What it actually produces
Measured, not estimated. A procedurally generated 2400×1600 (3.8 MP) test image built to carry photographic entropy: value noise across nine octaves down to near-pixel scale, a dense field of small high-contrast features, a smooth gradient, hard edges and light grain. It is not a photograph — it is reproducible, which a photograph is not. Each figure below comes from running that image through this page.
| Quality | Input | Output | Change |
|---|
| 40 | 375 KB | 126 KB | 67% smaller |
| 60 | 375 KB | 178 KB | 52% smaller |
| 82 | 375 KB | 374 KB | about the same size |
| 95 | 375 KB | 579 KB | 54% larger |
Your own files will differ — a photograph, a screenshot and a diagram compress very differently. The point of a fixed image is that the formats can be compared with each other, not that your result will match.
How do I get a JPG under a specific size?
Drop the file in, lower the quality slider, release it, and read the number. Repeat until it fits.
That sounds unsophisticated, and it is the correct method. There is no reliable formula from quality setting to file size, because the answer depends entirely on the image: a photograph of foliage and a photograph of a clear sky respond completely differently to the same setting. The readout is the only honest guide.
Every adjustment re-encodes from your original file rather than from the previous result, so you can move the slider as many times as you like without compounding the damage.
What the quality numbers actually do
They control how aggressively the encoder discards detail. They are not a percentage of anything, and the scale is not linear.
The table above shows the real curve, measured on this page. Two things in it are worth reading carefully.
Quality 82 produced a file the same size as the input. That is not a bug — the source had itself been encoded at quality 82, so asking for the same quality again reproduces the same amount of detail. To get smaller, you have to go below whatever the source used, and you usually cannot tell what that was.
Quality 95 produced a file 54% larger than the input. Raising quality above a file’s existing level does not improve the picture. It just spends more bits faithfully recording the compression artefacts already baked in. If a file looks over-compressed, no quality setting can undo it.
Can I compress without losing quality?
Not in any strict sense. JPG is lossy: smaller means less detail, always.
What you can do is lose an amount nobody will notice. Around quality 82, most photographs are indistinguishable from the original at normal viewing sizes. Below about 60 you will start to see softening in flat areas like skies, and haloing around hard edges.
If you need genuinely lossless size reduction, JPG is the wrong format for the job — that is what lossless PNG optimisation is for.
If the destination is a web page you control, compressing the JPG is rarely the best available move.
On the fixed test image used here, the same source came out 40% smaller as WebP and 73% smaller as AVIF at comparable visual quality — savings a JPG quality slider cannot reach without visible damage.
Compressing the JPG is the right call when the destination specifically demands JPG: an upload form, an email attachment, a print submission.
Questions
- How do I compress a JPG to a specific size, like under 2 MB?
- Drop the file in, then lower the quality slider and release it. The size readout updates after each re-encode, so you can move the slider down until the number is under your limit. There is no formula that maps quality to a target size — it depends entirely on the image — so watching the readout is the reliable method.
- Can I compress a JPG without losing quality?
- Not meaningfully. JPG is a lossy format, so making the file smaller means discarding more detail. What you can do is lose an amount that is invisible: at quality 82 most photographs are indistinguishable from the original on screen. If you need genuinely lossless size reduction, the image needs to be a PNG and optimised losslessly instead.
- What do the quality numbers mean?
- They control how aggressively the encoder discards detail, not a percentage of anything. On our fixed test image, a 375 KB source came out at 126 KB at quality 40, 178 KB at 60, 374 KB at 82 and 579 KB at 95. The relationship is steep at the top end and flattens lower down.
- Why did quality 82 produce a file the same size as my original?
- Because your original was probably already encoded at around that quality. Re-encoding at the setting a file already uses gains nothing — it simply reproduces the same amount of detail. To get smaller you have to go below whatever the source used.
- Why did quality 95 make my file bigger?
- Because you asked the encoder to preserve more detail than the source was saved with, so it spends extra bits faithfully recording the original's own compression artefacts. Raising quality above a JPG's existing level always costs size and never improves the picture.
- Does moving the slider repeatedly degrade the image?
- No. Every adjustment re-encodes from your original file, not from the previous result, so there is no generational loss no matter how many times you move it.
- Is my photo uploaded?
- No. The decode and re-encode both run inside this browser tab. The Uploaded readout beside the result measures the page's own outbound network traffic, so a successful compression reads 0 B.
- Is there a limit on file size or how many photos?
- No limit is imposed, because no bandwidth or storage is being paid for. Your device's memory is the only ceiling.
- Will compressing add a watermark?
- No. Nothing is added to the image, and no setting is held back behind a paid tier.
- Would converting to WebP or AVIF be better than compressing?
- For a web page, almost certainly. On the same test image, converting to WebP gave a file about 40% smaller than the JPG and AVIF about 73% smaller, at comparable visual quality. Compressing the JPG only helps if the destination requires JPG specifically.
- Does compressing remove EXIF data?
- Most metadata is not carried across, which strips embedded GPS coordinates and device details. Keep the original if you need them.
- What quality should I use for printing?
- 95 or above, or better, do not compress at all. Print resolves detail that a screen hides, and compression artefacts that are invisible on a monitor can be visible on paper.