How to resize an image without losing quality

Shrinking versus enlarging: the single biggest factor

Every pixel in a digital photo was captured once, at the camera's original resolution, and that's the only real detail that exists to work with. Shrinking a photo down (say from 4000 pixels wide to 1200) is a safe operation: there's more source detail than the smaller canvas needs, so the resize can select the best representative color for each new, larger pixel and the result looks sharp.

Enlarging works the opposite way. Going from 1200 pixels wide up to 4000 means the resize has to invent roughly ten times as many pixels as it started with, and it does this by interpolating (blending) between the pixels that actually exist. There is no way around this with an ordinary resize, because the detail that would fill those new pixels correctly was simply never captured. A modest enlargement, up to perhaps 25 to 30 percent larger, is usually still presentable; going much further starts to look visibly soft no matter which resizing tool or algorithm is used. Try the image resizer on a photo of your own to see the difference between shrinking and enlarging directly.

Keep the aspect ratio, unless you have a specific reason not to

Aspect ratio is the relationship between a photo's width and height, commonly written as two numbers like 4:3 or 16:9. When you resize with the aspect ratio locked, both dimensions scale by the same factor, so a person's face or a building's edges keep their correct proportions. Unlock it, and you can set width and height independently, but doing so stretches or squashes the entire image to fit: circles become ovals, faces widen or narrow, and the distortion is usually obvious at a glance.

There's a common mix-up worth calling out directly: resizing with the aspect ratio unlocked is not the same thing as cropping. Cropping removes part of the photo and keeps the remainder at its original proportions; a mismatched resize keeps every part of the photo but distorts all of it to fit the new shape. If a target size requires a different aspect ratio than your source photo (a square Instagram post cut from a widescreen photo, for example), cropping to the right ratio first, then resizing, gives a clean result; forcing the wrong ratio through a straight resize does not.

Choosing a file format for the resized result

PNG is a lossless format: once a photo is decoded, PNG encoding doesn't discard any further pixel information, and it supports transparency. That makes it the right choice for screenshots, logos, icons, and anything with sharp text or hard edges, where JPG's compression tends to introduce visible fuzzing around high-contrast boundaries. The cost is file size: an ordinary photo saved as PNG is typically several times larger than the same photo saved as JPG.

JPG applies lossy compression on top of the resize, controlled by a quality setting (usually 1 to 100). At quality 90 or above, the difference from an uncompressed original is very hard to spot with the naked eye, while the file size drops substantially compared to PNG. Lower settings (60 to 80) trade a bit more visible softness, especially in fine detail and sharp edges, for a smaller file still, useful when a strict size limit matters more than maximum fidelity. JPG has no transparency channel: any transparent area in a source PNG becomes a solid color (usually white) when saved as JPG, so keep transparent graphics in PNG.

DPI and print: pixels alone aren't enough on paper

A screen only cares about pixel count: a 1200 by 1800 pixel photo displays at whatever size the screen's own pixel density allows. Paper is different, because a printed inch has to be filled with a specific number of dots (commonly measured as DPI, dots per inch, though for a digital source it really means pixels per inch). The same pixel count printed at a higher DPI produces a smaller, sharper physical print; printed at a lower DPI, it produces a larger but softer looking print, because the same number of pixels is now stretched across more paper.

A widely used target for a sharp photo or document print is 300 DPI. To resize an image for a specific print size, multiply the physical dimensions (in inches) by the target DPI: a 4 by 6 inch print at 300 DPI needs 1200 by 1800 pixels; an 8 by 10 inch print at 300 DPI needs 2400 by 3000 pixels. Printing a smaller pixel count at a large physical size (stretching a low-resolution photo to fill a poster, for instance) runs into the same softness problem as enlarging on screen, for exactly the same underlying reason: the pixels being stretched across more space were never captured at that density.

A short, repeatable process

This same sequence applies whether the destination is a social media post, an email attachment, a printed photo, or a web page image; only the specific numbers (target size, DPI, format) change. Every step above runs locally in your browser; see the privacy policy for what, if anything, this site collects.

1. Start from the highest-resolution version of the photo you have access to, since shrinking is safe but enlarging is not.
2. Decide your target dimensions: pick a known preset (see the size presets page) or calculate a print size from physical dimensions times DPI.
3. Keep aspect ratio locked unless you've deliberately cropped first to match a different target ratio.
4. Choose PNG for graphics, screenshots, or anything needing transparency; choose JPG (quality 85 to 95) for ordinary photos where file size matters.
5. Preview the result before finalizing: check for stretching, obvious softness, or an unwanted white background where transparency should have been kept.

Common mistakes that make a resized photo look worse than it needs to

A short list of avoidable errors accounts for most complaints about a resize looking "off" (the size presets page lists exact, pre-checked dimensions if you would rather start from a known-good target than type your own):

• Enlarging a small photo well beyond its original size and expecting sharp detail that was never captured.
• Resizing with aspect ratio unlocked to force a photo into an unrelated shape, producing visible stretching.
• Saving a photo with transparency as JPG, which silently fills transparent areas with a solid background color.
• Choosing a very low JPG quality to save a small amount of file size, introducing visible blockiness for little benefit.
• Resizing repeatedly and re-saving as JPG each time, which compounds compression loss with every round trip; keep an original, uncompressed copy and resize from that each time instead of resizing an already-resized file.