Elysia Tools
Navigation
Workflow Playbook
Prepare an image for a fixed aspect ratio by preserving the source, adding controlled borders, reducing distracting noise lightly, and sharpening only at the intended delivery size.
Hubs
This workflow suits product tiles, social crops, screenshots, and other assets that must fit a known ratio without changing the main composition. Preserve the reviewed original first. Note the source dimensions, target dimensions, display size, and preferred border treatment before producing a candidate.
Use image-pad on a copy and calculate the border on each side from the target ratio. Transparent padding is useful when a later layout supplies the background; solid, mirrored, or repeated fills are alternatives when the surrounding design allows them. Inspect the corners and joins because a mathematically correct canvas can still look visually wrong.
Run image-noise-reduction only until distracting grain or specks recede. Bilateral filtering can be a reasonable choice for edges, while median or Gaussian treatment may suit different noise patterns. Watch labels, hair, skin, line art, and small UI text: once they become smeared, the pass is too strong.
When the file is already at its intended delivery dimensions, use image-sharpen with restrained settings. Review the original, padded intermediate, and final export at 100%, then inspect the image at the actual display size. Accept the candidate only when the border is clean and the final image looks clearer without halos or a promise that missing detail has been recovered.
For broader cleanup and resizing choices, continue with image-cleanup-resize-and-detail-enhancement. For geometry and canvas decisions that involve more than controlled padding, see image-geometry-transform-tools.
Workflow playbook
Duplicate the reviewed file before opening image-pad. Record the current dimensions, calculate the per-side border needed for the target ratio, and choose a fill mode without overwriting the original. Keep the original beside the padded candidate for every later comparison.
Run image-pad on the working copy with the planned top, right, bottom, and left values. Use transparent, solid, mirror, or repeat filling according to the brief, and confirm the subject stays in the same position and the output dimensions are exact.
Apply image-noise-reduction with the lightest useful algorithm and strength, then use image-sharpen only after the file is at its intended pixel size. Compare the original, intermediate, and final files at 100% and at the real display size; reduce the effect or reject the candidate when texture, halos, seams, or edge ringing become more visible.
Choose transparent padding when the next layout supplies its own background. Choose a solid, mirrored, or repeated treatment only when that edge behavior fits the composition; inspect the joins instead of assuming one fill works for every image.
Use a mild bilateral pass when edges, labels, or facial features must stay defined, median filtering for isolated specks, or a gentle Gaussian option for broad grain. Stop as soon as the distraction is reduced, because stronger smoothing can erase texture.
Apply image-sharpen after the image is already at the intended delivery dimensions. Prefer a restrained intensity and compare at 100% as well as the target display size; do not use a stronger setting to compensate for insufficient source resolution.