Why images get huge in the first place
Modern phone cameras capture 12–48 megapixels per shot, and many save in formats (or at qualities) tuned for editing flexibility rather than small file size. A single photo straight off an iPhone or a Samsung flagship routinely lands between 3MB and 8MB — far beyond what most upload forms, email providers, or websites want to accept.
The instinct is to drag a quality slider down until the number looks small enough. That works, but it's also the fastest way to introduce visible blotchy artifacts around edges and skin tones. There's a better order of operations.
The three levers that actually control file size
Compression isn't one setting — it's three, and they interact. Understanding the order fixes most 'compressed but ugly' results.
- Resolution — the raw pixel dimensions. Photos rarely need to be full camera resolution for screen or form use; downscaling before compressing often saves more size than quality reduction alone, with zero visible loss.
- Format — JPEG discards detail the eye barely notices (lossy); PNG keeps everything (lossless) and is far larger for photographs. Almost every 'compress this photo' task should output JPEG, not PNG.
- Quality percentage — JPEG's compression strength. 90–95% is visually identical to the source for most photos; 70–85% is where forms and email attachments usually land with no visible loss; below 60% is where blotchiness starts appearing in skies and skin.
The order that gets the best result
Resize first, then compress. A 4000×3000 photo squeezed down to a small file size at full resolution has to throw away far more detail per pixel than the same photo already resized to, say, 1600×1200 and then compressed. Cutting excess pixels before touching quality is the single biggest lever most people skip.
This is exactly how a target-size compressor should behave: rather than you guessing a quality number, the tool should search downward from high quality until the file just clears your limit — stopping at the highest quality that still fits, instead of overshooting into visible artifacts.
What 'losing quality' actually looks like
At healthy compression levels (75%+), quality loss is invisible on a phone or laptop screen. What you're watching for at aggressive compression is: blockiness in flat colored areas like skies or walls, color banding in gradients, and softened fine detail like hair or text. Faces and high-contrast edges are the most forgiving; smooth gradients are the least.
A quick decision guide
How to choose your approach
- Uploading to a form with a strict KB limit → use a target-size compressor and let it find the right quality automatically.
- Emailing a batch of photos → resize to roughly 1600px wide first, then compress each to under 1–2MB.
- Publishing on your own website → prefer WebP for the smallest file at equal quality, if your platform supports it.
- Archiving originals → don't compress at all; keep the source and only compress copies you actually share.