How to Resize an Image Without Losing Quality (2026 Guide)
Pixel sizes, aspect ratios, DPI and the presets every form seems to ask for — explained in plain language.
Forget DPI — pixels are the only size a screen understands
Almost every form, app or website that asks for an image size is talking about pixels. DPI (dots per inch) only matters at the moment of printing. A 600×600 px photo is 2×2 inches at 300 DPI and 8.3×8.3 inches at 72 DPI — the file itself is identical, and on screen nobody can tell the difference.
The conversion is simple: pixels = inches × DPI. When a form says “2×2 inch photo”, it means 2 × 300 = 600 px per side.
Common print targets converted to pixels:
- US passport / visa photo (2×2 in at 300 DPI): 600×600 px
- ID photo 35×45 mm at 300 DPI: 413×531 px
- Chinese 1-inch photo (一寸): 295×413 px
- Chinese 2-inch photo (二寸): 413×579 px
- Instagram post: 1080×1080 px
- YouTube thumbnail: 1280×720 px
Aspect ratio is what stops faces from stretching
Aspect ratio is simply width divided by height. A 4000×3000 photo is 4:3. If you type a new width but leave the height untouched, the image gets squashed — which is why good resizers offer a lock ratio option that recalculates the other side for you.
There are three honest ways to reach a target size:
- Fit inside — the whole image is scaled down until it fits within the box. Nothing is cut, but one side may be smaller than the target. This is what you want for most uploads.
- Fill and crop — the image is scaled until it covers the box, and the overflow is trimmed. Use this when the form demands exact dimensions (1080×1080, 600×600) and slight cropping is acceptable. Keep faces near the centre.
- Stretch — width and height are forced independently, distorting the image. Only useful for deliberate effects or for resizing charts, never for people.
Downscaling is safe. Upscaling cannot add detail.
Making an image smaller throws information away, but the result looks sharp because the remaining pixels are real. Making an image larger can only invent pixels by interpolation — a 400 px photo blown up to 1600 px will look soft no matter which tool you use. If you need a big print, start from the biggest source file you have; do not enlarge a thumbnail.
One practical consequence: if a form allows “maximum 600×600”, never submit a smaller image upscaled to 600 px, and never submit a huge image that the system should shrink for you — do the shrinking yourself so you control the crop.
Choosing an output format
- JPEG — the right choice for photographs. Smallest files, no transparency support. Quality 85–92 is visually identical to the original in almost every case.
- PNG — the right choice for screenshots, logos, charts and anything with transparency. Files are larger but every pixel is exact.
- WebP — typically 25–35 % smaller than JPEG at the same quality level, and by now supported by every current browser. Fine for websites; a few old desktop programs still refuse it, so stick to JPEG when a portal is involved.
Two habits that save time later
- Keep the original file. Always keep the unresized original as your master copy; re-resizing an already resized image loses quality each time.
- Resize first, then compress. If you have both a pixel limit and a file size limit, shrinking the dimensions is by far the most effective way to hit the size target — the quality slider then only makes small corrections.
Do it without uploading anything
The Image Resizer on this site follows all of the advice above: it reads the EXIF orientation so rotated phone photos come out upright, offers fit / fill / stretch modes, includes ready-made presets for passport photos and social media, and never sends your file anywhere — the resizing happens in your own browser.