Compress a PDF without losing quality
Free savings come from removing duplicate images, unused fonts and metadata — none of it visible. Anything beyond that re-encodes images and is lossy. The biggest quality killer is not compression itself but compressing the same file repeatedly.
Open the Compress PDF to Size tool
Every page promising "compress without losing quality" is selling a half-truth. Part of the job really is free. The other part cannot be.
Free: the waste
- Duplicate images. A logo stored once per page instead of once per document. Removing the duplicates changes nothing you can see.
- Unused fonts. Subsetting keeps only the glyphs actually used — often several megabytes of a font family nobody sees.
- Metadata. Author, application, revision history, embedded thumbnails. Pure overhead.
- Unused objects. Leftovers from editing: deleted pages still referenced, orphaned resources.
A file that is 40 MB because of duplication can drop to 5 MB with zero visual change. That is the case the marketing promises, and it is real — just narrower than advertised.
Not free: the images
Once the waste is gone, the only remaining bytes are content: image pixels. Reducing them means downsampling or re-encoding, and both are lossy by definition. The honest question is not "can I avoid loss" but "where does it stop being visible".
| Change | Visible? |
|---|---|
| 300 → 250 DPI | No, at normal reading size |
| 300 → 150 DPI | Only when zoomed or printed |
| JPEG quality 90 → 75 | Rarely; slight softening in fine detail |
| JPEG quality 75 → 50 | Yes — blocky edges around text and high-contrast lines |
| Below 100 DPI | Always, and often illegible on paper |
The habit that damages documents most
Repeated compression. Compressing an already-compressed PDF re-encodes the images a second time, and each generation loses more. People do this because the file was still slightly over a limit, so they run it through again — and the second pass does far more damage than the first.
The fix: always start from the original. Set the final target in one pass, and if you miss it, go back to the source rather than compressing the output.
A workflow that keeps quality
- Keep one untouched original, always.
- Do the lossless pass first: deduplicate, subset fonts, strip metadata.
- Downsample to what the destination needs — 150 DPI for screen, 300 for print.
- Hit the byte target in a single step, with a small margin, and stop.
The test that settles it
Open the compressed file at 100% and look at the smallest text and one photograph. That is the whole quality question. If both are fine, no amount of theory about lossy compression changes the answer for your document.
Questions
Is lossless PDF compression possible?
Partly. Removing duplicates, unused fonts and metadata is lossless. Shrinking the images themselves is not.
Why does my PDF look worse after the second compression?
Because each pass re-encodes the images again. Compression is generational: always start from the original file.
Which is better, lower DPI or lower JPEG quality?
Lower DPI, when the destination is a screen. It reduces the pixel count once and keeps each remaining pixel clean, rather than damaging every pixel.
Can I check quality before sending?
Yes — zoom to 100% on the smallest text and on a photograph. That covers everything compression affects.
Updated 2026-11-16 · written for this site, no trackers, no account.