A PDF that's unexpectedly large is one of the most common frustrations in document work. A one-page invoice shouldn't be 8 MB. A two-page contract shouldn't be larger than a high-resolution photograph. When that happens, it's almost always traceable to a specific cause โ and most of them are fixable.
Images Are Almost Always the Answer
In the vast majority of cases, a PDF is large because of embedded images. Text stored as actual characters in a PDF is extraordinarily compact โ a 100-page text document might be under 200 KB. A single full-page photograph at 300 DPI can be 3โ5 MB on its own. When a PDF contains multiple high-resolution images, every scan is stored as a photo, or graphics from design software are embedded at print quality, file sizes balloon quickly.
This is why scanned documents are the most common source of unexpectedly large PDFs. Every page of a scan is stored as a raster image โ there's no text, no vector graphics, just pixels. A 10-page scanned contract at 300 DPI in color can easily reach 40โ80 MB before any compression is applied.
Common Causes and How They Add Up
The table below covers the most frequent reasons a PDF ends up larger than expected, how much they typically contribute to file size, and whether compression can address them.
| Cause | Typical Size Impact | Compressible? | Best Fix |
|---|---|---|---|
| High-resolution embedded photos | Very large (can be 90%+ of total) | Yes โ dramatically | PDF compression tool |
| Scanned pages (image-based PDF) | Large (2โ10 MB per page) | Yes โ very well | OCR + compression |
| Full embedded fonts (not subset) | Moderate (1โ5 MB) | Partially | PDF Optimizer / font subsetting |
| Multiple copies of same image | Moderate | Yes | PDF Optimizer / deduplicate |
| Uncompressed vector graphics | Moderate | Partially | Re-export from source |
| Metadata, thumbnails, hidden layers | Small (but cumulative) | Yes | Sanitize / remove hidden info |
In most real-world cases, images and scanned pages are responsible for the lion's share of file size. Addressing those with a PDF Compression tool typically resolves the problem without needing to dig into the less common causes.
Why Scanned PDFs Are Especially Large
Scanners default to settings optimized for image quality, not file size. A typical office scanner set to 300 DPI in color produces about 3โ4 MB of image data per page before any PDF compression. At 600 DPI (sometimes used for fine text or documents that will be enlarged), that doubles again. Most of that resolution is unnecessary for on-screen reading โ 150 DPI is perfectly sharp on a monitor, and even 100 DPI is sufficient for most reference documents.
Running a scanned PDF through an OCR tool before compressing produces the best results. OCR adds a text layer to the document, which means the compression tool can then apply more aggressive image compression to the background scan while the text remains perfectly readable through the text layer. The combined effect can reduce a 50 MB scanned document to under 5 MB without any perceptible loss in readability.
Font Embedding and Why It Matters
PDFs embed fonts so that the document looks the same regardless of which fonts the viewer has installed. When a font is embedded in full, the entire font file is included โ including all characters, weights, and variations that the document doesn't actually use. A single full-embedded professional font can add 1โ3 MB to a file.
The fix is font subsetting, where only the specific characters actually used in the document are embedded. A document using only the Latin alphabet needs a tiny fraction of a full Unicode font file. Adobe Acrobat Pro's PDF Optimizer and Ghostscript both support font subsetting. For most documents exported from Word or Google Docs, modern export tools already subset fonts automatically โ but PDFs from older software or some design tools may not.
Hidden Content You Don't Know Is There
PDFs often contain more than what's visible on screen. Document properties, author information, revision history, comments, hidden layers, and embedded thumbnails all contribute to file size without showing up when you open the document. A PDF that went through multiple review cycles might have accumulated dozens of revision snapshots embedded in the file. Design PDFs from Illustrator or InDesign sometimes include the full editing data from the source application.
Stripping these hidden elements is called sanitizing or cleaning the PDF. Acrobat Pro's Remove Hidden Information tool and Document Properties cleaner handle this thoroughly. Some online compression tools also strip metadata as part of their optimization โ check the tool's description to confirm. For sensitive documents, sanitizing before sharing is also good practice from a privacy standpoint.
How to Fix a Large PDF
For most oversized PDFs, the fastest fix is to run the file through WukongPDF's PDF Compression tool. Upload the file, apply compression, and download the result. For image-heavy documents, this typically reduces file size by 60โ85% in under a minute. For documents where compression alone doesn't get far enough โ perhaps because the PDF is mostly text and metadata โ consider these additional steps:
- Convert to Word, clean up the document, and re-export as PDF. This regenerates the file from scratch and often produces a significantly smaller output than the original.
- Use Ghostscript or Acrobat's PDF Optimizer with font subsetting enabled.
- Remove embedded thumbnails and metadata through Acrobat's Remove Hidden Information panel.
- For scanned PDFs: run OCR first, then compress โ the combined approach gives dramatically better results than either step alone.
If none of these reduce the file enough, the document may simply contain a lot of content that doesn't compress well โ detailed vector illustrations, many unique high-resolution photographs, or content that was already heavily compressed before being placed into the PDF. At that point, the only meaningful option is to reduce the content itself, or accept the file size and find another way to transfer it (cloud storage link rather than email attachment, for example).