Skip to content
ForgePlug — Logo
All resources
PDF Guides

How to Reduce PDF File Size Without Making Documents Unreadable

PDFs bloat for predictable reasons — images, fonts, metadata. Here's what actually takes up space and in what order to attack it.

5 min read · By ForgePlug Team · Published August 15, 2026

A 40-page PDF that should weigh 2 MB often weighs 40 MB. The cause is almost never the text — text is cheap. The causes are, in order: embedded images at print resolution, embedded fonts, and accumulated metadata and junk objects from years of editing. Shrink those three things and you recover most of the size without touching a single word.

What actually takes space

  • Images — a single full-page photo at 300 DPI can exceed 1 MB; a scan of 50 pages at 600 DPI can exceed 200 MB. This dwarfs everything else.
  • Fonts — every embedded font family adds a few hundred KB; embedding five families across a document adds up fast.
  • Metadata & junk — thumbnails, edit history, duplicate objects, and XMP blobs can add hundreds of KB on files edited by many programs over the years.

Where to start

  1. Compress the images inside the PDF — this is 80% of the win on scanned or photo-heavy documents. Re-encoding JPEGs at a sane quality (around 80–85%) is visually near-lossless on screen.
  2. Remove or subset embedded fonts — if the document only uses a handful of glyphs, only those need to ship.
  3. Strip metadata — creator info, edit history, and previews are safe to remove for distribution.
  4. Only then look at the page count. Splitting a bloated document into chapters doesn't reduce total size, but it makes each part manageable to email and upload.

Watch for the scanner trap

A scanner saves photos of pages, not text — which means the 'text' in a scanned PDF is actually a full-page image on every page. If the document truly only needs to be readable, converting pages to JPEG at a moderate DPI and rebuilding the PDF will cut size by 10× or more.

What you can realistically expect

A text-only PDF with a couple of fonts might drop 30–60%. A photo-heavy brochure typically drops 60–80%. A scanned document drops 80–95% if you're willing to lower the image quality. The trade-off is always the same: smaller files mean softer images and sometimes weaker fonts. The goal is the smallest file that still looks right on the screens and printers your readers actually use.

Compress a PDF now

ForgePlug's Compress PDF runs entirely in your browser — upload nothing, watch the size estimate, and tune a quality slider until it looks right.

Open Compress PDF

Resolution is usually the bigger lever than quality

Scanners frequently default to 600 DPI, which is far beyond what a document needs. For text that will be read on screen and printed on an office printer, 200 to 300 DPI is ample; 150 DPI is often fine for something that will only ever be viewed on screen. Halving the resolution removes roughly three quarters of the pixels, because pixel count scales with area — a much larger reduction than any quality slider will give you, and with less visible damage to the text.

Colour mode compounds it. A black-and-white document scanned in full colour stores three colour channels per pixel to represent grey. Rescanning or converting to greyscale cuts the data substantially, and for pure line-art documents a bilevel scan is smaller again. If you control the scanning step, fixing the settings there beats compressing afterwards.

Where the artifacts show up first

Lossy compression discards fine high-frequency detail, and in a scanned document that detail is the edges of letters. Push too far and characters soften, develop a faint halo, and thin strokes begin to break up — which is why the smallest text on the page is the right thing to inspect. If the body text at 100% zoom still reads cleanly, the setting is safe regardless of what the file size suggests.

There is a practical consequence for scanned documents specifically: aggressive compression degrades OCR accuracy even when the page still looks acceptable to a human. If the PDF needs to remain searchable, or will be processed by anything automated, run the text recognition before compressing rather than after.

Two savings that cost nothing

Font subsetting embeds only the characters actually used rather than the complete typeface. A document using four weights of a font family can carry a surprising amount of unused glyph data, and subsetting removes it with no visual change whatsoever.

The other is accumulated cruft. PDFs edited repeatedly retain incremental revision history, orphaned objects, and old versions of replaced content — none of which appears on any page. A document that has been through several rounds of edits can shed a meaningful fraction of its size simply by being rewritten cleanly, before any quality is sacrificed at all.

Compression is not redaction

Content hidden behind a white box, cropped out of view, or covered by another element is still present in the file and recoverable — compressing changes nothing about that. To genuinely remove sensitive material, delete the page or use a tool that actually redacts rather than conceals.

More PDF Guides