Pixels, file size and quality: three different controls

Dimensions are not a file-size promise

Dimensions describe the pixel grid. A 2000 × 2000 image contains four million pixels regardless of whether its encoded file occupies 300 KB or 3 MB. File size also depends on the content, format and encoder. A detailed fabric photograph can need more bytes than a plain background at the same quality setting.

The quality percentage is an encoder input, not a visual quality score. An 85% setting does not guarantee a specific number of kilobytes or identical results in every browser. Dama displays the actual output file size after preparation; it does not promise a hard KB target.

Choose a format with the image in mind

JPG is a practical starting point for photographs when the destination accepts it. WebP may provide smaller files, but confirm destination support before choosing it. PNG is lossless at the encoding stage and can be useful for sharp graphics, although photo files can be much larger.

The PNG quality slider is disabled because the canvas PNG encoder ignores that parameter. All Dama exports are composited onto the selected background. This includes transparent PNG and WebP inputs; the tool currently does not preserve transparent backgrounds.

A controlled comparison

Keep the source and dimensions unchanged while comparing compression. Otherwise you cannot tell which control caused a visible change.

1. Export a representative photo at the starting quality of 85%.
2. Inspect text, fine texture, diagonals and high-contrast edges at normal viewing size and at 100%.
3. Export again at 75%, from the original file rather than the previous export.
4. Compare both actual file sizes. If edge artifacts or smearing become visible, use the higher quality version.
5. Only then apply the chosen settings to the full batch; check more than one result.

What re-encoding can change

Re-encoding a compressed source can compound visible damage. Keep originals separately and start each new variation from them. Enlargement adds pixels, not detail; it cannot recover a blurry label.

Canvas creates a new encoded image rather than copying the original metadata container. Original EXIF/IPTC metadata and color-profile behavior are not preserved as an archival workflow. Browser decoding, color management and interpolation can differ, so compare important product colors against the original.

If a browser cannot encode the chosen format and returns PNG, the download uses the actual .png extension and displays a fallback notice. Do not rename a .png file to .webp to make it acceptable to a platform.