Most image problems are not dramatic. They are small, repeated habits that quietly degrade quality, bloat files, or leak information you did not mean to share. Here are the four I see most often, and the simple fix for each.
Mistake 1: Over-compressing
Compression is your friend right up until it is not. Push the quality slider too low and you get the telltale signs: blocky squares around edges, muddy skies with visible banding, and a smeared, "underwater" look on fine detail. Worse, the damage is permanent — you cannot recover detail that has been thrown away.
The fix is to stop treating compression as a race to the smallest number. For photographs, quality around 75–85% is the sweet spot: big size savings with no visible loss. Compress once from the original, not repeatedly, because each re-save of a JPEG compounds the damage. If you need to edit later, keep a lightly compressed or original master and export compressed copies from it.
The Image Compressor lets you dial in quality and see the result, and it will never hand you a file larger than the original, so you can find the point where the file is small but still looks right.
Mistake 2: Using the wrong format
The classic error is saving photographs as PNG. PNG is lossless and brilliant for logos, icons, screenshots of text, and anything needing transparency — but for a photo it can be three to five times larger than a JPEG or WebP with no visible benefit. The reverse mistake is saving a sharp-edged logo as JPEG, which surrounds the edges with fuzzy compression artifacts.
A quick guide:
| Content | Best format |
|---|---|
| Photograph | WebP, or JPEG for maximum compatibility |
| Logo, icon, flat graphic | PNG or SVG |
| Anything needing transparency | PNG or WebP |
| Smallest possible photo | AVIF (falls back to WebP) |
The Image Converter makes switching trivial, and when you target AVIF it falls back to WebP automatically for files where AVIF is not the right fit.
Mistake 3: Upscaling a small image
You cannot add detail that was never captured. Taking a 400-pixel image and stretching it to 1600 pixels does not create sharpness — it just makes the existing blur bigger and softer. Basic upscaling invents nothing; it only interpolates between pixels you already have.
If an image is too small for its purpose, the real fixes are: find a larger original, re-export from the source at the size you need, or reshoot. Downscaling (making an image smaller) is safe and often improves apparent sharpness; upscaling is where people get disappointed. When in doubt, resize down, not up.
Mistake 4: Uploading private photos to unknown tools
This one is about privacy rather than quality, and it is the most consequential. Many online editors upload your image to their servers to process it. For a passport scan, a medical document, a screenshot with account details, or a photo of your family, that means handing a copy to a company you know nothing about.
There are two parts to protecting yourself.
Use tools that process locally. Every tool on this site runs entirely in your browser using Canvas and WebAssembly. Your images are never uploaded — they never leave your device — and after your first visit the tools keep working offline, which is proof in itself that nothing is being sent anywhere.
Strip metadata before sharing. Photos carry hidden EXIF data: the exact GPS location where a picture was taken, the date and time, and the device model. Post a photo online with that intact and you may be broadcasting your home address. Use the EXIF Viewer to see what a file is carrying, then strip it before you share.
The pattern behind all four
Each mistake comes from doing the easy thing without checking the result. Compress and look at it. Match the format to the content. Resize down, never invent pixels up. And before a private image leaves your hands, know what is embedded in it and where it is being processed. None of these takes more than a few extra seconds, and each one saves you from a problem that is hard or impossible to undo.