Start with the destination size
A 4000-pixel photo displayed as a small website card contains more pixels than the screen can show. Resize it first to the largest size where it will be displayed. Then compress the smaller image. This usually produces a cleaner result than squeezing a full-resolution file to an extreme target.
Reduce noise before reducing quality
Fine details are expensive to store. Cropping unused background, simplifying busy areas, and avoiding unnecessary sharpening all reduce file size without making the main subject look soft.
- Crop the image to its intended aspect ratio.
- Resize to a sensible display width.
- Use a moderate quality setting and inspect the result.
- Prefer WebP for modern web delivery.
Inspect the image at display size
Zooming to 400% is not a realistic way to judge a web photo. View the downloaded image at the size visitors will see it. Pay attention to faces, text, product labels, and smooth gradients, where compression artifacts are easiest to spot.
Keep the original
Compression creates a delivery copy. Keep the original camera or design file if you may need to print, re-edit, or export to another platform later. Repeatedly compressing the latest copy can accumulate quality loss.
Frequently asked questions
Is lossless compression always better?
Lossless compression preserves data, but it cannot always reach a small upload limit. A controlled lossy export often looks better than forcing a huge file into an impractical target.
Can I compress an image repeatedly?
Avoid it. Start from the original each time so previous compression artifacts do not accumulate.
What quality setting should I use?
Use a preview and start around 80–90%. The visible threshold depends on image content and display size.
Put this guide into practice: process the image locally without creating an account or uploading it.
Try target-size compression