How to Resize an Image: The Complete Guide

Resizing an image is one of the most common tasks in digital work, and one of the most misunderstood. Resizing incorrectly can make images blurry, distorted, or far larger than they need to be. Here is how to do it properly for every common use case.

Pixels, Resolution, and What They Mean

A digital image is made up of pixels: tiny squares of colour arranged in a grid. The dimensions of that grid, for example 1920 by 1080, describe how many pixels wide and tall the image is. A 1920 by 1080 image contains about 2 million pixels total.

Resolution in pixels per inch (PPI or DPI) describes how densely those pixels are packed when printed. A 300 DPI image at 4 inches wide contains 1200 pixels across that 4-inch width. The same 1200 pixels printed at 72 DPI would be about 16 inches wide.

On a screen, DPI/PPI mostly does not matter. What matters is the pixel dimensions. A 1920 by 1080 image displayed on a 1920 by 1080 monitor will fill the screen exactly regardless of what DPI value is stored in the file metadata. For print, DPI is critical.

Making an Image Smaller (Downscaling)

Reducing an image's dimensions is called downscaling, and it is a lossless operation in terms of image quality: you are simply keeping fewer pixels from the original. The image will look identical to the original, just smaller.

Downscaling is used for:

• Reducing file size before uploading to a website or sending by email
• Fitting an image within a specific display dimension
• Creating thumbnails or preview images
• Reducing the resolution of a photo taken at too high a resolution for its intended use

Most image resizing tools use one of several resampling algorithms to decide which pixels to keep and how to average adjacent pixels. Bicubic and Lanczos resampling produce the sharpest results for downscaling. Nearest-neighbour resampling preserves hard edges, which is useful for pixel art but produces jagged results for photographs.

Making an Image Larger (Upscaling)

Increasing an image's dimensions is called upscaling, and it is where things get complicated. When you make an image larger, you are asking the software to invent pixels that did not exist in the original. There is no way to recover detail that was not captured, so upscaling always results in some quality loss.

Traditional upscaling algorithms (bicubic, bilinear) produce blurry results when scaling up significantly. AI-based upscaling tools use machine learning to predict what the additional pixels should look like based on patterns learned from thousands of images. The results are dramatically better than traditional methods, often producing usable results even when doubling or quadrupling an image's dimensions.

If you need to upscale an image significantly for print, consider using an AI upscaling tool rather than a standard resize tool. The difference in output quality is substantial.

Maintaining Aspect Ratio

The aspect ratio is the relationship between an image's width and height. A 1920 by 1080 image has a 16:9 aspect ratio. A 1000 by 1000 image has a 1:1 (square) aspect ratio.

When resizing, you should almost always maintain the aspect ratio unless you specifically need a different shape. Resizing without maintaining aspect ratio stretches or squashes the image, distorting everything in it. Faces become wider or narrower. Circles become ovals. Buildings lean.

Most image resizing tools lock the aspect ratio by default and adjust both dimensions proportionally when you change one. If you need to fit an image into a specific width and height that has a different aspect ratio, you have two choices: crop the image to the new ratio before resizing, or use "fit" mode which adds padding (usually white or transparent) to fill the remaining space.

Common Size Requirements

Resize an Image Free

Resize any image to exact pixel dimensions or by percentage, directly in your browser. Your image is never uploaded to a server.