How PDF Compression Works (Without Losing Quality)

TL;DR

PDF compression mainly works by downsampling images, subsetting fonts, and applying Flate/JPEG/JBIG2 streams. Image downsampling is the biggest lever — a 600 DPI scan dropped to 150 DPI shrinks 16x with no visible loss on screen. Born-digital PDFs barely shrink because they're already small.

Why PDFs get huge

Most oversized PDFs are oversized for one reason: images. Scanned pages saved at 600 DPI, full-bleed photos in a brochure, or screenshots embedded at original resolution can balloon a 10-page PDF to 50+ MB. Embedded fonts, redundant objects, and uncompressed streams add the rest. Compression attacks each layer in turn.

Image downsampling — the biggest win

Compressors detect raster images, downsample them to a target DPI (usually 150 for screen viewing or 300 for print), and re-encode them in JPEG (for photos) or JBIG2/Flate (for monochrome scans). Going from 600 DPI to 150 DPI cuts image weight 16x with no visible quality loss on screen. This single step typically shrinks a scan-heavy PDF by 70–90%.

Font subsetting

PDFs embed fonts so the document looks identical everywhere. Embedding the entire font file is wasteful — most documents use only a fraction of its glyphs. Font subsetting strips the file down to just the characters actually used in the document. A typical multi-language font drops from 2 MB to 200 KB after subsetting.

Object streams and Flate compression

Inside a PDF, every page, font, and image is a numbered 'object'. Modern compressors bundle small objects into 'object streams' and apply Flate (the same algorithm as ZIP) to the whole stream. The result is smaller and faster to parse.

Lossy vs lossless compression

Lossless compression (Flate, LZW, run-length) shrinks files without changing a single pixel — typical savings 20–40%. Lossy compression (JPEG, JBIG2 lossy mode) discards information your eye won't notice — typical savings 70–90%. Most online compressors blend both: lossless on text and vector graphics, lossy on photos. The trick is picking a quality level that's invisible to humans but devastating to file size.

Why some PDFs barely shrink

If a PDF is already optimised, well-compressed, or full of vector text and small icons (like a born-digital report exported from Word), there's almost nothing left to squeeze. Don't be surprised if a 200 KB invoice only drops to 180 KB. The big wins live in scanned PDFs and image-heavy brochures.

How to compress safely

Three rules: (1) Start with the highest-quality original you have — never re-compress an already-compressed PDF, you'll just degrade quality. (2) Pick the right preset: 'screen' for email and web (around 100–150 DPI), 'print' for print shops (300 DPI). (3) Always check the compressed output before sending — open it on a phone and a desktop, zoom into a detailed page, and confirm text is still crisp.

Related tools you'll need

After compressing, you'll often want to merge multiple PDFs into a single send, split a PDF to email a section instead of the whole file, or convert PDF to JPG when you only need a preview image. All run in your browser, free.

Frequently asked questions

Will compressing a PDF reduce its quality?

It depends on the preset. 'Screen' presets re-encode images at lower DPI and lower JPEG quality — fine for email and web, visibly worse if you print. 'Print' or 'high quality' presets leave images near-original. Always preview after compressing.

Why is my PDF still large after compression?

Either the source was already optimised or it has a hidden large element — embedded video, a 50 MB attachment, or a giant XFA form. Use a PDF inspector (Acrobat 'Save As Optimized' or PDF24's analyzer) to see what's eating space.

Is online PDF compression safe?

Reputable services delete files within minutes. Avoid uploading confidential documents to free services without checking their retention policy. Desktop apps (PDF24, Adobe Acrobat) keep files local.

What's the smallest a PDF can be?

A pure text PDF can compress to under 5 KB per page. A scan-heavy PDF rarely drops below 50 KB per page even at aggressive settings without becoming unreadable.

Can I compress a scanned PDF?

Yes — and this is where compression shines. Run OCR first (so text stays searchable), then downsample images to 150 DPI and re-encode with JBIG2 (for monochrome) or JPEG (for colour). 80–95% reductions are normal.