
Use Enfuse's external weight-mask workflow when you need to revise which source contributes to a fused image. Reducing an input TIFF's alpha to gray is not an equivalent soft preference: Enfuse 4.2 treats input masks as binary eligibility, considering any nonzero value eligible. Saved weight maps are a different object and can be edited before a separate load-and-fuse pass.
Identify which mask you are changing
| Mask object | Documented role |
|---|---|
| Input image alpha | Zero excludes a source pixel; nonzero allows consideration |
| Generated weight map | Records fusion weights for that particular input |
| Loaded external weight map | Supplies weights without regenerating them from the images |
This differs from preserving partial transparency in an editor. The GIMP alpha-transfer guide describes such a transparency relationship; it should not be assumed to describe Enfuse's input-alpha interpretation.
If every input excludes the same location through black alpha, the output has a hole there. A gray input-alpha value does not instruct Enfuse to use that source at the same gray percentage.
Generate weights without requesting a final photograph
Sections 3.2 and 4 of the official 4.2 package documentation describe the two phases. In an otherwise empty working folder, an illustrative soft-mask pass is:
enfuse --save-masks frame-a.tif frame-b.tif
Without --output, Enfuse stops after saving masks. The default soft-mask naming template is softmask-%n.tif. Keep the emitted files associated with the corresponding source sequence. Adding --output would also request a fused image, so it changes what the first pass produces.
For a hard-mask workflow, the documentation also defines hardmask-%n.tif; do not mix those files with a soft-mask run and assume the names are interchangeable.
Edit the map without moving its coordinate system
Open a copy of the emitted grayscale maps in an appropriate image editor. Make the intended contribution change while retaining their geometry and offsets. Do not crop, resize, move the canvas origin or substitute a similarly named map from another stack.
The channel-dimension check is useful context for coordinate agreement, but external fusion also depends on the stored offsets specified by this workflow. Keeping the same visible subject is not sufficient.
Load the revised maps in a separate pass
enfuse --load-masks --output=controlled.tif frame-a.tif frame-b.tif
Use the same input sequence and the associated masks. The manual says loaded masks completely control this phase: Enfuse does not regenerate them. Changing a weighting option is therefore not a substitute for revising a loaded map. --save-masks and --load-masks are mutually exclusive, so they belong in separate invocations.
For a proposed acceptance check, compare the intended region with both originals and inspect the transition, holes and unchanged areas. Retain the unedited maps for comparison. This is a documentation-based procedure; the commands and edited maps were not executed here.
Related tasks
- GIMP Layer Mask: Alpha Channel or Transfer?
- Copy a Photoshop Mask Channel Without Overwriting the Destination Mask
Sources and scope
- Enfuse 4.2 official package manual: Enfuse 4.2 sections 3.2, 7.1 and --save-masks/--load-masks: input alpha zero/nonzero eligibility; two-pass grayscale weight files; unchanged geometry/offsets; no regeneration on load; stopping without --output; mutually exclusive flags and default softmask-%n.tif / hardmask-%n.tif.
Research checked on 2026-10-23 (Asia/Shanghai). The official SourceForge archive returned HTTP 200 and its bundled 4.2 HTML manual was extracted and read locally; this is not a claim that the public manual HTML endpoint was readable. Enfuse 4.2 sections 3.2, 7.1 and --save-masks/--load-masks: input alpha zero/nonzero eligibility; two-pass grayscale weight files; unchanged geometry/offsets; no regeneration on load; stopping without --output; mutually exclusive flags and default softmask-%n.tif / hardmask-%n.tif. Commands and comparison procedures are illustrative and were not 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: William Liu / Pexels, Pexels License. This is an illustrative source photograph, not an application screenshot or a generated test result.