Have you ever run a 10MB PDF through a compression tool only to see it reduce by a meager 2% (to 9.8MB)? While most PDFs shrink by 50%–85%, certain documents resist compression entirely. Here is the technical explanation behind why this happens.
Shannon Entropy & The Limits of Compression
According to Shannon's source coding theorem in information theory, data cannot be compressed losslessly beyond its information entropy. If a file is already encoded with maximum entropy (randomized byte distribution), no lossless algorithm can reduce its size further without discarding information.
The 4 Main Reasons a PDF Resists Compression
- Already Optimally Compressed: If the PDF was previously generated with modern tools using FlateDecode and optimized JPEG streams, there is no redundant metadata left to remove.
- Dense Vector Complexity (CAD / Blueprints): Architectural schematics and maps contain hundreds of thousands of discrete vector line coordinates (
l,m,coperators). Downsampling does not affect vector coordinates. - Encrypted or DRM-Protected Objects: Encrypted streams look like pure random noise to compression algorithms, making deflation impossible.
- Embedded Non-Image Attachments: PDFs can carry arbitrary file attachments (e.g. embedded video, audio, or 3D models) that ignore standard PDF image compression.
Workarounds for Stubborn Documents
If a PDF resists standard compression:
- Use PDFRack Extreme Compression to apply lossy DCT quantization.
- Toggle Grayscale Conversion to strip color channels.
- For extreme vector complexity, convert pages to high-res images using PDF to Image (300 DPI) and re-compile using Images to PDF.
Frequently Asked Questions
Why does my PDF become slightly larger after compression?
If an already compressed image is re-encoded with higher quality settings, the new byte stream can occasionally exceed the original size. PDFRack automatically detects this and retains the smaller stream.