PDF 6 min read
How to shrink a PDF without turning text into mush
A PDF can be large for two very different reasons. It may contain high-resolution images, or it may contain a lot of embedded fonts, redundant objects and duplicated resources. The right compression method depends on which problem you have.
Browser-based compressors usually rasterise pages: each page is rendered to an image and rebuilt as a PDF. That reliably reduces size, but text becomes pixels. It is fine for scans and posters, and a poor choice for contracts that must remain searchable.
Know which kind of PDF you have
If you can select text and the file is only a few hundred kilobytes, it is probably already efficient. Re-rasterising it may make it larger. If the file is tens of megabytes and full of photographs, image downsampling will help.
Open the document at 100% and zoom in. Blurry images and sharp text suggest mixed content. A scan will look like a photograph even at high zoom.
Choosing DPI and quality
For on-screen reading, 96–120 DPI is usually enough. For office printing, 150–200 DPI is a reasonable compromise. 300 DPI is for archival scans and print production, and it will not produce a small file.
JPEG quality around 75–85 is normally visually acceptable. Dropping to 40 saves space but introduces ringing around text edges. Because rasterised text is an image, that ringing is what makes it look “mushy”.
When not to rasterise
Do not rasterise documents that need to remain searchable, accessible or legally exact: invoices, contracts, signed forms, academic papers. Those need structural optimisation or a desktop tool that can recompress images while preserving text objects.
If you must share a smaller version, keep the original and label the compressed copy clearly. Do not overwrite the only copy.
Check the result
Open the compressed file, scroll every page and zoom in on the smallest text. Check that page order, orientation and signatures are intact. A small file that is unreadable is not a successful compression.