PDF Sanitization Guide: Compress, Flatten & Scrub EXIF

The Complex Anatomy of a PDF Document

Invented by Adobe in 1993 and subsequently standardized as ISO 32000, the Portable Document Format (PDF) is a PostScript-derived binary document container designed to guarantee pixel-perfect visual fidelity across any operating system or hardware printer.

A standard PDF consists of four distinct architectural sections:

1. Header: Declares the PDF version specification (e.g. %PDF-1.7).
2. Body: The content streams containing text strings, vector paths, embedded fonts, and raster images.
3. Cross-Reference Table (XREF): A byte-offset catalog allowing PDF readers to jump directly to arbitrary page objects without parsing the entire file linearly.
4. Trailer: Points to the document catalog dictionary and provides verification checksums.

The Hidden Security Hazard: Incremental Updates and Metadata

When you edit, redact, or annotate a PDF in standard desktop software, the software rarely rewrites the entire file. Instead, it appends an incremental update to the end of the file, updating the XREF table.

This means that black boxes drawn over sensitive text (such as Social Security Numbers, bank account numbers, or secret contract terms) may simply sit as transparent vector overlays above the original un-redacted plaintext! Anyone opening the file in a text editor or hex inspector can recover the original confidential data.

Key Sanitization Actions

• Scrubbing Author & System Info: Stripping the /Info dictionary, removing Author, Creator, Producer, and custom XMP metadata namespaces that leak organizational data.
• True Raster Redaction: Burning redaction directly into the underlying bitmap image rather than placing masking overlays.
• PDF Flattening: Collapsing multi-layer annotations, digital form controls, and sticky notes into a static base layer.

Meeting 100KB–200KB Official Portal Benchmarks

Many government recruitment, passport, and university portals impose strict upload limits (often requiring PDF files under 100KB or 200KB). Achieving this without unreadable blurring requires:

1. Selective Image Downsampling: Reducing image resolution from 300 DPI (print standard) to 150 DPI (screen resolution).
2. DCT Stream Compression: Converting raw uncompressed image streams to 80% quality JPEG streams within the PDF container.
3. Stream Deflation: Compressing text streams and font glyphs using the standard Flate (zlib) compression algorithm.