Why an image goes blurry when you make it bigger

An image goes blurry when enlarged because it has a fixed number of pixels, and making it bigger only spreads them out. Software fills the gaps by blending neighbouring colours, which softens edges. Up to about 150% the loss is usually acceptable; beyond that, find a larger original or a vector version instead.

What happens when you upscale

A 400 × 300 image contains 120,000 pixels. Resize it to 1600 × 1200 and the new image needs 1.92 million. The other 1.8 million have to be invented. Ordinary resizing invents them by averaging the pixels nearby. That is the best guess available, but an average of a black edge and a white background is grey, so every sharp edge becomes a soft gradient. Text and logos suffer most, because they are almost all edges.

The alternative — copying each pixel into a bigger block — keeps edges hard but produces visible staircases, which is what “pixelated” means. Either way, no detail is added that was not there.

How much can you enlarge?

There is no hard rule, but in practice:

• Up to about 150%: usually fine for photos viewed on screen. Resize Image warns when you go beyond this.
• 150–200%: noticeable softness on close inspection; often acceptable for a background or a small print.
• Beyond 200%: clearly blurry. Find another source.

Photos tolerate it better than screenshots, text, line art and logos, where any softness shows.

Why changing DPI does not fix it

DPI is only an instruction about how big to print the pixels. Changing 72 DPI to 300 DPI without changing the pixel count leaves the image identical; it just prints smaller. Raising the pixel count to reach 300 DPI at a larger print size is the same as enlarging, with the same blur. The DPI changer shows the pixel count a print really needs — for example, 1800 × 1200 for a 6 × 4 inch print at 300 DPI — so you can see whether the image you have is big enough.

Better options than enlarging

1. Find the original. A photo sent through a messaging app, or downloaded from social media, is usually a reduced copy. The original on the sender’s phone or camera is often four times larger in each direction.
2. Use the vector version of logos and icons. An SVG has no pixels; it is a set of shapes that can be drawn at any size. Ask for the SVG, then use SVG to PNG to draw it at exactly the size you need — it stays sharp at any width.
3. Display it smaller. On a web page or document, a 600-pixel image looks fine in a 600-pixel space. Use a layout that does not stretch it.
4. Recreate simple graphics. A small logo with text can often be retyped and redrawn faster than it can be rescued.
5. AI upscalers generate plausible detail rather than recovering real detail. They can look good for photos but may invent wrong text, faces or numbers, so do not use them for documents, ID photos or anything evidential.

Mistakes to avoid

• Enlarging, then compressing heavily: the blur and the artefacts stack.
• Enlarging in several small steps; it is no better than one step.
• Sharpening to hide the blur on text, which adds halos instead.
• Upscaling an ID or passport photo to meet a pixel minimum — it may be rejected as unclear.

Checklist

• Is there a larger original or an SVG?
• Enlargement kept under about 150%?
• Pixel count checked against the print size you need?

For the opposite problem — photos that were sharp but went soft once uploaded — see why photos look blurry after uploading.