Almost every PDF compressor on the internet promises to shrink your file “without losing quality”. Some of them mean it. Many are quietly converting your pages into pictures, which makes the file smaller and also makes the text impossible to select, search or copy. Before picking a tool, it helps to know what “without losing quality” can honestly mean for a PDF, because the answer depends almost entirely on what is inside your particular file.
Two different promises hiding in one phrase
Lossless compression makes the file smaller while keeping every piece of content exactly as it was. Nothing is re-encoded; the file is just stored more efficiently. Visually lossless compression does throw information away, but only information your eyes can't detect at the size you'll actually view or print the document.
Both are legitimate. The problem is that lossless compression usually saves far less than people expect, so tools that advertise big reductions are almost always doing the second kind, and some go further and damage things you care about. Knowing which one you need is the first decision.
What actually takes up space in a PDF
- Text and vector drawings. Tiny. A full page of text is usually a few kilobytes. Lines, tables and charts drawn as vectors are similarly small.
- Fonts. Embedded so the document looks the same everywhere. Most programs embed only the characters actually used (a “subset”), which keeps this to tens of kilobytes. Arabic fonts are larger than Latin ones because of the many letter forms, but still rarely the problem.
- Images. Photos, logos, scanned pages. In practically every oversized PDF, this is where the megabytes are.
- Leftovers. Metadata, unused objects, thumbnails, and old versions of the content that some editors append each time you save instead of rewriting the file.
A quick sense of scale: a single A4 page scanned at 300 dpi is 2480 × 3508 pixels, which is about 8.7 megapixels. That is as large as a photo from a phone camera, and it's one page. A twenty-page scanned contract is twenty photos.
What truly lossless compression can do
A lossless clean-up removes unused objects, recompresses internal data streams, merges duplicates and drops the old revisions left behind by repeated saving. It helps a lot on files that were edited and re-saved many times, and on PDFs from old software that stored data uncompressed.
On a normal file, though, the gain is small. We took a one-page document exported from a word processor (50 KB) and ran a full lossless clean-up on it. It came out at 49 KB. Modern programs already write compact PDFs, so there is simply nothing left to remove. If a tool claims to take 80% off an ordinary document without changing anything, be sceptical.
Where visually lossless compression comes in
Since images are almost always the bulk, the real savings come from two image settings:
- Resolution: how many pixels are stored per inch of page. Beyond what the screen or printer can reproduce, extra pixels add size and nothing else.
- JPEG quality: how aggressively each image is compressed. Around 75–85 is practically indistinguishable from the original for photos and scans; below about 60, blocky artefacts start appearing around edges and text.
| What the PDF is for | Image resolution | JPEG quality |
|---|---|---|
| Reading on screen, emailing, uploading to a portal | 120–150 dpi | 70–80 |
| Office printing | 200–300 dpi | 80–85 |
| Professional printing (brochures, books) | Leave as is — ask the printer | Don't recompress |
Choosing a method for your kind of file
If the PDF was created from Word or another program
Text in these files is real text, and you want to keep it that way. The only thing worth compressing is the images inside, and the best place to do that is usually before the PDF exists:
- In Word, click any picture, open Picture Format → Compress Pictures, untick “Apply only to this picture”, choose Print (220 ppi) or Web (150 ppi), and tick “Delete cropped areas of pictures”.
- Check File → Options → Advanced → Image Size and Quality. If “Do not compress images in file” is ticked, every photo is stored at full camera resolution.
- Export the PDF again. Our Word to PDF guide covers the export options.
If you only have the PDF, you need a tool that recompresses images while leaving text alone. Adobe Acrobat Pro's compress/optimise feature does this. A free alternative is Ghostscript, a command-line tool; this command resamples images to about 150 dpi and keeps text as text:
gswin64c -sDEVICE=pdfwrite -dPDFSETTINGS=/ebook -dNOPAUSE -dBATCH -sOutputFile=smaller.pdf original.pdf
Use /printer instead of /ebook for roughly 300 dpi if the document will be
printed.
Tools that redraw each page as a picture, including our own Compress PDF tool, are the wrong choice for text documents. We ran the same 50 KB word-processor PDF through our tool's lightest setting: it came out at 88 KB, larger than before, and its text could no longer be selected. Our tool warns you when it detects real text for exactly this reason.
If the PDF is a scan
Here the calculation flips. Every page of a scan is already a picture, so turning pages into compressed pictures loses nothing you had before. It just stores the same picture more sensibly. This is where big reductions are both possible and harmless.
To measure it, we built a five-page colour scan at 300 dpi, the default on many office scanners. It weighed 5.71 MB. Then we ran it through the three levels of our Compress PDF tool:
| Setting | Result | Reduction | Good for |
|---|---|---|---|
| Original scan (300 dpi) | 5.71 MB | — | — |
| Light (150 dpi) | 659 KB | 89% | Documents you may print |
| Balanced (120 dpi) | 259 KB | 96% | Screen reading, portals, email |
| Maximum (96 dpi) | 153 KB | 97% | Hitting a strict size limit |
On screen, the Light and Balanced results are hard to tell apart from the original at normal zoom. Maximum is still readable, but small print starts to soften, so use it only when a size limit forces you to.
Compress PDF Built for scanned documents. Pages become images, so text stops being selectable.
If it's a mix
A report with a few scanned pages appended is the awkward case. Compressing the whole file as images sacrifices the text pages to save space on the scanned ones. A cleaner route is to split the scanned pages out, compress only those, and merge everything back together.
Split PDF Pull out just the pages you need, or separate every page into its own file.
How to check that nothing important was lost
- Zoom to 200% on the smallest text in the document, such as footnotes, stamps and signatures, and compare it with the original.
- If it will be printed, print one page. Screens forgive more than paper does.
- Try selecting a sentence and searching for a word. If that worked before and doesn't now, the tool converted your text to images.
- Check the page count and that no pages are rotated or blank.
- Keep the original. Compression is a one-way street; you can't get the discarded detail back.
If a PDF carries a digital signature, any compression changes the file and the signature will no longer validate. Compress first, then sign, or send the signed file as it is.
The short version
- Truly lossless compression is safe but usually saves little on a modern file.
- Almost all the real savings come from images: lower their resolution to what you need and use sensible JPEG quality.
- For documents created in Word and similar programs, compress the images at the source and keep the text as text.
- For scans, converting pages to compressed images is harmless and cuts size by 90% or more.
- Always compare with the original before you delete it.