How PDF Compression Works Under the Hood

Someone asked me last week why their 120-page annual report shrank by 4% while a 12-page scan of signed contracts dropped from 48 MB to 6 MB. Same button. Same tool. wildly different outcomes.

That gap makes sense once you know what's inside a PDF — and what's actually being squeezed.

A PDF is a container, not a single image

From the outside, a PDF looks like one file. Inside, it's a structured bundle:

• Pages that reference content streams
• Fonts (embedded or referenced)
• Images (photos, logos, scanned page bitmaps)
• Vector graphics (lines, shapes, text outlines)
• Metadata (title, author, creation date)
• Optional extras (bookmarks, forms, attachments)

Compression tools target the heavy parts — almost always images — and tidy structural waste. They rarely rewrite body text character by character.

Where the bytes actually live

Open any bloated PDF in a profile mindset:

| Source type | Typical weight | Compression impact | |-------------|----------------|-------------------| | Phone scan saved as PDF | Very high | Large — entire page is an image | | Word / Google Docs export | Low–medium | Small unless many high-res images | | PowerPoint export | Medium–high | Moderate — slides often rasterize effects | | CAD / design export | Variable | Depends on embedded previews | | Mixed merge (scan + digital) | High | Images dominate savings |

If you've read When Should You Compress a PDF?, you know when to compress. This article is about what changes when you do.

Image compression inside PDFs

Most size reduction comes from re-encoding embedded images:

1. Downsampling — lowering resolution (e.g., 300 dpi → 150 dpi for screen reading)
2. Lossy JPEG recompression — merging similar color blocks, reducing fine detail
3. Color space changes — RGB to grayscale for monochrome scans
4. Stripping invisible layers — OCR layers or duplicate preview images some scanners add

Text rendered as real text objects stays sharp because it's vector instructions: "draw glyph A at position x,y." Compression doesn't pixelate that the way it pixelates a photograph of a page.

When entire pages are images

Flatbed scans and phone camera "document" modes often produce PDFs where each page is one big bitmap. Then compression behaves like photo compression — because it is photo compression wrapped in PDF clothing.

That's why contracts scan-compress dramatically and spreadsheets barely move.

Object and stream optimization

Beyond images, optimizers may:

• Remove duplicate font subsets when the same font was embedded multiple times after merging
• Drop unused objects left over from editing cycles
• Compress content streams with Flate (zlib-style) encoding — lossless packing of PDF instructions
• Remove thumbnails or embedded alternate images you never see

These fixes add up on messy files — especially after merging several PDFs — but they won't halve a clean two-page letter.

What compression usually leaves alone

• Searchable text layers (unless the whole page is an image)
• Hyperlinks and bookmarks (in most consumer tools)
• Vector line art at reasonable complexity
• Form field structure (though some aggressive optimizers flatten forms — test if you rely on fillable fields)

Step-by-step: compress with eyes open

1. Identify the source — scan-heavy vs digital export.
2. Merge first if combining files (Merge PDF), then compress once.
3. Open Compress PDF and start medium for mixed documents.
4. Download and check:
   • Smallest text on page 1
   • A photo or scan page if present
   • File size vs original
5. If email still rejects it, see email attachment limits before cranking to maximum.

Worked example: board pack

Input: 35 MB merged PDF — 10 slides exported from PowerPoint + 40 pages of scanned approvals.

What compression likely did:

• Recompressed slide embedded images moderately
• Aggressively downsampling 40 scan pages (where most bytes lived)
• Removed duplicate font entries from slide exports

Result: ~9 MB, readable on laptop, acceptable for email. Signatures still legible at 100% zoom but not print-archive quality.

Archive copy: Original 35 MB kept offline for records.

Common misconceptions

"Compress always improves quality"
No — it always trades detail for size on image-heavy content.

"PDF compression equals encryption"
Unrelated. Size reduction doesn't protect content. See PDF password protection for access control — and its limits.

"Print-to-PDF then compress fixes everything"
Print-to-PDF often rasterizes content into images, making files bigger and harder to search. Compress can't fully undo that decision.

Privacy and browser-based compression

When compression runs in your browser (how browser PDF tools protect privacy), the algorithm still reads image streams locally — nothing mystical about "upload-free" except that bytes may never leave your device. Sensitive docs still deserve trusted tools and retained originals.

Alternatives when compression isn't enough

• Split PDF — send chapters separately
• Re-scan at lower dpi (300 → 200 for internal reference)
• Replace scan pages with digital signatures where legally acceptable
• Share via link instead of attachment

Troubleshooting

Does compressing a PDF re-render every page? Most tools re-encode embedded images and remove redundant objects. Pure text pages often shrink only slightly because the text itself is already compact vector data.

Why do scanned PDFs compress more than Word exports? Scans are usually full-page bitmaps — huge image streams. Word exports are mostly text and vector shapes, which take less space to begin with.

Is ZIP the same as PDF compression? No. ZIP wraps the whole file. PDF compression optimizes objects inside the PDF structure — especially images and duplicate resources.

Limitations

When not to use this approach

Related reading

• When Should You Compress a PDF?
• How to Compress PDF Files Without Losing Quality
• Combining Scanned and Digital PDFs

Conclusion

PDF compression works because PDFs are assemblies — and most assemblies carry overweight photos disguised as pages. Understand that, and the compress button stops feeling random. You'll predict which files shrink, which barely change, and when to keep an untouched original instead of chasing megabytes.