Image conversion with limits explained before you click

Before you convert

Choosing a format is a question about your content

There is no image format that is simply better than the others, and any tool that tells you otherwise is selling something. The right answer depends on what the image contains and where it is going, and the difference is not small. In our measured test, the same 1200×800 photographic scene was 826,866 bytes as PNG, 68,917 bytes as JPG at quality 85, and 26,774 bytes as WebP at quality 85. On a flat 1200×630 graphic with sharp text, the ranking reversed: PNG was 33,657 bytes while JPEG at quality 70 was 49,875 — larger, and with visibly degraded text.

The reason is structural. JPEG compresses by discarding fine high-frequency detail, which photographs barely miss and text cannot survive. PNG compresses by finding repeated patterns, which flat graphics have in abundance and photographic noise has almost none of. So photographs belong in JPG or WebP, screenshots and diagrams belong in PNG, and anything needing transparency rules JPEG out entirely — the format has no alpha channel, and a converter must flatten transparent pixels onto some opaque colour permanently.

What conversion cannot do

Four limits come up constantly, and knowing them in advance saves both time and images. Converting a JPG to PNG does not improve quality: it stores the already-decoded pixels losslessly, artifacts included, and the information JPEG discarded is not in the file for anything to recover. Converting a PNG to JPG destroys transparency permanently, flattening it onto a background colour that cannot be removed afterwards. A multi-page TIFF cannot become one PNG, because a PNG holds a single image — the conversion is a choice about which page to keep. And tracing a PNG to SVG approximates pixel regions with fitted paths; it does not recover the original design file, its fonts, or its layers.

Each of those is stated on the relevant converter page before you convert rather than discovered afterwards, because the mistakes they cause are usually irreversible. The most expensive one is not any single conversion — it is replacing an original with a derivative. Keep the file that came off the camera or scanner, and treat every prepared copy as disposable.

When an upload form rejects your image

Forms collapse many different failures into one unhelpful message, so work through the constraints in a fixed order rather than compressing repeatedly and hoping. Check the format first, since it is the most common cause and the cheapest to verify — and note that renaming a file changes its label, not its contents, so a validator reading the signature bytes will still catch it. Then check byte size, then pixel dimensions, then aspect ratio, then metadata, then anything application-specific.

One free diagnostic: if the form refuses the file instantly with no upload progress, the check ran in your browser and concerns the extension or file size. If it uploads fully and then fails, the server decoded it, and the problem is more likely dimensions, aspect ratio, or colour profile. The Upload-Ready Image Builder handles format, byte ceiling, and dimensions together, searching for the highest quality that fits your limit rather than overshooting downward — and it reports what it changed instead of silently altering dimensions you specified.

Read the full rejection-diagnosis guide →