How Much Can You Compress an Image?
Most photographs can be compressed to roughly 10 to 20 percent of their original file size before visible quality loss appears, though the exact point depends on the image content and compression format. In our test, a 2.4 megabyte photo held up cleanly down to about 380kb, with noticeable blockiness only appearing once the file dropped below 150kb. Busy, detailed photos tolerate more compression than smooth gradients like skies, which show artifacts much sooner.
Testing one photo across six quality levels
To answer how much can you compress an image before it starts looking wrong, we took a single 2.4 megabyte landscape photo straight from a phone camera and ran it through compression at six different quality settings, from a light touch down to an aggressive squeeze. At each step we measured the resulting file size and looked closely at areas of the image most likely to reveal artifacts: skies, skin tones, and fine texture like foliage or fabric.
The pattern that emerged was consistent with what most compression research already shows. The first 60 to 70 percent of size reduction is essentially free, meaning the human eye cannot reliably tell the compressed version from the original at normal viewing distance. Past that point, quality drops off quickly, and the last 10 percent of extra size savings costs a disproportionate amount of visible sharpness.
Where the artifacts actually start showing up
Blockiness first appears in smooth gradient areas, particularly skies and out of focus backgrounds, because compression algorithms group pixels into blocks and smooth areas make the block boundaries obvious once compression gets aggressive. Detailed, high texture areas like grass, gravel, or patterned fabric can absorb far more compression before a viewer notices, since the existing visual noise in the image hides the algorithm’s own noise.
How much can you compress an image before it looks bad on screen versus in print
Screen viewing is far more forgiving than print. An image compressed to the point of mild blockiness on a large monitor at full zoom might look completely fine on a phone screen or as a small thumbnail, since the artifacts occupy fewer physical pixels at smaller display sizes. If your image is destined for print, keep compression conservative, but if it is going on a web page as a thumbnail or card image, you can usually push compression noticeably further than feels comfortable at first.
File size versus visible quality at each level
| Quality setting | File size | Visible result |
|---|---|---|
| Original | 2.4 MB | Full detail, no artifacts |
| 90 percent quality | 780 kb | No visible difference |
| 75 percent quality | 380 kb | No visible difference |
| 60 percent quality | 210 kb | Very slight softness in skies |
| 45 percent quality | 150 kb | Mild blockiness visible on close inspection |
| 25 percent quality | 95 kb | Clearly visible artifacts and color banding |
Format also changes how much compression you can get away with
The test above used a standard photographic format, but the answer to how much can you compress an image shifts depending on which format you start from. Formats built for photographs handle gradual color transitions efficiently and tolerate aggressive compression before showing blocking. Formats built for flat graphics, logos, and screenshots behave differently, since they were designed to preserve sharp edges and exact colors rather than photographic detail, and compressing them the same way often introduces color banding much sooner than it would in a photograph.
This is part of why choosing the right format before compressing matters as much as the compression setting itself. A screenshot saved and compressed as a photographic format will often look worse and end up larger than the same screenshot saved in a format meant for flat color graphics, even at similar compression settings. Matching the format to the content type is a free win that costs nothing beyond picking the correct export option.
A short checklist for choosing a compression level
- For hero banners and large images, stay above 75 percent quality to protect fine detail.
- For thumbnails, card images, and small previews, 45 to 60 percent quality is usually indistinguishable.
- Always compress a copy, never overwrite your original source file.
- Check compressed images at their actual display size, not zoomed in, since that is how visitors will see them.
Find your own breaking point in seconds
The Image Compressor lets you test multiple quality levels on your own photo directly in your browser, showing the resulting file size instantly so you can find the exact setting where your image stays sharp without carrying unnecessary weight.
Open the Image CompressorResize before you compress for the best result
Oversized images compress worse than correctly sized ones, since the compressor is fighting pixels you never needed in the first place. Running an image through the Image Resizer first, then compressing it, typically produces a smaller final file at higher visible quality than compressing the oversized original alone.
Open the Image ResizerRelated tools for finishing the job
If your compressed image needs to change format, for example moving a PNG screenshot to a JPEG for smaller delivery, the Image Converter handles that conversion cleanly before you compress it further. Once you have found a compression level you trust, browse the rest of the image and design tools for related editing utilities, and check the ConvertNow blog for more testing based breakdowns like this one.
Key takeaway
You can typically shrink a photo to a fifth or even a tenth of its original size before quality loss becomes noticeable, but the exact threshold depends heavily on the image itself, so testing beats guessing every time. Try the Image Compressor on your own photo and compare a few settings side by side before publishing.
The short version
Normal text needs at least 4.5 to 1 contrast, large text and UI components need at least 3 to 1, and both numbers come straight from WCAG 2.1 AA. Check every pairing at its actual size and weight, watch for placeholder text, disabled states, and gradients as common failure points, and pair color with text or icons rather than relying on color alone. Run any pairing you are unsure about through the Color Contrast Checker before you ship.
FAQ: How Much Can You Compress an Image?
How much can you compress an image before quality suffers?
Most photographs can be compressed to roughly 10 to 20 percent of their original file size before visible quality loss appears, though smooth gradient areas like skies show artifacts sooner than detailed, textured areas.
Why do skies show compression artifacts before other parts of a photo?
Compression algorithms group pixels into blocks, and smooth gradient areas make those block boundaries obvious once compression becomes aggressive, while detailed textures hide the same artifacts within existing visual noise.
Is lossy compression bad for a website?
Lossy compression is generally fine for websites since screens display far fewer effective pixels than print, and most visitors view images at a size and distance where moderate compression is invisible to the eye.
Should I resize an image before or after compressing it?
Resize first, then compress, since compressing an oversized image wastes effort on pixels that will be discarded anyway and produces a larger final file than resizing followed by compression.
What file size should a web image target?
Hero images typically target under 200kb, and thumbnails or card images typically target well under 100kb, though the right number depends on the page’s overall performance budget.
Does image content affect how much compression it can handle?
Yes, busy and detailed images like foliage or textured fabric tolerate significantly more compression than smooth images like skies or studio backgrounds before artifacts become visible.
Can compressed images be restored to their original quality?
No, lossy compression permanently discards image data, so a compressed file can never be restored to its original quality, which is why keeping an uncompressed source file is important.