Resize Image

A digital image only has pixels. Its width and height in centimeters or inches appear when it is printed, and they depend on how many pixels are packed into each inch of paper. That number is the DPI (or PPI). At 300 DPI, a picture 600 pixels wide prints 2 inches wide; at 150 DPI the same file prints 4 inches wide and looks coarser. The table above shows the pixel counts behind a few common print widths at 300 DPI.

So the unit you choose depends on where the image is going. For screens, websites and social profiles, work in pixels and ignore DPI, because a monitor shows one image pixel per screen pixel whatever the file says. For anything printed, or for a form that states a size in mm or cm, choose the physical unit and set the DPI the printer or the form expects. 300 is the usual value for photo prints; some online forms ask for 200 instead.

Typing a size in centimeters does not change how sharp the picture can be printed. If the original is only a few hundred pixels wide and you ask for a large print at 300 DPI, the tool has to enlarge it to reach that pixel count, and the print will look soft. The honest fix is a bigger source photo. When none exists, a lower DPI such as 200 or 150 keeps the pixel count closer to the original, and the print is still acceptable for many uses viewed at arm's length.

Percentage is the quickest choice when you just want a lighter picture to send: 50% halves both sides, which leaves a quarter of the pixels. Values above 100% enlarge the image, up to 400%. Enlarging is fine for a pixel-art sprite or when a form refuses a picture for being too small, but it cannot invent detail; for the second case, Increase Image Size can often raise the file size without touching the dimensions at all.

The Fit setting only matters when both sides are set or when you choose an aspect ratio. With one side set, the other follows automatically and nothing is lost. With both set, the original shape rarely matches, and you decide what gives way.

Fit inside is the default because it never removes anything: the picture is scaled until it touches the frame on one side. The result can come out narrower or shorter than the numbers you typed, which matters if a form checks both dimensions exactly. Crop to fill gives exact dimensions and trims the edges; Crop position picks which part stays, and Smart (subject) looks for the busiest area, usually a face. If you need to choose the framing by hand, Crop Image lets you drag the crop box yourself.

Pad with color gives exact dimensions without cutting anything by adding bars in the Pad color, white unless you change it. It is the way to put a wide product shot into a square listing without losing the product's ends. Stretch forces the picture into the frame and distorts faces and circles, so it is only useful for textures and backgrounds where nobody will notice.

Batches follow one rule for every file. If the folder mixes portrait and landscape photos and you set a fixed width and height, the landscape ones and the portrait ones are treated differently by the same Fit setting: with Crop to fill the portrait photos lose their top and bottom, while the landscape ones lose their sides. Setting only the width, or only the longest side through a percentage, avoids that surprise, because each photo keeps its own shape.

Each card shows a size from the platform's own help page and the shape it draws. Social sites recompress whatever you upload, so matching their dimensions mostly saves you from an unexpected crop: an Instagram feed post, for example, may be anything from 1.91:1 landscape to 4:5 portrait according to Instagram Help, and pictures outside that range are cut to fit.

To use one of these sizes, type its width and height, choose Crop to fill and pick the part to keep. For shape-only targets such as a square avatar or a 9:16 story, switch Resize by to Aspect ratio instead. The list runs from 1:1 square through 4:5, 3:4 and 2:3 portrait, 9:16 vertical, and 16:9, 4:3, 3:2 and 21:9 landscape, and the picture keeps as many of its pixels as the new shape allows.

Profile pictures deserve a second look after resizing. Most sites show them inside a circle, so the corners of your square are hidden; keep faces and logos away from the edges. Banners and cover photos are the opposite case: they are shown at different widths on phones and desktops, and the outer parts may be cut off on one of them. Placing the important part in the middle of the frame keeps it visible everywhere.

Resizing does not have to change the format. Same as input keeps JPG as JPG and PNG as PNG; you can also save as JPG, PNG, WebP, AVIF or GIF. Pick PNG or WebP if the picture has a transparent background, and GIF or WebP to keep an animation. Quality applies to JPG, WebP and AVIF and starts at 85, which keeps photos looking like the original.

Reducing a picture is the direction that works well: many pixels are combined into fewer, and edges stay clean. If the result looks slightly soft at the new size, that is usually the focus of the original showing at a scale where you can now see the whole frame at once, and a light pass with the sharpen effect in Image Effects can help. Going the other way, from small to large, always costs some crispness, which is why Never enlarge exists for mixed batches where only some files are bigger than the target.

A smaller picture is lighter as well, but sometimes not light enough. Limit file size adds a KB or MB maximum after the resize, which saves a second trip through Compress Image. Strip metadata is on by default and removes GPS position and camera data, and Auto orient turns phone photos upright before the new size is applied, so width and height mean what you see on screen.

Before you upload a resized picture to a strict form, compare its width and height with the form's wording once more. Some portals ask for a minimum, some for a maximum, and some for an exact size, and the right Fit choice differs for each: an exact size needs crop or pad, while a maximum is satisfied by fit inside.