Advanced PDF Compression Techniques: Beyond the Basics

Ghostscript offers four built-in compression presets that most users never look past: screen (lowest quality, smallest size), ebook (150 DPI images, good balance), printer (300 DPI, good print quality), and prepress (high quality, larger files). But beyond these presets, individual parameters give you fine control. `-dColorImageResolution=150` sets color image DPI independently from the preset. `-dGrayImageResolution=150` controls grayscale images separately. `-dDownsampleColorImages=true` enables downsampling for color images. `-dEmbedAllFonts=false` with `-dSubsetFonts=true` removes full embedded fonts and keeps only the characters used in the document, saving significant space in font-heavy documents. `-dCompressFonts=true` compresses font data. Combining these parameters allows you to achieve ebook-quality image resolution while stripping metadata and subsetting fonts for maximum size reduction.

1. 1Step 1: For documents where text quality matters but images can be lower resolution, use a mixed approach: set image DPI to 96 while keeping text rendering settings high. In Ghostscript: `-dColorImageResolution=96 -dGrayImageResolution=96 -dMonoImageResolution=300` — this keeps monochrome (black and white) images at full quality while compressing color images aggressively.
2. 2Step 2: Strip all metadata using `-dFastWebView=false -c '[ /Title () /Author () /Creator () /Producer () /Keywords () /CreationDate () /ModDate () /DOCINFO pdfmark'`. Metadata embedded in PDFs can add hundreds of kilobytes to large documents with many revisions.
3. 3Step 3: For scanned multi-page documents, apply OCR first using LazyPDF's OCR tool. Adding a searchable text layer slightly increases file size initially, but allows subsequent compression to be more aggressive because the document no longer relies solely on image fidelity for readability.
4. 4Step 4: Use split-compress-merge for mixed documents. Split pages into image-heavy and text-heavy sections, apply heavy compression to image sections and lighter compression to text sections, then re-merge. This preserves text quality where it matters most while maximizing compression on photo-heavy pages.

Many real-world PDFs contain a mix of content types: text-heavy summary pages, image-heavy illustration pages, and full-photograph product shots. Applying uniform compression across all pages wastes quality on text pages (which don't benefit from high resolution) while potentially over-compressing critical image pages. The selective approach: split the PDF into content groups using LazyPDF's split tool, apply different compression levels to each group (heavy compression for text-only pages, lighter compression for images you need to remain crisp), then merge the compressed groups back together. This technique can achieve an overall compression ratio approaching what screen-quality compression provides while maintaining noticeably better visual quality on key image pages.

A counterintuitive but effective technique: apply OCR to a scanned PDF before compressing it. Here's why this works. OCR adds a text layer to the PDF that describes the document's words. Once the text layer exists, compression tools can be more aggressive with the underlying image data — the text layer ensures the document remains searchable and readable even at very low image resolution. The OCR text layer itself is extremely compact (text data is far smaller than image data). Running OCR first, then applying screen-quality Ghostscript compression, often achieves final files 10–20% smaller than compressing without OCR, while actually improving the document's searchability. LazyPDF's OCR tool (powered by Tesseract.js) can process the document, then the compress tool handles the final size reduction.

When you need to compress dozens or hundreds of PDFs regularly — monthly invoice archives, weekly report collections, or document digitization projects — manual single-file compression becomes impractical. For Windows users, a PowerShell script can loop through a directory of PDFs and call Ghostscript on each: `Get-ChildItem *.pdf | ForEach-Object { gs -sDEVICE=pdfwrite -dPDFSETTINGS=/ebook -dNOPAUSE -dBATCH -sOutputFile="compressed_$($_.Name)" $_.FullName }`. For macOS and Linux, a bash loop achieves the same: `for f in *.pdf; do gs -sDEVICE=pdfwrite -dPDFSETTINGS=/ebook -dNOPAUSE -dBATCH -sOutputFile="compressed_$f" "$f"; done`. These scripts process each file sequentially; for parallel processing, tools like GNU parallel or Python's multiprocessing module can dramatically reduce total processing time for large archives.