Brand Image Tools: Padding, Dark Presentation and OG

Actual local example: one speck changes the crop

Two owned 320 × 160 inputs differ by exactly one painted pixel at (280, 20). Both contain the same blue badge and white i. The current padding function was executed directly in Chromium on each; these are its actual 1400-square PNG outputs.

The fixed 980-pixel content area fits the entire detected region. For these bounds, the main badge’s layout width is 980 pixels without the speck versus 160 × 980 ÷ 221, about 710 pixels, with it, before raster rounding. Remove a mark only when you know it is unwanted; an intentional dot deserves preservation. The function does not remove painted white backgrounds or decide the artwork’s meaning. This direct function run is not a full UI, platform-mask or delivery qualification.

Actual local mask checks at 64 pixels

These are actual wide and tall padding outputs resized to 64 square and clipped by a circle or a rounded square with 12-pixel corner radius. The broad badge remains inside these explicit masks, while tiny detail becomes less readable. They are local crop checks, not screenshots from a social platform; inspect the receiving app’s real mask and display size.

Understand the padded canvas

The tool detects content using alpha when present, otherwise non-white pixels. It crops to those bounds and contains the region within 70% of a transparent 1400 × 1400 canvas, preserving proportions. Transparent outer padding does not erase a white rectangle painted inside the source crop.

Worked margins for a wide mark

Illustrative arithmetic: detected content 400 × 100 has a 4:1 ratio. It fits as 980 × 245 on this canvas, leaving 210 pixels on each side and 577.5 pixels above and below before raster sampling. A square mark instead has 210-pixel margins on all four sides. These are layout calculations, not measured tool outputs.

Check detection before accepting the crop

White artwork on a white opaque background may be missed by the non-white test. Faint alpha or isolated specks can also change the detected bounds. Compare the crop with the source, including accents, fine lettering and intentional whitespace. If detection is wrong, prepare a deliberate crop in an editor; do not keep processing the same damaged output.

For a 1024-pixel white profile image

The former avatar preset used a white 1024-square canvas with 78% containment. This tool uses a transparent 1400-square canvas with 70% containment, so it does not reproduce the same pixels. If your selected profile placement needs 1024 × 1024, resize the inspected square as a separate step. If it requires white, apply a deliberate white matte in an editor; the resizer’s JPEG white-matte mode is an option only when JPEG is accepted. For an opaque PNG, save the matted image as PNG in an appropriate editor.

Preview the final circle or rounded mask in the receiving application. Keep the approved mark inside the actual crop and inspect tiny lettering at displayed size; a larger upload alone cannot make it readable.