Going from JPG to PNG is one of the few conversions people expect to improve an image, and it does not. It is worth being clear about that up front, because the disappointment is avoidable: PNG is a better format than JPEG in almost every technical sense, but converting an existing JPG cannot undo what JPEG already did to it.
Why the quality does not improve
JPEG throws away image data at the moment the file is saved. It splits the picture into 8×8 pixel blocks, discards fine detail inside each block, and usually stores colour at half resolution. That is why JPEG files are small, and why sharp edges — text, logos, screenshots — pick up the faint halos and blocky smudges people call artifacts.
PNG is lossless: it stores exactly the pixels it is given. So a PNG made from a JPG is a perfect, permanent copy of an already-damaged image. The artifacts are now baked in as real pixel values rather than compression side effects, and no later step can remove them. What you gain is that nothing gets worse from here on.
When this conversion is the right move
- You are about to edit the image repeatedly. Every JPEG save re-compresses and degrades a little more; PNG stops that generation loss dead.
- Something downstream demands PNG — a print shop, an app icon pipeline, a document template, a game engine asset folder.
- You need to add transparency. JPEG cannot store an alpha channel at all, so you must convert to PNG first, then cut out the background in an editor.
- You are layering text or graphics over the image and want the new elements to stay crisp.
Expect the file to get much bigger
A 2 MB photo commonly lands somewhere between 8 MB and 20 MB as a PNG. This is normal and not a fault in the conversion. PNG compresses by finding repeated patterns, and photographs — especially photographs full of JPEG noise — have very few. Flat graphics with large single-colour areas behave the other way round and can come out smaller than the JPG.
If size matters more than losslessness, WebP is usually the better answer: it is lossless when you need it and dramatically smaller than PNG on photographic content.
Things that catch people out
Metadata does not survive. EXIF fields — camera model, GPS coordinates, capture date, copyright — are dropped, because the conversion rebuilds the image from decoded pixels rather than rewrapping the original file. For most people that is a privacy win. If you need the location data or the shot settings, copy them out before converting.
Rotation is handled correctly. Phone cameras often store photos sideways with an EXIF orientation flag telling viewers to rotate them; browsers apply that flag when decoding, so your PNG comes out the right way up rather than on its side.
The converter runs entirely in this tab. Your JPG is read from your disk into memory, decoded, re-encoded as PNG and handed back — no server sees it. That also means very large files are limited by your device rather than by an upload cap.