Practical guide

How to Compress a PDF Without Ruining It

Most of a large PDF's size is usually images, not text. Knowing what makes your file big tells you which setting will shrink it, and which will only blur it.

Find out what makes the file large

Text and vector drawings take very little space. A 50-page report of plain text is often well under 1 MB. Size comes from embedded photos, scanned pages, full embedded fonts, and leftovers such as thumbnails, metadata history, and unused objects.

A quick test: if you can select the text and the document has few pictures, a structural clean-up is all it needs. If every page is a scan or a photo, only re-encoding the images will make a real difference.

Choose the right kind of compression

Lossless structural optimisation rewrites the file more efficiently. For example, it packs objects into compressed object streams (a PDF 1.5 feature) and removes non-essential metadata. Nothing visible changes, but the saving is usually modest.

Image re-encoding is where large savings come from. In ToolkitPoint's Compress PDF, the Web & Ebook preset re-encodes embedded JPEG images at 70% quality, up to 1600 px, and Screen & Email uses 50%, up to 1000 px. Text, fonts, and vector graphics stay sharp and selectable, because pages are not turned into pictures.

Check the result before you send it

Compare the old and new sizes, then open the new file and zoom to 200% on a photo and on small print. If the photos look blocky, step back to a gentler preset. If the size barely changes, the file was already efficient. A PDF of mostly JPEG scans can only shrink much further by lowering image quality.

Common mistakes

Primary sources and further reading

Open the related tool

Last reviewed: September 28, 2026