Resize Image Dimensions

How to resize an image — 3 steps

1. 1

   Add your images

   Drop in one file or a batch. Resizing runs on your own machine, so a folder of camera-resolution photographs is limited by your processor rather than an upload.

2. 2

   Set the dimensions

   Enter the width or height you need. Keep the aspect ratio locked unless you have a specific reason not to — unlocking it stretches the picture.

3. 3

   Download the resized copies

   Your originals stay as they were, so you can produce several sizes from the same source without ever degrading the master.

Making an image smaller and making it larger are not the same operation

Reducing an image is straightforward and almost free in quality terms. There is more information in the source than the target needs, so the software averages groups of pixels down to fewer pixels, and the result is a clean, sharp, smaller picture. This is why downscaling is the most effective way to reduce file size: halving both dimensions removes three-quarters of the pixels, and the file shrinks roughly in proportion with no visible penalty.

Enlarging is a different proposition, because the detail you are asking for was never captured. Upscaling can only estimate the new pixels from their neighbours, and estimation produces softness — a 400-pixel image blown up to 1600 will look blurry and slightly plastic no matter what algorithm is used. It does not become a higher-quality picture; it becomes a bigger file containing the same amount of real information.

The practical consequence is to always work from the largest original you have. Keep the full-resolution file as your master and generate the sizes you need from it, rather than resizing a copy that has already been reduced. Once detail has been discarded it cannot be recovered, and a sequence of resizes compounds the loss each time.

Pixels, DPI, and choosing a target size

One persistent confusion is worth clearing up. For anything shown on a screen, only pixel dimensions matter — DPI is a number stored in the file that says nothing about how the image displays on a website or in an app. Changing an image from 72 to 300 DPI without changing its pixel dimensions does not improve it in any way; it changes a tag. DPI becomes meaningful only at the point of printing, where it describes how densely those pixels are laid onto paper.

For screens, size to the space the image will occupy, then roughly double it to look sharp on high-density displays. A photograph shown 600 pixels wide in a page layout wants a file around 1200 pixels; going much beyond that costs download time for detail nobody can see. For printing, work backwards from the physical size: around 300 pixels per inch is the standard for photographic quality, so a 6×4 inch print wants roughly 1800×1200 pixels.

Aspect ratio is the last thing to watch. Locked, it keeps the picture's proportions and you specify one dimension while the other follows. Unlocked, you can force any dimensions you like — and anything that does not match the original proportions will visibly stretch or squash, which is unmistakable on faces and on anything containing a circle. When an image has to fit an exact shape it was not shot in, crop it to that ratio first and then resize; the result keeps its proportions and simply shows less of the scene.

Related tools