Crop image

Cropping is not resizing, and the difference matters

Cropping removes pixels from the edges. The pixels you keep are unchanged — same detail, same sharpness. Resizing changes all the pixels. If you need both, crop first and resize second, so the resize works from the largest possible source.

The box on screen is drawn at whatever size fits the page, but the crop is calculated in the original image's own pixels. The readout under the picture shows the real numbers, so a crop that has to be exactly 1,200 pixels wide can be dragged until it says so.

The ratio presets, and where each is asked for

• 1:1 — profile pictures nearly everywhere, and product photos on most marketplaces.
• 16:9 — video thumbnails, presentation slides, and most site hero images.
• 4:3 — what most cameras and phones shoot by default.
• 3:2 — 35mm proportions, and what standard photo prints are cut to.
• 9:16 and 3:4 — vertical video and phone-first layouts.

With a ratio selected, dragging a corner keeps it. Switching ratio adjusts the existing box rather than resetting it, so you can frame the shot first and then decide how it should be cut.

Using it on a phone

The box responds to touch the same way it responds to a mouse — drag the middle to move it, drag a corner to resize. There is no double-tap zoom to fight with, because the stage does not scroll.

A very large photograph is displayed scaled down to fit, which means the smallest adjustment you can make by dragging is a few source pixels rather than one. For pixel-exact work, crop on a larger screen.

What happens to the file

Nothing leaves your device. The image is decoded in this tab, drawn to a canvas, and the region you selected is copied out. Because the result goes through a canvas, it comes out without EXIF — no GPS, no camera serial, no colour profile. That is worth knowing in both directions: it is good for a picture you are about to publish, and it is a loss for colour-managed photography.

Cropping as a privacy tool

Cropping is the only reliable way to remove something from a photograph. Blurring, pixelating and drawing a black box over a face or a licence plate all leave the original pixels somewhere in the file in some cases, and blurred or pixelated regions have been reversed more than once — the transformation is often deterministic and the search space is small.

Cutting the region out entirely does not have that problem. The pixels are not in the output because they were never written to it.

If the thing you want gone is in the middle of the frame rather than at an edge, crop as tightly as you can and then paint over what remains in an image editor, saving as a new file rather than editing in place.

Crop first, then everything else

The order of operations matters more than people expect. Crop, then resize, then compress.

Cropping first means the resize is working from the largest available source for the region you actually want, so it has the most detail to average. Resizing first and cropping second throws away pixels you were about to keep and keeps pixels you were about to throw away.

Compressing last means the compressor is only being asked to encode the final pixels once. Every lossy encode adds artefacts, and encoding, resizing, then encoding again compounds them — the same reason a photograph forwarded through five chat apps ends up looking like it was faxed.

All three of those steps run in your browser here, so doing them in the right order costs nothing but the order you click in.