How PDF compression works
Why scanned PDFs compress differently from text PDFs, and which quality checks matter.
Reviewed by HSQ Tools ·
Find what takes up the space
A PDF can contain text, font data, images and other objects. A document made from page scans may be dominated by photographs of the pages. A text report with few images may already be compact. These inputs respond differently to compression.
Reducing a PDF’s bytes does not imply that every page has been simplified in the same way. Text and image objects can be handled separately. Keep the original for printing, archiving and any signed-document workflow.
What this compressor attempts
HSQ first tries rewriting embedded images using JPEG quality 55 with a target of 72 dpi. If this does not produce a suitable reduction, it tries image recompression and a structural stream-compression candidate. The service selects a smaller acceptable candidate or reports that no reduction was achieved.
Image rewriting can reduce small-print readability and introduce artifacts. The 72 dpi target is not a promise that the output suits print. The tool does not provide a quality slider for selecting a print-specific preset.
Check representative pages
Open the output and examine a page with small text, one with a chart and one with a photo. Zoom into scanned text and compare line edges. If a barcode or QR code is present, test that it can still be read.
For a large report where the receiver needs only a chapter, splitting may be more appropriate than degrading images across the whole document. Neither operation replaces checking page count, links, forms and any signatures after processing. Files are uploaded to the processing server.
Inspect a compression result systematically
Keep the source PDF, compress a copy, and open both at the same zoom. Compare the smallest scanned text, chart labels, barcodes and photographs. Also try selecting text and opening links: a fallback rewrite can affect more than the byte count.
A text-only or already optimized PDF may return no useful reduction. If an image-heavy output becomes hard to read, keep the original or ask for a better source export. Do not assume a smaller result is suitable for print or a signed-document workflow.