PDF tools

Text in a PDF is almost weightless. A three-hundred-page text document can sit comfortably under 2 MB. Size comes from images, and above all from scans: a page photographed or scanned at 600 DPI in full colour is a large photograph, and a document of them is a stack of large photographs regardless of how little ink is on the page.

Embedded fonts add to it — a PDF carries the typefaces it needs so it renders identically everywhere, and a document using several weights of a few families can carry several megabytes of font data. Files generated by design software often also retain hidden layers, metadata, and objects that are no longer visible on any page.

Compression resamples images to a sensible resolution, re-encodes them efficiently, subsets embedded fonts down to the characters actually used, and strips unreferenced objects. Text remains real text throughout, so the result stays selectable and searchable. A scan-heavy PDF commonly loses 70–90% of its size while still looking correct on screen and printing acceptably.

One caveat worth knowing: a scanned PDF is images of text, so it has no text layer at all. Nothing is searchable in it and PDF to Word cannot recover editable content from it, no matter how sharp the scan looks. Making a scan editable needs OCR, which is a different operation from conversion.