Why resizing beats compressing for oversized photos
Modern phone cameras produce 12 to 48 megapixel images, but screens display 2 to 4 megapixels. For any photo that will only ever be viewed on a screen, the extra pixels are pure freight: invisible on arrival, expensive to store and send. Downscaling to 1920 discards only what cannot be seen and cuts file size far more than quality reduction alone, without any compression artifacts. When both levers matter, resize first, then compress the result.
Format behavior during resize
Each image is re-encoded in its own format after scaling: JPG stays JPG (via the MozJPEG encoder), PNG stays PNG losslessly, WebP stays WebP. HEIC photos come back as JPG, since browsers can read HEIC but not write it, which conveniently also makes the resized copy universally shareable. EXIF orientation is applied before scaling, so sideways phone photos come out upright, and like every tool on this site the pixels never leave your machine, which you can verify by going offline after the page loads.
Common questions
What does 'longest side' mean exactly?+−
The larger of width and height. A 6000x4000 landscape photo resized to 1920 becomes 1920x1280; a 3000x4000 portrait becomes 1440x1920. One number handles both orientations in a batch, which is why professional export dialogs work the same way.
Which pixel value should I pick?+−
1920 covers full-screen viewing on almost every display and is the safe default for sharing. Use 2560 for high-DPI screens, 1200 for web content images, 800 for thumbnails and chat, and whatever a form explicitly demands when it names a dimension.
Does resizing reduce file size too?+−
Dramatically. Bytes scale with pixel count, so halving both dimensions roughly quarters the file. A 6000-pixel 8 MB photo at 1920 typically lands under 1 MB before any quality reduction. For hitting an exact kilobyte rule, combine with the compress-to-size tool.
Can this enlarge a small image?+−
No, deliberately. Upscaling cannot add detail; it only smears the pixels that exist and inflates the file. If an image is already smaller than your value it passes through at its original dimensions, re-encoded cleanly.