or quantitative noise reduction was measured here.
Related tasks
- How to Clean a Noisy Background Without Making Skin Look Plastic
- Why Does a Person Look Transparent After an HDR Photo Merge?
Sources and scope
- Enfuse 4.2 official package manual: Enfuse 4.2 section 8.2 Repetition: identically captured repeats reduce photon-shot noise by weighted averaging; defaults exposure 1.0 and saturation .2, other criteria zero; optional saturation-zero comparison. Sections 5.1.1-5.1.2 distinguish weighted averaging and hard-mask noise tradeoff.
Research checked on 2026-10-23 (Asia/Shanghai). Official SourceForge archive returned HTTP 200; bundled 4.2 HTML was extracted and read locally. Public standalone HTML was unavailable and is not reported as readable. Enfuse 4.2 section 8.2 Repetition: identically captured repeats reduce photon-shot noise by weighted averaging; defaults exposure 1.0 and saturation .2, other criteria zero; optional saturation-zero comparison. Sections 5.1.1-5.1.2 distinguish weighted averaging and hard-mask noise tradeoff. No software/image command, comparison or printing test was executed; the cover is decorative.
Prepared by Imagild Editorial, the company editorial team responsible for these guides. AI assisted research and drafting. This is documentation-based guidance with proposed examples, not a claim of hands-on testing or an independent product review. Report a correction.
Cover: Mariya Eskina / Pexels, Pexels License. This is an illustrative source photograph, not an application screenshot or a generated test result.