Why do some PDFs refuse text selection and behave like images?

Question

Weirdo · Accepted Answer

A scanned PDF page is often just a bitmap — no text objects. Your eyes see letters; the selection tool sees a picture. OCR lays invisible text over the scan; without that layer, copy and search have nothing to grab.

The PDF Association glossary, quoting ISO 32000-2 clause 8.2, lists five kinds of graphics objects a page can paint: paths, text, XObjects (often images), inline images, and shadings. Text objects carry character strings and glyphs. A scanner save often ships none of that — just a bitmap. The Association’s scanned-PDF note is blunt: instead of real, selectable text, the file contains images of each page. It looks like a document. As a file, it is a legal picture container.

OCR does not turn the scan into a born-digital page. It reads shapes in the bitmap and parks transparent characters on top. The glossary says Optical Character Recognition often places invisible text over a scanned image so you can select and extract it. Copy needs those characters. Skip OCR, and the viewer has nothing to highlight — the page never had text objects.

Why do some PDFs refuse text selection and behave like images?

The spec lets a page be only a picture

Selectable text is a hidden layer, not the picture