How to convert and resize images while preserving quality

7 min read

Resizing and converting look like simple operations, yet they are the most common source of blurry, stretched or oversized images. The good news is that nearly all quality problems come from a handful of avoidable mistakes. This guide walks through them in the order you will meet them.

Pixels, not centimetres

On screen, an image is defined by its size in pixels, for example 4000 × 3000. Most upload limits are written in pixels (“at least 1080 px wide”) or in file size (“under 2 MB”). These are different things: a smaller pixel size usually makes a smaller file, but the format and quality setting matter just as much.

DPI (dots per inch) only matters when printing. A 3000 × 2000 image printed at 300 DPI gives a 10 × 6.7 inch print; changing the DPI value alone does not change the pixels.

Keep the aspect ratio

The aspect ratio is the shape of the image — width divided by height. If you type a new width and a new height that do not match the original shape, faces and objects get squashed or stretched. Change one side and let the other be calculated, which is what Resize Image does when the ratio lock is on.

If a site demands an exact shape (for example a square profile picture), crop first, then resize. Resizing alone cannot change the shape without distortion.

Shrinking is safe, enlarging is limited

Reducing an image keeps it sharp because you are combining many existing pixels into fewer. Enlarging is the opposite: the software has to create pixels that never existed. Standard enlargement — the kind our Image Upscaler performs — smooths between existing pixels in several small steps to keep edges clean, but it cannot invent real detail. A 640 × 480 photo enlarged 4× will be bigger, not sharper.

Some services advertise “AI upscaling”, which generates plausible new detail. That detail is a guess, can alter faces or text, and usually requires uploading the image to a server. Our upscaler does not do that and says so.

The right order of operations

  1. Start from the best original you have — not a copy downloaded from a chat app, which is often already compressed.
  2. Crop if needed.
  3. Resize to the final pixel size.
  4. Convert to the final format last, and save once.

Saving a JPG, reopening it, editing and saving again repeats the compression each time. Doing the format conversion at the very end avoids stacking losses.

Choosing a JPG quality setting

JPG quality is a trade-off between size and detail. As a rough guide:

  • 90–95%: visually identical to the original for nearly all photos; good default for printing or archiving.
  • 75–85%: much smaller files with differences hard to see on a phone; good for email and web.
  • Below 70%: blocky artefacts appear around edges and in skies.

When converting with PNG to JPG, transparent areas are filled with a colour because JPG cannot store transparency — check the result if your image had a transparent background.

Worked example: meeting an upload limit

Suppose a job portal wants a photo under 1 MB and your phone photo is 4032 × 3024 and 3.5 MB. Resize the long side to 1600 px (the other side becomes 1200 px automatically), and keep JPG as the output. The result is usually well under 1 MB and still looks sharp on screen.

Common mistakes

  • Typing both width and height and distorting the image.
  • Enlarging a small image and expecting more detail.
  • Saving screenshots or text as low-quality JPG — use PNG instead.
  • Repeating edit-and-save cycles on the same JPG.
  • Confusing DPI with pixel size when a form asks for “high resolution”.

Tools mentioned in this guide

More guides